Advinv 15 June Macro As A Pdf

565KB ∙ PDF file

Download

Download

Also, here is this week’s macro market dashboard as a spreadsheet:

Markets Dashboard 5th June 2026

49KB ∙ XLSX file

Download

Download

The week ahead - watch those central bankers

This is a blockbuster week for monetary policy, with major central banks deciding on interest rates alongside crucial inflation and retail data. Kevin Warsh delivers his first press conference as Fed chair o…

Read more

Advinv 11 June 26 Links As Pdf

1.06MB ∙ PDF file

Download

Download

I’m making this letter available to all readers thanks to sponsorship from AXA XL:

SPONSORED CONTENT

The future of energy depends on building a more flexible, resilient energy mix. Here’s what that means for risk.

As global electricity demand rises, businesses need to navigate an increasingly complex energy landscape shaped by geopolitical volatility, cyber risk, natural catastrophe, and the rapid growth of AI. In this piece, Vicky Roberts-Mills, Global Head of Energy Transition at AXA XL explores how insurance, risk management and specialist expertise can help businesses scale new technologies, strengthen infrastructure, and support the transition to a more secure energy future.

Read the article HERE

Big Stories in Robotics and AI

NVIDIA premieres Isaac GR00T Reference Humanoid Robot. NVIDIA has unveiled the Isaac GR00T Reference Humanoid Robot. This is the industry’s first completely open humanoid reference design built to unify a heavily fragmented hardware and software landscape. The reference design utilises a c.6-foot-tall, 150-pound Unitree H2 Plus chassis coupled with Sharpa Wave tactile five-finger hands. This gives the robot a total of 75 degrees of freedom across its body and hands, with peak arm payloads of 15kg. In terms of brain power, it’s powered by the onboard NVIDIA Jetson AGX Thor T5000 (utilising the Blackwell GPU architecture), delivering 2,070 FP4 teraflops of AI performance.

Source

World-First AI-Designed “Super-Antigen” Vaccine Enters Clinical Trials. Researchers at the University of Cambridge successfully developed and initiated human trials for a world-first AI-designed vaccine. Rather than targeting a single specific strain of a virus, researchers fed the genetic codes of various coronaviruses into an advanced AI model. The AI successfully analysed and synthesised a synthetic “super-antigen.” Source

Yet another dexterous humanoid robot. A French startup, Genesis AI, introduced an advanced model (GENE‑26.5) paired with a highly dexterous robotic hand capable of domestic and fine‑motor tasks such as chopping vegetables, cracking eggs, solving a Rubik’s Cube, and playing piano. Source. See video HERE

Seed Innovations makes first big investment in robotics space. Seed Innovations (AIM: SEED), an AIM-quoted investing company, has invested US$1,000,000 (c.£740,000) in Feather Robotics, Inc. “Feather is a US-based robotics company developing modular humanoid robotic systems designed for deployment in real-world industrial environments. Its approach differs from a number of existing providers in that its systems are designed for deployment rather than research, with a focus on cost efficiency and practical application. The platform is modular in design, allowing it to be adapted across a range of industrial processes, and is intended to address mobile and intermittent tasks that are not well-suited to traditional fixed automation. Feather is also positioning its technology as a general-purpose hardware platform for physical AI developers and integrators, supporting wider adoption and scalability”. Source

Factoids

Those IPOs. A useful summary box from the folks at Alpine Macro.

IPO flow. Analysts at HSBC reckon that total US IPO issuance, including mega caps, could reach over USD250bn this year, making it the heaviest year for primary US issuance since 2021. That said, they see sufficient demand to absorb these flows. US ETFs have taken in around USD390bn year-to-date, a monthly pace of USD80bn (source: Bloomberg). Corporate demand is also substantial, with buyback announcements totalling USD805bn year-to-date, 17% or USD120bn above 2025 levels.

The history of S&P 500 sell offs. Since 1928, the S&P 500 has suffered a daily loss of at least 2.5% within three trading days of a record high only 12 times ... Only one instance led to a decline exceeding 10% from the prior high—the COVID-19 selloff, which was driven by an unprecedented global economic shutdown.

Source: Turning Point Market Research

How much is enough to retire on? Advocacy group Pensions UK has mapped out what retirement could cost across three spending levels, based on what people think is needed for different lifestyles, with examples of what each level might look like in practice. It estimates that a moderate lifestyle costs £32,700 a year for a single person and £45,400 for a couple, but suggests only around 23% of the working-age population would be on track to reach this level. Rising bills have pushed up the cost of retirement, the group says, adding to calls for action to boost retirement savings. You can read the BBC coverage here

Who owns UK equities? A report by Shore Capital breaks down the structure of UK equity ownership:

• Overseas Investors (c59%): International funds, sovereign wealth funds, and global institutions now own the clear majority of the UK stock market, with beneficiaries in the United States holding the largest single share among foreign countries.

• UK Institutional Investors (c31%): Domestic financial institutions—including pension funds, insurance companies, and investment trusts—constitute the second-largest ownership block. However, this is a steep decline from historic levels.

• UK Individuals (c10%): Direct share ownership by individual UK residents has steadily declined since the 1960s, though many individuals still hold exposure to UK equities indirectly through mutual funds and pooled nominee accounts.

Sharing the wealth – or not, as the case may be. New research from Handelsbanken Wealth reveals that only 56% of parents would prefer to see their children and grandchildren benefit from inherited wealth during their lifetimes. Only a third (33%) believe those to whom they gift money would support them financially later in life. One in four (25%) worry that gifting could trigger family conflict, while 21% fear that wealth could leave the family altogether

Laos has gone full on net zero. The Sunday Briefing reports that Laos has banned imports of fuel-powered vehicles until the end of the year, to push the transition to electric. It currently spends nearly 10% of GDP on fuel imports – a macroeconomic nightmare – but has the geography to become a major hydropower producer. Ethiopia, another poor country with a big bill for fuel imports and a massive hydroelectric dam, banned the import of internal combustion engine vehicles in 2024.

Is the CAPE measure useful?

Boring, pointy-headed types like me keep referring to Robert Shiller’s CAPE measure to understand how overpriced US equities are, taking the long view. The problem is that, by this measure, US equities have been overvalued for years, and the measure itself has very little predictive power for indicating which way markets might go in the short to medium term. But maybe we need to look again at the CAPE measure, according to a bunch of academics, including Rui Ma at La Trobe University, Ben R. Marshall and Nuttawat Visaltanachotiat at Massey University, and Nhut H. Nguyen at Auckland University of Technology.

They argue that the widely used CAPE ratio that everyone watches has a quiet flaw baked into it. When you build the traditional version (the “Aggregate CAPE”), you’re effectively weighting each company’s valuation by its earnings rather than by its market value. Trouble is, market returns are value-weighted, so there’s a mismatch. The fix the authors propose is simple enough: first work out the CAPE ratio for each individual stock, then blend them together using market-cap weights. They call this the Component CAPE ratio.

Using S&P 500 data going back to 1964, they show this matters in practice. The Component version is always bigger than the standard one, and the gap widens whenever valuations get more spread out across companies, which is exactly what’s been happening lately as a handful of giant tech names dominate. More usefully, it forecasts long-run returns noticeably better. Their out-of-sample R-squared improves by roughly 10 to 14 percent, and the result holds up across loads of robustness checks, different earnings measures, and longer 15 and 20-year horizons. The improvement comes mainly from cutting systematic forecast distortion rather than just reducing noise. For an investor switching between equities and T-bills, it also delivered better risk-adjusted outcomes.

You can read the full paper HERE

Small AI models are closing the gap

One of the most interesting recent arguments I’ve heard about AI is that small, probably more specialised large language models, or LLMs, are beginning to catch up with the large, well-known models such as Anthropic and OpenAI. This has profound infrastructure implications. Maybe you don’t need to have vast mega data centres full of Nvidia chipsets but could run AI models even in the home or on local servers.

According to technology expert Paul Kedrosky, evidence is mounting that these small LLMs are really catching up. He observes that in 2022, reaching 60% on MMLU, a test of academic reasoning, required a 540-billion-parameter model. By 2024, a 3.8-billion-parameter model hit the same threshold — a 142-fold parameter reduction in two years. The direction is clear, and the implications are wide-ranging. Edge computing, for instance, becomes real, i.e., models that fit on a phone or a sensor don’t need a data centre. That also means that inference shifts from hyperscaler to device, which is a different supply chain and a different competitive moat than anyone has built for – and the data centre API business model has a half-life.

“If a fine-tuned 7B model running locally can handle 80% of enterprise use cases, the case for paying per-token to a cloud provider shrinks fast. The most exposed providers are those whose moats are model quality rather than distribution, i.e., the capex-heavy frontier models.

· Compute becomes less scarce.

· A country or institution that couldn’t staff or fund a frontier lab can now run a capable model on modest hardware.

· The geopolitical assumption that AI power concentrates in a handful of large actors/countries gets harder to sustain.

· Distillation is an asymmetric gift.

· Small models are catching up partly because they’re trained on large-model outputs. which means the convergence depends on continued access to frontier models as teachers.

· If that access gets restricted, whether by licensing, export control, or commercial terms, the gap may reopen a little, but it won’t return to historical levels.

· The convergence isn’t uniform or clean.

· Contaminated benchmarks, distillation from large models, and saturation effects all flatter the small-model numbers.

Pnamure Liberum’s strategist, Joachim Klement, has also picked up on this research, and he thinks the market implications are huge.

“…the [small language] models run on a desktop PC, and even more so, the ones run on a smartphone are much more energy efficient than the ones run in the cloud. The inference per Watt of these SLMs is typically seven times larger than that of LLMs. And that means that when you encounter a task that can be solved on a desktop or even a mobile device, it is cheaper to do so locally than send it to a data centre. This has important implications for investors, in my view:

• We need many fewer data centres than we think. If we can already replace 70% to 80% of the tasks that are expected to run on LLMs with SLMs, the hyperscalers have simply no revenue growth in the future that is nearly enough to justify the capex. In fact, if this research is true and this trend continues, data centres may be the worst investment in the AI space one can make right now.

• While we continue to need enormous investments in semiconductors of all sorts, we do not need to invest in the most advanced Nvidia chips. The cheaper ones that run on desktop PCs will be enough. The best case for Nvidia is that it can replace its high-end data centre GPUs with its new chips for desktop PCs. What will that do to Nvidia’s margins and revenue growth going forward?

• We still need LLMs for the most advanced tasks, and companies like OpenAI, Anthropic and others will be able to ‘dumb down’ their models to an SLM and sell them instead of LLMs. But given the already fierce competition from Chinese providers like QWEN or IBM’s GRANITE, the profit margins for these models will be much smaller than for LLMs. So what does that mean for the valuation of these companies in their planned IPOs and their growth trajectory?

• While agentic AI is still better on LLMs, this may only be a temporary advantage, similar to what we have seen in reasoning tasks and single chat requests. If that is the case, the true winners of the AI boom will not be the providers of advanced hardware and data centres but the boring manufacturers of desktop computers like Dell and Apple.”

HgCapital on AI

HgCapital held a capital markets day to discuss investment trusts, the rise of AI, and its impact on SaaS and enterprise software businesses. The slides that accompanied the capital markets day (CMD) can be found here. Hg Capital argues that its portfolio businesses are less vulnerable to AI than many maintain, and that in fact many will benefit from it. It also emphasised that the fund has significant expertise in AI and how it will impact portfolio businesses, especially around agentic AI.

Here’s a fab summary of the event from fund analysts at JPMorgan:

. Hg claims long-standing expertise in AI and says it is implementing agentic AI across all its portfolio companies with a focus on product-led approaches in areas where deterministic outcomes are business critical. Hg views AI as the next software super cycle and thinks it knows what is necessary for software companies to win in an AI world.

AI as the new software super cycle: Hg views the two prior super cycles for software as the PC/internet era from the 1980s to 1990s and the software-as-a-service (SaaS) era in the 2000s and 2010s. Hg has completed a SaaS business model migration for 40+ portfolio companies. AI is the 2020s super cycle and Hg thinks we are at an early stage of this cycle.

Strong exit trajectory over the next 12 months not reliant on public markets: Hg thinks AI capabilities and thorough preparation have positioned its portfolio companies well and it says it has a large number of potential realisation candidates, including companies in each of its last two fund vintages. Hg does not rely on public markets and last sold a portfolio company via an IPO in 2014, having since conducted all its exits via private markets.

Hg itself not new to AI with a 10-year history: Hg has already adopted GenAI across c.60 portfolio companies and completed two AI-led exits, GTreasury and intererad. Hg is not new to AI having worked with large language models (LLMs) and machine learning software for a decade already. Hg has what it calls hg_catalyst, an Hg team of 100+ software engineers who share best practice across Hg portfolio companies. This includes focused work to implement ideas into products, rapidly accelerating design work. To date in 2026, Hg says that AI-focused innovation has added $260m to budgeted EBITDA for its portfolio companies and that there are 1,600 projects live across its portfolio. Across 60 portfolio companies, 100+ agentic products have been launched. Hg trialled LLMs before they were launched to the public and has a long-standing agreement in place with Anthropic for use of its AI tools.

Focus on implementing agentic AI within mission critical software: Hg’s focus is on companies that provide what it considers to be mission critical software where failure to deliver outcomes could mean regulatory breaches. Within these companies Hg aims to implement enterprise solutions that become the system of record (SoR) driving dependency and sticky customer bases. A SoR is a system that stores and manages data, e.g. an authoritative database for business data. The software company product is a user interface, user experience and the features of the software. Users interact with the SoR to input or look up data. Hg sees AI as hugely increasing its total addressable market (TAM) through replacing a large proportion of the human cognitive work that these software products produce by making them into systems of action (SoA) using agentic AI. A SoA works in real time to execute tasks and generate responses. Now the software company product is measured in outcomes its product produces with the software doing work itself rather than just holding business data. This is achieved by having agentic AI prompt LLMs with context and tailored prompts to deliver the outcome within the software package embedded in the business.

How portfolio companies can win: Hg believes embedded software companies already have a proprietary data advantage from transaction and behavioural data that is domain-specific and unique to business verticals and regulatory systems. This data will have been obtained over years of customer relationships. Hg thinks its portfolio companies can deliver a sustainable competitive advantage in an AI world by creating software that tailors LLM model outputs that are relevant to a business domain, that improve through usage and are rigorously evaluated to ensure consistent outcomes. Hg continues to look for the same characteristics in target companies that it already did but now doubles down its focus on backing top founders/CEOs and focusing on product led strategies with AI-driven cloud computing. Revenue models need to evolve to be based on measuring outcomes, particularly as the old SaaS pricing model is based on number of users of the software and agentic AI may replace many of those users. Hg contrasts a PE generalist approach, which might look at using AI to cut some costs with a 1-3 year earnings growth contribution to its specialist product led approach with a 10-20 year earnings growth potential.

Offsetting LLM industry monetisation post huge capex spending: Companies and their shareholders will want a return on the large capital expenditure that has gone into data centres, computer chips and other related elements all focused on rolling out LLM products. This may lead to cost increases for LLMs including higher token rates and perhaps tiered pricing, where the latest and greatest models command a premium rate or are released only to premier clients. Hg thinks it can offset some of this potential impact by building products that can segment queries sent to LLMs with those that are less intensive, targeted at older, cheaper, LLM models. Hg also has an agreement in place with Anthropic that gives its portfolio companies a competitive rate for accessing LLM models.

AI spend is focused on coding and not deterministic workflows: Hg has analysed public reported data from AI native start-up companies and found the vast majority of their revenue comes from coding tasks, replacing work that would have been done by software engineers, and as yet there is little earnings coming from areas such as accounting and other deterministic workflows, the area where Hg portfolio companies operate. Hg thinks that much of what has caused the public software companies sell-off in 2026 is an information vacuum. It thinks that additional disclosures would show AI progress such as pull requests, token u, sage (acknowledging the ‘token maxing risk’), products in market and other quantifiable data points. Hg points out that as a private company owner, it has access to a vast set of granular data on its portfolio companies that gives it conviction in their strategies. The irony here is that investors in HGT, a public company, get very little information other than the brief snapshot of top-down information HGT chooses to share, along with the “glint in the eye of management” as a measure of sentiment, to quote Hg itself.”

UK taxes: no more other people to tax

One of my pet obsessions is UK voters complaining about taxes. As a country, we seem obsessed with a simple message: we want more and better public services, we know we need to raise taxes, but we already pay too much tax, so can someone else pay for this extra spending? And money grows on trees, and my uncle is Santa Claus. I grow especially vexed when average taxpayers bitterly complain about their huge tax burden, ignorant of the fact that, in European terms, middle-income taxpayers are some of the least taxed!

This all encourages a kind of infantilism in our political class, with talk of trade-offs avoided and mad schemes punted that will somehow square the circle. The truth, regardless of your political opinions or bias, is that

a) We can spend more if we tax everyone more

b) Or we can spend less, accept underfunded public services and then cut taxes.

c) But we cannot cut taxes AND have more public services because most of the tax loopholes that don’t affect the average taxpayer have been used.

Evidence for my last point, C, comes from a regular in these letters, Dan Neidle, of the Tax Policy Associates organisation, a think tank focused on tax fact-finding. Note, Dan is a Labour Party supporter, but I think his argument has very little to do with that membership. His point is that we’ve run out of road with the argument that we can tax other people to help fund the things we all want from public services – or at least that’s the argument he uses in a brilliant deep dive I thoroughly recommend at https://taxpolicy.org.uk/2026/06/05/taxing-other-people-uk/.

He starts by making the observation that polls tell us that the public does support tax rises, as long as those tax rises don’t affect them.

“In 2025, only one country raised more in tax than the UK without charging the average worker more in tax, and that was Iceland, with a population about the same as Cardiff. By 2027, every country that raises more tax than the UK will tax its average worker significantly more heavily than the UK. If instead we look at taxes on a range of wages, the UK stands out – particularly for low-paid workers, whose tax in the UK is significantly lower than for any comparable country:

Running out of road

The strategy is now simply insufficient to cover the fiscal pressures the UK faces as a result of demographic change and spending pressures (in defence and elsewhere). I expect that’s why the Government felt it had to increase employer national insurance in the October 2024 Budget. This was very much not a tax on “other people”. It raises significant sums but, in the long run, the economic incidence of employers’ national insurance falls on employees.²³²⁴

And continuing the present course inevitably means taxing the average person. Our data suggests that, in 2029, one in three full-time workers will start paying higher rate tax. Because earnings rise and fall over people’s careers — and because people move up and down the earnings distribution — around two-thirds of full-time workers will pay higher-rate tax at some point in their working lives.²⁵

And just tax the rich which might raise £38bn based on TUC proposal

To put all these numbers in context:

• Defence spending: The planned increase in defence spending to 3.5% of GDP by 2035 is estimated by the IFS at £36bn a year in today’s terms.

• Housing benefit: DWP forecasts spending of £37.3bn in 2025/26.

• Triple lock: The cumulative effect of the triple lock costs about £12bn each year.

• Disability benefit: The Resolution Foundation says non-pensioner disability and incapacity benefit spending has risen by £24bn since 2019/20, and the DWP forecasts it will reach £77.1bn by 2025/26.

• NHS: in a recent spending review, the government had to allocate a £29 billion real-terms increase to the NHS’s annual day-to-day budget by 2028–29 simply to keep pace with demand and inflation.

• Debt interest alone is forecast by the OBR to cost £110bn in 2025/26.

So the £38bn figure is mostly illusory but, even if it were real, it’s insufficient to keep pace with current spending demands. Equally illusory are the demands that Amazon and other tech companies can fill the fiscal gap.³⁵

The inconvenient truth is that, whilst the very wealthy are indeed very wealthy, there are not that many of them – so the amount of tax that you can raise is always going to be limited in fiscal terms. And it will never fund (say) German levels of public services.

Running out of road

Taxing the rich will always be politically popular, whether it’s justified as progressive redistribution or political expediency. There is a reason politicians keep promising that somebody else will pay.

The difficulty is that Britain has already spent a generation taxing somebody else, and by international standards we have been remarkably successful. The UK now raises a relatively large amount of tax while imposing a relatively low tax burden on ordinary wages. But this has run its course.

A generation of frozen thresholds means that they are no longer just a tax increase for “other people”; the average worker can expect to be directly affected. The popular new ways of taxing “other people” – targeting billionaires and banks – raise very limited sums in the context of UK public spending. And that ominous uptick at the end of the UK “tax wedge” chart matches the uptick at the end of the “British attitudes” chart.

All this is, in part, a failure of the modern Left. The traditional social-democratic argument was never that we can fund a decent society by finding a few unpopular groups and taxing them ever harder. It was that good public services require broad-based taxation, and that people should be willing to pay for the civilisation they want to live in. Most of the contemporary Left has abandoned that argument. It still wants a European-sized state, but sells it with the fantasy that someone else can pick up the bill.

This has been mirrored by a failure of the Right. It is still just as eloquent at denouncing the tax burden, but has become unwilling to identify spending cuts large enough to make a serious difference. It will argue for welfare cuts, but rarely identifies specific reforms that save more than a tiny fraction of overall spending. And it is generally unwilling to touch the biggest and most popular areas of spending: pensions, the NHS, social care and defence. So the Right ends up with its own version of the same evasion.

And so we end up with political failure. The Left says we can have European spending without European taxes because “the rich” will pay – and so, when it finds itself in power, presides over deteriorating public services. The Right says we can have lower taxes without confronting the spending programmes voters most want to protect – and so, when in power, has raised tax.

Britain may have a viable path in either direction: higher broad-based taxes to fund a larger state, or lower spending to sustain a lower-tax economy. I don’t know which would be more successful, but I know that we cannot have both.

Politicians can keep pretending there are pain-free answers: cutting “waste” on one side, taxing “other people” on the other. That may fool the voters and win elections, but sooner or later, we have to choose. Higher taxes that most people will pay? Or spending cuts that most people will feel? The strategy of taxing “other people” has run out of road.”

Quite.

Does smaller government help boost growth?

The USA and its various states (and to a lesser degree the European Union) offer economists a valuable real-world experiment: do states with smaller state governments, looser regulations and smaller tax takes grow faster than their peers? This is the subject of a cracking little paper by analysts at Eurizon SLJ Capital, by Stephne Jen and Joana Freire. Note : their analysis is purely an economic one, and there may be other reasons why states might choose to have a bigger government and tax more. Its not purely an economic decision and political considerations are crucial.

“We examined the 50 US states and computed the relationship between the size of the state government’s budget – adjusted for the size of the state – and the state’s economic growth and employment growth: by looking at the blue and red states in the US, we control for country differences to bring into sharper relief the issue we want to focus on.

The first chart below shows that the states with the lowest government spending (20th percentile) have grown an average +17% (cumulative) since 2019, compared to +11% for the states with the highest government spending (80th percentile). Unsurprisingly, the states with smaller government budgets also enjoyed employment growth (+2.4%), almost double that of the states with bigger government budgets (+1.3%).

In sum, based on these observations, all else equal, small government seems better for economic growth and employment generation. This should not come as a surprise, as less government crowding out, intervention, taxation and regulations are consistent with a more vibrant private sector.

Source: Datastream and Eurizon SLJ

The chart above splits up the US into red and blue states. Unsurprisingly, red states tend to have higher economic and employment growth, because they also tend to have smaller government policies. While social and equity considerations should be included in this discussion, from a purely economic perspective, it seems that red states with small-government policies tend to grow faster and have had greater job creation than the blue states.

We conducted the same calculations for the 27 EU members, ranking them by government spending size and splitting them into quintiles. When we compare those in the first quintile with the fifth quintile, we see even bigger divergences in economic and employment growth than among US states. Specifically, since 2019, countries in the lowest quintile (small government) have enjoyed an average 27% cumulative GDP growth and 10.5% employment growth, while those in the top quintile (big government) have experienced 10.5% economic growth and 4.4% employment growth.

Source: Datastream and Eurizon SLJ (countries in Europe’s 20th percentile of government size include countries like Ireland, Malta, and Switzerland)

More HERE

What’s happening in rare earth elements markets?

It’s always worth keeping a watchful eye on the rare earths market. Here’s a useful summary from specialist fund managers Westbeck Capital, who run the Volta fund, which invests across the battery and net zero space.

“In May, China’s export-licensing regime started impacting the rationing of real production. Shin-Etsu, the largest rare-earth magnet producer outside China, reportedly stopped taking customer orders for any magnet containing dysprosium, unable to source heavy rare-earths after China extended its export restrictions to Japan in January. The arithmetic behind the squeeze is stark: of a ~300kt global permanent-magnet market, only ~20–25kt is produced outside China (~93% dependence). Since that ex-China sliver is overwhelmingly Japanese, the feedstock wall Japan just hit is effectively the entire ex-China industry’s.

Dysprosium oxide in Rotterdam ended May at ~$1,185/kg, >6x the China domestic price (~$185/kg), a multiple that has nearly doubled from ~3.5x a year ago and is widening from both ends: the ex-China price steps higher while China’s domestic price drifts lower. Two details matter for the read. The staircase shape of the Rotterdam line reflects discrete repricing events as licensed material grows scarcer, i.e. each step is someone paying up, not a market tick. And the falling China-local line means there is no global dysprosium shortage; there is a shortage outside the wall. Scarcity here is manufactured by policy.

Shin-Etsu’s halt is, in effect, a pull-forward for our portfolio name. NEO Performance Materials leads the race in heavy-rare-earth-free traction magnets. Its chemistry simply does not contain the input being rationed. Stranded ex-China magnet demand is being forced toward HRE-free chemistry, and NEO is one of the furthest down that road. “

More from Westbeck here

In my various journeys around the outer reaches of the investment world, I spent some time last week loitering around an ETF issuer called Leverage Shares. This is one of the smaller issues, focusing on short and leveraged products, as well as income products that use covered calls to generate options inco…

Read more

I’ve given my markets dashboard a spring clean and automated it within an AI engine – it features virtually all the same measures as before, with a simpler structure, courtesy of Claude. I’ve also hand checked all the measures to make sure there’s no hallucinations!!

You can download the spreadsheet here:

Global Markets Dashboard Final 6th June 2026

44.3KB ∙ XLSX file

Download

Download

Timing for t…

Read more

Advinv 7 June Links As A Pdf

1.33MB ∙ PDF file

Download

Download

I’m making this letter available to all readers thanks to sponsorship from AXA XL:

SPONSORED CONTENT

The future of global investment depends on managing political risk. Are you ready?

Political Risk, Credit & Bond insurance may be little known, but it plays a critical role in enabling investment, trade and infrastructure development across the world. In this piece, Rafael Docavo-Malvezzi from AXA XL explains how this market supports projects that improve energy access, water security and housing, and why managing geopolitical risk is key to driving sustainable growth.

To read the AXA XL article – click here

Big Stories in Robotics and AI

NVIDIA Cosmos 3: The World’s First Fully Open “Omnimodel” for Physical AI. NVIDIA has officially launched Cosmos 3, a frontier open-world foundation model designed to give robots, autonomous vehicles, and vision agents a “digital mind” to understand reality before interacting with it. Cosmos 3 is the world’s first fully open omnimodel that natively unifies vision reasoning, world simulation, and action prediction into a single system. Unlike prior models that required separate AI stacks for seeing, processing, and moving, Cosmos 3 processes text, images, video, ambient sound, and physical actions simultaneously. It utilizes a breakthrough Mixture-of-Transformers (MoT) architecture. It pairs a reasoning transformer with an expert generation transformer, allowing a robot to genuinely understand spatiotemporal relationships, motion, and physics with high accuracy. Source

“The big bang of physical AI is just around the corner... The Cosmos 3 family of open, frontier omnimodels gives developers a generational leap in ability to build robots that perceive, reason, plan, and act in the physical world.” Jensen Huang, Founder and CEO of NVIDIA

Factoids

US corporate earnings are still surging, led by the semiconductor sector.

Solar has spared Europe €10 billion in additional gas imports since the war in Iran began. In March alone, solar cut import costs by around €110 million a day as gas prices doubled. This is a big deal, representing real cost savings that support the new global paradigm for energy security. Renewables don’t just replace fossil fuels anymore; they now protect economies when energy markets get upended. SolarPower Europe

AVERAGE INVESTMENT TRUST DISCOUNT HITS SINGLE DIGITS. - Narrowest discount since 2022 for investment trusts. According to the industry association, the AIC, “the average investment trust discount has reached single digits for the first time in nearly four years. The average discount on 31 May 2026 was 9.6%¹, the first month-end since August 2022, when it had been less than 10%. The discount at the end of August 2022 was 9.4%. “

Monumental CAPE measure – usually nothing good happens if these levels. At today’s Shiller CAPE of 39 for the U.S. market, an analysis found no historical period since 1881 that was followed by attractive long-term real returns. Link

Never retiring, ever? According to the People’s Pension, the UK’s largest workplace pension provider, 2.2 million young adults – one in eight (12%) – feel engaging with their pension is pointless because they don’t think they’ll ever be able to retire. A third (36%) of young people say the financial services industry fails to communicate the benefits of saving for retirement well. Of these, one in five (20%) say financial services companies make pensions feel boring and irrelevant. As a result of the industry failing to connect with them, nearly half (47%) of Gen Zs admit they are not engaged with their pension

What UK investors are buying? Interactive Investor has just released its survey of the most-bought investments in May 2026 by its customers.

“Global funds continue to catch the eyes of investors, and AI demand grows. Global funds continue to be in demand in the passive space, as the Vanguard FTSE All-World UCITS ETF takes the top spot for passive funds in May. A mix of money market, global income, and technology-focused funds dominate the list of most-bought active funds. Micron Technology tops the list for the most-bought stocks list in May, followed by Nvidia.”

Optimistic investors. The average investor holds a record-high allocation to equities, and stock market valuations (a measure of investor confidence) are near all-time highs.

Iran: The stolen revolution

Quite the most depressing aspect of the Iran war is that, despite the stated intention of the US administration, its corrupt, quasi-militaristic, quasi-clerico fascist rulers have not been removed. In fact, they may even be more powerful and have sanctions imminent removal. Trying to understand what’s going on in Iran is incredibly difficult but Lawrence and Sam Freedman in their Comment is Freed newsletter have a stab at it by interviewing Bozorgmehr Sharafedin and Yeganeh Torbati, authors of a brilliant new book, Stolen Revolution: Betrayal and Hope in Modern Iran. It covers the country’s history since the 1978-79 revolution through the lives of six very different characters. It’s well worth a read, and confirms my deeply depressing observation: the guys ruling the place are stronger than ever!

Sam: So does the war that’s going on at the moment make it more likely that the regime will hang on for at least a good while longer, because it gives them more opportunity for repression and unites more people behind a nationalistic, anti-American, position? Or because it makes the economic position even more perilous, does it make the regime weaker and more likely that it’s going to fall sooner?

Bozorgmehr: The war has made the Islamic Republic weaker than before. But that doesn’t mean the regime is close to collapse anytime soon. We should not underestimate the resilience of totalitarian regimes. Economic hardship is one reason for revolt, but we’ve seen many examples, including in Iraq after the Gulf War, where people are very hungry, but Saddam had a grip of power. That didn’t lead to the fall of Saddam, despite all the sanctions and all the economic troubles.

I have no doubt that Iranians will stand up, they will try to push for change. A big part of our book is that, despite all the obstacles, the Iranians try to innovate and reinvent themselves and find new methods to stand up against the Islamic Republic and voice their demands. We will see a more energized democratic movement, and people will push for change after the war, but it’s very difficult to see how that will be translated into tangible changes in the near future.

Lawrence: One of the ways regimes like this fall is because of divisions within the elite, and there’s been some talk of that. The first layers of leaders have been killed. But you also get a sense that the Guards have just used the war to consolidate its power, and those actually trying to run the government, including the President, are a bit marginalized at the moment. Is that fair? At some point you have to govern the country. You can’t keep the internet off forever, and there’s not very much income coming in at the moment. I don’t think the Strait of Hormuz will provide a saviour in that regard. So are there any signs of divisions within the leadership, or are they all so much in it together that all they can imagine is clinging on to power as ruthlessly as they’ve done in recent years?

Yeganeh: I don’t see any signs of significant division within the Iranian elite. There’s always some contestation, and so there will be people who disagree with each other and maybe criticize, but only a really unified regime can carry out what they did in January. There were no defections. There was no one that we know of saying, “I’m not going to do this”. If there was it clearly wasn’t an obstacle. It’s true the elected government is more marginalized. The guards are more ascendant, although that’s just an acceleration of a long trend. But I don’t see significant divisions, and I think the war is helpful for the regime in that regard, because there is a larger enemy to worry about.

Bozorgmehr: The Guards are running the country and are in full control of the government and the negotiations. In that sense, I agree that the Islamic Republic is much more unified when it comes to governance, but there are huge obstacles ahead. For example, to give a sense of continuity, they made much about appointing Mojtaba Khamenei as the new Supreme Leader. But that will come with a cost. You can’t appoint the son of the previous Supreme Leader in a system that claims that it’s against hereditary forms of governance without pushback.

Also you can’t run a country from a bunker. Hassan Nasrallah, the former leader of Hezbollah, was living in hiding for years, but at least we could see some video broadcasts from him. We don’t see any, even audio messages, from Mojtaba. Nasrallah had established legitimacy before going underground. Mojtaba has started his leadership in the underground. He doesn’t have any legitimacy. These are the big challenges for the new leadership. Despite all the unity, it’s going to be very difficult for them to run the everyday affairs of the country and to keep their support base connected to them.

Lawrence: We’ve talked a lot about managing the Iranian economy, the politics, society. Another key feature of the regime has been the axis of resistance, Hamas, Hezbollah the Houthis, Assad’s Syria. As of last year that whole project had almost fallen apart. It has been seen as a complete waste of money in some respects, yet it’s part of the demands of the current regime in the negotiations that this can continue. Do you see a similar effort being funneled into Hezbollah and Hamas in the future, has this strategy run its course?

Bozorgmehr: I think even the Islamic Republic can see that was a bad investment. They invested a lot in the proxies as a deterrent, but when it came to the moment of the war, those proxies didn’t help. They were one of the main reasons why the war happened. This is why the Islamic Republic is so fascinated with the control of the Strait of Hormuz. The regime thinks it has discovered a new deterrent.

Sam: Has the unity of the elite been helped by the fact that there’s been so much emigration from Iran of educated people who in other countries have been so critical to overthrowing totalitarianism. Are there enough people left in the country to lead an opposition, or would people come back if there was a real moment of opportunity?

Yeganeh: There’s such a hunger among the Iranian diaspora to come back and rebuild the country. To help it catch up to where it should be economically and its integration with the world, I think that there’s an enormous store of energy there. But every time there’s been a possibility of change, those people have been under great threat from the government itself.

Bozorgmehr: We have had the waves of immigration in the history of the Islamic Republic. The first wave started at the beginning of the revolution, and those who were absolutely shocked with an Islamist regime left the country. The second wave of immigration was during the Iran-Iraq war, people who didn’t want their children to go to the battlefield started sending them abroad in order to not go to the military service. During the Reform movement and Green Movement, we had a new immigration wave by people who gave up hope for any change.

Now we see immigration by the children of the ruling elite as well, so that is another sign of collapse. We saw this with the Soviet Union, and it’s one of the signs that the current Russian government is weak. Because when even the children of the ruling elite do not believe in this idea that their parents are preaching, that shows the foundations are becoming weaker and weaker.

More HERE

Software businesses are not dead yet

I’m not remotely convinced that AI means the death of Saas businesses or enterprise software. I can see many of us using Claude to build our own tools, but the embedded nature of much Saas and corporate culture, plus the interconnectedness of much Saas and traditional software, suggests to me an alternative, augmentative future.

And as a recent note by market analysts at French investment firm SocGen argues, the market seems to have taken these AI concerns in its stride. Three months after the sell-off, the EPS growth cycle for the sector continues to strengthen, tracking +35% growth.

“Moreover, the software EPS revision ratio (a breadth measure of EPS) has surged to an 8-year high. Along with the strongest EPS season in 5 years, even software stocks have seen a surge in upgrades, with 24 of the 30 US software stocks seeing higher ’26 and ’27 EPS revisions”.

That said, some enterprise software vendors might be in trouble, and on that score, it’s worth reading an excellent Sparkline Capital report by Kai Wu: AI Disruption Moats and Value Traps.

Wu argues that traditional value metrics struggle when applied to firms facing disruption. Long-term winners of disruption often possess key intangible assets, both technological and complementary, not captured by traditional metrics. Intangible value, which incorporates these moats, has been effective at separating disruption’s winners from value traps. Applied to the software selloff, this framework reveals not a blanket bargain but a dispersion-rich stockpicking opportunity.

The report argues that the current AI-driven sell-off in software is not a simple buying opportunity but rather a classic disruption environment in which traditional value metrics are prone to misfire. Many software names now look optically cheap on earnings and cash flow, yet history shows that disrupted incumbents often appeared “cheap” right up until their business models broke, turning into value traps like Blockbuster and Borders.

Using a patent- and text-based disruption framework, the paper shows that waves of technology (internet, e‑commerce, cloud, now AI) consistently hurt the performance of traditional value in exposed sectors, even while value continues to work fine in insulated, less-disrupted industries. The recent underperformance of value is framed less as “value is dead” and more as “value is short disruption”: it keeps selling future winners like Amazon and buying doomed incumbents.

The proposed fix is “intangible value” – a value factor that explicitly accounts for intellectual property, brand, human capital, and network effects, not just balance‑sheet tangibles. This measure aligns with Teece’s complementary assets idea: the real edge is not just the core tech but the surrounding capabilities (distribution, ecosystem, trust, regulatory footing) that enable a firm to actually profit from an innovation. Backtests suggest that intangible value works in both insulated and technology-exposed sectors, and particularly well in “disruption scare stocks” – companies down 30%+ on disruption fears, where future returns are highly dispersed.

Applied to the current software selloff, the framework finds a mixed picture: the average beaten‑up stock looks reasonably attractive, but there is a fat left tail of genuine value traps and a smaller set of high‑quality survivors, often enterprise software names with strong brands, embedded workflows, and aggressive AI adoption. The “sweet spot” is software firms that both score well on complementary moats and are early AI adopters, as these are most likely to defend or grow their franchise in an AI world.

UK pessimism is overdone

Not everyone on Wall Street shares the gloom about the UK that is so prevalent at home. Global strategists at Morgan Stanley had this to say in a note earlier this week: “

“…. in the UK we think pessimism may be overstated. Rates have been rising and it’s fair to say that economic sentiment feels poor. Yet listen to this comment from Mark Allen, CEO of Landsec, one of the largest UK commercial real estate operators: “Occupancy is now up to a 20-year high and rents are growing at their fastest pace in nearly two decades.

Indeed, what if the UK is doing a bit better than the downbeat narrative suggests? 1Q growth was solid (+2.4% annualized), Manufacturing PMI is >50, productivity has surprised to the upside, the country’s fiscal trajectory is among the best in the G7 (really), and we remain materially below consensus on inflation versus market pricing. With policy already restrictive, we think the Bank of England will keep rates on hold rather than hiking as markets suggest.”

US consumer pessimism increases

The disconnect between how voters and consumers ‘feel’ and what markets and economists think (generally positive) seems to be growing, even in the US, where Bloomberg (subscription required) notes that the Michigan Consumer Sentiment Index hit a record low in May, falling ~10% month over month. Consumers foresee prices advancing 4.8% over the next year. Inflation and high gas prices have long been major causes of sentiment drops.

The University of Michigan’s final May sentiment index decreased 5 points to 44.8 from April. The gauge was weaker than all projections in a Bloomberg survey of economists, as well as the preliminary reading of 48.2. Consumers expect prices to rise an annualized 3.9% over the next five to 10 years, up from 3.5% in April and the highest in seven months. They also saw costs advancing 4.8% over the next year.

What are the gold technical measures telling us?

Plenty of us are trying to figure out what might happen next to gold prices – my own view is that we might see the range between $4000 and $4500 tested, but that, longer term, the outlook is positive because of fiscal debasement.

Analysts and fund manager Charlie Morris at ByteTree’s Atlas Pulse has dug into the technical measures and shares my optimism over the medium term, although he concedes gold is in a short-term correction phase:

“Shown over five years, it is easy to forget how gold traded below $2,000 pre-2024, reaching $5,500 this year, which is one hell of a move. For the medium-term trend, the 200-week MA (green) is upward-sloping, as is the 50-week MA (purple). The price is clearly below the 50-week MA but is yet to move below the 20-week min (red). The max remains at the all-time high, which was set 13 weeks ago. As a result, with the only chart fault seeing the price below the 30-week MA, the weekly ByteTrend Score is 4.

Gold Weekly ByteTrend Score – since 2021

Source: Bloomberg

The Long-Term Trend

The long-term trend, as measured by the 200-month MA, the 30-month MA, and the 20-week max min, remains in an uptrend. I have taken it back to 1970 and shown it in log, which makes the 1970s stand out. The monthly ByteTrend score is a firm 5. For that to downgrade, the price would need to fall below $3,326, where the 50-month MA resides.

Gold Monthly ByteTrend Score – since 1970

June

Source: Bloomberg

Outlook

I emphasise the short-, medium- and long-term trends to put things into perspective. Gold is in a correction within a medium- and long-term uptrend. We have been here before in 1974, 1980, and 2011. In the first instance, in 1974, gold soon recovered in an inflationary environment. In the early 1980s, gold peaked, not to make a new high for 27 years. In 2011, gold took nine years to make a new high.

It’s anyone’s guess what happens next, but the technical picture remains constructive. A fall below the 200-day MA would be unhelpful, but there is strong underlying support for gold during these uncertain times.

There’s one more thing worth mentioning. Gold’s correlation with the oil price comes and goes, often randomly. But currently, it is deeply negatively correlated. When oil goes up, gold goes down, making it a very good time to own both assets.”

Subscribe HERE

Advinv 4 June Ideas As Pdf

427KB ∙ PDF file

Download

Download

All in on the AI revolution: Manchester and London Investment Trust

Something extraordinary happened recently – Alphabet raised $80 billion in extra capital – presumably to finance its AI ambitions – and barely any of us noticed. The full $80 billion is less than 2% of Alphabet’s mammoth $4.6 trillion market cap…

Read more

Advinv 1 June 26 Monday Macro As Pdf

439KB ∙ PDF file

Download

Download

The Week Ahead: watching Broadcom and US labour data

The S&P 500 hit another record high on Friday, closing at 7,580.05. Surging corporate earnings and lower oil prices helped. Brent crude settled at $92.05 a barrel, the lowest weekly close since April 17. The odds that the US blockade of the Strait of Hormuz will be lifted by June 3…

Read more

Advinv 30 May 26 Links As Pdf

742KB ∙ PDF file

Download

Download

I’m making this letter available to all readers thanks to sponsorship from AXA XL:

SPONSORED CONTENT

The future of global investment depends on managing political risk. Are you ready?

Political Risk, Credit & Bond insurance may be little known, but it plays a critical role in enabling investment, trade and infr…

Read more

Advinv 27 May Ideas As A Pdf

529KB ∙ PDF file

Download

Download

Worth researching: the Algebris Financial Income fund

Something a little different this week, and focused on those investors looking for

a) A strategy that uses equities to generate both capital gain and some dividend income, while

b) mixing in some fixed income or credit exposure, all the while

c) focusing on a mo…

Read more

Advinv 23 May 26 Links As Pdf

1.16MB ∙ PDF file

Download

Download

I’m making this letter available to all readers thanks to sponsorship from AXA XL:

SPONSORED CONTENT

The future of global investment depends on managing political risk. Are you ready?

Political Risk, Credit & Bond insurance may be little known, but it plays a critical role in enabling investment, trade and infr…

Read more

Advinv 21 May 26 Ideas As Pdf

323KB ∙ PDF file

Download

Download

Shorting Oil? A mad idea?

At the moment, there’s a dominant market narrative suggesting that oil prices will continue to rise, possibly pushing well above the current $110 to a peak of $150 or even, whisper it, $200. The driver is obvious, and the impact is obvious, not least on Western inflation and EM fuel-dep…

Read more

Advinv 18 May 26 Monday Macro As Pdf

4.21MB ∙ PDF file

Download

Download

The week ahead

This week features critical inflation indicators in the UK, the highly anticipated FOMC minutes from the US Federal Reserve, and a massive wave of retail and tech earnings headlined by AI bellwether Nvidia. That said, the week ahead is actually very light on US data.

April’s FOMC minutes will b…

Read more

Advinv 16 May 26 Links As Pdf

1MB ∙ PDF file

Download

Download

I’m making this letter available to all readers thanks to sponsorship from AXA XL:

SPONSORED CONTENT

The future of global investment depends on managing political risk. Are you ready?

Political Risk, Credit & Bond insurance may be little known, but it plays a critical role in enabling investment, trade and infr…

Read more

Advinv 13 May Ideas As Pdf

575KB ∙ PDF file

Download

Download

Later in this letter, I’ll update readers on two of my core buys in the investment trust space – the listed VC fund Chrysalis, which reported disappointing NAV numbers, and the steady compounder and care home specialist Target Healhtcare, which produced solid numbers and remains a core holding, as does Chrysalis.

But first, …

Read more

Advinv 11 May 26 As Pdf

467KB ∙ PDF file

Download

Download

The week ahead

This week is another big week for the markets, dominated by the US April CPI report—a critical data point that could redefine interest rate expectations, especially as the measure is expected to show upward pressure from fuel costs. The big question is whether this “bleeds” into core services.

There’s a…

Read more

Advinv 9 May 2026 As Pdf

1.23MB ∙ PDF file

Download

Download

I’m making this letter available to all readers thanks to sponsorship from AXA XL:

SPONSORED CONTENT

The future of global investment depends on managing political risk. Are you ready?

Political Risk, Credit & Bond insurance may be little known, but it plays a critical role in enabling investment, trade and infr…

Read more

Advinv 7 May 26 As Pdf

268KB ∙ PDF file

Download

Download

This week, I’m returning to one of my favourite asset allocation themes of the last few years: Latin American equities. I first discussed the LatAm opportunity in detail in this deep dive note back in May 2025. At the time I wrote:

“Historically, there has been a very good reason why Latin America has been so under-o…

Read more

The week ahead

The highlight of the week is likely to be the April US Employment Situation report on Friday, which remains the primary gauge of Fed policy. According to Yardeni Research, “The unemployment rate is likely to tick down to 4.2% based on its relationship with initial unemployment claims, which have been falling i…

Read more

Advinv 1 May As Pdf

748KB ∙ PDF file

Download

Download

I’m making this letter available to all readers thanks to sponsorship from AXA XL:

SPONSORED CONTENT

The future of global investment depends on managing political risk. Are you ready?

Political Risk, Credit & Bond insurance may be little known, but it plays a critical role in enabling investment, trade and infr…

Read more

Advinv 29 April As Pdf

537KB ∙ PDF file

Download

Download

LSE makes buying corporate bonds easier

After the shambles that was the retail order book for bonds, or ORB, the LSE Group has finally realised that there’s real demand from private investors for mainstream corporate bonds on a standalone, bond-by-bond basis. The idea of the LSEG’s new initiative is simple. Take existing …

Read more