The oil market has already decided what it thinks.

Brent, which spent the spring climbing towards a hundred dollars on the closure, slipped under eighty this week — its lowest since early March — as the first Iranian tankers in two months began clearing the blockade.

The war-risk premium is unwinding on a signature that has not yet been written.

Two and a half thousand miles east, in Karachi, the unwinding has not reached the regasification berth. Since the closure began, Qatar — ninety per cent of the LNG Pakistan imports, a fifth of the world’s — has landed three cargoes there.

Three, in three months.

The grid covers the daylight on rooftop solar it did not have in 2022 and runs more hydro than it has in years, and then the sun goes down, and the load-shedding schedule takes over the hours the molecule used to hold.

The price has settled.

The trade has not.

That is the entire distinction, and Pakistan is living inside it — a country holding a contract for gas it cannot take delivery of, discovering, as the paper market congratulates itself, that a claim is not a cargo.

Pakistan secured the price, not the molecule

A long-term LNG supply contract is not gas in a tank.

It is a claim on future delivery at a formula price — a promise that, on a stated schedule, a named counterparty will load cargoes a buyer is obliged to lift.

After the 2022 shock priced it out of the spot market, Pakistan built its import strategy on exactly that promise, concentrating its supply onto Qatar until a single counterparty stood behind ninety per cent of the LNG it brought in.

The logic was the logic every importer reaches for after a spot-market burning: take the volume off the volatile market, lock the terms, and convert an uncertain price into a secured supply.

Read that the way a settlement desk reads a trade, and the flaw is visible before the strait ever closes.

Every trade has two legs — payment and delivery — and they clear independently.

Pakistan had locked the payment leg: the price formula, the volume schedule, the take-or-pay clause that binds the buyer to pay for contracted molecules whether it lifts them or not.

What no contract can lock is the delivery leg, because delivery does not run on the contract. It runs on a waterway twenty-one nautical miles wide at its narrowest, and the contract has no jurisdiction over the water.

The instrument secured the price of the molecule and left the arrival of the molecule to a chokepoint it could not write a clause against.

Force majeure is the escape hatch in the guarantee

What happens when the delivery leg cannot clear is not that the supplier fights to deliver anyway. It is that the supplier reaches for the clause that was in the contract all along.

In March, after missile and drone strikes set fires across the Ras Laffan complex, QatarEnergy declared force majeure on its deliveries.

Force majeure is the provision that suspends a party’s obligations when performance becomes impossible through no fault of its own; it is a legal declaration, not a request, and it accepts no liability for the shortfall.

Here is the part that should change how a buyer reads its own contract.

The same clause that frees Qatar from the duty to deliver frees Pakistan from the duty to take — and dumps it straight back into the spot market it had paid the contract to escape.

The guarantee Pakistan bought to remove its exposure to volatile prices contained, in its own text, an exit that opened in precisely the state of the world the guarantee was bought for.

A guarantee with an escape hatch is not a guarantee. It is a fair-weather promise — money-good while the molecule moves freely, and suspended the moment it does not, which is the only moment the buyer ever needed it.

The security was the exposure

Trace the loop and the inversion at the centre of it comes clear.

The belief is that a long-term Qatari contract is secured supply.

That belief shapes behaviour: a country acting on it concentrates ninety per cent of its imports on one counterparty and defers the things a diversified position would have demanded — the storage, the second and third supply relationships, the domestic generation flex — because the contract is understood to have covered all that.

The behaviour then alters the fundamental: the concentration is not the building of security but the quiet removal of the optionality that survives a delivery shock, so that the more secured Pakistan felt, the less hedged it actually was against the only failure that would matter.

And then the loop feeds back.

Cargoes arrived, year after year, each one confirming the belief — until the chokepoint closed and the contract paid out a force majeure notice instead of a ship, destroying the belief and leaving the country holding neither the molecule nor the flexibility it had traded away to hold the claim. The inflection was the March morning the notice arrived: the moment the security revealed itself as the exposure.

The cruelty of LNG specifically is that nothing stands between the failed claim and the dark terminal. When crude’s flow failed, I wrote here a week ago, a stock in motion masked it — strategic and commercial barrels walked down out of the salt caverns to hold the price steady while the throughput was gone. Gas has no such cavern.

There is no strategic reserve of molecules to draw against a suspended contract; when the claim fails, it fails to an empty berth, and the only replacement is the spot cargo bought at a distressed premium. Pakistan was awarded April delivery near eighteen dollars forty per million Btu and June delivery near nineteen — each above the Asian marker on the day it bought.

That gap is the premium a must-buy pays for being the buyer that cannot walk away; Bangladesh, in the same scramble, paid towards twenty-eight.

The premium is the running cost of holding a claim that would not convert.

The fail clears after sundown

The shortfall does not land evenly across the day, which is what makes 2026 survivable where 2022 was not.

Rooftop solar Pakistan had barely begun installing four years ago now carries a real share of the daytime load, and greater water releases have lifted hydropower availability from roughly eighteen hundred megawatts to over four thousand.

The grid is genuinely better defended than it was. Then the sun sets, evening demand climbs towards a summer peak above twenty-eight thousand megawatts, and the one source that was supposed to fill the gap between sundown and the night trough is the gas that is no longer arriving.

At spot prices, the arithmetic forecloses the obvious answer.

LNG-fired generation runs to around fifty-one rupees a kilowatt-hour before operating and capacity charges are even counted, so the grid does the rational thing and sheds the load rather than burn the cargo it cannot afford. The 2022 rhyme is exact in its lesson and opposite in its mechanism: when prices touched seventy dollars then, Pakistan’s contracted suppliers Gunvor and Eni diverted its cargoes to higher-paying European buyers and power prices rose a hundred and fifty-five per cent.

One supplier would not deliver because Europe paid more; another now cannot because the strait is shut.

Two routes to the same dark evening, and the long-term contract was protection against neither — a default by economics on one side, a force majeure by geography on the other, the buyer exposed identically to both.

The battery is the only good delivery

The one supply that is not a claim on a distant counterparty routed through a contested waterway is the supply already standing on the grid when the sun goes down.

Battery storage charges off the solar surplus Pakistan now wastes in the middle of the day and discharges into the evening peak — the exact hours the failed cargo was meant to cover. IEEFA’s accelerated case puts battery imports at sixteen-point-one gigawatt-hours by 2030, enough to cover fifty-four per cent of projected peak demand

On those numbers, the storage build is not a green aspiration bolted onto the energy mix; it is the closest thing to delivery security the country can actually take possession of.

A battery is only good delivery, though, if the grid can place what it holds.

Discharging gigawatt-hours into a distribution network at the evening peak without widespread smart metering, transformer and consumer monitoring, and feeder-level automation invites the voltage and frequency problems that strand storage exactly when it is needed — the same modernisation Pakistan has deferred for as long as it deferred its supply optionality.

The molecule has taught the lesson plainly enough.

Whether the grid is wired to learn it is the open question, and it is an engineering question with a budget line, not a hope.

What clears at the speed of the strait

Now that the memorandum is initialled and the price has come off, the thing visible from Karachi that was invisible in March is that the contract was never supply.

It was a price hedge with a delivery hole, and the hole is structural — it opens whenever the molecule cannot move, and the molecule moves through one strait that a willing supplier’s force majeure cannot reopen and a buyer’s signature cannot either.

The lesson Pakistan is being handed is the one every mono-source importer keeps declining to learn: a long-term contract converts price risk into delivery risk and files it under security.

The failure mode simply differs by who is trapped. Where a German take-or-pay book traps its holder into taking molecules that a falling domestic demand no longer wants, Pakistan’s traps it the other way — needing the molecule and holding only the claim. Same instrument, opposite failure, and in neither direction the security it was sold as.

The playout is already running at desk level.

The power buyer facing the evening gap re-marks the long-term Qatari claim against what it can actually take, finds it short, and chooses between lifting a spot cargo at a premium to its own marker or scheduling the load off — and each spot lift near nineteen dollars pushes the all-in cost towards fifty-one rupees a unit, which pushes the tariff, which makes the next evening’s load-shed the cheaper choice and quietly buries the true size of the gap inside the outage statistics.

The structural consequence sits one level up: the delivery-fail premium now belongs inside the valuation of any mono-source delivered-cargo contract. For Pakistan, and for every buyer watching Ras Laffan, the long-term claim reprices downward against physical optionality — storage, diversified routing, domestic flex — and the battery, not the contract, becomes the asset that clears at the peak.

Watch three gauges on that same channel.

The slope on Pakistan’s next term contract: at the fifteen-to-seventeen per cent oil-linked slopes of 2025, a signature now would lock the high price through the rest of the decade, the take-or-pay trap waiting on the far side of the delivery trap.

The battery-import run-rate against the sixteen-point-one gigawatt-hour path, and whether the feeder automation that turns storage into good delivery is funded or deferred one more year. And the evening peak this summer against twenty-eight thousand megawatts — the hour in which the whole thesis is settled, night after night, in megawatts served or shed.

The memorandum was signed on Friday in Switzerland, brokered in part by an Islamabad that needed the strait open more than anyone else at the table.

It will reopen the waterway on paper.

The only signature that delivers gas to Karachi is a Qatari cargo manifest, and the UAE’s own national producer dates the return of full Hormuz flows to 2027. That is the quiet correction underneath all of it: there was never a secured molecule, only a claim on one — and a claim clears at the speed of the strait.

Subscribe now

You cannot hurry a cure.

An ocean away, over the Gulf, the United States spent that same kind of fortnight in late winter firing the inventory those ovens exist to replace — more than five hundred ballistic missiles and two thousand drones came at American and allied positions in the first hundred hours of Operation Epic Fury, and the interceptors went up to meet them at a rate the line behind them has never once matched and was never financed to match.

The buffer emptied in days. The thing that refills it cures in weeks, qualifies in years, and earns nothing at all for the months it stands ready.

The interceptor stockpile is the same buffer, spent in the same crisis, and it carries the same fault on the opposite leg. The oil trade breaks on its funding leg, the molecules leaving faster than the schedule replaces them.

The missile base breaks on its refill leg, the motors arriving slower than any war will spend them — because the market that makes them has been built to do the one thing a surge cannot survive: to price the cost of what it produces, and to pay nothing for the standing capacity to produce more.

This is not a defence story wearing finance for the evening.

It is, almost line for line, an electricity market — and the specific electricity market that every power engineer has spent fifteen years arguing about.

The solid-rocket-motor base is a peaker.

The Pentagon buys it like base load.

The bill for that confusion came due in the first hundred hours over the Gulf.

You cannot surge a cure

The motor is the product of a sequence that refuses to be compressed.

The propellant is mixed, cast into the case around a mandrel, then cured — three to fourteen days in a heated oven while the grain hardens and bonds, a window the chemistry sets and the schedule cannot — before X-ray inspection hunts the voids and debonds that would alter the burn, and a motor pulled from the batch is fired to acceptance so the rest can ship.

Each step is rate-limited and fault-intolerant, which means the output is a flow with a hard ceiling and a war is a demand spike with no notice.

You do not meet a spike in a rate-limited flow by running the line faster; the cure will not be rushed and the X-ray will not be skipped. You meet it only by having built, in advance and at cost, more lines than peacetime demand can keep busy — capacity that sits idle until the spike arrives.

Capacity that sits idle until the spike arrives is a peaker.

Read the curing oven not as a piece of factory plant but as a peaking unit, and the whole problem rearranges itself into a question power markets have already named: it earns when it fires, it is needed precisely because it can fire on the worst day, and somebody has to decide whether to pay it for the long stretches it stands cold.

I have written before that energy desks keep mistaking a flow problem for a stock problem — pinning their hopes on drilled-but-uncompleted wells and shut-in barrels that turn out to refill nothing fast. The missile base makes the same error in reverse: it treats a capacity problem as a procurement problem, as though the answer to too little standing readiness were a cleverer way to buy the next motor.

Even the workforce is a multi-year capital good and not a switch — an Army study put the average munitions line worker at two years to full effectiveness and the energetics worker at seven.

Seven years is not a hiring plan you execute during a hundred-hour war.

It is a thing you either built a decade ago or do not have.

The base is a peaker, and the Pentagon buys it like base load

A peaker exists to run rarely and be available always, and that duty cycle is its economic curse: it cannot recover its fixed cost on the handful of hours it actually runs, so any market that wants peakers to exist has to pay them a separate charge simply for being there.

The solid-rocket-motor base has the peaker’s exact duty cycle — quiet for years, then everything at once — and the only thing the customer buys is the finished motor inside the finished missile.

Across a quarter-century, close to two-thirds of missile contract dollars went to completed systems; the motor itself, the subsystem common to very nearly every interceptor, drew under two per cent of the obligations.

The component that gates the entire ramp is a rounding error in the buy.

Price the base the way a generation desk prices a peaker and the trouble is instantly legible — you are asking a unit to be there at peak while paying it only for energy, and an energy payment can never cover the cost of standing idle the rest of the year.

The report reaches the same conclusion from the other direction, observing that capacity left sitting idle — even a surplus of it — is not the same thing as resilience.

A power engineer would not blink at that: it is the missing-money problem arriving by another road. And the missile peaker has it worse than any real one, because the real peaker at least sells into a live energy market between scarcity events, while the motor line has no second customer to keep it warm between wars.

The closest thing to one walked away in 2011, when the Space Shuttle retired and demand for motor propellant fell from about twenty million pounds a year to five, and the launch firms that replaced it chose liquid fuel.

The commercial analogue did not shrink; it left — and with it went the one buyer that might have kept a motor line lit through the quiet years, which leaves the entire cost of readiness resting on a single customer that has no line item for it.

An energy-only market has no money for the megawatt

Power markets have a name for this, and the name is the whole argument.

The missing-money problem is the structural failure in which energy prices alone do not give a generator enough revenue to cover the fixed cost of capacity, so capacity retires or never gets built, and the system runs thin on margin at exactly the moment it needs margin most.

Texas chose to run an energy-only market and lives with the consequence by design; the eastern grids bolted a capacity auction onto theirs precisely to pay for availability that energy revenue would not fund. The solid-rocket-motor base is energy-only by construction, paid per delivered motor and never for readiness — and the Government Accountability Office said the quiet part aloud years ago, finding that excess capacity “keeps SRM manufacturers from being cost competitive.”

Read that as the missing-money statement it is: the market actively penalises the firm that holds the very capacity the mission depends on.

Both sides are trapped inside their own fiduciary logic, and both are behaving correctly. The Pentagon answers to the taxpayer and is right to drive down unit cost; the prime is a public company answering to shareholders and is right to refuse idle lines that depress returns without a business case to carry them.

Each does its job.

The system the two of them jointly produce has no money for the megawatt — no line item, anywhere, that pays a motor maker to keep an oven hot and a crew current against a war that has not started. The capacity that wins wars is the capacity that loses the cost competition, and a market that settles only on cost will retire it every time.

Six suppliers became two because nobody bid the capacity

Run that logic forward and you get the base the United States actually has.

Start with the belief that does the damage: the customer assumes the capacity will be there when the war comes, because there is always latent capacity somewhere — idle lines, a mothballed oven, a supplier who will surge — and the producer assumes the demand is a cycle that will turn down, so standing capacity is a liability to be shed, not an asset to be held.

Those two beliefs shape behaviour exactly as you would expect.

The customer buys the cheapest motor at the smallest quantity that meets the plan, and the producer sheds every line that does not pay its way — which is what an energy-only merit order does in any market, retiring everything but the must-run incumbent and calling the result efficiency.

The behaviour then alters the fundamental it assumed away.

Between 2000 and 2015 the domestic motor base fell from six suppliers to two, Aerojet Rocketdyne and Orbital ATK, now folded inside L3Harris and Northrop Grumman.

The workforce thinned and aged, and because the energetics worker takes seven years to ripen, the knowledge did not so much depart as fail to be replaced.

The chain narrowed onto single nodes: one qualified domestic source of ammonium perchlorate, the oxidiser in nearly every motor flown; a rayon thread for the nozzle so commercially marginal that NASA once stockpiled two and a half million pounds of it to keep the line breathing. The pursuit of efficiency did not sit beside the fragility. It manufactured it.

Then the loop closes, and this is the part that matters.

When the spike finally arrived — a hundred hours, five hundred missiles — the capacity was not there, and the customer did the only thing a customer can do at the top of a scarcity event: it paid up. A billion-dollar equity stake in L3Harris’s motor business in January, a seven-year deal to lift Patriot-interceptor output from about six hundred a year toward two thousand, a wall of framework agreements behind them.

Every cheque confirms to the base what the base already believed — that the real money comes in spikes and not in a durable curve — which is precisely the belief that tells it not to hold capacity for the next time.

The supplemental is not the moment the cycle breaks.

It is the moment the cycle is rewarded.

A demand signal sent in emergencies trains the market to wait for the next emergency, and a signal whose form is a spike defeats whatever its content was trying to say.

The billion-dollar stake is a scarcity price

The strongest objection to all of this is that the cheque is the fix — that direct-to-supplier investment is, at last, the capacity payment, the government finally buying availability rather than output. It is a fair objection and deserves its best form.

The department took convertible preferred equity in L3Harris’s motor arm, is standing the business up as a separate Missile Solutions company, and plans to capitalise it through a public offering in the second half of this year; the under-secretary called it “fundamentally shifting our approach.”

Granted in full: this is the closest the United States has come to paying for motor capacity since the Cold War.

It still is not a capacity market, and the difference is the whole point.

A capacity payment is a standing obligation, repriced every auction, that a generator can finance a thirty-year asset against; a one-time equity injection that converts at an IPO is a grant, and a grant ends.

The framework agreements that sit on top of it still require annual appropriations, and the record there is not ambiguous — Congress has authorised multiyear munitions buys before and then declined to fund them, breaking the agreements in all but name.

The missile peaker is being handed a single payment inside a market that still has no mechanism to pay it again, and the next lean budget will do to munitions what every lean budget has done — treat them as the painless economy, trimmed a few interceptors at a time rather than by killing a whole airframe, the account that quietly underwrites everyone else’s overruns.

A peaker paid once does not stay built.

The power markets have already run this experiment where everyone can watch.

PJM’s capacity price climbed from under thirty dollars a megawatt-day to the low three hundreds across three auctions, and the most recent one still cleared below the reliability requirement — the market shouting that even a capacity payment at its administrative cap cannot conjure steel quickly enough when the load arrives faster than the queue, which in PJM’s case means data centres and the electricity an AI build-out demands.

• That is a market that pays for capacity and still falls short of what it needs.

• The missile base pays for none.

• One is rationing a shortage it priced years too late; the other has not started pricing at all.

Lay the pieces out together — the cure that will not be hurried, the single customer, the merit order that ground six firms to two, the cheque written at the very top of the spike — and the thing the report names as the most significant challenge to the whole enterprise, the cyclical funding, resolves into the missing-money problem under another name.

The solid-rocket-motor base is the cleanest live test of that problem anywhere outside an electricity market, and the United States is running the test in real time, having just discovered over the Gulf what an energy-only market always discovers eventually: that the capacity you declined to pay for is the capacity you do not have when the peak comes.

My read is that the equity stake is real money aimed at the wrong variable.

It buys motors; it does not buy a market that pays for availability, and without that the new Missile Solutions company will do exactly what a public prospectus obliges it to do — serve its shareholders, which means not carrying idle lines that the share price punishes. The Pentagon has bought a quarter of output and is calling it resilience.

The gauge that will settle it is the appropriations cycle, because the framework agreements convert into standing capacity only if the dollars come every year.

The first honest test is the 2027 budget and whether the seventy-odd billion requested for these missile programmes survives the markup or is decremented the moment the headlines move on; if it thins, L3Harris’s spun-out motor company — public by year-end — will re-mark its own capacity precisely as a power producer re-marks an uneconomic peaker, declining to build past the contract, and the next interceptor drawdown will find the same shortfall the last one did.

This reaches past the missile because the identical question is now the live one in energy itself: the gas peaker that must be there for the windless evening, the regasification berth that earns nothing until a strait shuts, the firm capacity a grid needs for a load growing faster than anything can be built to meet it — all of it asked to stand ready in markets built to settle output, not readiness.

Availability is repricing from a cost to be minimised into an option that has to be bought, in megawatts and in motors alike, and the buyers who work out how to pay for the standing oven rather than only the delivered grain are the ones who will have the grain when the next hundred hours arrive.

So watch the funding leg, on three gauges.

The oven will cure the next motor in two weeks,

the way it cured the last one and the way it will cure the one after it.

What no budget has agreed to pay for is the oven standing hot and empty between wars, waiting for an order the cycle will not promise — and a peaker no one pays is a peaker no one builds.

Subscribe now

At half past ten on Wednesday morning the print came.

340.3 million barrels — not at the line but more than six million barrels through it, an 8.9-million-barrel draw in a single week.

In the same hour, the screen read $88.62.

That is the whole of this week’s report, stated as a contradiction: the buffer crossed the floor it was built never to cross, and the price the floor was supposed to defend fell another five dollars and seventy cents into calm.

A level like 346.7 million barrels is not a price and it is not a fuel gauge.

It is a covenant — a term in the unwritten agreement under which this market has consented to stay calm, a number that means something the way a minimum-balance clause in a loan means something.

On Wednesday it was breached.

What the rest of the report records is not the breach, which everyone expected, but the response to it — and the response was that the market, holding every right a breach confers, chose to waive it.

The covenant was the 1983 line

A covenant is a promise with a number attached, and the number is chosen to trip before the borrower is actually insolvent, so the lender learns of the trouble while there is still something to be done about it.

The 346.7-million-barrel line did exactly that work.

It was the modern trough, the lowest the reserve had stood in the series, and beneath it lay a level last seen in 1983 — a forty-three-year floor that traders, ministries and newsrooms had each, separately, agreed to treat as the point at which the drawdown stopped being a policy choice and became a constraint.

Nobody legislated it.

It functioned anyway, the way an unwritten covenant functions: as the level whose breach would mean the buffer was no longer a buffer.

On Wednesday, it broke.

The strategic reserve drew 8.9 million barrels in the week to 12 June, to 340.3 million — through the floor, not down to it.

Set beside the commercial tanks, which gave up another 8.3 million in the same week, the country drew 17.2 million barrels of crude out of storage in seven days, close to two and a half million barrels a day.

Total crude inventory, the strategic and commercial books added together, now stands at 758.5 million barrels — nearly sixty-five million below where it sat a year ago.

Read that print the way a credit officer reads a covenant test, because the distinction is the whole story.

A breach is not a missed payment.

The reserve did not run dry on Wednesday; three hundred and forty million barrels are still in the caverns, more than enough to load a tanker tomorrow.

What broke was a term — the promise the level encoded — and a breached term is a technical default: it does not stop the borrower functioning, it transfers a right.

From the moment the line went, every barrel the system draws from here must be re-sourced at a market price with no cheaper backstop beneath it, and the holders of the calm — the desks and ministries short of volatility through inventory — owe that re-sourcing into a market they no longer control.

The default is real.

The only question the rest of the report answers is what the lender did about it.

The lender granted a waiver

The lender, in this market, is the flat price — the one mark with the standing to declare the breach a problem, because a repricing is how a physical shortage announces that it can no longer be papered over.

On Wednesday, the flat price declined to declare it.

West Texas Intermediate settled at $88.62 at Cushing, five dollars and seventy cents lower on the week and the lowest in nearly two months; Brent settled alongside it at $88.64.

In the week its strategic buffer broke a forty-three-year floor, the market marked crude down, not up — and marked it down, on the international screen, hardest of all.

That last detail is where the waiver becomes legible.

Brent did not merely fall; it fell faster than the domestic grade, shedding close to nine dollars on the week against West Texas’s five and seventy, until the premium Brent normally carries over WTI had compressed to about two cents by Friday’s mark — from roughly three dollars a week earlier, and from seventeen in April.

The benchmark nearest the missing barrels, the one priced off a strait that is still shut, fell the harder of the two in the very week the shortage behind it deepened. A lender granting a waiver does not merely decline to call the loan; it re-marks the credit as money-good and lets the borrower draw again.

That is what a flat price near eighty-eight dollars, with its international premium gone, does: it tells every risk committee and every ministry that the situation is improving, and it invites them to keep selling the calm that produced the mark.

The retail tape agrees and extends the waiver to the consumer.

Pump gasoline eased to $4.05 and on-highway diesel to $5.06 in the week to 15 June, each lower on the week.

Demand, which a true repricing would have begun to ration, did the opposite — total products supplied ran at 20.6 million barrels a day on the four-week average, three and three-tenths per cent above the same period last year.

The waiver reaches all the way to the tank in the driveway: nobody along the chain has been instructed by price to stop.

Cross-default in every other line

A covenant breach rarely sits alone in the agreement, because loan documents carry a cross-default clause: trip one covenant and you are deemed in default on the others, the lender’s logic being that distress is seldom confined to a single line.

The June report reads like a cross-default schedule.

The commercial crude tanks, having shed their 8.3 million, sit about six per cent below the five-year average for the season.

Distillate — the diesel and heating barrel — is roughly thirteen per cent under its own five-year average. Cushing, the delivery point that sets the WTI mark, drew to twenty million barrels, down nearly twelve per cent on the year, its cover thinning beneath the very contract it settles.

Every buffer the system carries against a crude shock is in breach of its own seasonal covenant at once — which is precisely what cross-default describes: not five separate problems but one, registered in five ledgers.

One line is not in breach, and it locates the breach exactly.

Propane stocks stand thirty-six per cent above their five-year average, some twenty million barrels fuller than a year ago.

The molecule in surplus is the one that owes nothing to the strait — fractionated out of domestic natural-gas liquids, produced at home at the wellhead, never loaded at a Gulf terminal.

The geography of the breach is written in the contrast: the buffers that depend on a barrel arriving by sea are all through their floors, and the one that depends on a barrel produced in West Texas is overflowing.

This is not a shortage of oil.

It is a shortage of oil that can reach the water, and the propane line is the control sample that proves it.

The default the waiver postpones

A waiver postpones a default; it does not retire it.

The covenant on the reserve existed to trip early — to force the repricing while there was still a buffer to manage the transition, the cushion the lender draws on during a workout.

Waive it, and the breach you have deferred is no longer the technical one that printed on Wednesday but the real one beneath it: the day a refiner needs a cargo and there is no barrel, at any price, that can reach its dock.

Nothing in this week’s report says that day has been pushed back.

The report says only that the warning meant to precede it has been spent.

The flows are where the postponement would have to show, and they show none.

The import slate records no Gulf recovery whatever.

Saudi arrivals ran at 125 thousand barrels a day on the four-week average, against three hundred and fifteen a year ago; Iraqi arrivals at 38 thousand against a hundred and eighty-seven, and in the current week Iraq delivered nothing at all. Nigerian and Libyan barrels are at zero.

Where the missing Gulf grades are replaced, they are replaced from the one direction that owes nothing to Hormuz: Venezuelan crude is arriving at more than four times last year’s rate, backfilling a closed strait with rehabilitated heavy.

I wrote midweek that recovery would announce itself as the Gulf’s two largest suppliers climbing back toward their year-ago combined rate; on this week’s slate that rate is about five hundred thousand barrels a day, and the actual arrivals sit nearer a hundred and sixty.

The break condition for the calm has not printed. The strait is still shut, in the only ledger that counts the barrels rather than the headlines.

Meanwhile the system runs on the buffer at close to full tilt.

Refineries took in 17.2 million barrels of crude a day and ran at 96.7 per cent of capacity, drawing the stock down to feed a demand the calm has kept whole.

This is what spending the warning looks like: the buffer emptying at maximum rate to sustain consumption no price has told to slow, into a supply picture that has not improved by a single sustained cargo.

The waiver buys exactly one thing — time at the current price — and it buys it by consuming the cushion that was supposed to be there when the real default arrives.

Each waiver deepens the breach

Now assemble the loop, because the waiver is not a neutral act of patience; it is the mechanism by which the breach it waives gets deeper.

It begins with a belief — that the falling price is de-escalation arriving, the spike over, the strait about to reopen.

The belief shapes behaviour: holding that view, every participant waives the covenant in the same motion — the fund trims its length, the ministry authorises another release, the refiner defers precautionary cover and runs its own tank, the consumer fills up unbothered, because a price near eighty-eight dollars instructs all of them that the danger has passed.

That behaviour alters the fundamental it claimed only to read: the waiving is the drawing — the trimmed length and the authorised release and the deferred cover are the 17.2 million barrels that left storage this week, and the breach goes deeper. And the deeper breach feeds back on the belief in the cruellest available form, because the extra selling lands on the flat price and pushes it lower still, and the lower price reads, to the next risk committee, as further confirmation that de-escalation is real, which licenses the next waiver.

The covenant was supposed to break this loop.

Tripping it was meant to be the event that forced the market to look at the buffer instead of the screen. By waiving it, the market has converted the one mechanism that could have stopped the loop into another turn of it.

The inflection is the day the cure period lapses — the day a barrel is actually called for and cannot be sourced, and the breach deferred week after week must finally clear at a market price.

It will clear into a thinner book than absorbed the original shock, because the cushion that would have softened it has been spent on the waivers, and into a Brent–WTI spread pinned near zero, which can compress no further and can only widen. A suppressed signal does not decay while it is suppressed.

It accrues, and it is paid out in full on the day the suppression stops.

The sellers’ case, in its strongest form, is that the waiver is correct — that the market is reading a real reopening the buffer cannot yet see, that a ceasefire framework needs only signatures and flows resume in weeks, at which point Saudi and Iraqi arrivals climb back toward five hundred thousand barrels a day, the reserve draw halts, and the breach is cured by restored supply rather than merely waived.

It is a coherent case.

But it is a bet on a binary the sellers do not control, made in a week when Iraq delivered nothing and the combined Gulf arrival ran at a third of the rate that would signal recovery — and a waiver granted on a default that has not begun to cure is not foresight.

It is the decision to spend the buffer before knowing whether it will be needed.

The next test is Wednesday

Put the week together and the report stops reading as a market that failed to notice its buffer break a forty-three-year floor.

It read the break precisely, and waived it.

The calm of the past month was, until Wednesday, the market’s interpretation of an ambiguous situation; since Wednesday it is something narrower and harder — a deliberate waiver on a default that has printed in the official record, granted not in ignorance of the breach but in full sight of it.

I think the waiver is the most dangerous answer the market could have given, more dangerous than panic would have been. A repricing on Wednesday would have hurt, but it would have spent the warning to do the thing the warning was for — to force the adjustment while a buffer still existed to absorb it. The waiver takes the hurt now avoided and stores it, at interest, for a day with no cushion left.

Granted the thesis, the playout runs through the physical desk first.

The refiner fed all spring by reserve barrels and Venezuelan backfill must, when the waiver lapses, replace stock supply with term supply — and that bid lands not on the flat price the screen is watching but on the deferred contracts and on Eastern physical, where the barrel nearest the strait is priced.

It steepens the backwardation that destocking flattened, re-firms war-risk on anything loading inside the Gulf, and re-widens Brent over WTI from the two cents it has compressed to — the export pull cannot intensify past parity, so the spread can only travel the other way.

Structurally, the residue outlasts the breach: war-risk migrates from a spot cost into a term-contract input, Gulf official prices keep ceding ground to Dated- and Murban-linked formulae, and the chokepoint premium the flat price has spent a month pretending away re-enters the curve where it cannot be waived — in the price of the barrel that has to physically move.

Watch the same gauges, all on the funding channel, all dated.

• Next Wednesday’s EIA print: if the reserve draw does not stop, the line crosses toward three hundred and thirty-five million and, on this pace, three hundred million before August — each print a further breach the market must choose again to waive or finally call. Table 8’s Saudi and Iraqi rows: a climb back toward their combined five hundred thousand barrels a day is the cure the sellers are betting on; another week near a hundred and sixty is the breach uncured.

• And the Brent–WTI spread: a move back through eight dollars is the sound of the lender finally calling the default the screen has spent a month refusing to acknowledge. The next OPEC monthly, on 13 July, will say whether OECD forward cover has broken below fifty-nine days — the same waiver, read one ledger up.

The barrels are still in the caverns; three hundred and forty million of them answered the test on Wednesday, and the price called it cured.

But a waiver is not a cure.

It is the lender agreeing not to look — and the buffer that funds the calm is now more than six million barrels below the floor it was built never to cross, with the next test due, as every test is, on Wednesday.

Subscribe now

The headline is the three per cent of long-term GDP the Fund says Europe could buy by deepening its banking and capital markets — a figure already circulating, already half-discounted as the kind of round number these exercises produce.

The number that matters sits two pages later, in the dry register of a gravity regression: forming a new cross-border bank-firm relationship inside the euro area is subject to an implicit tax of ninety-nine per cent relative to forming the same relationship at home.

Not a tariff written into the Treaties.

Not a capital control.

A wedge, estimated from the near-universe of euro area corporate lending, large enough that it rationalises why only one in a hundred bank-firm links and roughly five per cent of loan volume cross a national border in a union that has spent a decade insisting it is single.

Read that the way a customs officer reads a manifest.

A levy of ninety-nine per cent at the point of entry is not a frictional cost to be optimised around; it is a wall priced as a number. And the most telling thing about this particular wall is where it stands. The same regression finds that once a cross-border relationship exists, the friction collapses to almost nothing — an implicit tax of two-tenths of one per cent, an interest-rate difference of twenty-five basis points between a foreign lender and a local one.

The tax is not on the lending.

It is on the first crossing.

Europe has built a border that charges everything to arrive and nothing to stay.

The tax is levied at the threshold, not the transaction

This is the asymmetry the whole note turns on, and it is worth holding before the policy machinery starts.

A flat frictional cost — a tax on every euro of cross-border interest — would show up as a wide and persistent spread between domestic and foreign credit.

It does not.

The spread is twenty-five basis points, which on a corporate book is noise.

What shows up instead is a near-total suppression of new relationships, an entry barrier so steep that the relationships which do exist are confined to a handful of large firms with the legal departments to pay it once and the scale to amortise it.

The IMF builds the tax out of something it calls a policy-distance index: an average across eight categories — deposit insurance, macroprudential rules, microprudential rules, resolution, supervision, governance, bank entry, and corporate bankruptcy — of how far any two member states’ rulebooks sit from each other. The median distance is about a quarter, which the authors gloss with admirable bluntness: roughly one policy in four differs between any two countries you pick.

A German bank contemplating its first loan to an Italian firm is not pricing the Italian firm. It is pricing the cost of learning, and living inside, a legal and prudential environment that differs from its own across a quarter of the surface area that matters — and pricing the recovery it would actually collect if that firm defaulted into an insolvency regime its credit committee has never tested.

The distance is the tax base.

The bank pays it in legal uncertainty and compliance, and it pays the whole of it on the first crossing, which is exactly why the crossing so rarely happens.

Two collectors take a third each

Decompose the levy and it has named collectors.

The note runs the policy-distance index against the estimated frictions and finds that two of the eight categories do most of the work: harmonising corporate bankruptcy regimes and harmonising deposit-insurance schemes each account for about a third of the total effect on cross-border credit.

Bank resolution and macroprudential rules add smaller amounts; the remaining four categories are statistically indistinguishable from zero. The tax, in other words, is not a diffuse fog of twenty-seven bureaucracies. It is two tollbooths.

The bankruptcy tollbooth prices a real thing.

A lender’s loss-given-default is decided in an insolvency court, and a German creditor lending into an Italian or a French or a Spanish firm is exposed to the recovery value that that jurisdiction’s law and that jurisdiction’s courts will deliver — a value it cannot mark with confidence because it has no book of experience in that regime.

Different insolvency law does not merely add paperwork; it lowers the risk-adjusted expected recovery and raises the legal certainty premium, and a bank that cannot price recovery prices the loan out of existence instead.

The deposit-insurance tollbooth prices something subtler and more circular.

Because deposit guarantees are national, the stability of a bank is tied to the fiscal capacity of its home sovereign — and a bank whose backstop is its own state has every incentive to hold that state’s bonds and to lend to that state’s firms, because their fortunes are correlated with the thing standing behind its deposits.

The sovereign-bank nexus is not a side effect of fragmentation.

It is the mechanism by which fragmentation reproduces itself.

Hold that, because it is the loop

A fragmented deposit guarantee makes home assets look safer to the bank than they are, because home assets move with the sovereign that insures the bank’s own funding.

The bank, acting on that perception, concentrates at home.

The concentration deepens the correlation between the bank, its sovereign, and its domestic borrowers — it builds the very nexus that made home look safe.

And the built nexus then confirms the original belief: cross-border exposure now genuinely does carry a diversification benefit the bank has declined to take, while the home book genuinely is wired to the sovereign, so the prudent-looking choice to stay home has manufactured the risk profile that makes staying home look prudent.

The belief constructed the fundamental.

This is the inflection the note never quite names but its own regression draws: the implicit tax is not an exogenous barrier sitting on top of an integrated market.

It is the price of a fragmentation that each bank’s rational response to fragmentation keeps paying forward.

There is a second tax, on the equity side, and it is heavier

The note does not stop at banks, because Europe’s deeper problem is not the price of its bank credit but the absence of the financing that banks structurally cannot provide.

Young, innovative firms — short credit histories, no tangible collateral, the precise profile a bank’s risk model rejects — need equity, and Europe’s venture capital market is roughly a quarter the size of the United States’, with EU firms that do raise VC taking in about half what their American counterparts receive per deal.

Here the tax has a different structure.

Part of it is, again, a border levy: intra-EU cross-border VC flows are throttled by differences in legal regimes and withholding-tax procedures, and the note’s gravity model estimates that simplifying withholding-tax friction alone could lift cross-border flows by around a quarter, with legal harmonisation adding perhaps six per cent more.

But the larger levy on equity is not at the border at all — it is on the supply itself.

Europe’s households hold their savings in deposits and their pensions in systems that, by regulation, are tilted away from risk capital; the long-term money that funds American venture rounds simply is not permitted to flow the same way here. The note finds that closing half the gap to the best-performing systems in pension contributions and insurance assets could be associated with an increase in EU venture supply of around ninety-six per cent.

And the heaviest levy of all is the one that decides whether there is anything to fund: the real-sector barriers — thin public research, under-investment in tertiary skills, business-environment distortions, punitive top marginal rates — that determine how many investable projects exist before any financier looks at them.

Reduce those across the modelled pull factors and the note’s partial-equilibrium estimate for VC demand rises by something on the order of one hundred and seventy per cent.

The tax on equity, in short, is collected three times: once at the border, once on the savings that would supply the capital, and once — most heavily — on the ideas that would absorb it.

The repeal is partial, and the Fund says so

What the general-equilibrium model does with all this is the part a practitioner should respect rather than round off.

It does not simulate paradise.

It simulates a moderate repeal — cutting the regulatory distances that tax cross-border banking by about thirty per cent, which the note describes as fully harmonising one policy in three, and closing only half the gaps that suppress risk capital — and it deliberately leaves equity-market integration out entirely.

On those conservative settings, EU GDP rises by about three per cent in the long run: two points from the banking-tax repeal, one from the VC-tax repeal, with the venture layer small because the sector it operates on is small to begin with.

And a fourth point appears only in combination: financial reform and real-sector reform are complements, so undertaking both together yields a percentage point that neither delivers alone. The smaller economies and the youngest firms collect most of the rebate — start-up output more than doubles in the comprehensive scenario while mature firms give a little back, because the repeal reallocates capital toward exactly the borrowers the tax was hitting hardest.

This is where the reflexive reading earns its keep against the optimistic one.

A three-per-cent GDP gain priced off a thirty-per-cent cut in policy distance implies the full repeal is worth multiples more — the note is explicit that deeper effort yields larger gains. But the same loop that makes the tax self-reproducing is what makes the repeal so hard to legislate, and the political record is the evidence.

The European Deposit Insurance Scheme — the single measure the note’s own regression says would do the most, one of the two tollbooths that take a third each — was proposed in 2015 and remains, in June 2026, unbuilt.

The Commission’s Savings and Investments Union package, adopted in December 2025 and grinding through Council working parties since, attacks the equity-side levy — passporting, a Pan-European Market Operator status, the optional twenty-eighth corporate regime — but the mid-term review that would judge whether any of it bit is not scheduled until the second quarter of 2027.

Now put it on a desk

Grant the thesis — that the binding constraint on European corporate finance is a tax levied at the threshold of the first cross-border relationship, collected mainly through bankruptcy law and deposit insurance — and the transmission is concrete.

The first actor to move is not a bank; it is the legislator who either does or does not mutualise deposit insurance, because that is the toll the note says is worth a third of the banking gain and the one with the longest record of not being paid.

If EDIS arrives, a euro area bank’s home assets stop being privileged by the structure of its own backstop, the sovereign-bank nexus loosens at the margin, and a German lender’s credit committee can for the first time price an Italian loan against a recovery it has reason to trust rather than a recovery it cannot mark. The cross-border M&A that has fallen for two decades — the trapped capital sitting in ring-fenced national subsidiaries that European bankers themselves keep flagging — begins to clear, because the implicit tax on building scale across a border is what kept it trapped.

The structural implication for the market is a compression: the home-bias premium embedded in fragmented bank credit narrows, the basis between domestic and cross-border corporate funding tightens toward the twenty-five basis points the note already measures inside existing relationships, and the funding mix shifts — the note has bank market share falling to non-bank intermediaries as the venture layer thickens, a quietly enormous reweighting of who finances European growth.

So the gauge to watch is not the GDP estimate, which will not be settled for a decade and cannot be falsified in a quarter.

It is the EDIS file.

The note has effectively told the market which single repeal carries a third of the largest gain; the European political system has told the market, for eleven years and counting, how hard that repeal is to pass. The signal that the three-per-cent number is becoming real is movement on mutualised deposit insurance specifically — not the CMDI half-step already taken, but a funded European deposit insurance fund with a date attached — and the place it will first show up is the cross-border bank merger that fragmentation has been suppressing.

The tax was never voted into existence and it will not be voted out of existence in one motion. It will be repaid one tollbooth at a time, and the deposit-insurance booth is the one that has stayed shut the longest.

Europe is not short of savings — it is sitting on something near ten trillion euros of household deposits earning almost nothing.

It is short of the permission to move them across a border.

The IMF has now put a price on that permission, and the striking thing about the price is not its size.

It is that nobody ever legislated it, nobody collects it, and it costs the continent three per cent of its output anyway — a tariff with no treasury, levied at a border the single market was supposed to have abolished.

Subscribe now

The reserve is leaving at roughly nine million barrels a week.

Eleven days into June, with the Strait of Hormuz still closed to most commercial traffic, the front-month Brent screen reads near eighty-seven dollars — a near two-month low, printed not because the war resolved but because Washington floated a fourteen-point draft for weekend signature and the market sold the headline, far beneath the levels Dated cargoes touched in early April, a number that would read as de-escalation if the de-escalation had happened.

The war is not over.

The price is being sold.

This morning the OPEC Secretariat published its June Monthly Oil Market Report, and read properly — not as a forecast document but as an account statement — it is a ledger of who sold calm in May, and from which account they funded the sale.

The position it records is a single trade held in two books: short volatility — collect the premium of calm today, fund the position out of inventory, and bet the thesis matures before the funding does.

• The funds run it in paper and post their margin in dollars.

• The governments run it in molecules and post theirs in barrels.

Everything in the report follows from that position — who is carrying it, what it costs to run, and the date on which the funding leg breaks.

The sellers of calm

The paper book sits in the futures chapter.

Hedge funds and other money managers cut their combined net length in Brent and WTI by a quarter in May, liquidating the equivalent of about 122 million barrels, “betting on easing geopolitical tensions” — the Secretariat’s phrase — while the long-to-short ratio in Brent collapsed from seven-to-one to three-to-one in four weeks.

Set that liquidation against the calendar and the trade declares itself: net length falls through a month in which the strait stayed shut, Iranian output dropped by half a million barrels a day, and the producers’ own basket rose five and a half dollars.

This was not profit-taking on a resolved story.

It was premium collection on an unresolved one.

The wet book sits in the stocks chapter, and it is larger.

The US strategic reserve released 35.6 million barrels in May alone — eleven hundred thousand barrels a day, every day — taking it from 415 million barrels at the end of March to 357 million at the end of May.

These are not two stories.

They are one position, marked daily in two books — and like every short-volatility position, it is defined less by where the premium is collected than by where it cannot be, which is where this report stops being routine.

The premium is still being paid

Calm can be sold wherever the instrument is financial.

It cannot be sold where a cargo either loads or does not — and in those markets, the June report shows the premium still being paid in full.

Start with the anomaly the Secretariat prints without comment: its own Reference Basket rose $5.49 in May to average $114.55 — roughly eleven dollars above front-month Brent’s $103.71 — in the same month the visible benchmark complex softened.

The basket inverted over the benchmark because its heaviest components are now marks on cargoes that mostly cannot move: Arab Light rose $14.89 to $122.10, Kuwait Export to $118.65, Iran Heavy to $115.39 — prices set by the trickle that escapes, standing in for the flood that does not.

Where a Gulf grade does have a deliverable route, it prices like a different hemisphere: Murban, piped overland to Fujairah on the seaward side of the strait, fell $2.29 to $102.17 — twenty dollars under Arab Light, loaded a few hundred miles up the same Gulf but on the wrong side of the chokepoint.

One geological province, two prices, and the difference between them is the cost of being inside the strait.

The volumes behind the marks are in the supply chapter, and they are brutal.

• Kuwait produced 573,000 barrels a day in May — less than a quarter of last year’s rate.

• Iraq ran at thirty-seven per cent of its 2025 average.

• Iran fell another 546,000 barrels a day in a single month, to 2.33 million.

And the quietest seller of calm is one table further down, in a footnote: Saudi Arabia supplied 7.01 million barrels a day to the market in May while producing 6.56 million.

Read that footnote the way a volatility desk reads it.

The gap — 449,000 barrels a day in May, and roughly fifty-five million barrels cumulatively since March — is continuity manufactured from the Kingdom’s own tank farms: the producer of last resort is writing the same insurance as everyone else, against its own inventory.

Freight is where the un-sold premium stops being notional.

A very large crude carrier from the Middle East to Asia was assessed at WS587 in May — nearly nine times its level a year ago — with the clean-tanker equivalent at WS596.

The premium has not gone away.

It has moved to where the screen does not look — into the basket, the freight, and the footnotes — which raises the only question that matters about the part of it that was sold: what is funding the sale.

The funding leg

Every short-volatility book has a funding leg, and this one’s is storage — the burn rate of which is, unusually for a funding leg, published weekly by its own counterparty.

The April data in this morning’s report show OECD commercial inventories down 48.4 million barrels in a month, to 2,748 million — 54 million below the five-year average — with forward cover at 60.1 days and falling by 1.6 days a month.

The United States alone drew fifty-nine million barrels in May between its commercial crude tanks and the strategic reserve — close to two million barrels a day of buffer, from one country. Even the oil at sea is thinning: the report’s balance tables show oil-on-water down from 1,546 million barrels to 1,297 million over the first quarter — a quarter of a billion barrels of floating cushion consumed in ninety days.

The hole this buffer is filling is in the balance chapter, stated but never headlined: the call on DoC crude for the second quarter is 41.0 million barrels a day; DoC production in May was 33.13 million.

Nearly eight million barrels a day separate what the world’s consumption assumes from what the participating producers are delivering — a gap closed partly by refineries that physically cannot run (global crude intake is running some seven million barrels a day below January’s level, with Japan’s crude imports down sixty per cent in a month to 853,000 barrels a day and China’s at their lowest since 2021), and for the rest by tanks.

The external ledgers agree on the arithmetic: the EIA’s June outlook puts Middle East production more than eleven million barrels a day below pre-conflict levels and projects global inventory draws averaging 6.3 million barrels a day this quarter — rising, not falling, to 7.6 million in the third. Its weekly data showed a seventh consecutive US commercial crude draw last week, and the strategic reserve at 349.2 million barrels on 5 June, draining at close to nine million a week, one print above the 346.7 million floor of July 2023 — below which the reserve is at its lowest since August 1983.

Note what is not absorbing the shock: DEMAND

Consumption is being protected, not destroyed — Japan has resumed wholesaler subsidies and released reserves, India has cut excise duties and stood up a $6.2 billion stabilisation fund, Europe is layering targeted relief — so the report still carries a million barrels a day of demand growth for 2026 while supply runs eight short of the call.

Every one of those measures is itself a sale of calm to the consumer, written against the same buffer. The position is elegant and it is working — premium collected daily, funding drawn weekly — and it has exactly one flaw, which is that the funding leg now has a posted depletion schedule and the thesis leg has none.

Calm confirms calm

Here is the loop, and it is the reason the position has grown rather than shrunk as the buffer thinned.

It begins with a belief: the spike is over, de-escalation is coming — the conviction the Secretariat ascribes to the funds, and the working premise of every government still releasing. The belief shapes positioning: paper length is cut by a quarter, 172 million barrels of strategic releases are authorised in Washington and drawn at better than a million barrels a day, Riyadh ships above production, and the selling lands where selling lands — on the front of the curve, where Brent’s one-to-three-month backwardation flattened by $3.46 in May to $7.45.

Then comes the step Soros would recognise as the act of participation: the positioning alters the thing the belief claimed to measure. A softer flat price and a flattening curve are the market’s de-escalation evidence — they are what every risk committee, every ministry, every newsroom reads as the situation improving.

The suppressed signal slows precautionary buying (the report records refiners cutting spot bids and differentials easing — easing, note, because the buyers are starved of run capacity, not because they are satisfied), it makes each further release politically free, and it keeps consumption whole.

Which keeps the draw at six million barrels a day.

Which thins the buffer that funds the selling that produces the calm that justifies the selling.

For now the loop is self-reinforcing, and that is precisely what makes it dangerous: every quiet week marks the short profitable, and a short that pays daily invites size. It earns by the day and it loses by the event — and the event is not a headline but a level.

The inflection sits at the funding floor: the week the seller of last resort steps back, the premium that was being sold from storage must clear at a market price again, and it will do so through a thinner book — combined Brent and WTI open interest fell 3.8 per cent in May even as prices calmed, which means the repricing, when it comes, transmits through less liquidity than absorbed the original shock.

A suppressed signal does not decay.

It accrues — and it is paid out on the day the suppression stops.

The case for the sellers

The sellers are not fools, and their case deserves its strongest form.

Flows are genuinely recovering: strip Iran out and the Gulf’s May was positive — Saudi Arabia up 157,000 barrels a day month-on-month, the UAE up 87,000, Iraq up 75,000, Kuwait up 16,000 — a third of a million barrels a day of restoration in four weeks, with the US energy secretary saying strait transits and exports are rising, reports of an Omani-brokered tanker-toll arrangement, and OPEC+ openly weighing output increases.

The Secretariat itself cut 2026 demand growth by 200,000 barrels a day and moved it into 2027 — deferral, which softens this year’s draw — and its feature article projects non-DoC supply accelerating to 900,000 barrels a day of year-on-year growth in the second half.

JPMorgan’s base case has the strait reopening this month, with Brent averaging near a hundred dollars for the year. If restoration arrives in weeks, the shorts collect everything and this column is the mispricing.

But run the sellers’ own numbers against their own clock.

At May’s ex-Iran pace, the eight-million-barrel hole closes in roughly two years; the EIA’s projected draw accelerates into the third quarter; and JPMorgan’s same desk now expects visible inventories to enter what it calls a “stress zone” within weeks and to approach operational floors by September — the point, on its own framing, at which any new shock transmits straight into the flat price with no buffer in between.

The trade requires not recovery but restoration — a reopened strait, restarted fields, reinsured hulls — and that is a binary the premium-sellers do not control, in a month when the leading belligerent's output fell by another half a million barrels a day.

A short-volatility position whose funding expires before its thesis can mature is not a market view; it is a deadline wearing one — and the rest of this report, read closely, is simply the schedule on which it falls due.

Assemble the whole report and a picture emerges that none of its chapters states alone: this is not a market healing, it is a market underwritten — the premium income showing up as a civil flat price near eighty-seven dollars with the war still running; the funding drawdown showing up in every stock line in the book, all of them below average and falling; and the un-sold remainder of the premium standing exactly where cargoes still have to be physical — a producer basket eleven dollars over the global benchmark, freight at nine times last year, and an Eastern crude curve, Dubai’s one-to-three-month backwardation at $12.19, refusing to flatten to Brent’s $7.45 because the market nearest the missing barrels cannot be paid in reassurance.

I argued here in early May, off the Oxford Institute’s post-mortem, that the futures curve had lost the ability to price physical scarcity — Dated then stood thirty dollars over the screen.

The June report's May marks show that gap closed to four.

It did not close because scarcity ended.

It closed because calm was sold into it, by the cavern-load — and my read is that the flat price has now stopped measuring the war at all and started measuring the remaining depth of the buffer, which means the calm is not the opposite of the drawdown; it is the drawdown, quoted in dollars instead of barrels.

Granted the thesis, Tuesday morning runs like this.

• The first move belongs to Washington, because the floor forces a choice: slow the releases and the front of the Brent curve re-steepens immediately — the $7.45 spread converging up toward Dubai’s $12.19, not the reverse — or sell through the 1983 line and let the market re-price the absence of the backstop instead, which is the same repricing with worse optics.

• Either branch reaches the same desks next: Asian refiners who have run crude cover down to a fifth below last year must rebuild it into a steepening curve — buying backwardation, the most expensive way to restock — and their bids land on Eastern physical first, where the Oman premium over Dubai, already $8.57, is the gauge.

• The structural residue outlasts the war: war-risk freight migrates from a spot cost into a term-contract input, East-of-Suez term deals reprice with Hormuz optionality embedded, destination-flexible Atlantic barrels take a durable share of Asian slates — the report already shows US flows to Japan and Korea rising — and Gulf official prices keep ceding ground to Dated- and Murban-linked structures, the benchmark migration I tracked in May continuing under sedation rather than panic.

The margin call, in other words, is dated.

Watch three gauges, all on the funding channel.

The EIA’s weekly print next Wednesday, 17 June: below 346.7 million barrels, the reserve is at its lowest since August 1983, and the funding leg has publicly broken.

The next MOMR, dated 13 July: OECD forward cover printing below fifty-nine days confirms the burn; back above sixty-one and this thesis, not the sellers’, is the mispricing. And the pair of front spreads, Brent’s against Dubai’s: whichever converges toward the other names the winner — the paper market or the wet one.

At Bryan Mound, the caverns do not empty as the oil leaves; they refill, with brine — water rising, by the day, where the calm used to be stored.

Calm has an inventory, and from next Wednesday,

It will be at a 43-year low.

Subscribe now

Process the organic waste a livestock economy already throws away — manure, slurry, food scraps, the grass it cuts anyway — and Europe could, in principle, make a quarter of the gas it burns at home, from material it currently pays to get rid of.

There is a number that surfaces every time the European gas price spikes, and it is enormous.

The number is real.

It is also, almost everywhere it is quoted, the wrong kind of number — a measure of what exists, not of what can be delivered.

Confuse the first for the second and you will write targets, design subsidies and underwrite plants against a quantity the geography has no intention of supplying.

That is what Europe has spent four years doing, and the bill for the confusion is starting to arrive.

The pitch is perfect — on paper

The Dutch front-month gas contract settled near EUR 48/MWh in the first week of June — easing toward EUR 47 by week's end as a Gulf peace draft surfaced, still a quarter above a year earlier — and the reason was not weather and not demand but a stalemate in the Gulf — talks between Washington and Tehran going nowhere, a flare-up in the fighting, and a European storage stack sitting little more than a third full going into a summer it is supposed to spend refilling.

I have written here before about the way a chokepoint twenty-one nautical miles wide reprices a molecule the entire continent had stopped thinking about; this is the same lesson arriving through the gas curve rather than the oil one.

The premium in that forty-eight-euro print is a geopolitical premium, paid on a fuel Europe mostly imports, and every time it widens the same answer surfaces in the same rooms: stop importing the molecule and grow it at home.

Biomethane is the cleanest version of that answer.

It is upgraded biogas — raw biogas is roughly half to two-thirds methane, the rest carbon dioxide and traces, produced when bacteria break down organic matter without oxygen; strip the carbon dioxide out and clean it and you are left with a stream that is upward of ninety-five per cent methane, chemically a twin of fossil gas, injectable into the same pipes, burnable in the same boilers and turbines, with no modification to the network or the appliance at the other end.

It is made from waste a livestock economy already generates — manure, slurry, food waste, sludge, grass that would otherwise be cut and fed. It is dispatchable, it can be stored, and it is produced within a few miles of where it is consumed.

On an energy-security scorecard it is close to ideal: domestic, low-import, no critical minerals, the kind of project a policymaker reaches for precisely when the import bill spikes the way it did this week.

I spent the last stretch finishing a postgraduate qualification in renewable-gas development, and the single thing that separates the people inside this market from the people commenting on it is not enthusiasm for the molecule — everyone has that — but a sense of how far the quoted numbers are from the deliverable ones.

The figures that circulate are enormous and they are real.

They are also the wrong kind of number.

They describe what exists, not what arrives.

That distinction is the whole of this piece, and it is borrowed from the discipline that has spent a century being disciplined about exactly this gap.

The reserve is a rounding error

An oil and gas company is not permitted to tell its shareholders that the hydrocarbons in the ground are its reserves.

The hydrocarbons in the ground are a resource — the total physically present.

A reserve is the subset of that resource which is recoverable under current technology, at current prices, with current infrastructure and current law: proven, probable, possible, a descending ladder of confidence, audited, and a small fraction of the resource beneath it.

The conversion factor between the two is not geology.

It is economics, logistics, and access.

A field can hold a billion barrels and book almost none of them, because they sit too deep, too far, too sour, or too cheap to lift against the prevailing price.

Take the resource honestly, because it deserves to be taken honestly.

The International Energy Agency’s 2025 geospatial assessment — the first study to model the potential and cost of biogas at over five million individual locations worldwide, screening for sustainability so that nothing competes with food or clears forest — puts the sustainable global potential at close to one trillion cubic metres of natural-gas equivalent a year, drawn from organic waste streams that already exist.

That is roughly a quarter of world gas demand. Almost eighty per cent of it sits in emerging and developing economies, led by Brazil, China and India, where India’s potential alone is nearly twice its current gas consumption.

The feedstock base is genuinely large and genuinely under-used.

Now read the reserve.

Of that trillion-cubic-metre resource, the same report finds that around five per cent is being used today.

Biomethane production — the upgraded, grid-quality molecule, as opposed to raw biogas burned on site — runs at roughly ten billion cubic metres a year globally, two-tenths of one per cent of gas demand, growing fast off a tiny base.

The resource is a quarter of world gas demand.

The reserve, after decades of that waste accumulating and four years of the highest gas prices in living memory, is a rounding error.

5% —Five per cent.

The gap between those two numbers is the most important fact in the sector, and it is the fact outsiders skip. They hear the trillion and price the molecule as if the trillion were available supply.

It is not supply.

It is the seam in the rock.

Whether it ever becomes a reserve depends on a sequence of filters, each of which marks the resource down, and the first of them is the one the new IEA study was built to measure.

Five filters, each one a markdown

Before the filters, one definitional point, because it is the other thing outsiders get wrong.

Biogas and biomethane are not the same product, and a great deal of confused commentary treats a country’s biogas plants as if they were biomethane supply.

• Raw biogas is a low-value, impure, on-site fuel — burned in a generator for heat and power near the digester, because moving it anywhere is uneconomic.

• Biomethane is the upgraded, network-grade molecule that has paid the cost of cleaning and compression to earn its way into the pipe.

Ireland has around thirty operating biogas plants producing roughly five hundred and eighty gigawatt-hours of biogas for combined heat and power; it has two plants upgrading to grid-injected biomethane.

The first number is not a down-payment on the second.

They are different businesses with different economics, and the ladder from one to the other is the first place the resource starts shrinking.

Each filter is a markdown. Applied in order, they take the trillion-cubic-metre headline and a five-point-seven-terawatt-hour national target and resolve them into the small, awkward, subsidy-dependent quantity that actually reaches a customer.

None of the five is fatal on its own.

Stacked, they are the difference between a resource and a reserve.

The order matters, because the filters feed each other — the distance problem manufactures the cost problem, the cost problem sharpens the emissions problem, and the time problem caps whatever survives the other four.

Start with distance, because distance is where the geography stops being an abstraction.

Distance does not bend

The defining contribution of the 2025 IEA assessment is not the size of the resource.

It is the map.

By placing every tranche of potential against the actual location of roads, power lines and gas pipelines, the study turns an energy number into a logistics number — and the logistics number is unforgiving. Globally, only about three hundred billion cubic metres of biomethane potential lies within twenty kilometres of an existing gas pipeline.

In Europe, only around thirty per cent of the assessed feedstock potential — some fifteen billion cubic metres — sits less than ten kilometres from a transmission pipeline.

Everything beyond that distance has to either truck a compressed, upgraded molecule to a central injection point, or burn its biogas on site and forget the grid, or not exist.

This is the artery problem in reverse.

• Fossil gas arrives through a network built around a handful of dense, high-pressure sources, and the network is the cheap part precisely because the sources are few and large.

• Biomethane inverts the geometry: thousands of small, dispersed sources, most of them nowhere near the high-pressure spine, each one needing its own metered, quality-controlled injection point into a distribution grid that was never laid out to receive gas from the edges.

The average biomethane plant produces around three and a half million cubic metres a year and must connect individually; the grid was designed to push gas out to farms, not to pull it back in from them.

Ireland is the cleanest case study available, which is why the Irish research literature is unusually deep for so small a market. The ESRI’s 2026 systematic review of that literature — the most complete synthesis of the Irish evidence — finds Irish farms, pig and poultry units aside, generally small, dispersed, and frequently more than ten kilometres from the gas network, with ten kilometres itself the outer economic limit for hauling wet, low-density slurry to a digester.

The molecule and the pipe are in different places, and the distance between them does not respond to ambition.

You can watch the Irish system trying to engineer around exactly this filter, in real time, which is the surest sign the filter is binding.

Gas Networks Ireland has just brought a Central Grid Injection facility at Mitchelstown in Cork towards operation — a thirty-million-euro hub, designed to take up to seven hundred gigawatt-hours of biomethane trucked in from plants too far from the spine to connect themselves: the virtual pipeline, made of diesel lorries, built because the physical one does not reach. And in March the regulator approved an interim Reverse Grid Compression scheme — letting rural plants push gas back up into a higher-pressure tier rather than flare it off when the local network cannot absorb what they make — which the network operator estimates might unlock an additional four hundred and thirty-seven gigawatt-hours, around seven and a half per cent of the national target.

Both interventions are clever.

Both are also confessions: the resource is stranded by distance, and turning any of it into a reserve means paying, in trucks and compressors and regulatory workarounds, to defeat the few kilometres the map will not bend.

The feedstock is mostly water

Distance would matter less if the feedstock travelled well.

It does not.

The single most attractive Irish feedstock on paper — animal slurry and manure, abundant because the agricultural economy is built on livestock — is the worst feedstock to move, because it is mostly water.

The ESRI synthesis puts the theoretical Irish slurry-and-manure resource at around four terawatt-hours a year and the practically accessible portion at roughly half a terawatt-hour: an eight-to-one markdown, driven by exactly two things — the low energy density that makes hauling wet slurry more than a few kilometres pointless, and the long grazing season during which cattle are in fields and the slurry is never collected at all.

The resource is real.

Most of it is in the wrong state, in the wrong place, for the wrong part of the year.

Grass is the Irish exception that proves how wide the markdown can run. Sixty per cent of Ireland’s land is grassland, and grass silage is the feedstock with by far the largest theoretical energy — the ESRI review records one estimate as high as thirty-five-point-seven terawatt-hours a year from grass alone, an outlier so large it sits off the top of the chart.

The practical, economically accessible grass figure across the same body of studies is somewhere between 1-13 terawatt-hour (TWh).

That is a spread of more than a factor of twenty within a single feedstock in a single country — not because the researchers disagree about the grass, but because they disagree about what to subtract for yield, for the land you would have to free from the national cattle herd, for the distance to a digester, and for whether anyone would actually pay a farmer to grow energy instead of beef.

That chart — theoretical crosses scattered high, practical dots clustered low, for every feedstock side by side — is the resource-reserve distinction drawn as a picture, and it is worth putting in front of anyone who quotes the top of it. The dots are the reserve. The crosses are the seam in the rock.

It costs more than the gas it replaces

A molecule that reaches a pipe and is dense enough to move still has to be worth making.

Here the IEA’s location-by-location costing does the sector a hard favour.

On average, biomethane can be produced for around eighteen dollars a gigajoule — lower than end-user gas prices in many markets, but roughly five times the average cost of producing natural gas, and equivalent to oil at about a hundred dollars a barrel.

In Europe, the average is higher still, around twenty-two dollars a gigajoule, or about seventy-four euros a megawatt-hour at 2024 exchange rates.

Set that against the high-forties fossil print this market is paying — a print already inflated by a Gulf risk premium, and one easing as the premium bleeds out on peace headlines — and European biomethane is sitting something like one and a half times above the benchmark it is supposed to replace, in a week when that benchmark is unusually expensive.

The curve is not flat, and that is the part worth holding onto.

The cheapest tranches are genuinely competitive: landfill gas recovery comes in around eleven dollars a gigajoule, and roughly twelve per cent of the global resource — about a hundred and twenty billion cubic metres — sits in dense, high-yielding clusters with good infrastructure where production runs near fifteen dollars, fifteen per cent below the global average, because scale and proximity do the work.

The IEA’s reading is that nearly forty to fifty billion cubic metres of biomethane could be produced today at or below prevailing wholesale gas prices — four to five times current global output.

There is a real reserve in here. It is just an order of magnitude below the resource, and it lives in specific places — Brazilian sugar mills, dense feedlots, big landfills — not on the dispersed grassland the headline numbers lean on.

Ireland sits on the wrong side of that curve almost everywhere.

The ESRI synthesis puts the average Irish minimum selling price at around ninety-one cents a cubic metre — a mean near eight-and-a-half cents a kilowatt-hour, roughly double the post-crisis wholesale gas price — and finds biomethane production, across essentially the whole reviewed literature, not viable without financial support.

The capital cost runs between one-and-a-half and four million euros per megawatt, the digester alone above sixty per cent of it, and feedstock plus operating costs make up around two-thirds of the lifetime cost — which is the deeper reason the curve will not fall far: this is not a technology waiting for a learning curve, it is a logistics and biology business whose costs are mostly labour, haulage and feedstock that no amount of deployment makes cheaper.

At 30 euros a tonne, grass silage does not pay an Irish farmer as well as the beef the land already carries. The cost filter, then, does not delete the reserve.

It relocates it — away from the places the resource is largest, towards the few sites where density and proximity already align — and everywhere else it hands the gap to a subsidy, which is where the next filter and the politics begin to interact.

“Renewable” is a threshold, not a property

Here is the filter outsiders never see at all, because it hides inside the accounting.

Biomethane is not automatically low-carbon. Whether a given molecule counts as renewable — whether it is allowed to earn the obligation revenue and the certificates that make it bankable — depends on clearing an emissions threshold, and the threshold is demanding. Under the EU’s renewable-energy rules, biomethane used for heat must come in under roughly 14.5 grams of carbon-dioxide-equivalent per megajoule of energy, and for transport under about 33.

Miss the threshold and the same physical molecule is just expensive gas with a green story and no green premium.

What decides whether a plant clears the line is not the feedstock’s headline carbon story but the management of the process and the convention used to count it.

The largest single risk is fugitive methane: the IEA’s evidence suggests the average biogas plant leaks methane equivalent to around six per cent of its output, and because methane is a far more potent greenhouse gas than the carbon dioxide it is meant to displace, leaks of that order can undercut or entirely erase the climate benefit.

The fix is known — closed digestate storage, combustion of the off-gas during upgrading — but it is a management choice, not a guarantee, and there is, on the IEA’s own reading, too little measurement at existing plants to confirm the benefit across the fleet.

The Irish evidence sharpens this into something close to a ranking.

The ESRI synthesis finds that the threshold for heat is reliably met only by the mono-digestion of animal slurry — because counting the methane that the slurry would have emitted in an open lagoon hands the process a large avoided-emissions credit before it generates a molecule at all.

First-generation energy crops and mono-digested grass struggle to clear even the looser transport threshold, and only do so on optimistic accounting assumptions about land use and sequestration.

The reflexive sting is that the feedstock with the best carbon credentials, slurry, is the worst on density and distance — the very molecule that clears the emissions filter is the one stranded by the first two.

Which means “renewable gas” is not a property the molecule carries out of the digester. It is a status it earns, plant by plant, through capital spent on leak control and accounting choices defended to a certifier — and a molecule that fails the threshold has been marked all the way back down to the resource it started as.

Dated on the day it is booked

The last filter is time, and it is the one that makes the economics circular.

Across the Irish literature and the wider European debate, biomethane is framed not as a destination but as a bridge — the quick decarbonisation answer for the hard cases, heavy goods vehicles and high-temperature industrial heat, for a window of perhaps ten to fifteen years until battery-electric trucks, high-temperature heat pumps and hydrogen mature and take the load.

The ESRI synthesis is explicit that the reviewed studies expect biomethane’s role in power and then in heat and transport to be phased out as those alternatives phase in, leaving only a residual — trucked gas to off-grid corners the electrons never reach.

Read that as a project financier reads it, and the problem is immediate.

A biomethane plant is a multi-million-euro asset with a 15 - 20 years payback that is mostly operating cost, built to clear a cost curve only with a subsidy, to serve a use case the policy consensus expects to disappear inside its own depreciation schedule.

You are asked to commit patient capital to an asset that the same policy framework has pre-announced as transitional.

That is not a financing structure; it is a contradiction, and it shows up directly in the deployment data — investors want a long, certain offtake, and the bridge framing denies them one.

Europe has now begun, quietly, to act on that framing at the level of its own strategy. The Clean Industrial Deal, the EU’s 2025 statement of where the decarbonisation money should go, leans towards electrification and emerging technologies and largely passes biomethane over — and the sector’s own trade body has spent the year pleading for binding targets and legal certainty precisely because the soft, interim, non-binding status of the molecule is starving it of the long-horizon capital it needs.

The time filter, in other words, is not only an engineering fact about when electric trucks arrive. It is a belief about biomethane’s future that is shaping its present — and a belief shaping the fundamental it claims merely to describe is where this stops being a list of constraints and becomes a loop.

The map beats the target

Assemble the four years since the import shock and the shape of the thing becomes visible — the shape you could not see in 2022, when the target was set, because the deployment had not yet failed to arrive.

Start with the belief.

When Russian gas stopped flowing, Europe needed a domestic answer fast, and biomethane offered one with a number attached: thirty-five billion cubic metres a year by 2030, written into REPowerEU in 2022 as the bloc’s biomethane ambition.

Ireland followed in May 2024 with a National Biomethane Strategy targeting five-point-seven terawatt-hours by 2030 — ten per cent of national gas demand, more than three times the country’s previous target — on the stated basis that Ireland has one of the highest per-capita biomethane potentials in Europe.

The belief was that the resource was money-good: that because the waste existed, the supply would follow if the ambition were set high enough.

That belief shaped behaviour.

Targets were written against the resource number, not the reserve.

Blending obligations, capital-grant schemes, a Renewable Heat Obligation, certificate markets and decarbonisation accounting were all calibrated to the gross figure — thirty-five billion cubic metres, five-point-seven terawatt-hours — as if the trillion-cubic-metre seam and the deliverable reserve were the same quantity.

Capital and policy committed against potential.

The behaviour then altered the fundamental, and not in the direction the belief assumed. The thirty-five-billion target was set in 2022, when EU biomethane production was around three billion cubic metres. By the first quarter of 2025, installed capacity had reached about seven billion, and the European Biogas Association’s own investment outlook — the industry’s own pipeline, not a sceptic’s — projects roughly seven-point-three billion by 2030.

That is about a fifth of the target.

The constraint the targets ignored — distance, density, cost, the emissions threshold, the dating of the asset — was always there, marking the resource down to a reserve the policy had refused to price.

Then the feedback turned, and you can date the inflection precisely.

In late 2025 the IEA — the same agency whose geospatial work maps the resource — put Europe’s biomethane trajectory next to Europe’s target and judged the pace nowhere near sufficient to reach thirty-five billion cubic metres, projecting the bloc to land around ten billion by 2030.

That is the moment the loop reversed: a belief that ran self-reinforcing for three years — set the target, attract the capital, confirm the ambition — became self-defeating the instant the deliverable reserve was placed beside the resource-sized target and the shortfall was named out loud. The re-rating is already under way.

The Clean Industrial Deal passes the molecule over; the trade body begs for binding targets to replace the soft ones; the financing horizon shortens as the asset’s future is openly questioned. The number that was meant to pull supply into being is now, by being so visibly unmet, repricing the thing it was supposed to build — from scalable drop-in gas substitute to location-bound, subsidy-dependent, interim molecule.

That is the reflexive turn in full: the resource estimate even grew over the cycle — the IEA’s sustainable potential was revised up between its 2020 and 2025 assessments — while the deliverable reserve barely moved, so the gap the target was built to close got wider on paper precisely as the disappointment set in.

The belief did not measure the supply.

The belief tried to manufacture the supply, collided with the map, and is now being unwound by it.

What the reserve is — and the marks to watch

Now that all of it is on the table, the thing biomethane has been mistaken for comes into focus. It is not a wedge that scales with ambition.

It is a reserve — a real, bankable, defensible quantity of gas, an order of magnitude below the resource that is quoted for it, concentrated in specific geometries where density, proximity, the cost curve and the emissions threshold happen to line up: a feedlot beside a pipe, a landfill with a recovery system, a sugar mill with a by-product and a diesel habit, a cluster dense enough to share infrastructure.

My read, having spent a year inside the Irish and European version of this argument, is that the sector’s central error is not technical and not even economic — it is the category mistake of booking a resource as a reserve, and then writing policy, targets and investment cases against the booking.

So watch the marks, because they are dated and they are specific.

• In Ireland, the Renewable Heat Obligation has already slipped from its mid-2026 commencement toward 2027, held up by an EU standstill running to the twenty-ninth of June and a Commission objection to the biomethane multiplier the scheme leaned on — so watch whether it commences at all in 2027 or slips again, and whether the multiplier survives the redesign; and whether the RGGO sustainability-certification regime that the EU Gas Directive requires from the fifth of August 2026 tightens or loosens what counts as a deliverable molecule; whether the Mitchelstown hub and the Reverse Grid Compression scheme convert any meaningful slice of the stranded rural resource into injected reserve, or whether they confirm how expensive that conversion is.

• In Europe, the 2030 production print against thirty-five billion cubic metres — the IEA has already called the miss at around ten, and the gap between ten and thirty-five is the size of the category error, measured in gas.

  

Europe is paying a Gulf-risk premium this week for a molecule it imports, and reaching, as it always does when the premium widens, for the gas it could grow at home.

The waste is there.

The trillion cubic metres is there.

But the gas Europe can actually deliver is not the gas in the slurry tank — it is the far smaller quantity that survives the ten kilometres to the pipe, the cost curve, the carbon threshold and its own pre-announced obsolescence.

A resource is not a reserve.

The map said so before the target did, and

the target is now learning it the expensive way.

Subscribe now

The Strait of Hormuz has been effectively shut since 28 February — transits running at roughly six per cent of a pre-war cadence that topped a hundred ships a day, war-risk premia repriced from a fifth of one per cent of hull value to a full one per cent for seven days of cover, all twelve International Group P&I clubs having cancelled their war-risk extensions on seventy-two hours’ notice in the first week — and the crude and LNG that used to cross the Gulf now reroutes through the corridor Malaysia shares with Singapore, a waterway that already carried roughly a quarter of seaborne oil before anyone needed it to carry more.

Kuala Lumpur has played the moment the way it plays every moment: in late March the prime minister telephoned Iran’s president and by early April had walked seven Malaysian tankers through a closed strait on the strength of the call, thanked Tehran publicly, and has watched the national fuel assurance roll forward one month at a time since — to May, then end-June, then end-July, now ‘beyond July’ this week — cover renewed on a calendar, not secured on a structure.

A barrel changes hands in the high eighties — a two-month low printed this week, not because the war resolved but because Washington floated a weekend signature on a fourteen-point draft that would reopen the strait within thirty days, and neither the mines (six months to clear, on the American estimate) nor the war-risk architecture reprices on a draft — and Malaysia, net LNG exporter, premium-crude seller, friend of everyone, is still being paid, not punished, by a war it stayed out of.

That is the posture into which the International Institute for Strategic Studies has just dropped a thirty-four-page invoice.

Its June report on a Taiwan Strait contingency prices what happens when the next chokepoint event is not someone else’s — when the strait that closes is not the one Malaysia profits from but the one its flagship industry is plugged into. The maximum estimate for a year-long conflict is RM790.2 billion, or US$172.6 billion: forty-one per cent of GDP, 2.6 million jobs.

The number is not the finding.

The finding is the wiring diagram underneath it — the demonstration that the equidistance Malaysia has treated as redundancy is nothing of the kind.

Not a parallel circuit. A series one.

Equidistance was drawn as a parallel circuit

The doctrine is explicit and it is proud.

At the 2025 Shangri-La Dialogue, Anwar Ibrahim named Malaysia’s approach ‘active non-alignment’ — the deliberate refusal to choose between Washington and Beijing, held up as the configuration that maximises strategic space.

The economic logic beneath the diplomacy runs like a parallel circuit: keep a heavy-gauge connection to China, the largest trading partner; keep a second to the United States, the largest single investor among the three principals at roughly RM7.6 billion of annual inbound FDI; keep a third to Taiwan, which took about RM67 billion of Malaysian exports in 2024. If one leg browns out — a tariff war, a sanctions round, a quarrel that is not Malaysia’s — current reroutes through the others and the lights stay on.

The 2025 tariff shock seemed to prove the design: Putrajaya diversified toward the EU, the Gulf and Africa within months, and growth ran above five per cent into the autumn.

The load this circuit carries is enormous, which is precisely why the design matters. Malaysia’s trade ratio — exports plus imports against GDP — stood near a hundred and fifty per cent in 2022, down from its 2000 peak above two hundred but still among the highest of any mid-sized economy on earth.

A trade ratio of that size is a decision to run the national economy at high current all the time — to draw prosperity through the wires rather than generate it behind a wall — and the doctrine’s entire claim is that the wiring is redundant: many sources, many sinks, no single point of failure.

The IISS report’s contribution is to have traced them.

They do.

The board is one supplier deep

Start where the report starts, with the export ledger.

Electrical and electronic products are sixty-seven per cent of everything Malaysia sells to Taiwan.

But the export line is the shallow reading; the deep one is on the import side.

Malaysia’s semiconductor industry — which generates half of all E&E value added — buys sixty-eight per cent of its imported inputs as electronic components and boards, and Taiwan’s share of Malaysia’s E&E imports has climbed from 2.8 per cent in 1989 to 8.5 per cent in 2022, rising in almost every reporting period and accelerating after the 2018 trade war pushed Taiwanese supply chains south.

The wafer arrives from Hsinchu, the packaging and test happen in Penang and Kulim, the finished device ships to everyone — to China, to the United States, to Europe.

The China leg and the America leg of Malaysia’s supposedly parallel circuit are, at the level of the physical product, the same wire passing through the same node, and the node is Taiwan.

The report quantifies what a break at that node conducts.

Its econometric pass finds that halving E&E imports from Taiwan does not merely halve the matching exports back to Taiwan — a loss of RM32.6 billion — it cuts Malaysia’s E&E exports to every other destination by a further RM10.7 billion, because the products sold to third countries cannot be built without the inputs that stopped arriving.

That is the signature of series wiring: the fault does not stay in the leg where it occurs; it propagates through the whole board, dimming customers who have no quarrel with anyone.

Stockpiles soften the first weeks — firms learned inventory discipline in the pandemic and again in the tariff war — but those buffers were built for disruptions measured in months, and in a sector where the product itself goes obsolete on the shelf, a deep inventory is a depreciating asset, not a reserve.

This is the section of the report that should be read twice, because it converts a diplomatic abstraction into an engineering fact. Equidistance describes Malaysia’s relationships. It does not describe Malaysia’s circuit, and in a contingency it is the circuit, not the relationships, that carries the current.

Forty-one per cent is the fault current

The report runs two scenarios across one year, each with a minimum and maximum case.

The crisis scenario — an air-and-sea quarantine of Taiwan, traffic recovering to half its prior level and staying there — costs Malaysia between RM40.2 billion and RM79.5 billion, between 2.1 and 4.1 per cent of GDP, with up to 250,000 jobs gone, concentrated in E&E and in a tourism sector that contributes fifteen per cent of GDP and would lose its Chinese segment, twelve per cent of receipts, overnight. Painful, and historically legible — the Asian financial crisis took 7.4 per cent of GDP in 1998, the pandemic 5.5 in 2020.

A quarantine is a shock Malaysia’s institutions have rehearsed.

The conflict scenario is not. A full invasion, US and allied entry on Day 15, bilateral US–China trade down ninety-five per cent by year-end and global trade down by as much as half — wartime numbers, calibrated against the fifty-five to sixty per cent collapse Glick and Taylor estimate for the whole of the Second World War — produces Malaysian losses of RM464.4 billion to RM790.2 billion: twenty-four to forty-one per cent of GDP, 1.5 to 2.6 million jobs on the report’s employment elasticity of 0.38 per point of output.

The estimate is deliberately maximalist — it models no counter-cyclical response from Putrajaya, because no honest model can predict one — but the calibration that matters sits in the report’s own historical mirror: Glick and Taylor put the cost of the world wars to neutral countries at 7.2 and 11.8 per cent of GDP. Malaysia, neutral, loses up to forty-one.

Hold those two figures against each other, because the gap between them is the article’s argument in a single comparison.

Twentieth-century neutrality cost a tenth of the economy at worst; twenty-first-century neutrality, for a country wired like Malaysia, costs four times that — more than many combatants surrendered in the wars themselves.

Nothing about the legal status changed.

What changed is the trade ratio — the decision to run at a hundred and fifty per cent — which acts on an external shock exactly as a conductor acts on current: it does not attenuate the fault, it transmits it at full amperage.

The fault current is not proportional to your distance from the short.

It is proportional to the gauge of your wire.

The loop held until Day 15

Here is where the mechanism turns reflexive, and it is worth building the loop in full because the report supplies every step without naming it.

The belief: equidistance insulates — staying close to both blocs protects Malaysia from their collision. The behaviour the belief licenses: deepen both legs simultaneously, because if connection is protection, more connection is more protection — and so Taiwan’s share of E&E imports rises in every period, Chinese FDI compounds, the report observes that most of the Malaysian public still treats a Taiwan contingency as an abstraction, and no structural mitigation is undertaken.

The altered fundamental: the deepening itself rewires the economy — every additional point of integration is another strand in the conductor, so the belief in insulation manufactures, year by year, the exposure it claims to be insulating against. And the feedback: each calm year, each crisis profitably skirted — the tariff shock absorbed, the Hormuz closure monetised, seven tankers talked through a war zone on one phone call — confirms the doctrine and licenses the next increment of integration.

The inflection — the moment the loop turns self-defeating — has a date in the report’s own scenario: Day 15, when Washington enters the war and demands that every country in the region choose allegiance, basing rights and sanctions participation included.

Equidistance is an option that pays only while unexercised; the contingency it was bought against is the very event that cancels it.

And the Hormuz war, run live since February, has already shown the retail version of that cancellation: Malaysia’s safe passage through a closed strait was not produced by the doctrine, it was produced by the prime minister personally, call by call, seven hulls at a time — protection as a favour renewed at the discretion of a belligerent, which is the opposite of insulation. The parallel circuit confirmed itself for twenty years because nobody ever threw the breaker.

The scenario throws it.

The swap line is not a fuse

The strongest counter-argument lives in the report’s financial-sanctions section, and it deserves its full weight.

In the crisis scenario, the IISS judges, the Western response stays measured — no SWIFT expulsion, no freeze of People’s Bank of China reserves, because amputating an economy ten times Russia’s size from the dollar system would cost the amputators trillions.

Even in the conflict scenario, China’s remaining US$652.3 billion of Treasuries — pared six per cent in March alone as Beijing trims exposure — makes the PBoC, in the report’s judgement, too large for the system to fail. And Malaysia holds what looks like a dedicated fuse: the bilateral PBoC–Bank Negara swap line, in place since 2009 and renewed in 2021, through which ringgit–renminbi settlement could route around a sanctioned dollar system entirely, on CIPS rails, the way Indian banks built rupee vostro accounts for Russian trade after 2022.

If the financial leg holds, the argument runs, the real-economy damage is survivable — Malaysia trades on through the side channel.

The report’s own second-order analysis melts the fuse.

A swap line is plumbing; what moves through plumbing is decided by compliance officers, and a Malaysian bank facing US secondary-sanctions exposure will restrict renminbi settlement not because the channel is blocked but because using it endangers every dollar relationship the bank owns — the correspondent accounts, the letter-of-credit confirmations, the clearing access that carries Malaysia’s palm oil, LNG and refined-product exports through USD-based commodity houses. The fuse fails for the same reason the circuit fails: the dollar leg and the renminbi leg are not independent, because the institutions that would operate the alternative are themselves wired into the system the alternative is meant to bypass. De-risking is the financial sector’s series wiring.

The report expects letter-of-credit non-confirmation, settlement lag, counterparty premia and a working-capital squeeze on export SMEs — the fault current, again, arriving through the very channel built to interrupt it.

The reading on the meter

Assemble the whole board and the report’s deepest finding is not the forty-one per cent — it is that the cost of someone else’s war now scales with openness rather than with proximity or participation, which means the most globalised neutrals carry more war exposure than some combatants did a century ago, and carry it unmarked, because nothing in a trade ratio looks like a munition.

Malaysia’s number is Malaysia’s, but the wiring is the region’s: every economy that read integration as insurance has been running the same loop, and the IISS has simply been first to put a meter on one of them.

The doctrine buys voice, optionality, the seven-tanker phone call — genuine goods. It does not buy a single ohm of resistance between Hsinchu and Penang, and the report’s service is to have separated the two ledgers that ‘active non-alignment’ had allowed to blur into one.

The playout starts on the desks that already moved once.

When the 2025 tariffs landed, Malaysian Pacific Industries fell twenty-three per cent in the wake of one announcement — the calibration print for what Bursa Malaysia’s semiconductor complex does when the Taiwan node so much as flickers, and in a conflict the report expects that repricing to be deep and durable rather than sharp and mean-reverting.

The tariff print itself has already round-tripped — the stock trades within sight of its fifty-two-week high this week, the loss unlearned inside a year — which is not a rebuttal of the calibration but the loop in miniature: every repricing that reverses teaches the market to treat the next flicker as an entry, right up to the one that is not.

The ringgit, which gave up roughly two per cent against the dollar between late May and the eleventh of June on nothing more than Hormuz headlines — and recovered less than half of one per cent of it when the peace headlines arrived — trades near 4.06 with its safe-haven beta running entirely the wrong way: a currency that sells the war hard and buys the peace timidly has told you which risk it believes.

The structural implication lands in the energy market through the corridor itself: the Hormuz closure has already shown — as I argued during that war — that the insurance market shuts a strait before the first missile does, and the same architecture of war-risk premia, P&I withdrawal and sovereign backstops would migrate east to the Malacca–Singapore lane in any Taiwan contingency, putting a standing risk premium into delivered LNG costs for North Asia and repricing the option value of every storage and transhipment asset on the strait — including the Maharani Energy Gateway in Johor, whose free-zone gazettement decision is due from Kuala Lumpur before the end of 2026 and which is, in effect, a levered bet that chokepoint risk is now a permanent feature of Asian freight.

Two gauges will tell whether the loop is still running.

• The first is Taiwan’s share of Malaysia’s E&E imports: 8.5 per cent at the 2022 reading, rising for three decades — if the doctrine were genuinely converting into structure, the 2026 print would fall; the thesis here says it rises, because the loop rewards deepening right up to the inflection.

  

• The second is the report’s own early-warning indicator: sustained PBoC non-subscription to new Treasury issuance — holdings dropping decisively below US$600 billion in the monthly TIC data — the next print, covering April, lands on 18 June — would mark Beijing repositioning for the scenario in which the wiring is finally tested.

The corridor is being paid wartime economics this quarter — bunkering, transhipment, the war-risk premium embedded in every delivered cargo — by a war Malaysia neither joined nor priced, even as the flagship terminal at Port Klang turns away the boxes that same war diverts toward it.

The strait Malaysia profits from today and the strait its prosperity would break on tomorrow are the same wire — and a wire does not know which direction the fault comes from.

Subscribe now

The first came from the Oval Office.

President Trump told reporters that American forces have been running oil through the Strait of Hormuz at night — “We’ve been taking out millions of barrels of oil. Nobody knows it” — twenty-two ships on Tuesday alone, he said, lights off, past Iranian radar he says has been bombed blind, an operation he credits with holding crude below a hundred dollars in a war that would otherwise have it at two hundred and fifty.

By evening the figure had grown: two hundred ships, more than a hundred million barrels, the United States in control of the strait.

No tracking service has produced the ships.

The story arrived pre-armoured against that — dark vessels, blinded radar, a secret that exists only in the telling of it.

The second manifest posted at half past ten Washington time, as it does every Wednesday: the EIA’s Weekly Petroleum Status Report, the audited account of every barrel that entered, left or sat inside the American oil system in the week to 5 June.

It is not secret.

It is not capable of being secret — it is assembled from mandatory weekly surveys of every large refiner, importer, pipeline and terminal operator in the country.

And it names a different supplier for the market’s calm.

The tape, meanwhile, sits obediently inside the President’s frame.

Brent has swung between roughly ninety and ninety-four dollars through Wednesday’s session — whipping on Iranian strikes against Bahrain, Jordan and Kuwait and on American threats of strikes in reply — some twenty-seven dollars above last June, a long way under April’s panic prints near a hundred and seventeen. Petrol at American pumps has fallen four Mondays running, to $4.15 from May’s peak above $4.60.

A market exhaling.

The question this piece answers is what, exactly, it is breathing.

Two hundred ships by sundown

The claim’s load-bearing feature is not its size but its construction — every element removes a channel of verification.

The ships ran dark, which answers the transponder record; Iranian radar is destroyed, which answers Tehran’s silence; the operation is covert, which answers the Pentagon’s; and the counterfactual — two hundred and fifty dollars in the world without the convoy — answers the price itself, because every print below that level now testifies for the operation.

Even the silence of the marine-traffic monitors, who report nothing resembling a nightly twenty-two-ship lift, is converted by the story into evidence of tradecraft.

Run the arithmetic anyway.

Twenty-two laden tankers is somewhere between twenty and forty-odd million barrels depending on class — at the top of that range, a single night moving twice what the entire strait carried daily in peacetime — and steel of that displacement does not hide from synthetic-aperture radar, which images hulls through cloud and darkness alike; a tanker can switch off its transponder, not its existence.

And the figure would not hold still long enough to be checked: between the Oval Office in the morning and Truth Social in the evening, twenty-two ships became two hundred and the millions became a hundred million.

A logistics report does not grow ninefold inside one news cycle.

A position being added to does — which is the first reason to read the announcement not as intelligence about ships but as an instrument aimed at a price, a suspicion that the doctrine of convoys, properly understood, hardens into something close to certainty.

Convoys deter in daylight

The United States has run this exact operation before, and the way it ran it is the tell. Operation Earnest Will, 1987–88, reflagged eleven Kuwaiti tankers and escorted them through these waters in public — warships alongside, press embarked, the schedule an open provocation — because the mechanism of a convoy is not stealth, it is demonstrated will.

When the Bridgeton struck a mine on the very first run, it struck it in front of cameras, and the convoys continued, which was the point: escort deters only if the adversary watches it survive. A convoy that hides from its adversary protects nothing, because the protection is the visibility.

I wrote here on the sixth of May that the 2026 architecture could not support an Earnest Will posture — and the White House promptly supplied the proof, pausing Project Freedom’s visible escort lane through Omani waters after two transits; the reach lasted thirty-six hours.

Hold those two numbers against each other.

• The convoy you could watch managed two transits in a week before it stopped.

• The convoy you cannot watch manages twenty-two a night.

• Doctrine says deterrence is performed in daylight; the daylight version failed at scale; so the operation has migrated to the one theatre where it cannot fail — the theatre where it cannot be seen.

• A convoy relocated from the strait to the imagination has changed cargo: it no longer carries crude, it carries the risk premium.

The manifest posts at half past ten

Now open the other Wednesday document and ask the only question that matters: what physically supplied this market last week?

American refiners ran at 95.3 per cent of capacity and swallowed 17.0 million barrels of crude a day — the system’s intake runs on molecules, not on narrative, and every one of those molecules has a recorded origin.

Production supplied a near-record 13.8 million, including an 88-thousand-barrel re-benchmark. Imports supplied 5.9 million — down half a million on the week, just under six per cent below last year on the four-week run. And the balancing item supplied the rest: 15.1 million barrels came out of American crude storage in seven days, 7.2 million from commercial tanks and 7.9 million from the Strategic Petroleum Reserve — a combined 2.2 million barrels a day of supply that no wellhead produced and no strait transited.

Read the import slate — Table 8 of the report — the way a convoy officer reads a route card, because it records exactly which lanes are open.

Saudi arrivals on the four-week average have halved against last year — 146 thousand barrels a day versus 300. Iraqi arrivals have fallen by roughly two thirds. Nigerian and Libyan barrels have gone to nothing. And the gap has been filled from the one direction that owes nothing to Hormuz: Venezuelan deliveries are running at more than five times last year’s rate, 568 thousand barrels a day against 105 — sanctioned heavy crude rehabilitated mid-war to replace Gulf grades that stopped arriving.

If a covert lift were moving Gulf oil to world markets at scale, the largest refining complex on earth — running its Gulf Coast at 97.4 per cent — is a strange port for it never to call at.

The slate is instead the precise silhouette of a closed strait.

The ledger does record an American operation keeping the world supplied — it is just not a secret one. Net exports of crude and products ran at 5.3 million barrels a day over the past four weeks, more than double a year ago, the export pull that compressed Brent’s premium over WTI from seventeen dollars in April to about three now.

America is the visible swing supplier of this war, customs-cleared and counted — and the barrels it swings are increasingly coming out of storage, which means the manifest’s answer to the President is not that nothing is moving; it is that what moves is a stock.

Walking the floating stock out

Here is the strongest version of the claim, and it deserves its full weight, because a kernel of it is corroborated.

Energy Secretary Chris Wright said Tuesday that exports through Hormuz are "rising very meaningfully", and Lloyd’s List Intelligence reported last week that dozens of the ships stranded inside the Gulf since the strait closed have exited in recent weeks under quiet coordination with the US Navy.

The Gulf has held a trapped fleet since the war shut the strait at the end of February — by early May the count ran past 1,550 commercial vessels — and walking laden tankers out of that anchorage is real, dangerous and worth doing.

It is also, precisely, a stock release.

Those cargoes were lifted weeks ago; they have sat at anchor as floating storage; their exit delivers stored barrels to market exactly as a salt cavern delivers them — once. A hundred million barrels, the President’s own evening figure, is a queue, not a rate: cleared over some weeks it is a meaningful flow while it lasts and zero on the day the anchorage empties, and it refills only if loadings at Gulf terminals resume and repeat — which is what the import slate, the halved Saudi line, the vanished Iraqi line, says has not happened.

The pattern extends past America: China, the strait’s main remaining customer, is reported to be running its refineries off inventories rather than lifting fresh cargoes. Line the channels up — the cavern, the anchorage, the tank farm — and they are one object wearing three costumes: stored barrels in motion, dressed as supply, and the costume holds exactly as long as nobody asks what refills them.

Words as escort

This is where the announcement starts doing economic work, and the loop is worth building in full, because the entire trade rests on it.

It begins with a belief: covert American operations have quietly restored the strait’s flow — abundance is hidden, capacity is in reserve.

• The belief shapes positioning: desks trim the disruption premium, insurers hold rates rather than ratchet them, refiners defer precautionary cover and run their own tanks, and the strip settles into the nineties.

• The positioning then alters the fundamental it was pricing — because the calm is not neutral.

A calm market is the only market a 1.25-million-barrel-a-day reserve drawdown can sell into without announcing itself; a calm market lets commercial holders destock into backwardation; a calm market lands the queue’s cargoes without a price signature; and a calm market keeps demand growing — total products supplied is up 3.5 per cent on the year — because the price never instructed it to stop.

The buffers that manufacture the calm are consumed by the calm they manufacture. And the loop closes on itself: every spike that fails to arrive is booked as proof the convoy works — below a hundred dollars because of us — which deepens the belief, which extends the drawdown.

Self-reinforcing, for now.

The inflection arrives the week a public gauge proves the flow never returned — Gulf arrivals flat on Table 8, the reserve through a forty-year floor, the anchorage counted empty — at which point the same loop runs in reverse against a system that has spent its shock absorbers demonstrating it did not need them.

Verbal intervention of this kind has a distinguished history, and the history specifies the failure mode. Draghi’s 2012 pledge moved sovereign spreads without a bond being bought, because the capacity behind the words was unlimited and unaudited. The capacity behind Wednesday’s words is 349.2 million barrels — metered, finite, and published weekly.

Soros, whose framework this publication runs on, made the trade of his life against exactly this structure: a defended belief backed by countable reserves invites the market to count them, and in 1992 the counting took months because the Bank of England's reserves were opaque.

• Sterling’s reserves were a secret.

• The SPR’s are a table.

Which is why the life expectancy of this loop is set by arithmetic, not conviction.

Every stock has a floor

The residual strong claim — that flow itself is genuinely recovering, that the exits are the front edge of reopened, repeating transits — has a clean break condition, and it is not rhetorical.

New loadings would surface within weeks as recovering Gulf-origin arrivals: Saudi plus Iraq back toward last June’s combined rate near 455 thousand barrels a day on the four-week average, against roughly 200 now. They would surface as an anchorage queue that stops shrinking because terminals are refilling it.

Print either and this thesis is wrong in the best possible way.

Until then, the arithmetic runs the other way.

The reserve stands at 349.2 million barrels, and the past month’s draw has run near 1.25 million barrels a day — three times the year’s average pace, eight to nine million barrels a week — which means next Wednesday’s ledger, 17 June, prints the reserve through its mid-2023 trough near 347 million and into levels last recorded in 1983, unless the draw stops this week.

On the same pace the line crosses three hundred million before August. Around it, the cushions are already thin: commercial crude sits about five per cent below its five-year average, distillate thirteen per cent below with retail diesel at $5.21, and the five-year-range charts show the draw biting in the season stocks normally build.

Against a combined American stock drain of 2.2 million barrels a day, the only genuine new flow the system was offered this week was OPEC+’s 188 thousand barrels a day for July — one part flow to roughly twelve parts drawdown.

A market can run on stocks for a long time; that is what stocks are for.

It cannot run on them indefinitely, and the date at which it stops being able to is now the most important number in the complex — which is exactly the number the convoy story exists to keep anyone from computing.

The only convoy with a schedule

Assemble the morning’s two manifests and they turn out to describe the same operation from opposite ends.

There is an American effort keeping this market supplied and calm, and part of it is even maritime — but every one of its channels, from the Louisiana caverns to the Gulf anchorages to the Chinese tank farms, is the same object: stored barrels in motion. The convoy is real in the only sense that matters to a price and unreal in the only sense that matters to a balance — it moves oil, and it replaces none.

What the market has priced as the return of flow is the velocity of inventory.

The instrument that prices the convoy story’s expiry is therefore not flat price at all; it is the back of the Brent curve and the Brent–WTI spread, and a move in that spread back through eight dollars is the market saying aloud that the stock, not the strait, was the supplier.

The gauges are public and they are dated.

Wednesday the 17th, half past ten: the reserve line, on pace, prints its lowest level since 1983. Table 8, the Saudi and Iraqi rows: recovery toward last year’s rate confirms reopened flow; another month near current levels breaks the claim.

The Lloyd’s List queue count: the convoy story has the lifespan of the anchorage it is emptying. The ships, if they sailed, ran dark for one night.

Of Wednesday’s two manifests, only one posts again next week.

Subscribe now

An editorial note before the piece.

From this week, Kardamow widens its coverage of the physical energy markets.

Alongside the established weekday essays, the publication will run short weekly interpretations of the official data — beginning with the EIA’s releases, weekly through quarterly — reading each one for what it actually means for the markets it describes, and holding to one discipline throughout: the analysis gets simpler and more accessible, never thinner.

This piece, on the June Short-Term Energy Outlook published yesterday morning, is the first of that series.

Subscribe now

A crude oil storage tank is the least dramatic object in the energy business — a flat white cylinder behind a berm, holding half a million barrels of someone else’s optionality — and the only part of it that ever tells a story is the float gauge, which records, week by week, whether the world put oil in or took oil out.

Since the last day of February, when military action closed the Strait of Hormuz to all but occasional transits, the gauges have been telling the same story everywhere outside the Gulf: out, every week, in volumes the modern market has never recorded.

Yesterday morning, the US Energy Information Administration published the June edition of its Short-Term Energy Outlook, and for the first time put the full arithmetic of that drawdown into one document.

The arithmetic contains a puzzle, and the puzzle is the piece.

The EIA assesses that crude production shut-in behind the strait reached 11.3 million barrels per day in May — against pre-conflict regional output of 25.2 million barrels per day, and more than Saudi Arabia produces in total — and that the shut-ins are still rising as Iranian storage hits its physical ceiling, forcing wells to be shut in.

Yet the Brent spot price fell in May — by ten dollars, to an average of one hundred and seven — the first monthly decline since December 2025. The consensus explanation is the diplomatic one: reports that Washington and Tehran were nearing an agreement to extend the ceasefire and reopen the strait.

The consensus explanation is true and insufficient.

Prices did not fall because supply was restored — none was. Prices fell because the market has located a new marginal supplier, one that asks no price, fights no war and files no production target. It is the tank itself. And the entire structure of the June outlook rests on how long that supplier can keep delivering.

The barrel that arrived without a ship

Strip the report to its balance and the mechanism is plain.

The EIA estimates global oil inventories fell by an average of 6.3 million barrels per day across the second quarter, and forecasts the draw will deepen to 7.6 million in the third — every one of those barrels consumed by someone, none of them produced by anyone.

Set that against the supply side’s traditional shock absorber: OPEC surplus production capacity, the spare barrels a swing producer can bring on within months, which the same report puts at thirty thousand barrels per day for the second quarter — effectively zero, the lowest figure in the dataset, against a 2016–2025 average near two and a half million.

The institution that balanced the market for fifty years has, for the duration of this closure, nothing to offer. The drawdown is doing its job instead — and doing it at roughly three times the scale of the largest surplus-capacity cushion OPEC has carried in a decade.

Read that the way a supply desk reads it.

The marginal barrel clearing the market in June 2026 does not arrive on a ship; it arrives down a pipe from a tank farm in Rotterdam, Jurong or Cushing, and it differs from a wellhead barrel in two properties that the price has not yet fully internalised. A well’s deliverability is uncertain but renewable — it produces next month roughly what it produced this month. A tank’s deliverability is certain but terminal — its remaining capacity to supply is public, measurable, and falls by exactly the amount it delivers.

The world is currently buying its price stability from a producer whose decline rate is one hundred per cent of its output. A barrel drawn is a barrel spent. And the report’s most important pages are the ones that count what is left.

Fifty days is a countdown, not a cushion

The EIA now forecasts that commercial inventories across the OECD will fall to just under 2.3 billion barrels by December — below every reading since its dataset begins in 2003, and well under the previous five-year average of 2.8 billion.

On the measure that matters to a practitioner — days of forward demand the stock can cover — the forecast bottoms at fifty days by the end of this year, again the lowest on record.

The same agency, in its February outlook, completed before the first missile, expected the opposite: inventories building through the forecast toward more than seventy days of cover, some of the most comfortable levels since the post-pandemic recovery.

Twenty days of global demand cover — roughly two billion barrels of expectation — has been revised away in four monthly reports. That revision, not the spot price, is the truest measure of what closed on the twenty-eighth of February.

Two details in the report sharpen the countdown.

• The first: the EIA states plainly that it does not expect OECD inventories to recover to pre-conflict levels within the forecast period at all — meaning the buffer that absorbed this shock will not exist, on the agency’s own numbers, to absorb the next one before 2028 at the earliest.

• The second is geometrically elegant and operationally brutal: the tank works on both sides of the strait, in opposite directions.

Outside, it empties to feed demand; inside, it fills — Iranian storage is reaching its maximum limits, the report notes, which forces producers to shut in further volumes because oil with nowhere to go cannot be lifted from the ground. The same instrument that is keeping the consuming world calm is strangling the producing world’s ability to restart.

Storage is not a neutral cushion in this market; it is an actor on both sides of the chokepoint, selling on one side and suffocating on the other — and on both sides its influence grows as its remaining room shrinks.

Demand destruction is buying the tank time

The draw rate has a throttle, and the throttle is pain.

Track the EIA’s view of 2026 global oil demand across three reports: the February outlook forecast growth of 1.2 million barrels per day; May cut that to 0.2 million; June now forecasts an outright contraction of 1.1 million — a swing of 2.3 million barrels per day of expected consumption, erased in four months.

The mechanism behind the revision is not price elasticity working quietly through household budgets; it is administrative and abrupt — government initiatives to suppress fuel use, physical shortages, the curtailment of refined-product exports — concentrated in the Asian economies that drew most heavily on Gulf crude, with the report flagging that petrochemical feedstock demand may be falling further still in volumes too poorly reported to count yet.

The live data agrees with the revision: Chinese refiners have pulled back imports hard and are running on domestic inventory instead of cargoes, half a million barrels per day of new refining capacity has been deferred, and Saudi Arabia has cut its official selling prices into the weakening Asian bid — which gives the conceit its sharpest edge yet, because the tank is now the marginal supplier on both ends of the trade, Chinese stocks substituting for imports exactly as OECD stocks substitute for production.

Demand destruction, in other words, is performing the function spare capacity used to perform — balancing the market at the margin — except that spare capacity balances by adding barrels at a price, and rationing balances by subtracting consumers at a cost.

Every barrel of demand destroyed extends the tank’s life by exactly one barrel, which is why the report’s demand pages and inventory pages are the same argument in different units.

The American contribution belongs in this section too, properly sized: US net exports of crude and products hit a record 5.8 million barrels per day in April as Europe and Asia bid for replacement supply — but the report shows US crude production growing only 0.1 million barrels per day this year, so the export surge is overwhelmingly redistribution of existing output plus a draw on the American system, not new supply at the wellhead.

The market is being balanced, barrel for barrel, by the two most expensive mechanisms it possesses — rationed consumers and depleting stocks — while the cheap mechanism, the spare wellhead, sits inaccessible behind a closed strait.

That configuration is stable in the way a held breath is stable.

The calm is spending its own collateral

Now assemble the loop, because this is where the June report stops being a balance sheet and becomes a reflexive structure.

Begin with the belief: a memorandum largely negotiated since late May — a sixty-day ceasefire extension, the strait de-mined and reopened, a framework for nuclear talks — needs only signatures.

The belief shapes behaviour: traders sold the front of the curve through May, taking Brent from an April average of one hundred and seventeen dollars to one hundred and seven, and the EIA itself has built the belief into its base case — flows resuming in the third quarter, Brent at one hundred and five through June and July, eighty-nine by the fourth quarter, seventy-nine across 2027. And the market has run ahead of even that: as this piece publishes, Brent trades near ninety-three dollars — twelve below the June–July average the EIA released yesterday on the explicit assumption that the strait stays shut into the third quarter — which means the front of the curve has already spent the reopening before anyone has signed it.

The inflection — the point where the loop turns from self-reinforcing to self-defeating — has a date attached, and the date is movable in only one direction. At the forecast third-quarter draw rate, each month the reopening slips costs the system roughly two hundred and thirty million barrels of cover, against a buffer already heading for its 2003 floor.

The strongest counter-argument deserves its full weight: the loop may simply resolve. The agreement lands, transits resume on the EIA’s schedule, shut-ins fall from ten million barrels per day in the third quarter toward 5.7 million in the fourth, demand rebounds by 2.5 million into 2027, and the gauges turn.

That is the agency’s central case and it is coherent.

But observe what it requires — that the belief currently suppressing the price turns out to be correct — and observe the asymmetry if it is not: a market that has spent its inventory cushion discovering that the cushion is needed has no second cushion, because the report has already told us the first will not be rebuilt inside the forecast.

• If the consensus is right, prices grind down a forecastable slope.

• If it is wrong, the discovery arrives into the emptiest storage position of the century.

The payoffs are not symmetric, and the price of the hedge is being set by people reading the diplomatic wire rather than the float gauge.

Put the whole report on the table and it amounts to this: the June outlook is not really a price forecast — the price path is a conditional appendix — it is a depletion schedule for the one asset standing between an eleven-million-barrel supply hole and the price that hole implies.

The spot market’s composure since mid-May is not evidence that the shock is fading; it is a real-time reading of the system’s remaining capacity to ignore the shock, and that capacity is published, falling, and on the agency’s own numbers does not recover within its forecast horizon.

I think the consensus has the causality backwards — it reads the calm price as the disruption ending, when the calm price is the mechanism by which the buffer ends.

Here is how it plays out at the desk level

As days-of-cover prints fall through the summer, physical traders bid the prompt barrel over the deferred — holders of inventory in a visibly depleting system charge more to part with it now — so the backwardation at the front of the Brent curve steepens regardless of where flat price sits; Asian refiners short of Gulf crude keep bidding Atlantic-basin cargoes, lengthening voyages and holding freight and US export volumes at their records; — while the principal plumbing around the closure, Saudi crude rerouted down the East-West pipeline to Red Sea loadings, now runs past a second contested chokepoint, the Houthis having announced on Monday a ban on Israel-linked vessels in those waters — and the structural residue outlasts the conflict: days-of-cover becomes the statistic the market trades, the chokepoint premium threatens to become literal — Tehran is already floating a reopening priced with a toll — storage capacity itself reprices as strategic infrastructure rather than passive plumbing, and term contracts shift toward firm-supply and inventory-commitment clauses, because 2026 will have taught every offtaker that the cushion they assumed was collective was in fact exhaustible.

The signal to watch sits on exactly that channel: verified tanker transits through the strait, weekly, from July — the EIA’s eighty-nine-dollar fourth quarter requires shut-ins nearly halving between the third quarter and the fourth, and if measurable resumption has not appeared in the transit counts by the end of September, that forecast breaks upward, and the front spread will say so before the flat price does.

The gauge has been falling since the last day of February.

The most important producer in the world this year does not drill, cannot replenish itself, and publishes its own decline rate weekly — and the price of the market’s calm is whatever is still left in it.

Subscribe now

Mark Rutte, secretary-general of NATO, had told him in the spring that he deserved all the praise for an increase in allied defence spending that was well underway before he returned to office, and on another occasion likened his diplomacy to the work of a benevolent father.

Gianni Infantino invented a peace prize for football and named Trump its first recipient at a December ceremony in which the trophy itself appeared to be the point.

None of this is foreign policy in any sense the architects of the postwar order would recognise. It is tribute — and tribute, paid to a dominant power by states that need something from it, is the oldest evidence there is that the dominant power still holds the thing they need.

Read the receipts the way an auditor reads them, and the picture inverts.

Stephen Walt, writing in Foreign Affairs this spring, named the strategy behind it “predatory hegemony” — Washington using its privileged position to extract concessions, tribute, and displays of deference from allies and adversaries alike, in what it treats as a purely zero-sum world.

The label is exact.

What the label does not yet price is the asset being spent.

Not the cash in the till.

The goodwill above it.

The store was never the asset

Trump’s own image for American power is a retail one.

The United States, he said in April, is like a big, beautiful department store, and everybody wants a piece of it; the American consumer, a White House statement added, is what every country wants, and to put it plainly, they need our money.

The sentiment contains a category error that runs through the whole strategy.

He believes the store is the asset — the market, the consumer, the dollar, the carrier groups — and that the world owes Washington for the privilege of shopping there.

The store was always for sale.

Tangible book — gross domestic product, the military, the depth of the consumer market — is real, and large, but it was never what gave Washington its disproportionate pull on the behaviour of other states.

What gave it the pull was goodwill, in the precise accounting sense: the premium the world paid, in deference and dollar-holdings and alliance loyalty and acceptance of rules tilted modestly in America’s favour, above the identifiable value of those underlying assets.

Goodwill is the line an acquirer books when a franchise is worth more than the sum of its plant and inventory, because the relationships, the reputation, and the expectation of fair dealing carry a value of their own.

• It sits on the balance sheet as an intangible.

• It is the most valuable thing the United States owns, and it does not appear in any tally of ships or output.

Walt traces the arc by which it accreted:

Eisenhower played hardball over Suez in 1956;

Nixon closed the gold window in 1971.

Both were rough.

Neither was predatory, because both were bounded and aimed, and the franchise kept accreting goodwill through each.

That is the operative distinction — not that benign hegemons never squeeze, but that they squeeze rivals and spare clients, on the understanding that a reputation for not squeezing clients is itself the asset that makes the clients line up at this store rather than build their own.

A predatory acquirer strips it for cash

A predatory acquirer buys a franchise carrying real goodwill and strips it — raises prices, cuts service, breaks the implicit promises about how the business behaves, and books the proceeds as though the accumulated goodwill were a cash balance available for distribution this quarter.

Every instrument of Trump’s second term maps onto that operation with uncomfortable precision.

• The tariffs are the price rises, with the wrinkle that the cost of a tariff is mostly borne by the Americans who buy the imported good rather than by the foreign seller — the store raising prices on its own customers and recording the markup as a win.

• The linkage of military protection to economic demand — the suggestions that Washington might leave NATO, decline to defend Taiwan, or abandon Ukraine unless the terms improve — is the service cut, sold as cost discipline.

• The territorial appetite — the repeated designs on Greenland, the musing about Canada as a fifty-first state, the talk of retaking the Panama Canal zone — and the January operation that seized Venezuela’s president from Caracas and flew him to a New York courtroom are the broken promises about how the franchise conducts itself, the abandonment of the norms that were part of what the goodwill was paying for.

• And the monetisation — the Qatari aircraft, the Trump Organization’s development deals with governments seeking favour, the billions in tokens bought from his cryptocurrency venture around the time a Gulf state secured access to controlled high-end chips — is the insider selling the franchise’s goodwill into his own pocket while the books still carry it at value.

The fealty is the receipt.

The signed deals are real:

It all goes as cash now.

A predatory acquirer’s income statement looks magnificent in precisely the quarter he strips the asset — that is the nature of stripping.

Fifty-seven per cent is the first impairment charge

An auditor deciding whether goodwill must be written down runs a checklist of impairment indicators — the loss of key customers, the franchise no longer commanding its historical premium, a structural deterioration in the cash flows the goodwill was supposed to protect.

The indicators are already on the page.

Start with the dollar, the largest goodwill line Washington owns.

Its share of global foreign-exchange reserves fell below fifty-seven per cent in the third quarter of 2025, the first time since 1995, down from a peak near seventy-two per cent in 2001.

The honest reading matters here, because the move is gradual and partly mechanical: much of the recent slide reflects exchange-rate effects rather than active selling, the renminbi remains under two per cent of reserves and around three per cent of payments against the dollar’s near-half, and no rival currency is positioned to take the franchise whole.

What is happening instead is subtler and, for the conceit, more telling.

Reserves held in currencies outside the dollar, euro, and renminbi have risen to roughly a fifth of the total, now exceeding the euro; central banks bought gold at record pace, more than eight hundred tonnes across 2025; and China’s cross-border payment system now clears on the order of twenty-five trillion dollars a year across well over a hundred countries.

The customers are not leaving the store.

They are quietly building their own loading docks.

That is what early impairment looks like — not a competitor seizing the franchise, but the franchise’s customers reducing their dependence on it at the margin, one contract and one reserve allocation at a time.

China is the indicator the auditor cannot log as immaterial.

It finished 2025 with a record trade surplus near one and a fifth trillion dollars.

Its exports grew even as shipments to the United States fell, with the American share of Chinese exports down to roughly twelve per cent while exports to Southeast Asia and Africa rose by double digits and to the European Union by nearly seven per cent.

The Federal Reserve’s own staff now write openly about a “China shock 2.0.

The American tariff on Chinese goods ran from around twenty per cent up to a punitive peak near a hundred and twenty-seven and then settled near forty-eight after a truce — and the truce is the tell.

Trump cannot squeeze a peer the way he squeezes a client.

When India was hit with a fifty per cent tariff, Narendra Modi flew to China for the summit with Xi Jinping and Vladimir Putin; in December Putin came to Delhi, where Modi called the relationship a North Star and the two set a hundred-billion-dollar trade target for 2030.

India was not aligning with Moscow.

It was reminding Washington that New Delhi has options — which is the single most expensive sentence a goodwill line can hear.

The allies are running the same play more cautiously.

Mark Carney has made doubling Canada’s non-American exports within a decade an explicit goal, signed Ottawa’s first bilateral trade deal with Indonesia, opened talks with the Southeast Asian bloc, and mended fences in Beijing; the European Union has closed deals with Indonesia, Mexico, and the Mercosur countries and was near a pact with India.

None of this is a write-off.

Each is an indicator the auditor logs — and impairment indicators are cumulative, which is the entire danger of them, because no single one triggers the charge and the aggregate eventually does.

Even the instrument of extraction impairs

The sharpest development came after Walt’s essay went to print, and it cuts at the crowbar itself.

On 20 February 2026 the Supreme Court held, six to three in Learning Resources v. Trump, that the International Emergency Economic Powers Act does not authorise the president to impose tariffs at all — striking down both the April 2025 reciprocal tariffs and the fentanyl tariffs on Canada, Mexico, and China, on the plain ground that the power to tax belongs to Congress.

The roughly one hundred and seventy-five billion dollars collected under the voided regime became, overnight, a contingent refund liability with no settled mechanism to discharge it.

Within hours, Trump rebuilt a ten per cent global tariff under Section 122 of the Trade Act and signalled an increase towards fifteen; the administration then leaned on Section 301 investigations due to conclude this summer, and on the second of June a forced-labour probe proposed fresh duties of ten per cent or more on dozens of partners, twelve and a half per cent on China, Japan, India, South Korea, Brazil, and Switzerland among them.

The dissent’s warning that comparable tariffs would simply reappear under other authorities is being borne out. But every replacement is a worse instrument than the one struck down — Section 122 caps the rate, runs only a hundred and fifty days, and was itself struck by the Court of International Trade in May; Section 301 demands a separate investigation per target — and that is the point for the franchise.

The crowbar used to strip the goodwill was confiscated by Washington’s own court, and the improvised substitutes are narrower, shorter-lived, and more contested. When the legal basis of a hegemon’s principal coercive tool is found wanting at home, every counterparty abroad re-rates the durability of any deal signed under it.

The premium was the leverage

Set the loop out plainly, because it is the engine under everything above.

It begins with a belief: that aligning with the United States is safe, that Washington is the reliable, fair-minded power, the counterparty whose word is money-good.

• That belief shapes behaviour — states concentrate their dependence on it, routing supply chains through the dollar, parking reserves in Treasuries, outsourcing their security to the alliance, opening their markets on American terms. And here is the move that makes it reflexive rather than merely causal: that concentration of dependence is the leverage.

• The goodwill is not a sentiment sitting beside Washington’s position; the goodwill is the position.

• The premium the world paid, in deference and dependence, built the very pull Washington now possesses.

• The belief manufactured the asset. Then the asset is used against the people who created it.

• The leverage lets Washington turn predatory; it strips the goodwill for cash; the stripping destroys the belief that made aligning safe; the states that concentrated their dependence begin to diversify it away; and as they do, the dependence unwinds and the leverage evaporates.

• The same conviction that built the pull is the conviction the predation destroys, and the pull dies with it.

• The inflection — the moment the loop turns from self-reinforcing to self-defeating — was the second term.

The first could be priced as a temporary mispricing, an isolated episode, a glitch that would mean-revert once normal service resumed; allies hoped exactly that, and the hope is what kept the goodwill carried at par.

The second term wrote the hope off.

It demonstrated that the predation was the strategy and not the aberration, and a franchise the market re-rates from “temporarily mispriced” to “permanently impaired” does not recover its old multiple on the next good headline.

I argued in these pages, on the jet-fuel break, that confidence is a balance-sheet item; the lesson the molecule taught is the lesson here, one register up.

The velvet glove was not manners.

It was the knowledge that the fist drew its power from the glove, and that a fist used too openly stops being feared and starts being routed around.

The receipts are real, and that is the trap

The administration’s case is not empty, and the strongest version deserves stating without a thumb on the scale. The deals are signed and the concessions banked. Japan and South Korea did lower their exposure; pledges of investment were made; the bilateral goods deficit with China fell by something like a quarter to a third.

One can push the argument further: perhaps the goodwill was overstated all along — perhaps allied deference was always thinner than it looked, the rules-based order American self-interest in a fair-minded costume, and Trump is merely marking a soft asset to a more honest value, converting deference the United States was never going to realise into hard concessions it can bank today.

Grant every word of it, and the impairment logic is untouched, because the trap is in the timing. The cash arrives in the quarter you strip the asset; the impairment lands later, and it lands at once. The receipts are real precisely because the goodwill is still partly on the books — you can only strip what still has value left to strip — and each strip raises this quarter’s receipts while lowering next year’s carrying value, the two halves of a single transaction seen at two dates.

The deficit number is the cleanest illustration of the flattery: it fell, but much of the fall is Chinese goods rerouted through Vietnam and Mexico and relabelled at the final port, not dependence eliminated.

A lower mark that conceals the same exposure is the predatory income statement in miniature.

The receipt improves.

The position does not.

That is the thing visible only now, with the whole ledger on the table: the benevolent move and the predatory move were never two strategies between which Washington chooses. They are the same balance sheet read at two dates.

The benevolent hegemon was accruing goodwill; the predatory hegemon is realising it, and to realise goodwill is, by definition, to convert an asset you could have kept into cash you have already spent.

The error, I think, is not that Trump is too aggressive.

It is that he has misread which line he is drawing against. He believes he is spending down a cash balance the world owes America for access to the store. He is impairing goodwill — and the accounting that governs goodwill, the only accounting reputational capital has ever obeyed, permits no write-back.

Once it is written down, it cannot be marked up again on better news.

A future, more restrained president inherits the impaired value, not the original premium, and must re-earn it from a lower base over a longer horizon.

That is Walt’s “uphill battle to restore Washington’s reputation,” priced and put on the books.

The charge will not announce itself with a headline; an impairment shows up as drift, and there are three readings to watch for it.

1. The dollar's reserve share, adjusted for exchange-rate effects, is the cleanest — a fall on a constant-currency basis, distinct from the mechanical slide already booked, is active reallocation, and active reallocation is the crystallisation of impairment. The fourth-quarter print is the first to test it: the valuation effects that drove almost the whole earlier slide went quiet, and the share fell anyway. The next prints are the tape.

2. Whether the bilateral deficits that “improved” improve on a value-added basis or only on a relabelled-origin one.

3. The Section 301 determinations due this summer and the forced-labour duties floated on the second of June: if the replacements for the struck-down tariff regime keep arriving narrower and more contested than what they replace, the instrument of extraction is impairing in plain sight, on the public record, ahead of the goodwill it was built to strip.

A hegemon can sell its goodwill, or it can hold it.

What it cannot do is sell it and still keep it to spend — and the entry that records the sale is written in one direction only.

Subscribe now

The dollars are safe — whether they are yours is a different question.

A market fenced off from its own country

On 5 May, Zimbabwe moved its dollar-denominated stock exchange out from under the national securities regulator and handed it to a brand-new body created specifically to wall it off from the rest of the economy — and this month, in an essay for the Foreign Policy Research Institute, Charles Ray, a former US ambassador to Harare, asked the obvious question:

Can the Victoria Falls Stock Exchange revitalise an economy it has been deliberately fenced away from? The setup is unusual enough to slow down on.

VFEX is a stock exchange that sits physically inside Zimbabwe and refuses, by statute, to touch the Zimbabwean currency — every trade settles in US dollars, every listing raises offshore money, and the entire purpose of the structure is to give investors a place to stand that is legally separate from the country it is standing in.

• Ray’s verdict is cautiously optimistic, conditional on credibility, predictable rules, and deeper liquidity.

• Mine is narrower, and it turns on a single distinction that anyone who has worked at a settlement desk learns early on.

Money has two addresses

An asset’s physical location and its legal home are two separate facts, and the gap between them is where half of finance actually lives — a bond can sit in a depository in one jurisdiction while the law that governs it lives in another, custody is one question and control is a second, and the people who confuse the two are the ones who get hurt on the day the two pull apart.

VFEX is the cleanest example I have seen in years of a structure built entirely inside that gap: capital that physically sits in Zimbabwe while being treated, for every purpose that matters to an investor, as though it does not.

That is what a bonded warehouse is.

Goods enter the country, sit under seal, and the customs authority agrees to treat them as if they have not yet crossed the frontier — no duty, no tax, no domestic claim — until the day the owner chooses to clear them out.

A warehouse, not a market

Zimbabwe built the thing in August 2020, weeks after suspending trading on the Zimbabwe Stock Exchange over allegations that share transactions were feeding currency instability — and the timing is the whole story, because the warehouse was opened in the same season the authorities reached into the existing market and switched it off.

Statutory Instrument 196 of 2020, published in the Zimbabwe Government Gazette on 14 August 2020, set the terms:

• Securities listed in Victoria Falls trade and settle in US dollars or convertible currency, capital must arrive from offshore or from free funds, and proceeds may be held in approved accounts at internationally recognised banks. Dollars in, dollars out, dollars in between — the currency that wrecked three decades of Zimbabwean savings is barred at the door.

• The warehouse was built for a particular tenant — exporters, miners, tourism operators, firms whose costs and revenues are already dollar-linked and for whom raising equity in a collapsing local currency means watching the proceeds evaporate in real terms before the project breaks ground.

• Settling in dollars aligns the financing currency with the investment currency, and that alignment is the genuine economic logic underneath the politics.

The incentives are the warehouse’s rate card: a five per cent dividend withholding tax, exemption from capital gains withholding for foreign investors, reduced exchange-control friction, and trading costs below those of the Harare board it sits beneath — VFEX is a subsidiary of ZSE Holdings, in which the government holds a controlling 32 per cent.

Hold that last number for a second.

The body that owns the building is the state, and the 5 May reshuffle moved the warehouse out from under the Securities and Exchange Commission and placed it under a new Victoria Falls International Financial Services Centre — Statutory Instruments 62 and 63 — which Ray reads, fairly, as Harare trying to build a firewall between the enclave and the chaos next door.

Kardamow’s logo is a Goya portrait of a boy standing between a caged bird and three cats that have not yet moved — the painter’s study of the frail boundary between safety and the thing waiting on the other side of it.

• A bonded warehouse is that boundary written into statute.

• The cats are the currency regime.

• The seal on the door is an instrument the same government can rewrite.

Every parcel has cleared on time

Give the warehouse its due, because the operational record is real and it is the strongest argument its managers have:

Every trade since inception has settled on time, market capitalisation has climbed to roughly $1.65 billion, up nearly 29 per cent over 2025, and Econet InfraCo’s listing this March added close to a billion dollars on its own.

Caledonia raised twelve million, Karo Resources thirty-six — twenty-five of it from local investors. Turnover more than doubled between 2023 and 2024, from $26.3 million to $57 million.

The custody works.

The parcels arrive, they are logged, they are kept, and nobody has lost one.

But on-time settlement is custody, and custody was never the binding risk.

CLEARANCE is.

The question that decides whether VFEX is an asset class or a holding pen is whether you can get your dollars across the frontier when you want them — repatriation on demand, at par, without a discretionary delay dressed up as a liquidity measure — and that is precisely the thing a settlement record cannot prove, because it has never been tested by a government that needed the dollars more than it needed the promise.

Then look at what is actually in the warehouse:

A single listing this March, roughly the size of half the entire board, a handful of mining houses, an exchange Ray himself describes as fifteen names in a frontier market. Turnover is throughput, not depth — a warehouse can be busy without being liquid — and what VFEX holds is closer to a strongroom stacked with two or three oversized crates than a market with many buyers and sellers crowding a narrow spread.

The keyholder ran the fire once

Here is the structural problem, and it is the reason a publication built on reflexivity exists to trace structures like this one:

The warehouse protects investors from Zimbabwe’s currency discretion, and the warehouse is operated by the authority that holds that discretion.

The protection from the arsonist is a promise made by the arsonist.

SI 196 was created by a statutory instrument; it can be amended by another one, on a morning, with no market consultation — and the precedent that proves the state will reach into a functioning market when it decides the market threatens the currency is the 2020 ZSE suspension that created VFEX in the first place.

The disease and the cure share one point of failure.

This is the loop, and naming it adds weight rather than decoration.

The warehouse’s value rests on the belief that the seal holds; the seal holds because confidence stays high enough that Harare gains more from honouring it than from breaking it; and that confidence is itself produced by the belief that the seal holds.

The structure and the belief manufacture each other.

Knock out the belief — one redenomination, one “temporary” repatriation freeze, one instrument issued in a fiscal emergency — and the bond is just ink, the dollars are just dollars sitting in Zimbabwe, and the frontier that was supposed to be legal turns out to have been imaginary all along.

The frontier is supposed to disappear

Now the genuinely strange part, the one Ray’s essay gestures at without quite drawing out:

Zimbabwe is succeeding at the thing that would make the warehouse pointless. The ZiG, gold-backed and held to single-digit inflation — 4.1 per cent in January, the first reading below ten per cent since 1997 — sits under a ten-month IMF staff-monitored programme, with reserves built from $276 million in April 2024 to $1.2 billion by December, and a stated goal of becoming the country’s sole currency by 2030.

A bonded warehouse exists because the frontier outside it cannot be trusted.

If the ZiG project works, the frontier becomes trustworthy and the warehouse loses its reason to exist; if it fails, the seal is only ever as good as the same authorities’ restraint. There is no version of the next five years in which the warehouse’s rationale is stable.

The American dollars that would fill it are held back by a different lock entirely.

ZIDERA — the 2001 sanctions law — is on course to be repealed by H.R. 5300, introduced in the House in September 2024 and still awaiting a Senate vote as of this writing — its repeal arrives chained to a condition: Zimbabwe must remit the Global Compensation Deed arrears — an inflation-adjusted $3.5 billion, the figure the Zimbabwean government itself cited in the 2020 settlement framework — in hard currency, within twelve months, on a country carrying more than $14 billion in external debt.

Ray’s advice to American investors is the right frame — treat VFEX as a governance and operations problem, not an asset-selection one — and Washington’s interest is real, with China entrenched as Africa’s largest creditor and the US hunting for commercial footholds on the continent.

The warehouse can have a flawless settlement record, a firewall, and a brand-new financial-services centre, and still stand half-empty because the largest pool of compliant dollars in the world is told to stay on the other side of the wall.

Custody is not control

Strip away the conceit, and Ray is right about the thing that matters most:

VFEX is a serious institutional attempt to solve a genuine constraint — the near-impossibility of raising long-term capital in a currency nobody trusts — and it has executed that attempt with a clean settlement record, real listings, and a governance structure that, on paper, fences the enclave off from the storm.

That is custody, and the custody is excellent.

1. But the binding question was never whether the dollars are safely held;

It is whether they are freely yours — whether you can clear them across the frontier on the day you want them, whether the rulebook that protects them survives the next fiscal emergency, whether the authority that wrote the seal can be trusted not to break it.

2. Those are questions of control, and control sits exactly where it always sat:

With a government that has already shown, once, that it will reach into a working market the moment it decides the market threatens the currency.

VFEX can become a credible channel for hard-currency capital — but only to the precise degree that Harare chooses, day after day, to make it one, and a choice is not the same thing as a protection.

Every trade has settled on time.

That was never the question.

The question is, in the afternoon, someone in Harare decides the dollars inside the warehouse are worth more when opened than when sealed, and discovers it takes one statutory instrument to find out.

Subscribe now

The Iran ceasefire holds on paper.

The Strait of Hormuz holds open in practice — but only because the US Navy resumed escort duty through the channel two weeks ago after a missile-launch incident off Bandar Abbas, and only because Tehran’s tactical interest in showing the channel can be navigated outweighs, for now, its strategic interest in showing the channel can be closed at will.

Tanker traffic remains roughly 60 per cent of pre-war volume. Qatari LNG is still routing Cape of Good Hope, adding fourteen days and a hard premium to every cargo that would otherwise transit the strait.

European TTF gas closed yesterday at €42 per megawatt-hour, still nearly double the 2019 baseline and still pricing a continent that no longer treats Russian pipeline supply as an option.

None of this is normal.

All of it is what “normal” looks like fourteen weeks into a contested ceasefire that nobody fully trusts.

I wrote here ten days ago about jet fuel as the cleanest read on the Hormuz shock — the product that broke first, the inventory that does not reverse on a price signal, the substitution architecture that worked in spring partly because it took the supplier set by surprise.

This week I want to sit on the other side of the same story, from a Berlin desk, because everything Germany has built since 2022 to escape gas dependency runs through the assumption that the Hormuz–ceasefire distinction — open in fact, fragile in price — continues to hold, and the architecture Germany has built since then to escape from gas is now, by deliberate policy choice as of February 26 this year, on a trajectory that will re-anchor it back to gas.

The Institute for Energy Economics and Financial Analysis published two reports this month, one on Germany’s continued gas dependency, one on the hydrogen strategy that was supposed to fix it.

Read together — and they have to be read together, the policy logic only makes sense as a single loop — they describe a country that has spent four years and somewhere north of €70 billion of committed public exposure to leave gas behind, and is on track to spend the next twenty years and another €45 billion arriving back at gas through a different door. Roughly €1,000 per German taxpayer, on a path the architecture itself now defends.

This piece is about that round trip.

The financial instrument that holds it together — Germany’s hydrogen amortisation account — is the part nobody outside the trade press is yet pricing properly.

So let’s walk the loop, starting where Germany actually still is: importing gas.

Two years on, still buying gas

Germany’s pre-2022 gas position was the densest pipeline network in Europe — 55 bcm a year of Russian pipeline supply landing through Nord Stream 1, Yamal-Europe, and the Ukraine transit corridor, alongside Norwegian, Dutch, and domestic production — and the entire infrastructure investment case underneath that position assumed Russian molecule continuity over a multi-decade horizon.

The invasion broke the assumption and forced the fastest infrastructure deployment of any major European economy: five floating storage and regasification units commissioned in twenty-four months, long-term LNG offtake agreements signed with US, Qatari, and Mozambican suppliers stretching twenty years out, state-owned Deutsche Energy Terminal stood up to operate the federal terminals, and roughly €17 billion of public exposure committed across the buildout — according to Environmental Action Germany and the Center for American Progress, once you net out the charters, the fixed works contracts, and the loss-retention floors negotiated with European Commission state-aid clearance.

The speed was real — The diversification was not!

Germany imported 22.7 bcm of LNG in 2025 — about 23 per cent of total gas consumption — and 92 per cent of that LNG came from the United States, up from 82 per cent in 2023, on a trajectory that is the opposite of diversification: the share of the dominant counterparty has grown in every reporting period since 2022, and the cargo classes outside the US share — Qatari, Mozambican, Trinidadian — have shrunk in both absolute and relative terms even as total import volumes have tripled.

US LNG is the most expensive cargo class for European buyers — Henry Hub plus liquefaction plus shipping plus regas plus terminal fee — and the all-in delivered-ex-ship cost into northwest Europe sits structurally above piped Norwegian supply and Qatari long-haul, a gap that has widened, not narrowed, as the Iran war has elevated insurance and freight premiums on every non-American cargo class while leaving the transatlantic corridor largely untouched.

Germany is, in 2026, paying a premium for the privilege of being mono-source on a transatlantic supply chain that runs through US export terminal capacity, US politics, US pipeline economics, and the willingness of US producers to keep drilling at a Henry Hub strip the market has been short for two years.

There is a sentence in the IEEFA gas report that summarises the position with unusual precision.

Germany may have simply replaced one dependency with another.

It is the kind of line you write when the diplomatic costs of saying so explicitly are no longer worth the politeness of avoiding it.

Read that against the geography.

The 8 per cent of German LNG that is not American — Qatari, Mozambican, Trinidadian — comes either through Hormuz (Qatar) or via shipping lanes whose insurance premiums are still anchored to the Hormuz risk surface, and Qatar’s seaborne molecule has been substantially rerouted since the Iran war began, which means that if the ceasefire breaks and Hormuz tightens further, Germany does not get more options — it gets fewer, because the marginal cargo that disappears is the cargo from the supplier set Germany has been quietly thinning, not the cargo from the supplier set Germany has been quietly thickening.

The diversification cushion is structural fiction.

The mono-source is the structural reality.

Thirty-six per cent is a number

The infrastructure that absorbs the gas Germany imports is materially overbuilt against the gas Germany actually uses, and the overbuild is becoming harder to disguise as time passes and capacity utilisation refuses to climb toward the levels the project economics were sized for.

Average utilisation of German LNG terminals in 2025 was 36.3 per cent, against an EU-wide average of 50.8 per cent — a fifteen-percentage-point gap that puts Germany among the worst-performing terminal portfolios in the European peer set. Two FSRUs have already been redeployed elsewhere in Europe — Mukran lost its second FSRU in early 2025 to demand and pricing competition from western North Sea terminals, ReGas cancelled the Energos Power contract citing low utilisation, DET sits on €4.06 billion of EU-cleared state aid to absorb operator losses with a further €900 million of headroom approved — and none of those terminals is closing.

All of them carry fixed payment obligations that the public balance sheet now backstops, all of them count toward the import capacity Berlin will cite when the next political conversation about energy security takes place, and all of them are, by 2025 utilisation, doing roughly a third of what they were contracted to do — an imbalance that is structural rather than transitional, because the demand collapse on the consumption side is being driven by sectoral shifts that are not going to unwind even if gas prices retreat further.

The reason is straightforward.

Gas demand inside Germany has collapsed, and the collapse is structural enough that it is unlikely to reverse in the near term even at lower prices.

From a 2019 baseline of roughly 91.5 billion cubic metres per year, total German gas consumption fell to around 73 bcm in 2024 — a 20 per cent drop concentrated in the sectors where electrification or efficiency could substitute for the molecule.

• Industry demand fell 23 per cent as energy-intensive production migrated, as process redesign captured efficiency at the margin, and as the most exposed plants — fertilisers, chemicals, glass — curtailed runs or wrote off capacity entirely.

• Household demand fell 23 per cent as heat pumps displaced boiler replacements, as building-stock retrofits accelerated, and as the behavioural changes from the 2022 price shock proved stickier than anyone expected.

• Power generation demand fell 19 per cent, commercial demand 24 per cent, and the pattern is consistent across every sector where electrification or efficiency has been allowed to compete.

The collapse in demand was the genuine policy success of the post-2022 period.

The infrastructure deployment was sized as if it would not happen.

The twenty-year lock-in

Underneath the operational utilisation problem sits a contractual problem that is harder to walk away from, harder to defer, and harder to explain to a public that thinks the LNG terminals it paid for were a security investment rather than a structural commitment to a supply chain.

By 2030, German utilities — RWE, EnBW, Uniper, SEFE, BASF — have firm contracted LNG offtake volumes of roughly 26.2 billion cubic metres a year.

That is nearly three times the actual LNG that arrived at German terminals in 2025, and it is being layered on a domestic demand trajectory that has been declining since 2019 and is forecast — even by the German government’s own central case — to keep declining through the decade as the electrification trajectory we just looked at compounds.

• The contracts run 15 to 20 years.

• Most are take-or-pay or take-and-pay structures with limited destination flexibility.

This is the architecture that decides Germany’s gas position through to 2045, and it does so in a way that is essentially indifferent to what happens to domestic demand, because the contractual obligation to lift the cargoes does not care what the inland market looks like — the lifters can resell into European or Asian markets if Germany itself does not need the molecules, that’s the operational logic that is already keeping Mukran open even as German terminal utilisation falls, but the contractual exposure to the molecule, and to the price formula underneath the contract, stays with the German counterparty for the contract life regardless of where the cargo ends up.

The architecture is doing two things simultaneously, and both work against the strategic direction Berlin says it wants.

1. It is holding gas in the German energy mix at a structurally higher floor than the underlying demand profile would justify — not because demand wants gas, but because gas wants demand, and contracts that have been paid for find a way to put their molecules to work.

2. And it is creating an institutional constituency — the contract counterparties, the network operators, the regulator, the federal ministry that signed off, the unions in the gas value chain — with a direct fiscal stake in defending gas demand wherever it can be found: power generation, industrial heat, hydrogen reforming, eventually.

Gas finds a use because gas has been paid for.

This is the substrate. Now the hydrogen layer sits on top of it.

What Berlin got right by accident

Before the architecture, the data point that is doing the most quiet work in Germany right now → HEAT PUMPS

Between 2022 and 2025, German households installed roughly 1.1 million residential heat pumps, the 2025 sales figure crossed the gas boiler line for the first time at 47 per cent versus 44.7 per cent, and the Q1 2026 European heat pump market grew 25 per cent year-on-year with Germany leading the volume — pulling the European total to 575,000 units in three months and pulling gas boiler sales to their lowest quarterly run-rate in fifteen years.

IEEFA published a separate study in March quantifying the displacement effect — without those heat pump deployments, German LNG imports between 2023 and 2025 would have been 16 per cent higher, costing an additional €1.3 billion — and the savings replicate in subsequent years, because replacing a gas boiler with a heat pump produces compounded savings across the equipment lifetime rather than a one-off displacement in the year of installation.

The heat pump deployment is, on the numbers, the single largest contribution to German gas demand destruction over the period. It is also the policy intervention that was least championed and least funded — the previous coalition’s GEG buildings law set a 500,000-unit annual target that produced political backlash and was partly walked back, the market hit 300,000 units in 2024 before recovering in 2026, and the recovery happened not because federal subsidy machinery worked but because the underlying electrification economics had shifted enough that heat pumps were winning on running cost regardless of incentive.

The point is not that heat pumps solved the problem.

But the largest single piece of progress on German energy security between 2022 and 2026 came from a direct-electrification technology — one that needs no new gas, no new hydrogen, no new pipelines, and headlined none of the strategies the government wrote during the period — and the gap between where the policy attention sat and where the demand destruction actually came from is the cleanest single piece of evidence available that the policy framework is optimising for the wrong target.

Hold that for a second.

It matters for what’s coming next.

The exit that was supposed to work

Hydrogen was framed as the exit.

The original 2020 strategy positioned green hydrogen as the molecule that would let German industry decarbonise without de-industrialising — steel, chemicals, refining, heavy transport, eventually power generation in the dunkelflaute gaps that renewables cannot fill on their own — and the 2023 update raised the 2030 electrolyser target from 5 to 10 gigawatts on industrial demand projections that have, in the intervening eighteen months, been systematically revised downward by both the industry that was supposed to buy the molecule and the ministries that were supposed to deliver it.

The hydrogen core network — 9,040 kilometres of pipe linking domestic production and import terminals to industrial consumers by 2032 — got formal regulatory clearance in October 2024 at a headline construction cost of €18.9 billion, the European Commission cleared €3 billion of German state aid in July 2024, and the architecture looked from the outside like a textbook industrial policy package, anchored by EU climate targets, financed by a state-backed credit mechanism, validated by Fraunhofer modelling, signed off by every major TSO and utility in the country.

The architecture was sized for an end state.

By 2045, hydrogen was supposed to serve roughly 360 to 500 terawatt-hours of annual demand under the federal government’s 2024 system development strategy.

• The economy ministry’s normative scenarios go higher, up to 605 TWh.

• The exploratory scenarios go lower, down to 71 TWh.

Pick a number; the official ranges span almost an order of magnitude, which is itself the most honest admission the official documentation contains, because an order-of-magnitude scenario range is not really a forecast — it is a confession that the modelling cannot bound the answer to within a useful margin.

The IEEFA report does the work of pricing the demand sectorally rather than just accepting the official ranges, and the sectoral work tells you which way the gap closes.

• Heating is gone — heat pumps win, decisively.

• Most of road transport is gone — battery electric beats hydrogen on cars, light commercials, increasingly on long-haul, and the residual aviation and shipping cases mostly run on derivatives (e-methanol, e-ammonia, e-fuels) rather than on hydrogen gas.

• Low-temperature industrial heat is gone — direct electrification wins on capex and on operating cost.

• Power generation is partly gone — the dunkelflaute problem has solutions other than burning hydrogen at 60 per cent efficiency in a turbine, and long-duration storage, additional interconnection, and demand response all beat hydrogen-fired power on dispatch economics whenever they are allowed to compete.

What remains is high-temperature industrial heat, the existing 60-TWh-per-year grey hydrogen pool already used in refining, ammonia, and methanol, and a residual contribution to power generation that the official scenarios overweight.

That residual is real.

It is also, on the cleanest accounting available, well below the bottom of the official scenario range — IEEFA’s estimate of 2045 demand sits at roughly 170 TWh, at or below the low end of every official band.

And the cost question lands on top of the demand question with no relief.

Even where domestic hydrogen demand survives the sectoral analysis, German production costs sit structurally above the import alternatives — RFNBO-compliant green hydrogen produced in Germany costs more than the equivalent molecule produced in Iberia, North Africa, or the Gulf and then shipped as ammonia or methanol into German terminals — and the cost gap is wide enough that, on any honest reading of the project economics, the system-level case for domestic electrolyser scale-up rests on a strategic-industry argument rather than a competitive-economics one.

The core network was sized for the upper end of demand.

The cost analysis says the upper end is the part most likely to be lost to imports.

This is the gap the financial architecture is now absorbing.

The account is the architecture

Germany’s hydrogen Amortisationskonto — the amortisation account — is genuinely novel financial engineering, and it has to be understood as financial engineering rather than as a subsidy, because the mechanism is not a transfer to the network operators but a regulated credit facility with a defined repayment path and a defined backstop.

The mechanics?

TSOs build the network privately, network fees recover costs over the long run, but early demand is too low to support cost-recovering fees, and high fees would suppress the very demand the network needs to grow — so the regulator caps the fee at €25 per kilowatt of booked capacity per year (set by BNetzA on 14 July 2025), and the difference between the capped fee and the true cost-recovering fee accumulates each year on a credit balance. KfW, the state development bank, funds the balance through a €24 billion credit line to H2 Amortisationskonto GmbH, a special purpose entity owned by the network operators via Trading Hub Europe, and the first payment from the account to the TSOs went out on 25 March 2025 at roughly €172 million, with the next scheduled for March 2026.

• The account has a closing date and a clearing rule, and both are statutory.

• The account must be settled by 31 December 2055.

If a residual balance remains on the closing date, the federal government covers at least 76 per cent of it. The TSOs absorb the remaining 24 per cent.

That is the architecture in a paragraph.

Now the implications.

The architecture is broadly defensible if — and only if — hydrogen achieves system-wide relevance in the German energy mix by roughly 2040. That is the conditional the architecture rests on, and that is the conditional everything else in this piece is about.

Fraunhofer’s modelling for BNetzA, which produced the €25 fee cap, embeds in the most ambitious scenario about 30 gigawatts of hydrogen turbines plus another 11 gigawatts of hydrogen-fired CHP by 2040 — a fleet that exceeds Germany’s entire current gas-fired power capacity, and that the regulator’s central case requires to converge from the mid-2030s onward — and the Fraunhofer report contains one sentence the institute did not have to put on paper but did: long-term optimism is required for the functioning of the core network financing.

That is the institute that built the regulator’s model telling the regulator, in 2025, that the model works only if the demand arrives — and the same regulator’s monitoring report would concede, in September of the same year, that there is “hardly any market demand” for hydrogen. The Federal Court of Auditors would conclude in October that the strategy has “fallen behind plan” and that risks to public finances are material enough to require fundamental revision.

The institutional admission is on paper, and the architecture is unchanged.

One variable. It’s breaking.

Inside the €50-billion-plus regulated cost base, against the €24 billion credit ceiling, against the €25 fee cap, against the 2055 wall, there is exactly one variable that decides who pays the bill.

Utilisation

The variable is breaking on both sides.

On the supply side, Germany targeted 10 gigawatts of domestic electrolyser capacity by 2030, the ministry’s own monitoring report calls that figure “hardly achievable” given the current project pipeline, about 1.2 to 1.3 gigawatts of projects have reached final investment decision, and the German Energy Agency reckons fewer than 5 gigawatts has a high-to-medium probability of being delivered by 2030. The international supply pipeline is thinner still — Equinor cancelled its Norway-to-Germany blue hydrogen export project in September 2024 citing absence of long-term commitments from European buyers, RWE pulled out of the $10 billion Hyphen green hydrogen project in Namibia in September 2025 citing the same reason, and these are not edge cases but the companies with the deepest pockets and the deepest strategic interest, exiting projects they had publicly championed eighteen months earlier.

On the demand side, the official scenario range is widening downward. The economy ministry’s September 2025 monitoring report widened the 2030 demand range to 71 to 262 TWh from a previous floor of 95 to 130 TWh, the Federal Court of Auditors used the demand uncertainty as a primary argument for revising the strategy, and the Fraunhofer modelling that supports the regulator’s central case is now bracketed by official scenarios that no longer support it — the polite way of saying the central case has fallen below the floor of its own predecessor.

Run the scenarios through the account.

• Under the Fraunhofer reference case — moderate uptake, demand reaching 70 to 90 per cent of network capacity by 2041 — the account peaks around 2040 at roughly €25 billion and clears to zero by 2055 by design.

• Under an IEEFA gradual case — utilisation rising steadily but reaching only 75 per cent by 2050 — the account ends 2055 with a €14.6 billion residual balance, of which the state covers €11.1 billion.

• Under an IEEFA limited case — utilisation peaks at around 20 per cent of network capacity by 2037 and stagnates — the account ends 2055 with a €45.7 billion residual balance, of which the state covers €34.7 billion.

That is the architecture working as designed.

The architecture was designed to socialise utilisation risk during the ramp-up phase, the ramp-up is not happening, and the risk is therefore socialising — exactly as the architecture’s structural logic dictates it should, exactly along the timeline the legislators who passed it would have wanted to defer if they had been forced to confront it explicitly.

The lever that stops being a lever

The regulator has one lever left in theory — raise the fee cap — and in practice the lever has a marketability ceiling that converts it from a regulatory instrument into a self-defeating mechanism above a certain price point.

Fraunhofer identifies €35 per kilowatt as the level above which the ramp-up fee “is no longer marketable” — above that price, users defer conversion decisions, cancel bookings, exit the network entirely, and the fee feeds back into the utilisation it was meant to recover from — which is to say, the lever stops being a lever above a certain price point because the lever is itself one of the determinants of the variable it is trying to control, and that feedback effect is structural rather than psychological, it is built into the unit economics of every potential user weighing a conversion decision against a counterfactual that no longer requires hydrogen.

To clear the account by 2055 under the IEEFA gradual case requires a fee of around €34 per kilowatt — just below the marketability threshold, with no margin for error. Under the IEEFA limited case, the fee required exceeds €100 per kilowatt — three times the level at which the market exits.

This is the sentence I would put on the wall.

Delayed utilisation is not self-correcting.

Once the revenue shortfalls accumulate in the account, subsequent improvements in throughput are insufficient to unwind the accumulated balances, because the balances grow on financing charges faster than the regulator can recover them through any politically feasible tariff. The 76 per cent state share is the residual.

Migration

Stack the exposures.

• €34.7 billion of unrecovered hydrogen pipeline costs under the limited case.

• €5.1 billion of extended LNG support — the terminal economics that get worse if hydrogen does not arrive to displace the molecule.

• €4.8 billion of capacity payments for hydrogen-ready gas plants that operate longer on natural gas than the policy envelope assumed.

Roughly €45 billion of additional public funding requirement under the IEEFA limited case versus the Fraunhofer reference case.

About €1,000 per German taxpayer.

The number is not the most important thing about the number.

The architecture is.

Note what happens across the scenarios, because the pattern is the most important diagnostic in the entire piece.

Total system cost — taxpayer plus consumer plus private-capital contributions — barely moves: about €100 billion under rapid uptake, about €116 billion under limited uptake, a 16 per cent total cost difference across an order-of-magnitude demand range, because most of the system cost is fixed asset construction and financing that does not scale down with utilisation.

What changes is the distribution.

• Private capital exposure is capped at the state-aid retention floors negotiated with the European Commission when the architecture was approved.

• Consumer exposure compresses as low utilisation means fewer paying users absorbing the cost base.

• Taxpayer exposure expands to absorb the residual, because the residual has nowhere else to go once the contracts have been signed and the construction has been committed.

This is not an architecture failure.

This is architecture working as designed!

The architecture was the bargain Germany struck with the TSOs and the state-aid framework to get the infrastructure committed to in the first place — and the price of that bargain is the public balance sheet absorbing the demand risk, exactly as the bargain specified, exactly along the timeline the bargain made probable.

February the twenty-sixth

On 26 February this year, the Bundestag passed the Hydrogen Acceleration Act.

The bill had been delayed for months under intense lobbying from utilities, industry associations, and the regional state governments whose CCS opt-outs the legislation would override, and the final text did two things that matter for everything in this piece.

• First, it designated hydrogen infrastructure — pipelines, electrolysers, import terminals, storage, ammonia crackers — as being of “overriding public interest,” a legal-procedural status that compresses environmental permitting, expedites planning approvals, narrows judicial review, and removes the federal-state opt-outs that had blocked CCS pilots in Schleswig-Holstein and Brandenburg under the previous Storage Act. The status is the same one that was used to expand offshore wind through the 2023 Renewables Acceleration legislation — and the legislative lineage is deliberate, signalling that Berlin is treating hydrogen infrastructure with the same procedural urgency it treated the renewables build-out.

• Second, and this is the point that changes the architecture, the Bundestag’s economy and energy committee added a clause expanding the bill to include blue hydrogen. Blue hydrogen — produced from natural gas with carbon capture and storage — was designated of equal status with green hydrogen for permitting purposes, which means reforming capacity plus CCS plus CO₂ transport plus geological storage all now sit inside the “overriding public interest” envelope, alongside electrolysers and pipelines.

The legislation cleared a path.

It did not solve a problem.

Blue hydrogen is cheaper than green hydrogen today — roughly €4 per kilogram in Europe against more than €6 per kilogram for green, against a €3 grey hydrogen baseline that industry already pays for current production from natural gas without capture — but blue hydrogen requires CCS at commercial scale on gas-fired reforming, which Europe does not have today. Only five operational CCS projects on the continent, capturing roughly 2.7 million tonnes of CO₂ per year between them, none of them attached to gas-fired power generation.

The technology readiness level for CCS on gas-fired power generation is 7 out of 11 on the IEA scale — pre-commercial demonstration. Capture costs sit around €150 per tonne all-in against an EU Emissions Trading System price in the €70 to €80 range, which means every tonne captured is loss-making at current carbon prices and requires either ETS reform or direct subsidy to break even. And capture rates from international reference projects — Sleipner, Snøhvit, Northern Lights, Porthos — have all underperformed advertised levels in operational practice.

The economics do not work yet at the scale Germany is now asking them to.

The legislation works — The economics get fixed politically.

Blue means gas

Walk the next round of the loop.

Germany commits to blue hydrogen to fill the pipeline that green cannot fill on the timeline the financial architecture requires, blue hydrogen production locks in long-dated gas demand — every kilogram of blue hydrogen requires roughly 3.3 kilograms of natural gas as feedstock plus another kilogram-equivalent for the reforming and capture energy — and the CCS infrastructure gets built at very large public expense on the assumption of cheap long-term gas supply and high carbon capture rates, both of which the global track record makes unlikely.

The power generation backbone underneath the strategy gives you a quantitative read on the size of the new gas commitment, and the read is sobering. The federal cabinet’s November 2025 decision committed to 10 gigawatts of new hydrogen-ready gas-fired power capacity (revised down from the 20-gigawatt proposal the previous coalition floated in early 2025), gated to come online from 2030, and the premise of the “hydrogen-ready” framing is that the new plants burn natural gas for the first decade and convert to hydrogen as supply becomes available — which means that on the cabinet’s own timeline, the new build will burn 10 gigawatts of additional gas through the 2030s on top of the existing 32-gigawatt gas-fired fleet that has not yet been retired.

That long-term gas supply runs back through American LNG, through the lock-in contracts that already commit Germany to 26.2 bcm of US-dominated offtake by 2030, through the Hormuz risk surface for the non-American 8 per cent — and the loop closes here, on this page, where the second floor of the trap rests directly on the first floor that was supposed to be the exit from the first floor.

The hydrogen pivot that was supposed to be the exit from gas dependency rebuilds itself, layer by layer, into a new gas dependency — better-funded, more politically protected, more difficult to exit, and now permanently embedded in the financial architecture that defines who pays the bill.

The IEEFA report frames the endpoint cleanly and I think the framing is correct.

BDEW — the German energy association — proposed on 3 March consolidating the country’s patchwork of hydrogen support mechanisms into a single legislative framework, a Hydrogen Act, an H2G.

Framed as regulatory certainty.

Functionally, the legislative embedding of the assumption that hydrogen requires sustained policy support across the entire system for the foreseeable future — and once that embedding is in place, the question shifts from whether to deploy hydrogen at scale to how to coordinate the incentives required to sustain the deployment.

The policy question becomes the financial structure becomes the political settlement.

The settlement is the trap.

This is what reflexivity looks like in a regulated infrastructure programme.

The institutional substrate

The constituency layer is worth pulling out and pricing on its own, because the architectural rigidity I have been describing does not just live in pipe and contract — it lives in the corporate concentration of who owns the gas position and in the disclosure frameworks that govern how those owners justify their capital allocation to their shareholders, their regulators, and their European supervisors. Germany’s five largest power generators — RWE, LEAG, EnBW, STEAG, Uniper — account for nearly half of the country’s coal and gas-fired capacity, and EnBW alone dominates the upcoming gas-fired buildout, which means the policy framework’s defence of the gas position is, in practice, the policy framework’s defence of a handful of utility balance sheets that carry material political weight in the Bundestag and material fiscal exposure to the hydrogen-demand path that is not arriving

EU taxonomy reporting is the cleanest read on the same lock-in from the disclosure side. Across the same utility set, the share of taxonomy-aligned revenue from fossil-gas-linked activities remains material, the share of taxonomy-aligned capital expenditure being directed toward gas infrastructure — including “hydrogen-ready” gas plants counted as transitional activities — has grown rather than shrunk across the last two reporting cycles, and the decarbonisation lever-sets the same utilities publish in their transition plans depend, almost without exception, on the hydrogen demand path the rest of this piece has shown is not arriving

When utility decarbonisation plans depend on a hydrogen demand path the regulator’s own monitoring report says is not arriving, those plans become a forward-dated contingent liability — not for the utilities, which can adjust their books in line with reality, but for the federal government, which is the underwriter of the architectural assumption the plans rest on.

• The amortisation account is the financial expression of that underwriting.

• The institutional substrate is the political infrastructure that ensures the underwriting will not be quietly unwound.

The loop closes at Hormuz

Walk all the way back to where this piece started.

• Brent at $99.

• Qatari LNG still routing the long way.

• The Iran ceasefire fragile.

• European TTF still elevated.

• Germany 92 per cent dependent on US LNG for its waterborne supply, 26.2 bcm contractually committed through 2030.

Now layer in the blue hydrogen pathway.

• Add ten gigawatts of new hydrogen-ready gas-fired power capacity, gated to come online from 2030.

• Add a national CO₂ pipeline network financed under the November 2025 Carbon Dioxide Storage and Transport Act.

• Add reforming capacity with CCS attached. Add €6 billion of 2026 funding earmarked under the Reiche package for electrolysers, CCfDs, and CCS support.

• Add the H2G framework consolidating the assumption of permanent policy support across the system.

What does Germany’s structural exposure to a future Hormuz event look like?

Worse than it did in 2022 — measurably, demonstrably, and in ways the architecture itself is now actively defending. The pipeline gas dependence on Russia has been replaced by a more complex, more institutionally entrenched, more fiscally backstopped gas dependence — partially on the US LNG corridor, partially on the global gas market through the blue hydrogen pathway, partially on the carbon capture infrastructure that allows the blue pathway to claim climate compliance, and partially on the regulatory architecture that ensures all three exposures stay on the public balance sheet rather than the private one.

• Every layer of the exposure is now public.

• Every layer is now defended by a constituency that has a fiscal stake in defending it.

• Every layer makes the next exit costlier than the last one would have been.

The next Hormuz event, or the next Suez event, or the next North Sea event, will land on a Germany that has, by deliberate policy, increased rather than reduced its structural exposure to gas-market shocks — through the architecture it built to do the opposite.

The reflexivity framework names this.

Imperfect understanding shapes commitment, commitment shapes the fundamentals the next round of perception will have to handle, and the participants act on the next-round perception they themselves have shaped — not because they are confused about which round they are in, but because the rationality of each next move is decided by the commitments the previous round has already made unrecoverable.

• The amortisation account is the financial expression of that loop.

• The Hydrogen Acceleration Act is the legislative expression.

• The blue hydrogen pivot is the operational expression.

They are not separate decisions — they are the same decision, repeated.

What Berlin could still do

The exits are visible, the policy framework keeps walking past them, and the gap between what is technically available and what is politically accessible is widening rather than narrowing.

The IEEFA hydrogen report ends with a structured set of recommendations covering demand-side targeting, infrastructure resizing, derivatives-first import pathways, and electrification prioritisation, and the recommendations are essentially the same as the four exits I would name from the position I sit in.

Derivatives, not pipelines → German industry’s hardest-to-electrify demand — chemicals, ammonia, e-methanol, green hot briquetted iron — is for the products, not for hydrogen gas, and the derivatives can be imported as liquids through existing port infrastructure rather than transported through new pipeline networks. Hanna Schumacher at the economy ministry conceded in October that imports are “more cost-effective overall,” the terminals at Stade, Wilhelmshaven, Hamburg, and Brunsbüttel are already being designed ammonia-ready, and Uniper signed a framework agreement with Thyssenkrupp Uhde on ammonia cracking in November 2025 — but the derivatives pathway sits subordinate to the pipeline pathway in the current strategy, when it should sit at least parallel to it, and arguably ahead of it.

Electrification, not policy support. The heat pump deployment has done more than any single hydrogen initiative to reduce German gas demand, direct electrification has the same potential in low-temperature industrial heat and in additional transport segments, and every euro of CCfD funding directed at hydrogen blending in sectors where electrification is competitive is a euro not directed at the electrification that would actually reduce gas exposure.

Technology-neutral capacity planning → The 10 gigawatts of hydrogen-ready gas plants is the answer to a dunkelflaute problem that has other solutions — long-duration storage, demand response, interconnection with Iberian solar and Scandinavian hydropower, vehicle-to-grid as the EV fleet scales, baseload geothermal at the margin — and the capacity mechanism could be designed to let those compete on cost rather than reserving the role for thermal generation that needs hydrogen to make its decarbonisation pathway work. The IEEFA gas report’s scenario analysis of a 2045 power mix with minimal gas and hydrogen generation shows what the alternative looks like at the system level, and the alternative is not utopian — it is engineered, costed, and dispatch-feasible.

Gate the pipeline to confirmed offtake → The Federal Court of Auditors said it in October, the ministry’s monitoring report acknowledges the demand uncertainty, the regulatory mechanism exists — BNetzA can require the network development plan to be revised based on demand evidence — and it has not done so. The next NDP, scheduled for first BNetzA approval in June, is the immediate inflection point on whether the policy framework starts pacing the pipeline to the demand or continues pacing the demand subsidies to the pipeline.

None of these exits is being taken.

The Hydrogen Acceleration Act is the most consequential signal of which way the policy framework is moving — toward expansion, not toward gating. The €6 billion 2026 funding envelope is another. The H2G consolidation proposal is a third.

The line on the wall

Germany has done two things at once that should not be possible to do at once, and the fact that it has managed to do them both is what makes the position structurally difficult to walk back. It has spent four years building the most ambitious energy security and decarbonisation architecture in Europe. And it has, by deliberate policy choice as of February this year, set that architecture on a trajectory that re-anchors the country to the gas dependency the architecture was supposed to escape.

The bill for the contradiction lands on the public balance sheet, in stages, over a thirty-year window that does not require any single year’s payment to look catastrophic in order for the cumulative payment to be enormous: about €45 billion of additional public funding requirement under the demand-light path the data points toward, on top of the €17 billion already committed to LNG infrastructure, on top of the €24 billion KfW credit line backstopping the amortisation account, on top of the €6 billion of 2026 hydrogen funding, on top of the capacity payments to the 10 gigawatts of hydrogen-ready gas plants.

The components are separately defensible.

The composition is not.

The architecture is not failing.

The architecture is doing what the architecture was designed to do — allow a generation of infrastructure to be committed to ahead of confirmed demand, shifting risk forward in time and onto the public balance sheet, in exchange for getting the buildout started — and each successive policy iteration tightens the loop rather than loosens it.

• The Hydrogen Acceleration Act tightens it.

• The CO₂ transport act tightens it.

• The H2G proposal tightens it.

The amortisation account is the financial through-line that records every tightening as a balance entry on the federal credit line, and the balance entries are what the public will eventually be asked to settle.

The line that should sit on the wall: the greater fiscal risk is not the pipeline that fails. The greater fiscal risk is the pipeline that works just well enough that nobody is allowed to walk away from it. That is the trajectory Germany is on.

That is the round trip the public is paying for.

And the next time Hormuz closes — or the next time the US administration decides that European industrial competitiveness is a strategic vulnerability worth squeezing — the round trip will land on a Germany that has, through every defensive move it has made since 2022, made itself more exposed to gas shocks rather than less.

The pipeline is in the ground.

The account is open and growing.

The 2055 wall is fixed.

The architecture is doing exactly what it was designed to do.

That is the problem.

Subscribe now

The first is the weekly tonnage record for jet fuel held at the Amsterdam-Rotterdam-Antwerp hub, with every annual vintage from 2015 through 2025 traced in grey behind a single red line for the current year, and that red line tells a story you cannot un-see once you have noticed it: it begins the year somewhere around a thousand kilotonnes, well inside the top half of the decade-long band, and then declines for twenty consecutive weeks without a single sustained pause until it lands at five hundred and sixty kilotonnes by mid-May, sitting below every other annual line in the dataset, including the 2022 trough that was carved out during the worst of the post-pandemic refining squeeze.

Extrapolated forward at the rate of the past two months, the line crosses mechanical storage minimums sometime in early autumn.

The second picture is not really a chart but a transcript reel — the rolling parade of European airline and airport executives who have spent the past fortnight giving interviews that converge almost word-for-word on the same reassuring sentence.

Michael O’Leary at Ryanair has almost zero concerns about fuel supplies across Europe; Sebastian Ebel at TUI tells reporters that the discussion on fuel is a little bit artificial, as we do see no shortages for the next weeks, and I would also see no impact in the summer at all, except prices; Carsten Spohr at Lufthansa confirms supply through mid-July, beyond which his visibility, by his own admission, decreases somewhat; Gary McLean at Dublin Airport says he is hearing no concerns from anyone.

Five separate executives, five variations of the same reassurance, delivered into the most critical fortnight of the European summer booking window.

Both readings cannot be straightforwardly correct, and what follows is an attempt to work out which time horizons each of them is actually describing.

The crude side of the crisis has finally begun to mean-revert — Brent settled yesterday at $104.52, down four per cent on the week, with Marco Rubio briefing reporters about some encouraging signs and Pakistani mediators flying into Tehran for what feels like the fifth round of the same conversation — but the product side is not behaving anything like the same way, and the product whose dislocation is now telling us the most about what comes next is jet fuel, because jet fuel was always going to be the place where a Middle East supply shock would surface first, hardest, and with the longest tail to recovery.

I made the macro version of this argument last week, in a piece on Piero Cipollone’s Milan speech and Fitch’s quietly dramatic revision of its Hormuz closure assumption to five months.

The Three Scenarios took the position that the variable the ECB cannot fit inside its baseline is the regime change in how the price gets determined, rather than the height of any particular spike inside a fundamentally stable framework, and this piece is the same argument expressed at the product level rather than the policy level.

The forward curve the airline CEOs are implicitly defending is a price-shock baseline; the data on the ARA inventory chart suggests we are inside a structural break of which the price shock is only the most superficial symptom; and the cleanest place to read that break, in real time, is the jet fuel curve.

Built to fail first under exactly this kind of shock

The mechanism is not exotic, although it is genuinely under-discussed in the popular coverage.

Every barrel of crude oil yields somewhere between nine and twelve per cent jet fuel and kerosene depending on the configuration of the refinery and the grade of crude going through it, and the reason that range stays so narrow is purely physical: middle distillate occupies a specific boiling-point window between gasoline and diesel on a distillation tower, and you cannot meaningfully shift the cut points without redesigning the unit itself.

American refiners have pushed their average jet yield this spring from a long-run baseline of around ten per cent up to roughly twelve and a half — a meaningful operational flex even though the number sounds modest — and Antonio Mestre at Repsol told the Financial Times that his Spanish system had added twenty to twenty-five per cent more jet output for the summer by reconfiguring units and delaying scheduled maintenance, describing the move as unusually large for a system where operators typically have limited flexibility. The phrasing is worth reading carefully: what the European system has already done is close to the operational ceiling of what European refining can do.

The demand side gives the structure even less room to move, because aviation cannot run on diesel or gasoline or LPG in any meaningful sense — there is no engine-level fungibility between fuel types, and an aircraft without jet fuel in its wing tank does not generate revenue. Sustainable aviation fuel is technically a drop-in substitute, but the volumes are not serious; IATA’s 2026 estimate puts global SAF supply at 2.4 million tonnes against conventional jet demand of approximately 280 million, which makes SAF rounding error rather than alternative supply.

The demand curve for jet fuel is therefore nearly vertical in the short run, and the only adjustment mechanism on the demand side is flight cancellation — which is precisely what every European carrier is now publicly trying not to do.

Now layer the geography on top of those two rigidities.

The Middle East Gulf was, in 2025, the single largest net exporter of jet fuel anywhere in the world, sending roughly four hundred thousand barrels per day into international markets, of which approximately sixty per cent flowed to OECD Europe; the IEA’s May 2026 report documents those exports collapsing from the 400 kb/d baseline to just 70 kb/d in April, an eighty-three per cent reduction in two months, with European arrivals from the Gulf dropping from 330 kb/d in March to 60 kb/d in April.

A single chokepoint — the Strait of Hormuz — was the routing path for nearly three-quarters of the marginal jet fuel flowing into the global system, and that chokepoint has been effectively closed since the last day of February, with the US naval blockade of Iranian ports formalised on 13 April and still in force, freezing the deadlock in place rather than resolving it.

Hold those structural facts together for a moment.

You have a product whose demand curve is nearly vertical, a supply curve where the marginal barrel originates from a single militarily contested maritime chokepoint, a refining system with single-digit yield flexibility, and an alternative blending stream that exists only in rounding-error volumes — and once you accept the structure, it becomes obvious that jet fuel was always going to be the product that broke first when Hormuz closed, because every other refined product has at least one source of meaningful slack in its chain.

• Crude oil draws from multiple basins.

• Diesel substitutes across grades and origins. Gasoline reblends from naphtha and components.

Jet has no comparable flexibility on either side of its curve, which means that when Hormuz closes, Jet is the place where the price signal arrives first, hardest, and with the least available shock absorber to take the edge off.

And that is precisely what the price record shows.

Northwest European jet fuel prices climbed from roughly seven hundred dollars per metric tonne in mid-February to a peak of just over eighteen hundred in late March, with the crack spread between jet and dated Brent widening from a normal range of fifteen to twenty dollars per barrel out to a peak of approximately eighty-nine in early April; by the week ending 15 May the global jet fuel price had retraced to $162.55 per barrel, down ten per cent on the week and thirteen per cent on the month, with the crack settling at $52.74 — a retracement that looks orderly on the surface until you notice the crack is still up one hundred and fifty per cent year on year.

That residual hundred-and-fifty per cent is the structural premium the substitution has not been able to clear, and on the analysis that follows, it will not clear until physical Hormuz flows resume in something approaching normal volumes.

Eighty days of plumbing have kept the planes in the air — at four prices nobody is naming

The substitution story of the last eighty days is impressive market plumbing, and it has been doing more heavy lifting than most of the headline coverage credits.

The map at the front of the LSEG tanker tracker tells the cleanest version in four traced routes — the Maetiga from Antwerp down to the Seychelles, the Nord Ventura hauling product from Baton Rouge across the Pacific to Melbourne, the Minerva Zen running west to east from Perth Amboy to the Isle of Grain, and the Velos Fortuna taking a similar US Atlantic-coast cargo south to Walvis Bay — and read together, those four routes describe a rewiring in which US Gulf and Atlantic refining have effectively become the global swing supplier of jet fuel to Europe, Africa, and Oceania for the duration of the closure.

The numbers underneath are substantial.

West African exports, driven almost entirely by Nigeria’s Dangote refinery, have more than doubled from a three-month average of around 70 kb/d to roughly 145 kb/d in April; American refinery yield to jet has stepped up by two-and-a-half percentage points, translating into roughly 150 kb/d of incremental supply once export logistics are factored in; the IEA estimates European refinery yield maximisation alone has added roughly 100 kb/d of jet, or about twenty per cent of pre-war Middle East import volumes; and Eugene Lindell at FGE NexantECA puts the strategic reserve contribution at thirty-four per cent of the projected European deficit for 2025, with the remainder filled by redirected US and Nigerian flows.

Stack the arithmetic and the page-level totals for Europe almost manage to balance, which is precisely the basis on which the airline CEOs are speaking when they project summer confidence, and it is the reason the IEA’s May report assumes European supply will not break in the manner the April warning had suggested it might within six weeks.

Jet fuel did not run out in April, did not run out in May, and the airlines are correct on the literal reading when they say the molecule has continued to flow into their airfields — the price did the work it is meant to do, and at $1,800 per tonne every refiner anywhere that could produce middle distillate was producing it at maximum yield, exactly as József Váradi at Wizz Air put it when he observed that that kind of market price mobilises forces.

But the substitution did not arrive at no cost, and the four costs that have been paid are not being aired in the airline press cycle nearly as openly as the supply numbers themselves.

The first is freight and ton-mile, where the geometry is straightforward: a barrel of jet fuel from Perth Amboy to the Isle of Grain covers approximately four thousand nautical miles, a barrel via Suez from the Persian Gulf to Rotterdam covers about six thousand, and a barrel routed around the Cape of Good Hope covers eleven thousand, with clean product tanker rates now running at multiples of pre-war levels and that freight premium living quietly inside the jet fuel price the airlines are actually paying, even at the moments when wholesale Rotterdam quotations come off their highs.

The second cost is what happens to the rest of the barrel once a refiner has decided to maximise jet yield, because refining inside any given unit is a zero-sum allocation problem, so when you push jet output upward, something else has to give — typically diesel.

The Middle East gasoil shortfall, with exports falling from 1.4 mb/d in 2025 down to roughly 700 kb/d in April, is now feeding into African and European diesel markets, with diesel cracks in Northwest Europe widening alongside the jet crack and Eastern European trucking economics coming under visible pressure; the substitution did not create new molecules at the global level, it merely re-allocated existing ones inside the European slate.

The third cost is the one sitting in plain sight on the inventory chart and, in my view, the most analytically important of the four.

The substitution has kept European airports fuelled and the refining system running at high utilisation, but it has not stopped the European jet fuel inventory pool from drawing down — ARA stocks have gone from a thousand kilotonnes in January to 560 in mid-May, and the i6 Group data showing European airport stockpiles up sixty-two per cent year on year in April sounds like resilience until you trace the source, because those airport barrels were pulled from the regional wholesale pool that is itself shrinking.

The airports have not added molecules to the European total; they have repositioned existing molecules from tank farms into airfield kerb-tanks, and the jet fuel the airlines will draw on in July and August has been pre-allocated rather than augmented.

The fourth cost is hedging exhaustion, operating on a longer timescale than the other three.

Morningstar data shows European airline 2026 hedge coverage running between fifty-five per cent at the low end (Wizz Air) and eighty per cent at the high end (Ryanair), with Lufthansa at seventy-seven, IAG at sixty-two, and EasyJet at roughly seventy, and those hedges, laid down at pre-war forwards in the $700-800 per tonne range, are now meaningfully in the money and providing genuine cash protection through summer.

The complication is what happens beyond: the 2027 hedge book is being priced at post-shock forwards as we speak, the 2026 book runs progressively thinner as contracts roll off, and the spot price will, in any scenario short of complete Hormuz reopening with prompt arbitrage to pre-war levels, still be elevated.

The airlines are not running out of fuel — what they are running out of is cheap fuel, supplied entirely by an earlier hedging decision that someone else is now paying for.

Beijing’s decision in February has already locked in the next shock

The Chinese piece of this story should bother the world’s energy planners more than it apparently does, and the structural numbers are the cleanest way in.

China produced 1.28 million barrels per day of jet fuel and kerosene in 2025 — an all-time record well above the pre-COVID peak — while domestic consumption, including the bonded-area volumes that fuel international flights departing Chinese airports, came in at roughly 1.05 million b/d, and the arithmetic difference of around 230 kb/d of structural net surplus ought, in any normally functioning market, to have been available to clear into the stressed Atlantic basin once the price signal arrived.

Chinese jet exports peaked at 470 kb/d in 2025, the highest reading on record; the capacity exists, the molecules exist, the supply chain exists.

What actually happened is instructive in ways that reach past the immediate crisis. February 2026 produced a reading of 583 kb/d of Chinese jet fuel exports — the second-highest single-month figure on record — suggesting Beijing was initially prepared to let the surplus clear into the international market.

Two weeks into the war, however, once US and Israeli air strikes had intensified and Hormuz closure had hardened into something other than a temporary disruption, China imposed an immediate ban on the export of gasoline, jet fuel, and diesel, with Kang Wu of the Oxford Institute for Energy Studies characterising the decision as a huge turnaround of the policy.

March exports fell to 380 kb/d, April fell further, and although the ban has since been partially eased, China’s net jet exports for April finished essentially flat to a year earlier — no sign of the surge the structural numbers and price signal would have predicted under normal arbitrage logic.

The IEA’s April export breakdown lets you see the same regional swing at system level: OECD Asia net jet exports fell from 80 kb/d a year earlier to 10 kb/d in April, Russian exports collapsed from 40 kb/d to zero, and the two regions that should have been arbitraging hardest into the Atlantic shortage were instead either holding their barrels at home or losing them to parallel disruptions.

Goldman Sachs’s days-of-cover sensitivity on OECD European jet — the chart on page nineteen of Della Vigna’s note that has been circulating around every European energy desk since it dropped — explicitly models five scenarios in which 0, 25, 50, 75, or 100 per cent of disrupted volumes are substituted, with the 100-per-cent line keeping European cover near the IEA’s 23-day critical threshold through 2027, the 25-per-cent line falling through zero by autumn, and the actual outturn, given China’s policy and Russia’s collapse, tracking close to the 50-per-cent line — better than the bear case Goldman modelled, but well short of the smooth-substitution path that airline forward curves are pricing.

Three reasons stack to explain why China made the choice it made, and the third is the one that will outlast the crisis itself.

1. The first is domestic demand: Sinopec’s China Energy Outlook 2060 projects seven per cent growth in Chinese jet fuel demand between 2024 and 2030 — one of the most aggressive numbers in the institutional canon for any oil product — and Beijing has every plausible reason to preserve optionality on its own aviation sector.

2. The second is feedstock optionality on the upstream side, since China imports a meaningful share of its crude through Hormuz itself, and holding refined product domestically is a sharper hedge against further upstream disruption than holding crude would be.

3. The third reason is the merger, and this is where Kang Wu’s OIES paper does some real analytical work.

The Sinopec-CNAF integration, completed in early 2026, creates a vertically aligned platform running from refinery output through bonded coastal logistics into airport into-plane refuelling, placing Sinopec to capture the entire jet fuel value chain at the moment when that chain has become a strategic asset rather than just a commercial supply network; to have exported aggressively into a global shortage would have meant handing capacity to the international spot market exactly when Beijing’s instinct was to consolidate it domestically.

China did not behave as a swing supplier because China is no longer trying to be one — it is trying to be a vertically integrated domestic monopoly with selective export optionality calibrated to its strategic interests, and the merger is the institutional expression of that pivot.

The lesson generalises beyond China: the next jet fuel shock will encounter a supplier set that has just learned, in real time, exactly how much strategic value sits inside a refined-product molecule once you choose to hold it at home rather than ship it out.

Mid-July is the only horizon anyone is willing to stand behind in public

Look at the ARA chart for a second time, because it is doing more work than the headline read makes obvious: the 2026 red line opened the year at the top of the historical band, declined continuously for twenty weeks, and currently sits below every comparable year, with linear extrapolation placing ARA inventories at operational minimums by September.

Operational minimums are not zero — they are the level below which the physical logistics of the storage system stop working, where you can no longer blend, transfer, and dispatch product at required pressures because tank bottoms cease to hold enough volume — so what the chart is telling you is not the date Europe runs out of jet fuel; it is the date European storage infrastructure starts to fail mechanically as a buffer system.

The Goldman days-of-cover sensitivity tells you what the country-level distribution underneath that aggregate looks like, and the UK is the country exhibit for reasons that ought to make British readers slightly more uncomfortable than they currently appear.

UK commercial jet fuel stocks entered the crisis at around twelve days of cover, already the lowest among any major European economy; Goldman’s central case has UK commercial cover dropping below ten days by mid-summer, with the exhibit four chart showing UK stocks crossing the 23-day critical threshold by April and breaching ten days in July.

Even when strategic reserves are formally activated — and the IEA collective release of 400 million barrels announced on 11 March is one of the largest in the institution’s history — the UK trajectory still drops sharply, because strategic stocks carry a two-to-four-week deployment lag from declaration to physical delivery, and during that lag the commercial inventories do all the work of clearing demand at the airfields. Strategic cover is theoretical until the molecules reach the wing tank, weeks after the political decision is taken.

This is the variable airline CEOs are not folding into their public statements, and it is not realistic to expect them to.

Lufthansa is saying supply is secured through mid-July — roughly seven weeks from where we are sitting — and Carsten Spohr’s own acknowledgement that visibility decreases somewhat beyond that is the closest any sitting executive has come to flagging the second-half problem this earnings season; the five-month Hormuz closure assumption that anchored The Three Scenarios puts physical flow resumption no earlier than end-July, the IEA’s May report assumes gradual resumption from June onwards, and the gap between supply secured for the next seven weeks and resumption of substitute flows that do not require existing inventory cushions to be drawn down further is precisely the gap that breaks if either side moves wrong by even a few weeks.

A second clock is running on a parallel track to the inventory clock.

Bernstein analysis cited in CNBC observed that airline fuel hedges are front-loaded and thinning fast, and most European airline 2026 hedge books, laid down at pre-war forwards in the $700-800 per tonne range, are now in the money but rolling off progressively through the year, so by the fourth quarter the cash protection thins materially. By the time anybody starts writing the 2027 book the relevant forward curve is the post-shock curve — Goldman has 2026 Brent at $90 base case, JP Morgan at $100 central path, Morgan Stanley at $200 tail — and jet derivatives will price accordingly.

The i6 Group estimates an airline P&L impact of $2.6-3.9 billion through end-April that could rise as high as $12.3 billion by late August if the conflict remains unresolved; that is the 2026 figure, with the 2027 figure structurally larger.

So when O’Leary tells Sky News he has almost zero concerns about fuel supplies, what he is correctly saying is that physical fuel supply is not the binding constraint for Ryanair through summer 2026 — Ryanair has 80 per cent hedge coverage, only about 8 per cent of its current cost base is floating with the spot market, and the network is structured to absorb modest capacity cuts on marginal routes without losing strategic position.

Ryanair is not the airline that breaks.

The airline that breaks is the one with low hedge coverage, a higher fuel cost share, and a less flexible network — Wizz Air at 55 per cent sits at the painful end of that distribution — or, on the country side, the carrier most exposed to UK jet inventory dynamics, which means British Airways, Virgin Atlantic, and EasyJet in its London-anchored routings.

The right question is not whether Europe will run out of jet fuel; it is which airline, on which route, at which airport, requests slot relief first. Slot relief is the operational concession a regulator grants when the physical flow breaks at a single airfield and the carrier asks permission to cancel without losing its operating window.

Nobody has had to ask yet.

Somebody will, and the date when they do is the date the second-half problem becomes visible in equity prices.

The chorus is structural, not coincidental

There is a second-order phenomenon that deserves naming because it is doing real economic work in this crisis, even though it lives in the soft layer of corporate communication rather than the hard layer of inventory data: call it the performance economy, and what you are watching, when you read the airline CEO press statements alongside the airport stockpile communications alongside the tour operator marketing, is a coordinated chorus performing the same reassurance into the same booking window.

We are fine.

The discussion is artificial.

We have not heard of any concerns for the summer.

The repetition is structural, not coincidental.

Airline revenue is unusually path-dependent.

A reader who decides today to book a July holiday is locking in a ticket purchase that pays the carrier in May, hedges its Q3 load factor, and removes consumer-level optionality entirely in the airline’s favour; a reader who instead decides to wait — let me see how the fuel situation develops, perhaps Croatia in late August — moves the cash flow to the right and adds optionality back to their own balance sheet at the airline’s direct expense.

Airlines run single-digit operating margins, and even a modest shift in booking pace, in the complete absence of any actual fuel shortage, is enough to crater a quarter. EasyJet’s H1 2026 numbers, a £540-560 million headline loss, already incorporated weaker forward bookings tied to fuel-price uncertainty, with the carrier flagging customers were buying later than in previous cycles — and the CEO commentary you are reading this fortnight is the airlines’ marginal lever on the only variable they can still influence: the consumer’s willingness to commit to summer travel under uncertainty.

This is not a criticism of the executives concerned.

The performance is not deceitful — Carsten Spohr is not lying when he says Lufthansa has secured supply through mid-July, because the supply through mid-July is, on the documentation he has in front of him, genuinely secured.

Each assumption is non-trivial, none are statistically independent, and treating secured as unconditional is the rhetorical move that allows a chief executive to get through an earnings call without precipitating the booking collapse that would validate the bearish supply case.

The same pattern shows up in airport stockpile data: the i6 Group report describes European airport jet fuel inventories as up sixty-two per cent year on year in April, presented as a single striking number, while the Italian airport shortage flag from spring — where some operators reported delayed lifts — has been quietly resolved through emergency cargoes and is no longer in front-page commentary.

The transition between these two regimes is not gradual — it is a phase change, triggered by a single discrete event such as a major airport requesting slot relief, a regional refinery reporting an unplanned outage, or a second-order disruption removing a load-bearing assumption — and up to the moment that event arrives, the performance economy dominates the public discussion. After it arrives, the inventory clock binds suddenly and visibly, and the CEO commentary changes register inside a single news cycle.

The market is currently pricing the smooth-substitution path through end-2026 without that triggering event; implied volatility on jet derivatives sits below where the inventory data says it ought to; the forward curve assumes physical normalisation on a faster timeline than the Goldman sensitivity would support; and that gap is the trade — a long-volatility position on jet fuel cracks running through Q3, sized against the probability that any one of the load-bearing assumptions fails before the resolution arrives.

How the broken product propagates into the rest of the book

The product that broke first will also be the product that clears last, and the asymmetry carries cross-asset consequences worth mapping back into the wider portfolio context — because the jet curve is currently the cleanest single read on a set of macro variables that have not yet repriced.

European inflation prints will run hot through Q3 regardless of where the ceasefire negotiations end up, but the propagation mechanism is not the direct weight of jet fuel in the HICP basket, which is small; it is the diesel and middle-distillate complex, the freight footprint, and the transport-services component that picks up the elevated cost structure once it works through ticket pricing into the general price level. The Cipollone scenarios I worked through last week placed the adverse path with cumulative inflation 1.5 percentage points above the December projection — the current jet fuel data is consistent with that range, but the diesel and gasoil data is pushing toward the severe path — and the Governing Council, when it meets in June and again in July, is going to discover that the variable it cannot place inside its baseline is the variable that determines whether the current rate path remains plausible. On the data today, the path is harder to defend than it looked when Cipollone spoke in Milan three weeks ago.

Sovereign spreads in import-dependent economies should be pricing the rationing tail more aggressively than they currently are, and the UK is the cleanest example given Goldman’s flag on commercial stocks dropping below ten days of cover by July.

Gilts traded straight through the April crisis peak without meaningful repricing of the tail risk associated with physical jet fuel rationing, and the assumption embedded in that behaviour is identical to the smooth-substitution assumption that the airline forwards are pricing — a first formal request for slot relief at Heathrow, or a Department for Transport acknowledgement that contingency rationing protocols are being prepared, would change gilt pricing inside a single session.

Airline equity is the obvious read, but the more interesting position is the dispersion within the sector rather than the direction of the basket as a whole.

Ryanair holds 80 per cent hedge coverage and runs a network engineered to absorb marginal-route abandonment, which means it will absorb the shock and emerge with structurally better market share once the system clears; Wizz Air at 55 per cent sits at the opposite end of the same distribution; and the right trade is the spread between them rather than a sector-level long or short, with Lufthansa positioned in the middle but carrying disproportionate exposure to the Frankfurt-Munich long-haul network — the most price-elastic segment of the European matrix and the one most likely to give way under sustained ticket-price escalation.

Refining margins are arguably the cleaner long position underneath all of this.

European complex refiners with hydrocracker flexibility — Repsol, Galp, Eni — captured Q1 and Q2 results that one quoted refiner described to the Financial Times as among the highest in the company’s history, and that is unlikely to be a six-week phenomenon even if Hormuz reopens on the IEA’s central timeline, because crack normalisation requires physical flows to resume, plus inventory rebuild after flows resume, plus unwinding of the strategic reserve releases currently propping up the system. The IEA estimates rebuilding will require roughly an additional 1 mb/d of supply over the next three years on top of underlying demand growth — a structural margin floor that deserves a higher equity multiple than the sector currently carries.

The SAF trade is the position the airline industry is structurally not yet able to price, and it connects back to the Sinopec-CNAF merger as the institutional pivot of the whole story. China is positioning itself, through the merger and the SAF capacity build alongside it, to be the marginal SAF exporter into the European mandate window over the back half of the decade — the European SAF blending mandate ramps progressively, the CBAM definitive phase began 1 January 2026, the structural demand floor is in place, and the supply chain that fills it has not been determined. The OIES paper argues Sinopec is the only Chinese NOC with operational SAF capacity today, with the Zhenhai unit complemented by three new facilities totalling 630,000 tonnes per year against European producers that are slower, smaller, and more fragmented.

That is a five-year story compressed into a fifteen-month implementation window, and it is the structural consequence of the eighty-day shock we are currently inside.

• Hormuz exposed the vulnerability.

• The Sinopec-CNAF merger expressed the strategic response.

The crisis is accelerating a rewiring already underway, and the rewiring will outlast the crisis by years.

The sentence I would put on the wall

The airline CEOs are not wrong when they say supply is secured for the summer — they are correct at the time horizon that matters for their summer earnings season; they are simply not addressing the time horizon that matters for their 2027 hedge book, their long-haul economics in the back half of the year, or the European inflation print that lands while their passengers are still on aircraft returning from holiday.

The two pictures I started with — the ARA inventory series in red and the airline CEO press circuit in soundbite — are both factually correct readings of their respective subjects; they describe different parts of the same physical system at different time horizons, and the synthesis that connects them is that confidence and crisis can coexist for exactly as long as substitution flows keep running and strategic reserves keep clearing.

The trade is the gap between secured and secured assuming, and that gap is wider than the market is pricing, narrower than the bear case is claiming, and timed in a way that does not match the duration of either prevailing narrative.

Jet fuel is doing in 2026 what no other oil product can do, because no other oil product is built the way jet fuel is built — the vertical demand curve, the rigid supply chain, the chokepoint geography, the absent substitutes — and the combination of those four structural features is what makes jet the cleanest expression of the Hormuz shock and the slowest part of the global energy system to recover from it.

The crude leg is mean-reverting, and Brent will fall mechanically when Hormuz reopens, but the product leg is path-dependent: the inventory drawdown does not reverse on a price signal, the hedge book exhaustion does not reverse on a ceasefire announcement, and the substitution flows that worked in spring 2026 worked partly because they took the global supplier set by surprise — the next shock will encounter suppliers who have learned, in real time during this crisis, exactly how much strategic value sits inside the molecule once you are holding it at home rather than shipping it out.

That is the structural break I argued was hiding inside the baseline assumption in The Three Scenarios last week, and it is the break the airline CEOs are correctly refusing to articulate. Both behaviours are rational given the incentives the actors face, both are visible in the underlying data once you know where to look, and the market that closes the gap between them is the market that prices the next four months of European energy correctly.

Somebody is going to have to ask for slot relief eventually.

The date when they do is the date the gap closes.

Subscribe now

That is the tape.

The narrative wrapping it — Bazilian, Amoah, Johnston and Matisek’s The Impacts of the Iran War on Coal in The National Interest on 12 May, every commodity-desk note since the Ras Laffan strikes on 18 March, every Reuters wire on Asian utility behaviour — is that coal is back, Asia is pivoting, the transition is delayed, and the IEA’s December 2025 Coal 2025 report (which called peak coal at 8.85 billion tonnes in 2025 and a 3% decline to 8.58 Bt by 2030) is the forecast about to be overrun by physical reality.

That reading is wrong.

Or, more precisely, it is the kind of wrong that happens when you watch the price column rather than the price level.

The Newcastle rally is telling you the story, not confirming it

• Newcastle closed February at $116/tonne, topped at $146.50 on 20 March, and ground back to $131 by Friday — a 31% rally from pre-war levels, or 13% if you anchor against the 2025 average of $106.

Compare with 2022, when the Russian invasion of Ukraine drove API2 European thermal coal up 78% from pre-war levels and peaked it above $400 on the back of what was a coal supply shock layered onto a gas supply shock; the 2026 move, by contrast, is a gas-only shock with coal substituting in around the edges.

The structural relationship between the two fuels makes the asymmetry visible.

On a like-for-like USD/MWh basis, TTF gas has spent the past two years trading at a 2-3x premium to Newcastle thermal coal, with both moving inside a tight envelope — until the Iran war, when gas broke out and coal followed only modestly.

The live tape since December tells the same story across three benchmarks.

ICE-Newcastle thermal rallied sharply on the Ras Laffan strikes in mid-March and has now retraced toward pre-war levels; ICE-Rotterdam thermal did the same with European-specific friction holding it slightly higher; China coking coal — driven by domestic steel demand rather than thermal substitution — has decoupled entirely and continues to bid.

That price differential is not a footnote — it is the entire signal.

If the market believed this was 2022 redux the forward curve would be pricing $250-plus, but the June Newcastle contract on ICE has open interest of roughly 1,200 lots and Q3 calendar spreads show the curve in mild backwardation, not the structural contango that would print if buyers expected sustained substitution through next winter.

Bloomberg Intelligence’s Alon Olsha put the mechanism cleanly to Mining Weekly on 18 March: the 2022 rally drew on three structural conditions — European coal capacity intact, simultaneous coal and gas supply shocks, Russian rerouting — that are largely absent in 2026. European coal capacity has been retired in the intervening four years, North Asian utilities now operate under long-term LNG contract structures that cap how hard they can lean into spot substitution, and the plants that can swing to coal in Japan and Korea are already running closer to their operational ceiling than they were four years ago. The market is pricing a bounded substitution event, not a regime change.

Newcastle at $131 — not $250 — is the trade thesis disclosing itself in real time.

What the IEA already said two months before the missiles flew

Coal 2025, published by the IEA on 17 December 2025, is the report nobody is re-reading in light of the war, and it should be on every commodities desk this week.

The Demand chapter projects a near-perfect plateau — peak in 2025, falling 266 million tonnes to settle at 8.58 Bt by 2030, a 3% decline over five years that is best read as a flat line with a downward tilt — and even ASEAN, the fastest-rising region at more than 4% per year, adds only 127 Mt over the entire forecast horizon.

The shape of the curve, set against the twenty-five years of expansion that preceded it, is the entire structural argument.

The compositional argument was already there.

India and ASEAN drive the incremental demand, China holds flat at around 4.95 Bt with internal mix-shifting toward chemicals and away from cement, the US under federal coal-supportive policy contributes a 2025 bump of 37 Mt — the single largest absolute increase globally — before resuming its 6%-per-year decline through 2030, the European Union loses 153 Mt, and Indonesia’s production declines for the first time since COVID due to shrinking export demand.

The IEA’s own decomposition of the 2024-2030 trajectory shows this with the regional pluses and minuses lined up cleanly:

• India +225 Mt and ASEAN +127 Mt on the upside,

• EU -153 Mt and Rest-of-World -179 Mt on the downside,

  • netting to the 266 Mt decline.

That December framing is the bookend for everything happening since 28 February. The 2026 substitution into coal is not a refutation of the plateau forecast — it is an acceleration of demand inside the plateau, with Japan and Korea drawing harder on coal in 2026 pulling forward demand that was already coming and concentrating it in the 2026-2027 window.

The integral under the curve does not change. The shape does.

This is the read the National Interest piece misses.

Bazilian and co-authors treat the pivot to coal as evidence that energy security still depends on fuels capable of withstanding geopolitical disruption — a fair characterisation of the disruption itself, but a category error when extended to long-arc demand, because the Iran War is not building a new generation of coal plants in Vietnam or relicensing retired German lignite. It is running existing capacity harder for several quarters, which is precisely what dispatchable infrastructure exists for.

The plateau is doing its job.

The substitution is bounded by contracts and physics, not conviction

This is where the operational layer gets interesting.

LSEG’s coal-switching threshold for Japanese utilities sits at $10.24/MMBtu of spot LNG, Korea’s at $10.45, and JKM closed Thursday at $17.11 — both thresholds crossed weeks ago, comfortably, and South Korea’s April coal-fired increase of 39.7% YoY is exactly what an overwhelming economic switching case looks like converted into megawatt-hours. The problem is that it is also the outer bound of what the system can physically deliver, not a steady-state run-rate the market should be extrapolating from.

1. The first binding constraint is not price but plant utilisation: South Korean coal plants are running at higher utilisation than at any point since 2019 per KPX data, Japanese coal plants likewise, the maintenance-deferral window is short, and by Q4 2026 several of these plants will have to come down for inspection regardless of what the LNG strip is doing. Andre Lambine of S&P Global Commodity Insights put it correctly at the Singapore industry event last week — “the longer this war continues, the more switching we will see” — but the upper bound on switching is set by plant availability, not by price differential.

2. The second binding constraint sits in the LNG contract book. Most North Asian buyers procure on long-term Brent-indexed contracts plus a smaller spot tranche, which means that JKM at $17 is what shows up on Bloomberg but the blended cost of an average North Asian buyer’s LNG portfolio sits materially lower — somewhere around $12-13/MMBtu depending on the Brent linkage formula — and the switching arithmetic ultimately gets done on the marginal cargo, not the average. Utilities switch on incremental volumes, which is the spot piece, while the bulk of their gas burn rolls on under contract.

3. The third binding constraint shows up in freight, where Panamax rates on Australia-Japan have run a persistent premium of $5/t and more over Indonesia-South China routes through 2023-2025, with Richards Bay-Rotterdam Capesize freight tracking a similar pattern — meaning that the freight cost structure underneath any war-driven substitution into seaborne coal was already elevated long before the Iran war hit, and is now compounding as bunker costs and Panamax rates have risen further since February.

Bloomberg Intelligence’s high-end estimate of 40-60 million tonnes of potential gas-to-coal switching across Europe and Asia is the upside scenario, and the realised number will more likely land between 20 and 30 Mt — meaningful at the regional level, but a small fraction of the 8.85 Bt headline market.

Indonesia is the lever the IEA had already identified

The interesting structural read sits on Indonesia.

Coal 2025 projects Indonesian exports declining over the forecast horizon as domestic metallurgical demand — nickel smelters in particular — absorbs more output, with the 2026 government production target cut to 600 Mt from 790 Mt the previous year (a cut set before the war began), and Indonesia’s coal reference price mechanism, introduced in 2025 to let Jakarta set the floor on export contracts, was likewise designed for a different scenario.

Now the geometry inverts.

The war creates a demand pull on Indonesian coal at exactly the moment Indonesia was tightening export supply, and while Newcastle is the headline benchmark, for Vietnam, the Philippines, Bangladesh and India, the binding constraint is the volume Jakarta is actually willing to release this year.

The National Interest piece flags this correctly as a coercion vector — relocating dependence from one supplier to another rather than eliminating it — and if Jakarta tightens further to protect the domestic grid, the import-dependent set pivots to Russian, Australian and South African coal at a structural premium of 15-25% per tonne owing to higher calorific content and longer voyage times.

The calorific differential is what enforces the premium.

Indonesia FOB 4,200 kcal/kg coal has traded around $50/t through 2024-2025, against Newcastle 6,000 kcal/kg at around $110/t and Australian low-volatile coking coal at $200+ — and the gap is not arbitrage, it is energy density priced through the burn.

The market that should be watching this is sovereign credit, not coal equity.

Pakistan’s external position is already stretched, Bangladesh has gone to the World Bank twice this year for energy-sector emergency facilities, and the Philippines’ energy emergency declaration on 25 March was a fiscal signal as much as a power-market one — these are the countries for whom coal substitution is a Tier 2 problem: AFFORDABILITY, not availability.

Indonesia’s own production economics add a further layer of constraint.

Open-pit thermal coal mining is diesel-heavy and tyre-constrained, and the input-cost basket facing Indonesian producers — diesel, explosives, steel, mining tyres — has not normalised after the 2022-2024 inflation cycle in those lines, meaning that any war-driven uplift to export realisations is partly captured by elevated cost-of-production rather than passing through to producer margin.

The reflexive misreading sits in equities

This is where the reflexivity frame earns its keep, and where the cleanest mispricing of the year currently sits.

Coal equities have rallied harder than coal physical, with Yancoal up 41% since 28 February, New Hope up 38%, and Glencore — only partially a coal play, with most of its lift from energy trading — up 22% on a heavier weighting; Bloomberg Intelligence’s modelling published on 1 April projects Glencore’s 2026 earnings rising 14% YoY to $20 billion in a prolonged-conflict scenario, with the trading desk capturing the kind of arbitrage that doubled marketing earnings in 2022.

The reflexive loop runs like this: war breaks out, LNG spikes, coal substitutes, coal equities re-rate on a 2022-redux mental model, the re-rating draws further flows from generalist energy funds rebalancing into coal exposure, sell-side analysts ratchet earnings forecasts, and the price action — equity, not commodity — looks like confirmation of the regime-change thesis to capital allocators reading the share-price tape as fundamental signal. The misperception of a coal renaissance gets capitalised straight into the bid stack.

But the commodity curve is not playing along — Newcastle at $131 rather than $250, open interest thin, the forward strip in mild backwardation, the physical signal saying bounded substitution while the equity signal says regime change — and the gap between the two is where the structural mispricing lives.

I made a related point earlier this month on the asymmetry around Hormuz-scenario pricing, where the market was anchoring crude on the median path and ignoring the Fitch-implied tail; that asymmetry has now partially closed in oil, but it has not closed in coal.

Kardamow is not the publication that gives trade recommendations, but the structural read here is that coal equities are pricing a duration of substitution longer than the LNG market and the IEA forecast jointly support, and the interesting tail is what happens to Yancoal, New Hope and Whitehaven in the event of a sustainable Hormuz reopening at any point in H2 2026.

The first Qatari cargo cleared on or around 10 May, the IRGC’s redefinition of the Strait as a “vast operational area” — Mohammad Akbarzadeh’s phrase, reported on 12 May — keeps the channel tightly conditional, but Project Freedom’s brief launch and pause on 4-6 May reads as a political pressure release toward eventual normalisation; if the strait normalises before Northern Hemisphere winter, JKM retraces, the switching arithmetic unwinds, and the equity-level coal rally becomes the most exposed position in the energy complex.

The plateau is doing exactly what it was forecast to do

Strip out the narrative and the cleanest read of the last eleven weeks is this: the IEA Coal 2025 forecast is being validated in the worst possible way for the coal-renaissance thesis. The plateau is the insurance policy — the disruption arrives, the plateau absorbs it, and the forecast (derisively re-titled by some commentators as the “permanent floor” reading) is not refuted by the substitution event but operationalised by it.

The price discipline of thermal coal itself confirms the regime.

Newcastle thermal at $131 sits below the 2017-2019 South China CFR average and well below the 2017-2019 Newcastle average — and the 2024-2025 trajectory across both Newcastle and ARA was a steady grind lower toward those pre-2022 levels, exactly the pattern of a commodity returning to a structurally lower-volatility regime before a temporary shock pulled it off.

The volatility picture confirms the same read.

Across 2023-2025 the standard deviation of weekly European thermal coal prices fell from roughly 2.5 in 2023 to under 1.0 in 2025, while Brent and TTF volatility stayed elevated and effectively unchanged — meaning the coal market was structurally settling while gas and oil were not, and that the calm thermal coal regime of late 2024 into 2025 was the IEA forecast already showing up in the price discipline of the underlying market.

That is the structural picture.

Newcastle at $131 is the cyclical picture. The fuel that physical reality reaches for when LNG stops working is not the same fuel that demand-curve dynamics scale into the next decade — two different roles entirely, and the IEA report distinguishes them implicitly through the divergence between aggregate demand (a plateau into 2030) and net capacity additions (which continue to retire faster than they build, particularly in OECD markets). The National Interest piece reads the cyclical and extrapolates it into the structural.

The commodity market is, for now, not making that mistake.

The equity market is.

Cross-asset close: where the tape goes from here

Three things to watch from this desk over the next eight weeks.

• First, JKM relative to Brent-linked LNG contract pricing — when the spot premium narrows to under $4/MMBtu, the marginal switching incentive starts to compress, and the coal-burn rate normalises before the equity tape ever registers what is happening underneath.

• Second, Indonesian export quota guidance from MEMR, where any tightening signal compresses Newcastle’s correlation with JKM and puts the import-dependent set under proper duress, shifting the trade from coal price into sovereign credit at exactly the moment Pakistan, Bangladesh and the Philippines have least room to absorb it.

• Third, Glencore’s H1 trading-desk earnings release on 5 August, where the marketing line is the clearest disclosure on whether the 2022-comparison capital flows reflect actual basis spreads or just sentiment.

The Iran War has not undone the IEA forecast — it has stress-tested it, and the plateau, so far, is holding the shape the report drew in December. The interesting alpha sits not in being long coal (that trade is crowded and well-known) but in identifying the moment the reflexive equity rally outruns the commodity it claims to track.

That moment is closer than the tape is admitting.

Subscribe now

Piero Cipollone gave a speech in Milan on Tuesday last week that I have been turning over for seven days.

The venue was the 2026 Sustainable Development Festival.

The audience was not the usual one for an ECB board member — environmental policy people, climate finance people, the sort of room where the words energy transition land softer than they would in Frankfurt.

The press picked up one line.

“The current situation seems to be drifting away from our March baseline projections, which increases the likelihood that we may need to adjust our policy rates.”

Bund 10s lifted four basis points before lunch.

The market read it as a rate-hike signal.

It is a rate-hike signal, and it is also something else.

The speech is built around three scenarios — baseline, adverse, severe — and a closing argument about decarbonisation as the structural answer to energy supply shocks. The scenarios are the architecture. The decarbonisation argument is the moral. The architecture undermines the moral, and Cipollone almost says so himself, and then walks away from it.

Two days later, Fitch published its updated oil and European gas price assumptions. The revision was small in language and large in content — Fitch lengthened its assumed closure of the Strait of Hormuz from one-to-two months to about five, and in doing so the institution caught the tape.

That is the actual story of last week.

Cipollone laid out the framework, and Fitch revealed why the framework is already obsolete.

Three scenarios, one shape

The three scenarios are spelled out cleanly, and it is worth reading them as a unit.

• Baseline — published March 2026, with HICP peaking at 3.1% in Q2 2026, decelerating to 2.7% in H2, and returning to target through 2027. Oil and gas prices follow the futures curve, and the Hormuz disruption is treated as a temporary supply shock with a mean-reverting price impact.

• Adverse — oil peaks at $119/bbl in Q2 2026 and gas at €87/MWh, both declining thereafter, with cumulative inflation 1.5 percentage points higher than the December projection and cumulative growth 0.8 percentage points lower. Sharper, more persistent, but the variable still mean-reverts — it just takes longer.

• Severe — oil peaks at $145/bbl and gas at €106/MWh, both declining at a slower pace, with cumulative inflation 6.3 percentage points higher through 2028.

Read them together and you see what the architecture assumes.

All three scenarios have the same shape — a price spike followed by a return path. The peak height differs. The decline rate differs.

The shock has a beginning, a middle, and an end, and the question for the Governing Council is which of three curves the variable will trace.

The scenarios bound the future by varying intensity, duration, and propagation — the three nouns Cipollone uses repeatedly, and which deserve attention because they describe what is inside the model and therefore what is outside it.

What is inside the model is a price shock.

What is outside the model is a regime change in how the price is determined.

Fitch walked the assumption

Two days after Cipollone spoke, Fitch published.

The text is dry — The change is not.

Before this revision, Fitch was working off a “one-to-two month” closure assumption. That is the same window the ECB’s March baseline implicitly priced — short enough that the futures curve does the work, short enough that you can call it a supply shock and look through it.

After the revision, Fitch assumes the strait will “begin reopening around July” and that the closure period will run “about five months.”

The path Fitch now sees is Brent at $100-110/bbl in May through July, then falling to $70/bbl by September.

That September number is not a forecast of de-escalation.

It is the residual price after OPEC produces to maximum capacity, non-OPEC supply growth runs at 3 mmbpd, and oversupply hits the market post-reopening.

The agency assumes 5% of global oil demand needs to be destroyed during the closure window to balance the market — a level Fitch itself calls “possible, albeit severe.”

The institution is no longer arguing about whether the shock is transitory.

It is arguing about how long the transit takes and what residual premium survives. The variable being revised is not the price.

It is the duration assumption that makes the price model tractable.

This matters because Cipollone’s adverse scenario is built on the same assumption Fitch has just walked.

The $119 oil peak and €87 gas peak in Q2 2026, declining thereafter — that path requires the closure to behave the way Fitch previously assumed it would.

Once the closure runs five months instead of two, the adverse scenario stops being adverse. It starts being baseline. And severe stops being severe. It starts being adverse.

The ECB updates its scenarios in June. I would be very surprised if the architecture survives intact.

Tuesday’s tape

Yesterday’s tape belongs in the same paragraph as the scenarios, because tapes do not respect architectures.

• Brent settled around $107 on Tuesday.

• WTI around $99.

• TTF gas trading at €44-46/MWh after a 4.7% one-day move higher on Monday.

Aramco’s CEO told the Milken Conference last week the market is losing roughly 100 million barrels of supply per week and that prolonged disruption could push normalisation into next year.

President Trump said the ceasefire is on “massive life support” after rejecting Tehran’s latest peace proposal as “totally unacceptable.”

Reports surfaced of the White House weighing a return to military operations.

Netanyahu said the conflict with Iran is “not over.”

Iranian forces fired drones at the UAE. The US sank Iranian small boats in response. Fresh attacks near Qatar.

None of that lines up with a return-to-target inflation path that has Q3 2026 at 2.7% and 2027 at 2.0%.

The tape is pricing tail risk.

The architecture is pricing mean reversion.

Scenarios are a sound tool for bounding uncertainty around a process whose statistical structure you understand, and a poor tool for pricing political variables that produce binary outcomes.

The Hormuz closure is the second kind — either the ceasefire holds and the strait gradually reopens, in which case Fitch's July assumption stands and we converge to a path between baseline and adverse, or the ceasefire breaks, the US re-enters kinetic operations, and the closure becomes structural, in which case we are above severe with no return trajectory the scenarios contain.

The probability distribution over those two states is a coin flip with a thumb on the scale, and the thumb is in Tehran and Washington, not in Frankfurt.

The architecture shapes the variable

This is where reflexivity walks back into the room.

When the ECB publishes three scenarios that all share the same shape — peak, decline, return to target — it is doing more than analysis.

It is communicating an implicit constraint set, and the constraint set says: whatever the shock looks like, we will treat it as bounded.

That communication does work in the real economy.

• Firms set prices on the assumption that input costs will decline.

• Wage bargainers anchor on the assumption that the spike is temporary.

• Treasuries set discretionary spending profiles on the assumption that the deficit impulse is one-off.

• Banks tighten credit standards within the bounds of the projected downturn rather than the possible one.

The architecture shapes the variable it is trying to model.

This is the part Cipollone almost gets to. Late in the speech, he says that

"Our dependency on fossil fuels multiplies these effects”, and that "the current energy crisis thus underscores the pressing need to further reduce our reliance on fossil fuels not only because of climate risks but perhaps even more clearly because of energy security risks that are likely to be with us for some time."

That sentence is the entire moral case — decarbonisation as price stability infrastructure, sustainability as the precondition for stability.

But the scenario architecture undercuts that conclusion in the same speech.

• If the worst plausible outcome ends with declining energy prices in 2027 and a return to target by 2028, then fiscal authorities have no urgent reason to accept the political cost of accelerating the transition.

• The architecture tells them the shock will pass, the moral tells them they should act as if it will not, and both cannot be operative — the architecture wins, because the architecture is what gets quoted in finance ministry memos.

The ECB’s scenarios are not a passive measurement of the inflation outlook.

They are an active input into the political economy of the response, and the response, as a result, is asymmetric to what Cipollone is calling for.

Two hedges, one buried

Cipollone hedges the architecture in two places, and the hedges are worth reading next to each other.

The first hedge is the line that moved the front end —

“The current situation seems to be drifting away from our March baseline projections”

— and it is doing more than signalling a hike. It is saying that the scenario architecture itself, two months after publication, has lost its anchor variable.

The baseline is no longer the central tendency, and that is not normally how baselines work.

The second hedge is buried later in the speech.

“Europe could start running out of jet fuel and kerosene reserves by the end of May, potentially leading to material restrictions on the activity of several industries akin to those seen during the COVID-19 pandemic.”

That sentence is doing enormous work — it admits that the next eighteen days could produce a physical-input shortage equivalent to a pandemic-scale industrial restriction, which is not what the adverse scenario describes, let alone the baseline.

It is the severe scenario, with a different transmission mechanism, on a two-week horizon. Cipollone says it once, in passing, between paragraphs on inflation expectations and bank lending, and he does not return to it.

The architecture cannot accommodate a two-week physical shortage that operates outside the price channel, so the sentence sits there, unintegrated, as a footnote that contradicts the structure of the rest of the speech.

That is the most important sentence in the document.

The kerosene line is what tells you the speaker knows the architecture is not sufficient — the scenarios do not contain the possibility he raises, he raises it anyway, and then he goes back to the scenarios.

Where Cipollone is right, and where he isn’t

Cipollone is right about decarbonisation, and the structural case is correct.

Europe's fossil fuel dependence is a price stability problem, not just a climate problem, and the speech is at its strongest when it makes that argument. The data — 32% reduction in energy intensity since 2015, 48% of EU electricity from renewables, 23% from nuclear — is a real achievement and a real foundation, and the Italian 42% figure on AI labour-displacement fear, the labour-share contraction risk, the wage-compression channel — all of it is underpriced in the institutional discussion.

He is also right that monetary and fiscal responses to supply shocks are costly even when calibrated well, which is a true sentence and a brave one for a central banker to say out loud, and the implication — that the right answer is to remove the dependence, not to respond to the disruption — follows directly.

What he is wrong about is the framework for thinking about how the shock evolves. The scenarios are not three points on a probability distribution — they are three illustrative paths along the same shape, and the shape is the assumption. The shape is mean reversion of a politically determined variable, and that assumption is not in the data — it is in the architecture.

The right framework for the current moment is not three scenarios but two states.

• State one: the ceasefire holds, the strait gradually reopens through summer, Fitch’s revised path runs, and we land somewhere between the new baseline and the new adverse.

• State two: the ceasefire breaks, the closure becomes structural through year-end, and we are above severe with no return trajectory the model contains.

The probability split is a political variable, and the ECB does not price political variables — it uses scenarios as a substitute, and the substitute is failing.

Three windows

Three windows over the next eight weeks.

• Window one. Ceasefire status, day by day — Iran’s response to the one-page memorandum, US naval blockade posture, attacks on UAE/Qatar/Kuwait shipping. The single most important variable for every European asset price is whether the closure runs five months as Fitch now assumes, or eight months, or twelve.

• Window two. ECB June projections, where the architecture either updates or it doesn’t. If the new baseline incorporates a five-month closure as the central case and the adverse scenario moves to a structural-closure assumption, the ECB is catching the tape and the policy response becomes legible. If the baseline still assumes a one-to-two month closure with mean reversion, the gap between architecture and reality widens, and the rate path will be revised meeting by meeting in a way markets cannot front-run.

• Window three. Jet fuel and kerosene reserves at end-May. Cipollone flagged this as a potential trigger for pandemic-scale industrial restrictions, and if the headline materialises, the inflation expectations channel that has so far held — the line about long-term expectations being "well anchored" — comes under pressure no scenario in the speech describes.

Across all three, the same question.

The architecture assumes the variable mean-reverts, but the variable is not stochastic — it is political. The scenarios will get revised, and the revisions will look like the Fitch update: small in language, large in content.

The institutions catch the tape, they never lead it.

That is what last week was about. Cipollone published the framework, Fitch revised the assumption two days later, and the tape on Tuesday this week told you that even the revised assumption is contingent on a ceasefire the President of the United States describes as being on massive life support.

Three scenarios. One variable.

Is the variable not in the model?

Subscribe now

“He who controls the past controls the future. He who controls the present controls the past.” — George Orwell

A 15-metre missile model stood at the centre of Hall 3 of the Istanbul Expo Centre on Tuesday, 5 May. Anadolu’s wires called it Yıldırımhan — Lightning Khan, after an Ottoman sultan. The technical placard at its base listed a 6,000-kilometre range, a 3,000-kilogram payload, and four liquid-fuelled rocket engines burning nitrogen tetroxide.

Defence Minister Yaşar Güler stood beside the mock-up and told the cameras it would only ever be used as a deterrent. He added that nobody should have any doubt that Turkey would use it without hesitation if the deterrent failed.

Turkey does not have nuclear warheads.

Nuclear warheads are what 6,000-kilometre missiles normally carry.

The Yıldırımhan, as currently presented, is a conventional weapon designed to fly halfway around the planet on liquid propellant — a class of system that does not really exist anywhere else in the inventory of any other state.

That is the headline. The story underneath it is older.

For roughly seventy-five years, the West has read Turkey through a binary lens.

Either Ankara is inside the tent — NATO frontier state, customs-union counterparty, EU candidate, Incirlik host — or it is drifting out of it, towards Moscow, Beijing, the SCO, the BRICS, the Sunni-Islamic bloc nobody can quite name.

Whichever side of that read you sit on, the lens itself is wrong.

Turkey is not drifting. Turkey is not ambiguous.

What the binary lens misses is the toolkit.

Erdoğan does not operate from a single doctrine.

He operates from four — Turkish Islamism, Ottomanist civilisational thought, Davutoğlu’s central-state geometry, and the Mavi Vatan maritime frame — and he reaches into them selectively, as a composer reaches for instruments. Each tradition supplies vocabulary, symbols, and constraints.

None of them is the policy. The composite is the policy.

And the composite, once you read it correctly, has a name.

Ahmet Davutoğlu wrote it in 2001. Merkez ülke. Central state.

A country whose value is precisely that it sits at the seam between systems and refuses to be absorbed into either one — because the seam is where the rent is.

This article is about what that has looked like in operational terms over the last ten weeks.

• The Iran war broke out on 28 February.

• The Strait of Hormuz has been effectively closed for seventy-two days.

• Brent settled at $101.29 on Friday.

• The IEA estimates that 14 million barrels per day of global supply remain dislocated.

• The US naval blockade of Iranian ports entered its fifth week on Wednesday.

NATO will meet in Ankara on 7-8 July. Erdoğan is preparing his second Gulf swing of the year. The Saudi-Pakistan Strategic Mutual Defence Agreement, signed on 17 September last year, is being renegotiated to admit a third party.

And seven days ago, Turkey unveiled an intercontinental ballistic missile.

None of these things are independent. All of them are the same trade.

Let me show you.

The Author and His Tools

There is a temptation, when writing about Turkish foreign policy, to reach immediately for either ideology or pragmatism — to argue that Erdoğan is either a true-believing Islamist or a transactional opportunist, and to frame everything else as decoration.

Both readings are wrong.

Both readings miss the structure.

The Middle East Institute’s April backgrounder, written by Jonathan Breibart with Gönül Tol, gets at the structure cleanly. Erdoğan, the piece argues, operates inside a hybrid ideological toolkit — Turkish Islamism, Ottomanist nostalgia, Davutoğlu’s civilisational doctrine, the Mavi Vatan maritime frame, selective Westernism — and he draws from each tradition not as a believer but as a composer. The traditions supply vocabulary and symbols. He selects what fits the political imperative of the moment.

I want to push that argument one step further than the MEI piece does.

The composer’s freedom is not unbounded.

Erdoğan can reach into the Kısakürek tradition for a leader-centric populist register, into Erbakan’s Milli Görüş for the language of Muslim industrial self-reliance, into Davutoğlu for the merkez-ülke frame, into the Eurasianist-leaning Mavi Vatan for the encirclement narrative — but he cannot use them in any combination he likes.

The toolkit imposes a constraint structure.

• Erbakan’s self-reliance argument, once you commit to it, requires a defence-industrial base.

• Davutoğlu’s central-state argument, once you commit to it, requires that you not be absorbed by any single bloc.

• Mavi Vatan, once you commit to it, requires that you contest the maritime claims of Greece and Cyprus.

Each tradition the composer draws from binds him further.

This is what makes the toolkit argument interesting.

The output of Erdoğan’s coalition-management exercise — the AKP plus the ultranationalist MHP plus the Eurasianist networks embedded in the security services after 2016 — looks improvisational from the outside.

It is not → It is path-dependent.

It is the trajectory you get when the composer commits to four traditions at once and then has to honour the implicit promises each of them makes.

• Selçuk Bayraktar’s drone empire is Erbakan’s industrial self-reliance argument, monetised.

• The 2019 maritime delimitation deal with Libya’s Tripoli government is Mavi Vatan, executed.

• The Astana Format with Russia and Iran is Davutoğlu, executed.

• The Hamas dialogue, the Muslim Brotherhood backing in Egypt and Tunisia in 2011-13, the diplomatic pivot to Doha during the GCC blockade — all of it is Kısakürek’s authentic-Muslim-public framing operating in the foreign policy domain.

Each move is legible to one tradition; the composite is only legible if you read all four at once.

The reason this matters for what follows is simple.

When Western analysts argue that Erdoğan is bluffing — that the BRICS application is a signalling exercise, that the SCO dialogue partnership is performative, that the trilateral defence talks with Riyadh and Islamabad will not survive contact with NATO obligations — they are reading the moves through a single tradition.

Neither will.

The toolkit is the doctrine.

The composer is committed to all of it.

The Hybridity Premium

Jeffrey Mankoff and Max Bergmann’s CSIS brief from December — Strategic Ambiguity — frames Turkey as a hedging middle power that monetises its position at the seam.

I want to take that frame and price it.

A hedge has a cost.

Hedging premiums in any market exist because counterparties are willing to pay for the option not to commit.

• In a unipolar order, hedging Turkey-style does not work because there is only one sun, and gravity points in one direction.

• In a multipolar order, hedging works precisely because each pole pays for the option of not losing you to the other one.

That is the fee Turkey collects.

Not a discount on F-16s.

Not a Saudi central-bank deposit.

The fee is the structure of the relationship — the fact that Washington, Brussels, Moscow, Beijing, Riyadh, Tehran, Doha, Abu Dhabi, Cairo, and Jerusalem all maintain a working channel with Ankara because the cost of letting that channel close is higher than the cost of indulging the ambiguity.

Look at what each customer is paying.

• Washington is paying with F-16 modernisation kits, the Sweden-NATO accession trade Erdoğan extracted from Biden in early 2024, and Tom Barrack’s posture as ambassador — a posture that the Foundation for Defense of Democracies describes, accurately, as actively encouraging behaviour the United States would punish in any other ally.

• Brussels is paying with the migration deal originally cut in 2015 and quietly extended in every iteration since, plus the customs union nobody has the political appetite to terminate.

• Moscow is paying with discounted gas through TurkStream and Blue Stream — Russia currently supplies 40-42% of Turkey’s natural gas imports, anchored by extended contracts running through end-2026 — plus Rosatom’s $20 billion Akkuyu nuclear build, whose first reactor goes critical this year.

• Beijing is paying with $40 billion in trade volume, AIIB financing, and a tolerated silence on the Uyghur diaspora.

• Riyadh paid in February with two solar plants, $2 billion in investment, and a free-trade-agreement framework with the GCC.

  

• Doha pays with central-bank liquidity, the Al Udeid relationship, and the closest thing Erdoğan has to a personal patron in the Gulf.

• Tehran pays with conditional restraint on PKK-linked activity in northern Iraq and a tacit agreement to manage the Kurdish file as a regional condominium.

• Jerusalem pays — even now, even with relations near the worst point in fifty years — by maintaining the back-channel that the Mossad has not severed.

The fee is not symmetric. Some customers pay more than others.

The point is that all of them pay something, all of them keep paying, and the income stream is the strategic position itself.

• The customs union is not a contradiction with the BRICS application.

• The S-400 is not a contradiction with NATO’s Article 5.

• The Hamas dialogue is not a contradiction with the seat on Trump’s Gaza Board of Peace.

Each of these pairings is the product itself. The product is unambiguity for everyone except the person reading the contract from outside.

What Mankoff and Bergmann correctly note — and what every honest analyst since Davutoğlu has had to concede — is that the strategy has a structural ceiling.

Whether it holds through the next decade is the open question this piece will close on. But for now, on 11 May 2026, the hedge is paying.

The income is real — The customer list is intact — The toolkit is solvent.

The Iran war added a customer.

The War That Arrived

I have been writing about the Strait of Hormuz since the first week of March.

• The Same Room (3 May) walked through Bassam Fattouh and Andreas Economou’s OIES paper on the regime break.

• Kilian Picks Up the Pencil (17 April) framed the Dallas Fed inflation pricing of the closure scenarios.

• The Battle-Tested Stack (8 April) covered the AI-targeting architecture inside Operation Epic Fury.

None of those pieces were primarily about Turkey.

This one is.

Gönül Tol’s 16 April Middle East Institute commentary — Turkey Sees Short-Term Pain, Long-Term Gain in the Iran War — is the cleanest single document that anyone has written about how Ankara reads the war.

I want to engage with it directly, because the Tol piece identifies four openings Turkey is moving on, and I want to argue that all four are versions of the same trade — and that the trade is reflexive in the strict Soros sense.

Reflexivity, briefly, because it matters here.

Soros’s claim is that in markets with thinking participants, the participants’ perception of the fundamentals shapes the fundamentals themselves.

There is no equilibrium between perception and reality.

There is feedback.

Perceptions shift the world the perceiver is trying to perceive, and the world then shifts the perception. Bubbles inflate not because investors are wrong about prices but because investors believing prices will rise causes prices to rise.

The thinking participant changes the system by participating in it.

The Iran war is doing this to Turkey’s regional position right now.

The first opening Tol identifies is the energy-hub ambition.

Erdoğan has been pitching Istanbul as the gas crossroads of Eurasia for the better part of two decades.

• Until 28 February, the pitch was structurally hollow because the alternative routes — the Saudi East-West pipeline, the UAE’s ADCOP to Fujairah, the Qatari undersea grid — all bypassed Turkey, while the Strait of Hormuz handled twenty million barrels a day and there was no reason to need overland.

• The closure has rebuilt the case from the demand side. With Hormuz dark and the Gulf monarchies actively shopping for non-maritime crude routes, the Iraq Development Road from Basra and the Kirkuk-Ceyhan reactivation suddenly have customers willing to underwrite the financing.

• Energy Minister Bayraktar’s October pitch in Moscow about Azerbaijani gas through TANAP and TurkStream’s surge to 10 bcm in H1 2025 was forward-looking.

The post-28 February pitch is reactive — Turkey is the pipeline that already exists, and the Gulf needs pipelines.

The second opening is defence exports.

Tol notes that Saudi state defence firm SAMI is now in localisation talks with Baykar and Aselsan. Turkey’s largest-ever defence export deal, signed during Erdoğan’s 3 February Riyadh visit, was for Bayraktar Akıncı UCAVs and is being expanded to include warships, tanks, missile systems, and partnership in the Kaan fighter programme.

The UAE relationship is moving into joint projects between Baykar and EDGE.

Qatar is anchored by the Turkish base.

The Gulf customers are not shopping.

The Gulf customers are panicking — the Iran war demonstrated that NATO-grade Patriot systems and US-supplied F-15s did not deter Iranian missile strikes on Gulf cities, ports, and oil facilities, while the same Iranian salvo could not get through Turkey’s Hisar/Steel Dome envelope reinforced by NATO airspace.

The buyers are reading that delta as a marketing document.

The third opening is the Iraq Development Road — the multimodal freight corridor from Basra to Turkey to Europe, originally conceived as a counter to the US-backed India-Middle East-Europe Economic Corridor announced at the 2023 G20.

The IMEC bypasses Turkey by design.

The Iraq Development Road runs through Turkey by design.

The war has not closed IMEC, but it has made the Red Sea route look unattractive to insurers and shippers — which means the corridor competition is no longer between two viable routes; it is between one disrupted route and one Turkish-anchored alternative whose construction has been accelerated by both Iran’s weakening hand inside Iraq and the visible costs of maritime exposure.

The fourth opening is Istanbul as a financial centre.

Tol’s case here is the most speculative — that the war has shaken the Gulf monarchies’ carefully cultivated image as safe havens, and that fintech, Islamic finance, and insurance firms operating in Dubai are reconsidering Istanbul as a backup.

I am sceptical of this one.

Turkey’s politicised courts and erosion of the rule of law are real, and capital does not move into a venue where contract enforcement depends on the political mood of the executive. But the directional point holds: the Gulf’s branding as the region’s neutral financial hub took damage on 28 February, and Istanbul does not need to win the relocation race outright — it only needs to be in the conversation.

Now apply the reflexivity framework.

Turkey’s perception that the war creates a hub opportunity is itself causing capital, infrastructure, and diplomatic attention to flow toward Ankara.

• The flow is shifting the underlying economic and strategic geography.

• The shifted geography then justifies further perception of opportunity.

• The Saudi solar investment is signed because Erdoğan is positioning Turkey as energy-relevant. Turkey is positioning Turkey as energy-relevant because the Saudi investment is signed.

• The pipeline financing happens because the corridors are being built.

• The corridors are being built because the financing is happening.

This is the loop.

It does not stabilise.

It runs until something external breaks it — either a Hormuz reopening on terms favourable to the Gulf, in which case the case for Turkey collapses partially, or a settlement that leaves the disruption permanent, in which case the case becomes structural.

Brent at $101.29 on Friday, with the IEA estimating 14 mb/d disrupted, with the US blockade in week four, with Iran still reviewing the latest US memorandum on terms — none of those facts point to a near-term reset.

The reflexive loop has weeks of runway, possibly months.

What Erdoğan understood — and what most Western capitals did not — is that the moment the war started, Turkey stopped being one of several regional middle powers competing for relevance.

Turkey became the only one whose physical geography, infrastructure base, and political channels were positioned to absorb the displaced flows.

The opportunity did not come from Turkish strategy.

The opportunity came from Hormuz closing.

Turkish strategy is the response — and the response is creating its own gravity.

What Doesn’t Get Signed

The Middle East Eye reporting from 13 January, by Ragip Soylu and Levent Kemal, lays out the trilateral defence-pact discussions between Ankara, Riyadh, and Islamabad. Bloomberg confirmed in its 9 January story that Turkey was likely to join the Saudi-Pakistan Strategic Mutual Defence Agreement, which had been signed on 17 September last year and which contains language framing aggression against one party as aggression against all — language inevitably compared to NATO’s Article 5.

The MEMRI report on 21 January characterised the trilateral as advancing toward a formal “Islamic NATO” framework. Then, in the first week of February, Reuters carried the line that Turkey would not be joining the SMDA, and Erdoğan went to Riyadh on 3 February as if nothing had happened.

I want to argue that the most informative thing about this episode is not what got signed but what did not.

The SMDA in its bilateral form is a Saudi-Pakistani arrangement.

Pakistan’s Defence Minister Khawaja Asif told Geo TV on 3 October that the pact extends Pakistan’s nuclear umbrella to Saudi Arabia. Saudi officials have walked that claim back, then walked it forward again, then said it sits unmentioned in the text.

Both commitments are exclusive.

You cannot have a NATO-style mutual-defence umbrella and also be neutral.

Turkey, having participated in months of trilateral discussions, ultimately did not sign because signing would have collapsed the hybridity premium.

A binding mutual-defence commitment to Saudi Arabia and Pakistan would have foreclosed Turkey’s optionality with Iran, with the Houthis, with Egypt’s parallel defence-coalition initiative under the 1950 Joint Defence and Economic Cooperation Treaty, and — most importantly — with Greece, whose maritime claims sit at the centre of Mavi Vatan and whose defence relationship with India is precisely the Indian relationship Pakistan would expect Turkey to position against.

What Foreign Minister Hakan Fidan told TRT in mid-January is the operative line. “Our President Recep Tayyip Erdoğan’s vision is for an inclusive platform that creates wider, bigger cooperation and stability.”

Platform, not pact. Cooperation, not commitment. Inclusive, not aligned.

The vocabulary is exact.

Cinzia Bianco at the European Council on Foreign Relations read the substance correctly: what was being formalised was a triangular defence and security partnership in which Saudi Arabia would act as a conduit for Pakistani-Turkish cooperation on intelligence, defence-industrial work, and interoperability — without ever crystallising into Article 5-equivalent language.

That is the model Turkey will sign.

That is the model Turkey will not.

The same logic explains why Turkey expressed interest in Egypt’s proposed Arab defensive force under the 1950 treaty in September of last year, why the Gulf states (mainly Qatar and the UAE) blocked it, and why Cairo and Ankara have been quietly rebuilding their bilateral channel since. The blocked Egyptian initiative would have done what the SMDA does — formalise commitments, foreclose optionality. Turkey would have lost from joining it for the same reason it lost from joining the SMDA.

What Turkey will instead build, and what is now visibly under construction in Murat Yeşiltaş’s framing at the SETA Foundation, is a regional defence cooperation mechanism — multilateral, modular, bilaterally-stacked, where each participant retains its full optionality with non-participants and the architecture is held together by capability sharing and political signalling rather than binding promises.

This is the platform Fidan referred to.

This is what the Damascus war-room initiative in March was trying to be.

This is what the Erdoğan-Sisi-MBS-Sharif foreign ministers’ meeting on the sidelines of the OIC on 19 March in Riyadh was scoping. This is what the SETA Foundation calls “deepening military cooperation… not as a NATO-like pact, but as a defence cooperation mechanism that could eventually include Pakistan and Saudi Arabia.”

The American Thinker piece from earlier this month — written from a Washington-hawkish perspective — read this architecture as a Sunni Muslim NATO with Pakistani nuclear backing, a 500-million-person bloc that would reduce US leverage and force a counter-alignment with Israel, the UAE, Bahrain, Kuwait, Cyprus, Greece, and India.

That read is wrong in form but right in direction.

The architecture is not a bloc.

It is a platform.

What does not get signed is, in this case, the entire point. Signing would foreclose. Not signing keeps the option live and allows the platform to deepen.

This is hedging at the level of treaty architecture.

The platform is the tool.

The unsigned text is the feature.

Mach 25, Six Thousand Kilometres, Somalia

The IISS paper by Sıtkı Egeli and Arda Mevlütoğlu, published in late April, is the cleanest single document on Turkey’s missile programmes I have seen.

It walks the ladder from the Sakarya artillery rocket through the J-600T Yıldırım, the Bora and the Tayfun family, the Cenk MRBM, and the cruise-missile portfolio centred on the SOM, Atmaca, Akata, Çakır, and Gezgin.

The paper closes by noting that Turkey’s guided-weapons sector is now “nearly unrecognisable” from where it stood at the start of this century, and that further ballistic, cruise, and hypersonic propulsion R&D is under way.

Seven days after the paper went online, Defence Minister Güler walked into Hall 3 of the SAHA expo and unveiled the Yıldırımhan.

The IISS paper does not mention the Yıldırımhan.

The IISS paper closes its discussion of long-range ballistic systems with the Cenk programme — a notional 2,000-kilometre MRBM whose first flight test had not been confirmed at the time of writing. The Yıldırımhan jumps the ladder. Six thousand kilometres. Three thousand kilograms of payload. Mach 25. Liquid nitrogen tetroxide propellant, four-engine cluster, mock-up size suggesting the Cenk’s airframe scaled by roughly a factor of two.

Ankara has no domestic test range that can accommodate this missile.

Egeli and Mevlütoğlu flag this point in the paper — Turkey’s existing Black Sea ranges are bounded by less than 1,000 kilometres of clear arc to the western tip of its coast, which is why the Cenk programme has been delayed and why Ankara is investing $350 million in a Somalia spaceport with completion targeted for 2026-27. The Yıldırımhan, by Bloomberg’s 6 May reporting, will be test-fired from the Somali site starting in autumn. The state minister of foreign affairs in Mogadishu, Ali Mohamed Omar, characterised the partnership as primarily civilian — satellites, communications, climate monitoring, education. The Bloomberg sources characterised it as a missile range.

Both characterisations are correct. The technology overlaps exactly.

I want to argue four things about why this missile, at this moment, in this hall, matters.

The first is what it tells you about Erdoğan’s view of Iran.

Foreign Minister Fidan said in February that Turkey could be drawn into a regional nuclear competition because of concerns about Iran’s atomic ambitions.

He did not say Turkey would acquire nuclear weapons.

He said Turkey could be drawn in.

The Yıldırımhan is the conventional answer to that statement. A 6,000-kilometre conventionally armed missile with a heavy bunker-busting warhead is the kind of system that lets a non-nuclear state hold strategic targets at risk without actually building a bomb — the same logic that drove Iran’s own conventional ballistic missile programme over the last fifteen years.

The second is what it tells you about Erdoğan’s view of NATO.

The 7-8 July Ankara summit will be the second NATO summit Turkey has hosted, after Istanbul 2004. The 2025 Hague summit committed allies to the 5%-of-GDP defence-spending target by 2035.

Turkey will arrive in July as one of the few member states already within striking distance of that figure — its $10 billion 2025 defence-export number, the SIPRI ranking of 11th in global arms exports for the 2021-2025 period, and its position as the only non-US NATO member fielding a credible long-range conventional strike portfolio give Erdoğan a procurement-pitch presentation rather than a scolding-from-Brussels presentation.

The Yıldırımhan unveil two months before that summit is a stage prop.

It is the alliance’s strongest non-American long-range strike capability arriving on the host’s table.

The third is what it tells you about the F-35 question.

Ankara was ejected from the F-35 programme in 2019 over the S-400 procurement, and Tom Barrack’s pitch at the Antalya Diplomacy Forum — that the S-400 issue would soon be “resolved” and that bringing Turkey back into F-35 was “no problem” from Trump’s perspective — has been the leading edge of the Trump-administration position on the question.

The Foundation for Defense of Democracies has argued, correctly, that re-entry would be dangerous and that Turkey’s ties to Russia’s defence industry remain an unresolved liability.