Neura Robotics raises up to $1.4B Series C at $7B valuation
→ Tether, Qualcomm, Amazon, NVIDIA, Bosch, Schaeffler, EIB participate with conditional milestones
→ Why European robotics reaching mega-round scale signals competitive credibility alongside US/China leaders

Theker raises €80M, Europe’s largest robotics Series A
→ Barcelona-based startup builds reconfigurable robots: swappable hands, arms, form factors vs fixed humanoid design
→ Raising 2x target, growing from dozens to 120 employees, expanding showrooms across Europe, US, Asia

Robotiq ships TSF-85 tactile sensor digital twin on NVIDIA Isaac Sim
→ 97% SSIM accuracy on hyperelastic model, 90% on elastic, trained on 53,400 real-world maps
→ What 2.3M cycle durability with zero signal drift reveals: models maintain consistency through demanding factory use

Standard Bots raises $200M Series C at $1B valuation
→ AI-native robots taught by demonstration, targeting 10% US industrial deployments by next year, 100% US manufacturing by 2027
→ Why 150% job posting increase + 15% employment growth in robot-adopting plants challenges “job loss” narrative

Neura Robotics raises up to $1.4 billion in Series C! 💰

German company Neura Robotics has raised a Series C round worth up to $1.4 billion to develop its humanoid robotics technology, with U.S. tech giants Tether, Qualcomm, Amazon, and Nvidia all participating. The round also featured European industrial companies Bosch and Schaeffler, alongside the European Investment Bank. The company hit a valuation of around $7 billion, according to a source familiar with the matter.

“The future of AI will not only live on screens,” said David Reger, founder and CEO of Neura Robotics. “It will move, interact, learn and work beside us in the real world.” The full funding is contingent on Neura hitting certain milestones based on company performance, the source added.

Investors have piled into robotics startups in recent times as attention turns to deploying AI in physical systems that can interact in real-world environments. Robotics companies have raised $55.8 billion so far in 2026, according to Dealroom, a record figure nearly double the previous record raised last year. The majority has been raised by companies in the U.S. and China, but new European robotics companies are emerging too, including SoftBank-backed Agile Robots and U.K.-based Humanoid.

“Many believed globally relevant AI infrastructure companies could only emerge from Silicon Valley,” Reger said. “We believe the next generation of AI leaders can emerge anywhere in the world where there is enough vision, engineering talent and execution speed. With this financing, Neura is firmly among the global leaders in the robotics race, alongside the best in the US and China.”

Read more here!

Theker just raised $85M to build the factory robots! 🦾

Theker, a Barcelona-based AI robotics startup, has raised $85 million in what it’s calling “Europe’s largest ever robotics Series A.” The company aims to go beyond robots trained for a single task. “If you always have to put the same cookie in the same box, that works perfectly, but most processes aren’t like that,” co-founder Carla Gómez Cano told TechCrunch. Theker is designed for that messier reality.

Unlike humanoid robots designed around a fixed form, think Boston Dynamics, Theker’s machines are built to be reconfigured. Their hands, arms, and overall form can be swapped out or resized depending on the task, whether that’s sorting packages, packing clothing, or handling bottles and cans in a warehouse. Inditex, Zara’s parent company, signed on as an early backer, signaling where Theker’s ambitions start—the broader goal is to move beyond retail into heavier industrial settings like manufacturing, where the complexity and scale of manual tasks is even greater.

The Series A came less than a year after a record seed round, led by American VC firm CRV and backed by a mix of traditional and strategic investors, including Samsung and Aglaé Ventures, the investment vehicle tied to LVMH chairman and CEO Bernard Arnault. Gómez Cano said Samsung is not a client yet but the two are in advanced discussions—a potential trifecta of customer, supplier, and investor that would give the startup both revenue and credibility in manufacturing at scale.

To demonstrate it can deliver, Theker has a showroom in central Barcelona and plans to open others as it expands across Europe, the U.S., and Asia. The company has already received 15,000 job applications and could grow from dozens to up to 120 people by the end of the year. Raising twice its target has reinforced the startup’s conviction in keeping its HQ in Barcelona, a growing robotics hub, and in Europe’s tech ecosystem more broadly.

Read more here!

Robotiq releases tactile sensor digital twin on NVIDIA Isaac Sim! 🫆

Robotiq has released the digital twin of its TSF-85 tactile sensor in NVIDIA Isaac Sim, the first industrial-grade tactile sensor digital twin shipping on a commercial collaborative gripper. Tactile sensing promises to accelerate robotics, but its adoption has been limited by the lack of data from industry-ready hardware and accurate simulations. Model builders can now train contact-rich manipulation policies in simulation, then run them on the same physical sensor designed to operate reliably on the factory floor.

Most tactile sensors rely on a deformable contact interface, the very property that gives them sensitivity is also what makes them difficult to simulate. The TSF-85 digital twin is built around that constraint. It generates synthetic tactile maps through a custom Isaac Sim UI panel, visualizes them in real time, runs data generation at the simulation refresh rate, and exports to HDF5 for downstream training pipelines.

The simulation method is documented in two peer-reviewed publications. The first, published in Frontiers in Robotics and AI (2025), draws on a dataset of 53,400 real-world tactile maps to train, validate, and test each simulation pipeline, achieving up to 97% Structural Similarity Index Measure (SSIM) for the hyperelastic model and 90% SSIM for the elastic model on 12 unseen objects. A companion paper at ICCRT 2025 releases an open dataset of 46,200 real and synthetic tactile samples.

Simulation accuracy matters, but it only helps if the real sensor remains stable over time. The TSF-85 has been tested through 2.3 million cycles at maximum gripper force, with no significant variation in tactile signal output. This means models trained on its data can continue to rely on consistent signals for edges, shapes, textures, and geometry even after demanding real-world use.

Robotiq has been the go-to components provider for both academic research labs and industrial production for more than a decade, exactly the bridge Physical AI requires. The digital twin supports NVIDIA Isaac Sim 5.1 and is available now on GitHub.

Read more here about it!

Free Robotics Resources 👨🏼‍🏫

This course gives a solid intro to modern robot learning, mixing classical methods with recent large-scale models.

Students begin with imitation learning, reinforcement learning, and policy optimization, then move on to topics like Vision-Language-Action (VLA) models and foundation models for robotics.

Access it here for free!

Standard Bots raises $200 million to bring jobs back to the US! 🇺🇸

Standard Bots has announced a $200 million Series C at a $1 billion valuation, with the company claiming to be the largest manufacturer of AI-native industrial robots in the United States. The pitch is counterintuitive: robots will make American manufacturers more competitive, and that will unlock more jobs. “We’ve gone from 20 million manufacturing workers in 1979 to only 13 million today,” their CEO said.

Standard Bots makes AI-native robot arms and industrial humanoid robots that can be taught by demonstrations rather than code—you show the robot the job, it watches, learns, and starts doing the job itself. The pitch is that this collapses the cost and complexity that have kept advanced automation out of reach for most American manufacturers, especially small and mid-sized shops. Customers range from Sunoco, Lockheed Martin, Amazon, NASA, and the U.S. Army to hundreds of smaller manufacturers. “AI-native robots are the essential power tool of the 21st century,” said CEO Evan Beard. “The quickest way to get to full autonomy is through deployments, collecting real-world data, and iterating as fast as possible.”

The company argues investment in robotics is key to closing the gap with China. According to the International Federation of Robotics, China installed 295,000 industrial robots in 2024, 54% of the global total, more than the rest of the world combined, and its operational robot stock now exceeds 2 million units, roughly 4.5 times that of second-place Japan. The Americas accounted for just 9% of new deployments. Standard Bots says it will deliver 10% of new U.S. industrial robot deployments by next year, designs almost all its own parts including its own actuators, and by 2027 plans to manufacture everything—from metal in to robots out—right here in America.

The $200 million Series C was led by robotics fund RoboStrategy alongside existing backer General Catalyst, a steep markup from the company’s previous $63 million raise in 2024. “Standard Bots stands out because they’ve solved one of the hardest problems in industrial automation: making robots that are not only powerful, but actually usable on the factory floor without specialized programming,” said Andrew Kang, CEO of RoboStrategy. The big question is whether adding more robots really does mean adding more jobs, Standard Bots cites a 2025 University of Minnesota study finding roughly 150% increase in job postings and 15% increase in employment in plants that adopt robots compared to non-adopters. Whether that brings the jobs back, or just the work, is the question the next few years will answer.

Read more here about it!

OpenAI revives robotics division after 5-year pause
→ Hiring full-stack engineers targeting infrastructure robots for data centers, power grids, factories
→ Why near-term focus signals clarity: not consumer/warehouse, but skilled worker support building AI boom’s physical infrastructure

Genesis AI releases Genesis World 1.0 simulation framework
→ 89% hardware correlation, 45% smaller reality gap, 200+ hour real-world task runs in <30 minutes simulation
→ What separation of training (real) from evaluation (simulated) reveals: avoids simulator overfitting, enables systematic variable sweeps

Robotiq launches IQ for automatic robot integration
→ 5-minute Fit Check replaces weeks of manual quotes and site visits for palletizing decisions
→ Why deployment-ready workcells with validated designs in actual factory environments change SMB economics

Helsing launches Area 9, unveils RX-1 robotics platform
→ European defense tech incubator building full robotic stack in-house: hardware + software layer for autonomous research
→ What partnerships with ETH Zurich and INRIA Paris signal: developing unpredictable terrain navigation for debris-laden environments

OpenAI just made robotics its next frontier! 🤖

OpenAI is actively hiring robotics engineers as it revives its robotics ambitions, with an initial focus on robots that can help build data centers, power grids, and other critical infrastructure. On May 31, Sam Altman posted on X that OpenAI Robotics is hiring, seeking “exceptional full-stack hardware, ops, systems, and ML engineers to help us program and manufacture robots that are useful for society.” Greg Brockman, OpenAI’s co-founder and president, followed up to confirm the division is “making rapid progress towards building AI that can help people in the physical world.”

The move marks OpenAI’s strongest robotics push since shutting down its robotics team in 2021. The near-term focus is specific: robots to support skilled workers building infrastructure, not household helpers or warehouse sorters, but the workforce that will need to wire, pour, and assemble the data centers, power grids, and factories that the AI boom demands.

This is not OpenAI’s first attempt at robotics. The company ran a robotic hand project called Dactyl from 2017, most notably solving a Rubik’s Cube one-handed in 2019, before disbanding the entire robotics team in 2021. Since then, it has maintained a foothold through minority investments, the OpenAI Startup Fund led a $23.5 million round for 1X Technologies in 2023 and joined the $675 million round for Figure AI alongside Microsoft, NVIDIA, and Jeff Bezos. However, Figure subsequently ended its collaboration with OpenAI in February 2025 to build its AI models entirely in-house.

OpenAI’s robotics hardware lead Caitlin Kalinowski, who joined from Meta in November 2024, resigned on March 7, 2026, citing concerns about the company’s Pentagon deal and the lack of guardrails around domestic surveillance and lethal autonomy.

Read more here about it!

Genesis AI releases Genesis World 1.0! 🎮

Genesis AI has released Genesis World 1.0, a comprehensive simulation framework that treats simulation not just as a data generator, but as the fundamental infrastructure layer for robotics development. The platform includes Nyx (a photorealistic rendering engine built for robotics), Genesis World (a unified physics simulator supporting rigid and deformable bodies), and Quadrants (a GPU-accelerated compiler). The key insight: simulation should accelerate the robot development cycle itself, turning model evaluation from a real-world bottleneck into a compute problem.

Genesis’s approach prioritizes evaluation first. Rather than training on simulated data, they train on real-world data but evaluate in simulation—keeping training and evaluation workflows separate. This avoids overfitting to simulator artifacts. Their simulation correlates at 89% with on-hardware rollouts, and their reality gap is 45% smaller than competing simulators. In tests, a single evaluation that would take 200+ hours of continuous robot operation in the real world runs in under 30 minutes in simulation.

The platform enables systematic evaluation across hundreds of tasks with hundreds of episodes each, revealing exactly where policies fail. They can sweep variables—object shapes, lighting angles, camera trajectories—across multiple axes simultaneously, something nearly impossible on physical hardware. This lets researchers validate design choices before committing to large-scale pretraining, saving significant compute.

Going forward, Genesis is focused on three directions: scaling post-training through millions of simulated training iterations, building hybrid simulators that merge classical physics with learned models, and enabling self-evolving physical AI where agents improve end-to-end across both simulated and real-world loops. The north star: keeping robotics development from being bottlenecked by real-world wall-clock time, moving progress at the speed of compute instead.

Read more here!

Robotiq launches IQ to automate robot integration! 🔥

Robotiq has launched IQ, an AI platform that automates the entire robot integration process, attacking one of industrial robotics’ most stubborn problems. Integration has always been manual, unpredictable, and expensive. Site visits, weeks waiting for quotes, custom engineering every single time. IQ aims to change that entirely.

The platform starts with a 5-minute Fit Check, a structured assessment that tells you whether palletizing is a fit for your floor, what deployment looks like, and whether the financial return makes sense, including in 1-shift operations. No site visit required. No engineering hours upfront. No waiting weeks for a quote that may not go anywhere. Just an honest read on your opportunity before you commit to anything.

Once fit is confirmed, IQ guides the entire project, built on lessons from over 1,000 Robotiq deployments. It captures the reality of your operation, connects your team with the right people at the right step, and generates a validated workcell design simulated in your actual factory environment. Cycle time, reach, payload, all confirmed before you commit. The result: a deployment-ready palletizing workcell with predictable performance from day one.

Robotiq calls this “Automatic Integration”, the idea that deploying a robot workcell should be as predictable and repeatable as ordering one. For the thousands of small and mid-sized manufacturers who have been priced out of robotics by integration complexity and unpredictability, this changes the equation entirely. IQ was unveiled at RUC 2026 to selected partners and opens to everyone June 4th.

Read more here about it!

Robotics maps - London 🇬🇧

The robotics startups ecosystem is growing rapidly in London. Here’s a curated map of most of London-based companies working in the physical AI space. If I missed a company you know of, let me know! :)

Helsing officially launches Area 9, pushes into robotics! 🛡️

German defense tech giant Helsing has officially unveiled Area 9, a new research division designed as a pie-in-the-sky incubator for European defense tech, think Moonshot Factory or Skunk Works, but for Europe. As part of the launch, the company also announced RX-1, a robotics research platform developed by Area 9. “Until now, robotics research in Europe meant depending on platforms built elsewhere,” said Helsing’s Chief Scientist Antoine Bordes, who is leading the project. “We built RX-1 so that doesn’t have to be true anymore. Designed here, owned here, iterated here.”

Area 9 was first announced last year, with the innovation team initially working on AI agents for the battlefield, including Centaur AI. The Centaur agent was publicly unveiled in June 2025 when Helsing and Saab announced it had flown the latter’s JAS 39 Gripen E fighter. Centaur will also serve as the AI brain for the company’s CCA analog, the CA-1 Europa, unveiled last fall. Now the company is moving into the physical realm with robotics.

RX-1, in its current form factor, looks like a robot dog. Helsing plans to build all parts of the robotic stack—the robot itself plus the software layer that researchers can develop and test different autonomy capabilities on. The platform is more research-intensive and less currently-fieldable, but Helsing says it will make RX-1 available to research institutions and robotics laboratories across Europe. The company has already teamed up with the robotics group led by Marco Hutter at ETH Zurich and INRIA Paris to further develop RX-1.

The company’s goal is to develop an autonomous robotic platform that can navigate real-world, unpredictable terrain—like debris-laden environments—built entirely in Europe. “Area 9 exists to find the problems that look impossible and break them open,” Bordes said. “That’s why we made RX-1.”

Read more here about it!

Humanoid’s 1,000+ robot deal with Schaeffler signals 100K total shipments
→ Seven-digit actuator purchase (1M+) at 50% coverage implies 100K robots across all clients over 5 years
→ Why dual role matters: Schaeffler as both aggressive adopter (thousands of units) and preferred supplier

Boston Dynamics reveals Atlas fridge-lifting via reinforcement learning
→ Atlas trained millions of simulation hours with varied weight, friction, grip, positioning conditions
→ What proprioception demonstrates beyond cameras: sensing weight, balance, resistance for real-time unstable load adaptation

Locus Robotics acquires Nexera for AI-driven grasping
→ NeuraGrasp’s patented soft membrane + AI + vision adapts dynamically to shape, texture, material, porosity, weight variations
→ Why mobile manipulation at enterprise scale is the frontier: efficiently grasping millions of SKU types with speed/precision

Rivian spinoff Mind Robotics raises $400M, total hits $1B
→ $900M raised in 2 months at $3B+ valuation as Scaringe targets factory automation with “robotics with human-like skills”
→ Why “Project Synapse” launched: existing startups not fully equipped for industrial work

Humanoid’s 1,000+ robot deal with Schaeffler! 🏭

Two-year-old British startup Humanoid has signed a deal to integrate “a four-digit number of humanoid robots into live manufacturing operations” at German industrial giant Schaeffler’s factories starting later this year. The bigger news, however, might be that Humanoid has committed to buying a “seven-digit number of actuators” from Schaeffler over the term of the agreement, suggesting Humanoid intends to ship a staggering 100,000 humanoid robots across all of its clients over the next five years.

Four months ago, Schaeffler and Humanoid announced they’d put “several hundred” humanoid robots into factories over the next five to six years. On Wednesday, the companies upgraded the number significantly, signing an agreement to integrate at least 1,000 – and potentially many more, humanoid robots into Schaeffler’s manufacturing operations by 2032. This makes it one of the largest disclosed humanoid robot rollouts to date and arguably the most aggressive escalation between a humanoid OEM and an industrial customer.

The supply agreement indicates Schaeffler will become Humanoid’s preferred supplier, covering more than 50% of the company’s demand for joint actuators for its wheeled-based platforms through 2031. This partnership is expected to translate into the supply of a “seven-digit number of actuators” – at least one million. The wheeled HMND 01 Alpha offers close to 30 active degrees of freedom, and the “Schaeffler-relevant” actuator count per robot is probably roughly 18-22 joint actuators. At 50% coverage, one million actuators means 100,000 robots over the next five years.

Schaeffler is simultaneously one of the most aggressive industrial adopters of humanoids and a preferred component supplier for the platforms it’s buying. The company is committed to buying thousands of humanoid robots from Humanoid, at least 1,000 from Switzerland’s Hexagon Robotics, plus a bigger partnership with Germany’s Neura Robotics for 4,000-6,000 units. Schaeffler CEO Klaus Rosenfeld said the company is now engaged with roughly 45 humanoid robotics players worldwide and expects its humanoid robotics division to secure an order book in the hundreds of millions of euros by 2030.

Read more here about it!

Boston Dynamics reveals how Atlas robot lifts a.. fridge! 🧊

Boston Dynamics has revealed how its Atlas humanoid robot learned to lift and carry heavy industrial objects using reinforcement learning and large-scale simulation training. In a newly released technical blog, the company showed Atlas rotating its torso 180 degrees, squatting to pick up a mini-fridge, and carrying it across a lab floor while adjusting to shifting weight inside the object. The behavior was developed within weeks of Atlas’ public debut earlier this year.

The latest demo marks a shift from choreographed robot movements toward adaptable industrial behaviors designed for factories, warehouses, and construction sites. Boston Dynamics said Atlas is being developed as a “general purpose tool for physical work.” Rather than relying mainly on cameras, the robot uses proprioception, internal body awareness, to sense weight, balance, grip, and resistance while moving objects, allowing Atlas to adapt in real time to unstable loads and changing conditions.

Boston Dynamics trained Atlas through reinforcement learning, where the robot repeatedly practiced the same lifting task in simulation under different conditions. The company varied factors such as object weight, floor friction, grip strength, and fridge positioning to force the robot to adapt. “Atlas practiced the moves for millions of hours in simulations in parallel on GPUs,” the company said. The process starts with a reference trajectory, then Atlas is rewarded for completing tasks correctly while external disturbances are introduced.

One major advantage of the new Atlas platform is its reduced “sim-to-real gap”—a longstanding robotics challenge where behaviors trained in simulation fail in the physical world. Boston Dynamics said Atlas’ simplified hardware architecture makes accurate simulation easier, using only two actuator types across its body with symmetrical arms and legs.

Read more here!

Locus Robotics acquires Nexera Robotics! 🪬

Locus Robotics, the leader in Flexibility-First Warehouse Automation, has announced the acquisition of Nexera Robotics, a Vancouver-based robotics company specializing in advanced robotic grasping. The integration of Nexera’s proprietary NeuraGrasp end-effector technology into the Locus Robotics physical AI platform significantly expands the company’s autonomous mobile manipulation capabilities and broadens what Locus Array can handle across end-to-end fulfillment workflows.

NeuraGrasp combines AI-driven grasping intelligence, onboard sensory inputs, computer vision, and a patented soft membrane structure to adapt dynamically to the physical characteristics of each item. This enables a single gripper to conform to variations in shape, surface texture, material, porosity, and weight, creating reliable grasps across high-variability inventory found in real warehouse operations.

Locus Robotics CEO Rick Faulk stated that the frontier of warehouse robotics today is AI-driven mobile manipulation at enterprise scale, and being able to efficiently grasp millions of SKU types with both speed and precision is where the next decade of value gets created. Nexera CEO Roy Belak said they built NeuraGrasp to solve the manipulation challenges that have held robotic picking back for years, and joining Locus gives them the platform, scale, and customer base to bring this breakthrough technology into high-velocity fulfillment environments.

Read more here about it!

Robotics maps - Europe 🇪🇺

Alpine Tech Corridor is the undisputed deeptech capital of continental Europe. The data from the European Deeptech Report 2026:

Access the report here:

Rivian spinoff Mind Robotics raises another $400M!

Rivian’s spinoff company Mind Robotics has raised another $400 million, just two months after raising $500 million, as it works to develop industrial robotics that can further automate factory operations. The funding round, first reported by the Wall Street Journal, was led by Kleiner Perkins, with participation from the venture arms of Volkswagen (partnered with Rivian on a software joint venture) and Salesforce. The new funding brings total raised to more than $1 billion and values Mind Robotics at more than $3 billion.

Rivian CEO RJ Scaringe, who is chairman of Mind Robotics, told TechCrunch in March that he created the company because he felt other startups were not fully equipped to automate industrial work. He started the project—initially known as “Project Synapse”—as an effort to build “robotics with human-like skills.” Mind Robotics had previously raised $115 million from Eclipse after it was created in 2025.

Scaringe also helped create and spin out a micromobility company called Also, which has raised more than $300 million to date. The rapid fundraising pace, $900 million in just over two months, demonstrates strong investor confidence in Mind Robotics’ vision to bring advanced industrial robotics to factory operations and manufacturing environments.

Read more here about it!

Unitree debuts GD01 manned transformer mecha
→ 500kg bipedal-to-quadruped machine starting at $574K, world’s first mass-produced transformable mecha
→ Why 5,500+ humanoid shipments in 2025 demonstrates China’s manufacturing scale advantage: lower costs, faster scale-up

Genesis AI unveils GENE-26.5 for human-level manipulation
→ Foundation model demonstrates 20-step cooking, Rubik’s Cube solving, piano playing with proprietary hand mirroring human form
→ What 1:1:1 mapping glove with tactile e-skin reveals: 100x cheaper hardware, 5x greater data collection efficiency vs teleoperation

Apptronik stacks C-suite with Waymo, Boston Dynamics, Amazon vets
→ Chu (CPO), Garell (Services), Shah (Software), Perry (Marketing), Birtz (People Ops) join as race shifts from demos to fleet support
→ Teasing new humanoid beyond Apollo (6ft, 55 lb lift, 22hr/day, 42 DOF with DeepMind Gemini)

SoftBank creates Roze AI for automated data center construction
→ New company deploying autonomous robots to build server farms with $100B IPO target by H2 2026
→ Why automating infrastructure creation signals meta-level bet: building the physical AI boom’s foundation

China's Unitree debuts real-life transformer! ⚙️

Unitree Robotics has unveiled the GD01, a manned “mecha” capable of transitioning between bipedal walking and four-legged mode, bridging the gap between science fiction and reality. The GD01, resembling an Autobot from a Transformers movie, is a high-strength alloy machine designed for civilian transport. It weighs 500kg with a pilot on board (roughly the weight of a grand piano) and carries a starting price of 3.9 million yuan (US$573,674).

A demonstration video shows the GD01 carrying a pilot in its torso-mounted cockpit, walking like a humanoid robot and knocking over a brick wall with its hand. The machine then reconfigures its chassis, shifting into a four-legged crawl. Unitree described the robot as the world’s first mass-produced “transformable mecha.”

The GD01 expands the Hangzhou-based company’s portfolio at a time when Chinese manufacturers are significantly outstripping their US rivals, helped by lower production costs and faster manufacturing scale-up. Chinese companies accounted for nearly 90% of global humanoid sales in 2025, according to research firm Omdia. Unitree shipped more than 5,500 humanoid robots last year, demonstrating the company’s strong position in the rapidly growing robotics market.

Read more here about it!

Genesis AI introduces a model for more dexterous robot manipulation! 🪬

Genesis AI has unveiled GENE-26.5, claiming it gives robots “human-level physical manipulation capabilities.” The San Carlos-based company also announced a system to “unlock unlimited amounts of data” and train GENE-26.5 at scale, combining two proprietary components: a new data engine and a human-scale robotic hand that enables direct skill transfer from humans to robots.

GENE-26.5 is an AI foundation model purpose-built for robotics, designed to absorb massive amounts of data and environments to enable robots to perform complex, long-horizon tasks with human-level dexterity. Demonstrations show the system performing fluid, human-like dexterity for complex tasks including cooking a 20-step meal, preparing a smoothie, conducting high-precision lab experiments, wire harnessing, solving a Rubik’s Cube with continuous in-air manipulation, single-handed multi-object grasping, and playing the piano at a human level.

The company has addressed the “embodiment gap” with proprietary hardware that mirrors the human hand in form and function, paired with a data-collection glove equipped with tactile-sensing electronic skin. When worn by a human, the glove allows for a 1:1:1 mapping between the glove, the human’s hand, and the robotic hand, allowing humans to seamlessly provide GENE-26.5 with high-quality data at scale that translates into robotic skills. The glove is 100 times cheaper than typical hardware cost and demonstrates up to five times greater data-collection efficiency than traditional teleoperation methods.

Genesis AI’s data engine taps into egocentric video data from humans wearing cameras and massive amounts of human-based internet videos. The company has also developed a simulation system with a rendering engine and hyper-realistic physics to create accurate representations of actual conditions.

Read more here!

Apptronik teases new robot, stacks C-suite with Waymo, Boston Dynamics, Amazon vets! 💼!

Apptronik, the Austin-based humanoid robot company that closed a $935 million Series A at roughly $5.3 billion valuation in February, has stacked its executive ranks with tech company veterans from Waymo, Boston Dynamics, Amazon, iRobot, and Paramount+. The new hires include Daniel Chu as Chief Product Officer (from Waymo), Kevin Garell as SVP of Services & Support (from Boston Dynamics), Chirag Shah as VP of Software (from Amazon), Dave Perry as VP of Marketing (from Paramount+), and Justin Birtz as VP of People Operations (from Cellino/iRobot).

CEO Jeff Cardenas framed the leadership refresh as representing the urgency of the moment, stating “we are at a defining moment in robotics where the technology has finally met the magnitude of the mission.” He added a significant statement: “We have the talent, the technology, and the momentum to bring AI-powered humanoid robots to the world at scale.” The race is shifting from who can demo a robot folding laundry to which companies can support fleets of humanoid robots in production environments doing real work.

Apptronik also teased a new humanoid robot, saying the company is preparing to unveil it soon. This sounds like a new robot, not just a new version of Apollo, its existing humanoid that stands roughly six feet tall, can lift up to 55 pounds, operates up to 22 hours a day, has 42 degrees of freedom, and is increasingly powered by Google DeepMind’s Gemini Robotics models. Apollo is currently deployed in commercial pilots with Mercedes-Benz, GXO Logistics, and Jabil.

Read more here about it!

Robotics maps - Pittsburgh 👨🏼‍🏭

Pittsburgh is a good place to launch a robotics company too.

SoftBank is creating a robotics company that builds data centers! 🏗️

Japanese multinational SoftBank is reportedly creating a new company called Roze AI designed to automate data center construction in the U.S. Roze would seek to make data center construction more “efficient” by deploying autonomous robots to help build server farms, among other things. The Financial Times originally reported the news, with additional details from The Wall Street Journal.

In an interesting twist, SoftBank is already prepping Roze for an IPO, and some executives want it to happen by the second half of 2026. The desired valuation might be $100 billion, according to the Financial Times. The move comes as tech companies race to build out infrastructure that can further drive the automation boom, with SoftBank now positioning itself to automate the creation of that infrastructure itself.

Other recent ventures have also envisioned using AI and automation to make the industrial sector more efficient. For example, Amazon mogul Jeff Bezos has co-founded a startup called Project Prometheus that plans to buy firms in major industrial sectors and modernize them using AI. However, SoftBank has been known to back some dark horse startups, notably sinking hundreds of millions of dollars into Zume, an AI-driven pizza delivery startup that went belly up in 2023.

Read more here about it!

Schaeffler expands to 1,000+ humanoid deployment by 2032
→ AEON’s successful 2025 pilot leads to multi-site rollout starting late 2026 with automated part inspection
→ Why wheel-based locomotion and unique sensor suite demonstrate high-precision manipulation in multi-machine stations

Google DeepMind launches Gemini Robotics-ER 1.6
→ Reasoning-first model achieves 93% success on instrument reading using agentic vision (visual reasoning + code execution)
→ What multi-view understanding and spatial reasoning improvements unlock: pointing, counting, success detection, gauges/sight glasses

ABB Robotics launches PoWa cobot family
→ 7-30 kg payloads, 5.8 m/s top speed bridging gap between traditional cobots and industrial robots
→ Why 20% annual cobot market growth through 2028 signals demand for speed/payload without operational rigidity

Skild acquires Fetch Robotics assets from Zebra
→ Hardware-agnostic Skild Brain integrates with warehouse platforms to transform task-specific automation into unified fulfillment ecosystem
→ What phased rollout reveals about data flywheel catch-22: semi-structured (factories/warehouses)

Schaeffler to deploy a over a thousand of humanoids! 🤯

Hexagon Robotics and Schaeffler have announced the expansion of their strategic partnership to accelerate the rollout of AEON humanoid robots across factories worldwide, following the successful completion of a joint pilot program in 2025. Schaeffler plans to deploy at least 1,000 Hexagon Robotics humanoids across its global production system by 2032. Schaeffler will also provide its high-precision actuators to AEON, further strengthening the technological alignment between the two companies.

Working with Schaeffler allows Hexagon to deliver demonstrable business value across various factory environments and scale operations. During the pilot phase, AEON used its unique sensor suite and wheel-based locomotion to demonstrate high-precision manipulation capabilities in operating a multi-machine station to load, unload, and inspect parts in real-world production environments.

Building on the successful pilot, the companies will now move to deployment across multiple sites, with additional applications such as automated part inspection targeted for rollout starting at the end of 2026. A key focus of the partnership is the generation and use of real-world production data to enable faster deployment of AEONs across new use cases and production settings.

The cooperation shows how Schaeffler combines hardware excellence with industrial application, not only developing key components but also systematically integrating robots from Hexagon into its own plants with the clear aim of boosting competitiveness through physical AI.

Read more here about it!

Google DeepMind launches new model for physical AI! 🔍

Google DeepMind has introduced Gemini Robotics-ER 1.6, a significant upgrade to its reasoning-first model that enables robots to understand their environments with unprecedented precision. By improving spatial reasoning and multi-view understanding, the company is bringing a new level of autonomy to the next generation of physical agents. The model specializes in visual and spatial understanding, task planning, and success detection, and is now available to developers via the Gemini API and Google AI Studio.

Gemini Robotics-ER 1.6 shows significant improvement over previous versions, specifically enhancing spatial and physical reasoning capabilities such as pointing, counting, and success detection. The model unlocks a new capability: instrument reading, enabling robots to read complex gauges and sight glasses—a use case discovered through collaboration with partner Boston Dynamics. The model achieves 93% success on instrument reading tasks using agentic vision, which combines visual reasoning with code execution.

Pointing is a fundamental capability used to express spatial reasoning, relational logic, motion reasoning, and constraint compliance. The model advances multi-view reasoning, enabling better understanding of multiple camera streams and their relationships, even in dynamic or occluded environments. Success detection serves as a critical decision-making engine that allows robots to intelligently choose between retrying a failed attempt or progressing to the next stage of a plan.

Read more here!

ABB Robotics launches new PoWa cobot family! 🆕

ABB Robotics has launched its new PoWa collaborative robot family, featuring higher payloads and performance than its previous offerings. The new family addresses a long-standing gap in the market between traditional cobots, which often lack the speed and payload needed for industrial applications, and conventional industrial robots that typically work best in specialized, large-scale automation environments. ABB estimates the global cobot market will grow by 20% annually through 2028.

They said that cobots are growing significantly faster than traditional industrial robots, driven by demands from both small and midsized companies starting their automation journey as well as large enterprises. These customers are seeking higher speeds and payloads, but also greater ease of use and compact designs. Established manufacturers want to automate heavier, fast-cycle applications without the complexity and operational rigidity of traditional industrial robots.

PoWa extends ABB Robotics’ cobot offering with industrial-grade performance, including six different payload categories from 7 kg (15.4 lbs) to 30 kg (66.1 lbs), and can reach a top speed of up to 5.8 m/s (13 MPH). Purpose-built for compact environments, the line is ideally suited for applications such as high-speed machine tending, palletizing, screwdriving, and arc welding, enabling manufacturers to automate heavier and faster processes.

The new cobots are easy to use through programmable buttons on the arm-side interface and no-code programming, and are compatible with an extensive ecosystem of third-party accessories. PoWa can be unboxed and operational within an hour, enabling plug-and-play with a wide range of tools.

Read more here about it!

Robotics maps - New York City 🗽

New York City is a good place to launch a robotics company, apparently.

Skild acquires Fetch Robotics assets from Zebra! 🦓

Skild AI has acquired the robotics division of Zebra Technologies (formerly Fetch Robotics) in a move to bridge the gap between foundation AI and industrial hardware. The acquisition will help Skild deploy its “omni-bodied” intelligence layer across the global logistics sector. By integrating its hardware-agnostic AI “brain” with Zebra’s warehouse platforms, the acquisition aims to transform task-specific automation into a unified, autonomous fulfillment ecosystem. Financial details were not disclosed.

Skild CEO and co-founder Deepak Pathak told that “the Fetch Team is the main reason for the acquisition as they bring years of deployment experience.” Skild plans to support the existing Fetch/Zebra installed base, sell new Fetch robots, and integrate the Skild Brain into Fetch products to enhance the robots’ capabilities and open up new advanced manipulation solutions for the warehouse. The Zebra Symmetry platform, which coordinates and orchestrates robot work, will be deeply integrated with the Skild Brain.

The acquisition signals a pivot away from “brittle,” task-oriented robotics toward a unified intelligence layer capable of controlling diverse forms of machines. Skild is focused on building out its “data flywheel” to train AI world models. The primary differentiator is the Skild Brain, a foundation model designed to operate any machine that moves and adapt to physical changes without needing retraining from scratch.

Skild AI uses a phased rollout strategy to resolve the “catch-22” where robots need data to improve, but only capable robots can be deployed to gather data. Phase 1 focuses on semi-structured environments like factories and warehouses, Phase 2 moves into dynamic environments like hospitals and hotels, and Phase 3 aims for general-purpose robots in fully unpredictable environments like private homes. Key partners include ABB Robotics, Universal Robots, and Mobile Industrial Robots, allowing Skild to scale quickly by utilizing the hardware footprint of some of the world’s largest robotics providers.

Read more here about it!

Agibot ships 5,000 humanoids in 3 months, 10,000 total
→ Took 2 years for first 1,000 units, 1 year to reach 5,000, now 5,000 in just 3 months → Why China’s 90% global humanoid manufacturing share signals maturation: supply chain standardizing, pivot from viability to scalable value

Generalist unveils GEN-1 model: 99% success vs 64% baseline
→ Tasks completed 3x faster using only 1 hour robot data per task, trained on 500K hours real-world data
→ What pretraining on low-cost wearable human activity data reveals about avoiding expensive teleoperation datasets

Sanctuary demonstrates zero-shot in-hand manipulation
→ 21-DOF hydraulic hand reorients lettered cube 10 consecutive times without dropping, no palm support
→ Why finger abduction and proprietary hydraulic actuation matter for human-level dexterity in industrial applications

Manna raises $50M Series B, total hits $110M
→ 250K+ regulated commercial UAV flights across Ireland, Finland, Texas over 7 years with proven economics
→ What <3 minute delivery, 85% CO₂ reduction, and NPS 86 demonstrate about suburban retail unit economics

Agibot shipped 5,000 humanoids in the last 3 months! 🤯

Shanghai-based Agibot has announced it shipped 5,000 humanoid robots in the last three months alone, hitting 10,000 total robots shipped. For context, it took the company two full years to ship the first 1,000 units and another year to scale to 5,000. Now the company has surpassed that three-year total in just three months.

As the supply chain matures and manufacturing standardizes, there’s a pivot from small-scale, niche applications to robust, large-scale commercial demand. The widespread deployment of Agibot’s robots is no longer about seeking technical viability, but about delivering scalable value and driving the adoption of embodied AI.

Independent research released in January indicated that Agibot currently has the lion’s share of the global humanoid robot market. At that point, Omdia estimated the company had shipped just under 5,200 units, with second place Unitree at 4,200, and the first American humanoid robot manufacturers: Figure, Agility Robotics, and Tesla at around 150. Almost 90% of all built humanoid robots were manufactured in China, and with this new news, that percentage will only go up.

According to Agibot, these robots are not just being sold and used in China’s domestic market, with “a substantial number” active in Europe, North America, Japan, Korea, Southeast Asia, and the Middle East. The company is also seeing “a transition from initial pilot projects to repeated, large-scale rollouts.”

Read more here about it!

Generalist introduces GEN-1 model for physical AI! 🧬

Generalist AI has announced its GEN-1 general-purpose AI model for robotics, improving average success rates to 99% on tasks where previous models achieved 64%. The model also completes tasks roughly three times faster than current approaches and requires only one hour of robot data for each of these results. Founded in 2024, the San Mateo-based company is building embodied foundation models for general-purpose robots and asserts that GEN-1 “unlocks commercial viability across a broad range of applications.”

The latest release came just five months after the company released its GEN-0 model, which demonstrated that scaling laws exist in robotics. GEN-1 further scales GEN-0’s foundation and uses algorithmic advances to start mastering simple tasks. The company trained the model from scratch on its dataset of half a million hours of real-world data. While Generalist was optimistic about the AI model’s progress, it noted that GEN-1 can’t solve all tasks and some would require higher than 99% success rates to be useful in real settings.

Previous general models in robotics that surpass 90% success have depended on enormous teleoperation datasets that are expensive and difficult to scale. Instead, for GEN-0 and GEN-1, the base foundation model is trained without any robot data. The model uses data from low-cost wearable devices on humans doing millions of activities, and Generalist says this pretraining can lead to high levels of mastery without requiring large teleoperation or simulation datasets.

GEN-1 includes pre-training innovations that improved compute efficiency, advances in post-training techniques, learning from experience, multimodal human guidance, and new inference-time techniques.

Read more here!

Sanctuary’s hand demonstrates zero-shot manipulation! 🪬

Sanctuary AI has demonstrated its approach to training dexterous manipulation policies for its robotic hands. In its latest video, the company’s hydraulic hand autonomously manipulated a lettered cube, continuously reorienting it to match a specified goal. The system successfully achieved the target orientation 10 consecutive times without dropping the cube, highlighting robust in-hand manipulation capabilities. The manipulation takes place entirely at the fingertips without palm support, requiring the hand to simultaneously and stably grasp the object while making progress toward the orientation goal.

Sanctuary AI claims the high number of degrees of freedom in its proprietary robotic hands (21-degrees-of-freedom) enable finger abduction and sophisticated in-hand manipulation that most current robotic hands cannot do. The Vancouver-based company also uses hydraulic actuation, which provides increased strength, speed, and control. Combined with compact hydraulic valves, this advancement offers a promising path toward achieving human-level dexterity for industrial applications and forms the foundation for precise insertion, tool use, and other real-world dexterous tasks.

Simulating anthropomorphic hands remains a significant challenge due to their complexity, contact dynamics, and high degrees of freedom. The success of this transfer highlights Sanctuary AI’s ability to build highly realistic training environments.

Read more here about it!

Universities by number of alumni-founded startups! 👨🏼‍🎓

There’s another reason to study the ETH and EPFL models closely as both universities are also European leaders in spawning startups.

Source of this stat (Dealroom)

Manna raises $50M Series B to expand in the US! 🚁

Manna Air Delivery, a global leader in consumer drone delivery, has raised $50 million in Series B funding to scale its proven operations in the United States and Europe, bringing total funding to $110 million. Manna now operates one of the most active consumer drone delivery networks in the world, with more than 250,000 regulated commercial UAV flights completed.

As a UAV delivery pioneer, Manna has operated in six locations across Ireland, as well as in Finland and Texas over the past seven years, delivering items including books, medical supplies, food, and clothing. The company’s UAV operations are designed to make communities more accessible while making last-mile delivery greener, safer, and more efficient. Recently, Manna announced a partnership with Uber, adding to existing partnerships with Deliveroo, Just Eat, and DoorDash.

CEO and Founder Bobby Healy said the company has proven the technology and economics of autonomous delivery and is now focused on scale. Manna plans to launch up to 40 new bases in the United States as it builds the infrastructure for the next generation of logistics. Orhan Gazelle, Managing Director at Schooner Capital, said Manna is building one of the most compelling real-world applications of autonomous technology, demonstrating that drone delivery can operate safely, efficiently, and at scale with the potential to fundamentally reshape last-mile logistics.

For customers, Manna’s UAV technology delivers orders in under three minutes while reducing road congestion and cutting CO₂ emissions by up to 85% compared with road-based delivery. Deliveries are lowered safely to customers’ gardens or driveways using recyclable packaging and biodegradable tethers. Popular with families, people with mobility issues, and those working from home, Manna has achieved a Net Promoter Score of 86, placing it among the highest-rated services in consumer technology. Manna was one of the first drone delivery companies globally to demonstrate positive unit economics for suburban retail delivery.

Read more here about it!

Amazon acquires Fauna Robotics for consumer humanoids
→ Former Meta/Google engineers’ Sprout ($50K, 3.5 ft tall) joins Amazon after Disney and Boston Dynamics signed as early customers
→ Why M&A strategy signals shift from warehouse-only focus: Rivr for doorstep delivery, Fauna for home/business

Niryo unveils Nate modular robotics platform
→ Open interfaces for AI integration with 3-month iteration cycles vs rigid traditional automation
→ What €40K starting price vs €70K comparable systems reveals about cost and integration time as real blockers

Agile Robots partners with Google DeepMind
→ Gemini Robotics foundation models integrating with 20,000+ installed robotics solutions worldwide
→ Why iterative deployment cycle creates AI flywheel: real-world data improves models, models expand capabilities, unlocking broader deployment

Lucid Bots raises $20M Series B, operator network hits $75M revenue
→ Deployments surge from 100 to 700 units in 5 months, Sherpa drone with power tether pre-orders hit $1M
→ What 2-5x job completion speedup and <2 month payback periods demonstrate about operator economics

Amazon acquires humanoid maker Fauna Robotics! 🤖

Amazon has acquired Fauna Robotics, a startup that builds “approachable” humanoid robots for consumers and businesses, the company confirmed Tuesday. Terms of the deal weren’t disclosed. An Amazon spokesperson said the company is excited about Fauna’s vision to build capable, safe, and fun robots for everyone, and together with Amazon’s robotics expertise and decades of experience earning customer trust in the home, they’re looking forward to inventing new ways to make customers’ lives better and easier.

Fauna Robotics was founded in 2024 by former Meta and Google engineers. Earlier this year, the New York-based company launched Sprout, a $50,000 bipedal robot that’s 3 feet, 6 inches tall, weighs 50 pounds, and is designed to be “approachable and human-friendly” as well as “genuinely accessible” to software developers. The company signed up Disney and Hyundai’s Boston Dynamics as early customers. Fauna’s roughly 50 employees will join Amazon in New York, and the company will operate as Fauna Robotics, an Amazon company.

Amazon has spent more than a decade investing in robotics, primarily for warehouse operations. It acquired Kiva Systems for $775 million in 2012, which became the foundation for Amazon Robotics. The company has turned to M&A again to beef up its robotics expertise, last week, Amazon acquired Rivr, a Swiss robotics company developing machines for “doorstep delivery.”

The company has experimented with home robots before, launching a squat, roving personal robot called Astro in 2021, priced at $1,600 and available only by invitation. By acquiring Fauna, Amazon is entering the increasingly crowded humanoid robot market.

Read more here!

Founder of Fauna Robotics with its humanoid bot called Sprout

Niryo introduces Nate, a modular robotics platform for industrial Physical AI 🔧

At Global Industrie 2026, Niryo presented Nate (Niryo to Automate), a new modular robotics platform designed to make AI-driven automation easier to deploy in real factories. Instead of fixed, hard-to-modify systems, Nate is built as a set of interchangeable building blocks: robotic arms, actuators, grippers, and software modules that can be combined depending on the application. The key idea is not a single robot, but an infrastructure layer for industrial robotics.

The platform offers modular hardware for faster configuration and reconfiguration, open interfaces for easier integration of AI models (vision, control, optimization), and short iteration cycles enabling new robot designs in roughly three months. This directly targets a real issue in manufacturing: traditional automation is too rigid for high-mix, fast-changing production.

Nate is already being used to build pre-configured robotic cells called “Smart Machines” for common factory tasks like machine tending, end-of-line packaging, and sorting and quality control. Each system combines a 6-axis robot arm, gripper and workstation, application software, and optional AI modules. The goal is simple: reduce integration time and engineering effort while still allowing flexibility.

Nate is positioned as a bridge between robotics and AI systems. AI models for vision and control can run directly on the platform, systems can adapt to part variability (not just repeat fixed motions), and make real-time adjustments based on production conditions. On the hardware side, Niryo introduced new actuators and a 6-axis NR Series robot with target payloads up to 20 kg, enabling palletizing and heavier tasks.

Niryo claims systems start around €40,000, compared to roughly €70,000 for similar setups. This is important because cost and integration time are still the main blockers for automation, not capability alone. Nate fits into a broader shift from closed industrial robots to open, modular platforms, from fixed automation to adaptable, AI-driven systems, and from long integration cycles to faster deployment and iteration.

Read more here about it!

Agile Robots to deploy Google DeepMind foundation models on its humanoid! 🤝

Agile Robots has announced a partnership with Google DeepMind, Alphabet’s AI research lab, to combine DeepMind’s Gemini Robotics foundation models with Agile Robots’ scalable industrial robotics platform. Founder and CEO Zhaopeng Chen stated that Agile Robots has already installed over 20,000 robotics solutions worldwide, proving intelligent automation at scale. The huge opportunity lies in autonomous, intelligent production systems that can transform entire industries, and integrating Gemini Robotics models positions the company at the cutting edge of this rapidly growing market.

Agile ONE is designed to work safely and efficiently alongside people and other systems in industrial settings. By bringing together Agile Robots’ hardware and AI systems with DeepMind’s Gemini Robotics foundation models, the teams hope to improve performance through robot deployment, data collection, model training, and iteration.

The collaboration is structured as a series of development and deployment phases focused on improving robotic capabilities in real-world environments. The teams will work in an iterative cycle, continuously testing, learning, and improving system performance. The partners will first focus on high-value industrial use cases in sectors with acute and growing demand for adaptable, reliable automation, including industrial and manufacturing tasks where reliability and scale are critical.

Jointly training, deploying, and testing robots will create a scalable “AI flywheel,” where data from real-world operations will improve the models, and the models will then expand robotic capabilities, unlocking broader deployment.

Read more here about it!

Robotics maps #8 (Shanghai, China 🇨🇳)

Fantastic graphic about humanoid ecosystem in the Shanghai area prepared by RoboTuo.

Lucid Bots raises Series B funding as operator network hits $75M in revenue! 🚁

Lucid Bots, an autonomous exterior cleaning company, has raised an oversubscribed $20 million Series B round co-led by Cubit Capital and Idea Fund Partners, bringing total funding to $34 million. The Charlotte-based company will use the funding for commercial operations, expanding domestic manufacturing capacity, and accelerating the rollout of Lucid Refresh, its full-service subscription platform for exterior cleaning operators. The company has experienced rapid growth, increasing deployments from 100 to nearly 700 units in less than five months.

Lucid Bots has launched several new features for its Sherpa drone, which now supports spraying of sealing/coating additives and paint. The system can be equipped with a power tether for unlimited flight time, eliminating the battery to make drones more efficient in the air and able to lift more payload. The power tether option has garnered nearly $1 million in pre-orders. The company also introduced Lavo AI, an autonomous ground-based power-washing mobile robot designed for sidewalks, driveways, and parking lots.

Lucid Bots’ systems and operator network have generated more than $75 million in exterior cleaning revenue. The Sherpa platform can reduce job completion time by two to five times, with operators reporting payback periods of less than two months. With nearly 1,000 robots in active operator hands and 93% of new business from inbound interest, the company generated as much revenue in 2025 as in its first seven years combined. The customer base spans independent operators to enterprise accounts including Disney, Sunbelt Rentals, and commercial facilities managers across the U.S.

Read more here about it!

Lucid Bots drones for exterior cleaning tasks

Amazon acquires RIVR for stair-climbing delivery robots
→ Zurich-based startup’s doorstep delivery tech gets Amazon’s infrastructure to scale faster
→ Why Amazon Industrial Innovation Fund + Bezos Expeditions seed investment signaled strategic interest early

RoboForce raises $52M, total hits $67M
→ Robo-Labor platform targeting dirty, dangerous jobs across solar, data centers, mining, logistics, manufacturing, shipping
→ What foundation model + physical AI agents demonstrate about coordinating complex tasks reliably

Unitree files for $610M Shanghai IPO
→ China’s leading robot maker targeting STAR board listing to fund AI model research and manufacturing expansion
→ Why domestic IPO signals growing market confidence in robotics sector beyond VC rounds

Gecko Robotics scores $71M Navy deal for ship repair
→ Why software-only approach fails: AI decisions only as good as infrastructure data inputs
→ Supporting 80% fleet readiness by 2027 as U.S. races to modernize amid China shipbuilding gap

Amazon acquires RIVR, maker of delivery robots! 📦

Amazon has acquired Rivr, a Zurich-based autonomous robotics startup known for its stair-climbing delivery robot, in a deal that signals the e-commerce giant’s interest in doorstep delivery. Terms of the deal weren’t disclosed.

CEO and co-founder Marko Bjelonic said the acquisition will “accelerate our vision of building General Physical AI through doorstep delivery, bringing robotics and AI closer to real-world deployment at scale.” In plain terms, Amazon’s resources should help Rivr get its robots onto more doorsteps, faster. Last year, Rivr launched a pilot program in Austin with Veho, a package delivery company.

Rivr got Amazon’s attention long before its pilot program. The Amazon Industrial Innovation Fund and Bezos Expeditions invested in Rivr as part of a $22.2 million seed round that closed in 2024. The startup, which had raised a total of $25 million, was last valued at $100 million. The acquisition brings together Amazon’s massive delivery infrastructure with Rivr’s specialized robotics technology for navigating stairs and complex terrain.

Read more here!

The delivery bot from RIVR

RoboForce grabs $52M to unleash AI robots on dirty, and dangerous jobs! 🏗️

Industrial robotics startup RoboForce has secured $52 million in an oversubscribed funding round, bringing its total capital raised to $67 million. The round was led by YZi Labs, a $10 billion fund, with participation from Jerry Yang alongside existing investors including Myron Scholes, Gary Rieschel, and Carnegie Mellon University.

Founded in 2023 , the company is positioning its technology to address the growing gap in industrial labor, where physically demanding work is becoming harder to staff. Its platform focuses on Robo-Labor, autonomous machines designed to handle high-risk and physically intensive work across utility-scale solar installations, data centers, mining, logistics hubs, manufacturing sites, and global shipping operations.

At the core of RoboForce’s technology is a robotics foundation model combined with a stack of physical AI agents that control robot behavior. These agents interpret real-world data and coordinate actions across the robot’s systems, enabling machines to carry out complex physical tasks reliably. The platform continuously improves through a data-driven learning loop, where information from deployed robots feeds back into the system and combines with high-fidelity simulation to refine policies and behaviors. RoboForce is working with NVIDIA to power its computing and simulation infrastructure.

Read more here about it!

Chinese robot maker seeks $610M in Shanghai IPO! 📈

China’s leading robot company, Unitree Robotics, has filed for an initial public offering in Shanghai, moving closer to a domestic listing as investor interest in the robotics sector builds. Unitree plans to raise about 4.2 billion yuan ($610 million) on the STAR board, according to a prospectus released on Friday.

The proceeds will be used to fund research into AI models for robotics, develop new products, and expand manufacturing facilities. Unitree is known for its high-performance quadruped robots and humanoid systems, having recently unveiled models like the As2 quadruped that can carry up to 143 pounds and run at speeds approaching 11 mph.

The IPO filing signals growing market confidence in the robotics sector and represents a major milestone for Unitree as it scales production and development. The company has positioned itself as a leader in affordable, high-performance robotics for both research and commercial applications.

Read more here about it!

Unitree Robotics CEO and founder with a H1 humanoid robot

Robotics maps #7 (Hangzhou, China 🇨🇳)

Fantastic graphic about humanoid ecosystem in the Hangzhou area prepared by RoboTuo.

Gecko Robotics brings its robots to U.S. Navy ship repair in latest deal! 🚢

Gecko Robotics, a Pittsburgh-based robotics startup, has announced a $71 million deal with the U.S. Navy to cut down ship repair time as the government races to reindustrialize America’s aging defense systems. The company says its robots, capable of flying, swimming, and climbing critical infrastructure, use cameras and sensors to condense a three-month process down to as little as two days. Gecko also claims the robots can assess necessary maintenance 50 times faster than other manual techniques.

Co-founder and CEO Jake Loosararian told CNBC that this technology was never possible before, and it’s the reason why it’s taken 18 months to get a destroyer out of dry dock. He emphasized this is no longer acceptable and that Gecko will support the Navy’s goal of 80% fleet readiness by 2027, streamlining ship production so soldiers can focus on fighting and other threats. Loosararian stressed that software alone is not enough, the ability to use AI to predict and make decisions is only as good as the data inputs, representing a fundamental shift never before achieved by any robotics company in the military.

The U.S. is increasing its reliance on defense technology startups like Gecko as it seeks to modernize dated military systems amid rising geopolitical tensions. These companies are disrupting traditional defense contractors with innovative AI and autonomous tech solutions. Since taking office, President Trump has prioritized scaling and restoring U.S. shipbuilding capabilities, which have long lagged behind China, with the administration releasing a multipage plan last month to resurrect the struggling sector.

Over the years, Gecko has teamed up with mining, manufacturing, energy, and defense businesses to improve aging equipment and slash repair times, including partnerships with L3Harris Technologies, independent power operator NAES, and mining giant Freeport-McMoRan. Gecko was last valued at $1.25 billion in a $125 million funding round in June and ranked No. 30 on last year’s CNBC Disruptor 50 list.

Read more here about it!

Gecko Robotics wall-climbing robots are helping to maintain ships

Travis Kalanick launches Atoms for specialized robotics
→ Uber founder rolling CloudKitchens into new company targeting food, mining, and transportation with wheelbases vs humanoids
→ Why Pronto acquisition signals industrial autonomy as “main jam” over consumer robotaxis

Hexagon study reveals global robot anxiety divide
→ 18,000+ surveyed across 9 countries: UK adults most worried (52%), South Korea least (29%)
→ What robot density correlation demonstrates: China and South Korea’s high exposure drives lower worry, greater excitement

BMW’s 20-year automation journey reaches Physical AI era
→ 800+ Smart Transport Robots, 10-month Figure pilot producing 30K X3s, new Munich Competence Center for global rollout
→ Why imitation learning from human demos (including failures) matters more than hard-scripted programming

RealSense demos autonomous humanoid navigation
→ Dense 3D depth perception enabling localization, mapping, collision avoidance for legged robots in unstructured environments
→ What Intel spinout’s vSLAM + NVIDIA CuVSLAM reveals about shifting contact points and non-linear motion complexity

Travis Kalanick launches new company called Atoms focused on robotics! 🤖

Uber founder Travis Kalanick has launched a new company called Atoms focused on robotics that will operate in the food, mining, and transportation industries. Kalanick is rolling his existing ghost kitchen company, CloudKitchens, into Atoms. The company’s website says it will build a “wheelbase for robots,” and Kalanick said in a live interview with TBPN that his company will apply this wheelbase to “specialized robots”, not humanoids.

Kalanick emphasized that humanoids have their place, but there’s a lot of room for specialized robots that do things in an efficient, industrial-scale kind of way. To support the mining business, Kalanick revealed he’s on the precipice of acquiring Pronto, the autonomous vehicle startup focused on industrial and mining sites created by his former Uber colleague Anthony Levandowski. Kalanick is already the “largest investor” in Pronto and told TBPN that “the industrial thing is sort of like, probably, our main jam.”

Earlier Friday, The Information reported Kalanick was getting back into self-driving vehicles with “major backing” from Uber, and that he has reportedly told people he “wants to be more aggressive in rolling out self-driving technology than Waymo.” Atoms’ website makes no mention of Uber. Kalanick demurred on the idea of using Atoms robots to move people, at least in the near-term, noting that “once you crack movement in the physical world, there’s lots of people who want access to that.”

Kalanick resigned from Uber in 2017 after a confluence of crises including complaints of sexual harassment and discrimination. Before that, Kalanick had created a self-driving division at Uber in 2015, with Levandowski playing a big role after being lured from Google. Uber was sued by Google for stealing secrets related to its self-driving car project, and Levandowski was criminally charged and sentenced to 18 months in prison (later pardoned by President Trump). In a rare interview in March 2025, Kalanick expressed regret that Uber had abandoned developing its own self-driving cars.

Read more here!

ex-Uber CEO is founding a physical AI company - Atoms

Hexagon’s Robot Generation study reveals global divide in robot anxiety! 🌐

Hexagon has released a global study examining how attitudes toward robots vary sharply across generations and cultures, surveying over 18,000 adults and children across nine countries.

The research uncovers a striking truth: more than cost or technology, it’s people who ultimately determine the pace of robotic adoption, and these societal differences will set the industrial automation trajectory.

Cultural exposure and public attitudes dictate whether countries lead or lag in adoption. Hexagon CTO Burkhard Boeckem notes that trust is built through experience and clear boundaries, when people understand what robots are for and what they are not, confidence follows.

The study reveals that worry about robots is highest among adults in the UK (52%) and lowest in South Korea (29%). Children are 10% less worried about robots than adults, and comfort is highest when people imagine robots in factories or warehouses (63%). Hacking is the top concern for adults (51%), ahead of replacing humans at work (41%) and trust (26%). China is the most emotionally positive nation regarding robots, while Switzerland is the least emotionally positive.

China and South Korea have some of the highest robot densities in the world, and in these markets, greater exposure translates into noticeably lower levels of worry and much greater excitement. This points to a clear relationship, as robot density rises, so does familiarity, and with it, more positive sentiment. Exposure and attitude reinforce and accelerate each other.

The study shows robot anxiety is context dependent, highest where robots are least visible, and falls when people can see robots working safely alongside humans.

Read more here about it!

Top 10 Lessons from BMW Group Event! 🔥

1. BMW has pursued a 20+ year automation journey, evolving from 1970s industrial robot arms to sophisticated physical AI systems capable of autonomous decision-making in complex manufacturing environments.

2. Rather than manufacturing robots themselves, BMW focuses on developing the intelligence layer — creating manufacturing-specific AI knowledge that enables robots to understand and navigate BMW’s unique production requirements.

3. BMW has deployed over 800 Smart Transport Robots (STRs) across its global plants, capable of autonomous navigation without fixed routes and delivering parts just-in-sequence throughout the production network.

4. Through a partnership with Boston Dynamics, BMW’s Spot maintenance robot uses visual, thermal, and acoustic sensors for predictive maintenance, feeding real-time data directly into digital twin creation.

5. BMW’s 10-month humanoid robot pilot at its Spartanburg plant, conducted in partnership with Figure AI, supported the production of over 30,000 BMW X3 vehicles using a two-robot tandem system capable of full 10-hour shifts.

6. BMW has launched a new Physical AI Competence Center in Munich built around three pillars — industrialization, research and technology, and network partnerships — designed to replace isolated pilot projects with a structured global rollout strategy.

7. A new European partnership with Hexagon Robotics introduces the Aeon humanoid robot to BMW’s Leipzig plant, with permanent pilot operations targeting battery assembly beginning in summer 2026.

8. BMW trains its robots using imitation learning, where humans demonstrate tasks through teleoperation across multiple repetitions — including failure scenarios — shifting away from traditional hard-scripted programming toward adaptive, learning-based behavior.

9. Foundation models for BMW’s manufacturing robots are trained on three input streams — text/speech instructions, visual camera data, and robot sensor feedback — generating sequential task actions in a manner analogous to how language models produce text.

10. Leveraging a partnership with NVIDIA and millimeter-accurate digital twins of all its facilities worldwide, BMW embeds AI agents into virtual planning environments to support real-time collaborative decision-making across robotics, logistics, and ergonomics on a global scale.

Me and the Hexagon Robotics bots at the BMW Group event

Robotics maps #7 (USA 🇺🇸)

Fantastic graphic about humanoid ecosystem in the US.

RealSense unveils autonomous humanoid navigation! 👁️

RealSense demonstrated autonomous humanoid navigation with LimX Dynamics at NVIDIA GTC this week, claiming it as the first demonstration of its kind. The company emphasized that as robots including humanoids move from novelty to everyday reality, perception is no longer just about performance—it’s about safety. CEO Nadav Orbach stated that if robots are going to work safely beside humans, perception must function as the robot’s visual cortex, enabling accurate localization, collision avoidance, terrain understanding, and stable motion in unstructured environments.

Spun out from Intel last year, the Cupertino-based company delivers depth cameras and vision systems for autonomous mobile robots, humanoids, industrial automation, healthcare, and access control. The demonstration shows how dense 3D depth perception enables legged robots to localize, map, and navigate autonomously, safely, and predictably. The system uses RealSense depth cameras and vSLAM (visual simultaneous localization and mapping) integrated with odometry from NVIDIA CuVSLAM, with NVIDIA Isaac Lab accelerating development and serving as a high-fidelity digital proving ground for reinforcement learning.

RealSense explained that wheeled robots move on a relatively predictable plane, but humanoids and quadrupeds are far more complex, moving through full 3D space with shifting contact points and non-linear motion requiring stable foot placement and environmental awareness. Encoder-only odometry and 2D lidar used for wheeled robots lack the full 3D awareness needed for stable, safe motion—a gap that has constrained deployment and forced many legged robots to rely on teleoperation, supervision, or tightly controlled environments.

The partners demonstrated how RealSense’s dense depth perception fused with NVIDIA visual odometry and cuVSLAM provides scene understanding, localization, mapping, and navigation. This enables safer robotics through accurate localization and mapping, collision avoidance with people and moving objects, fall prevention through 3D terrain awareness, and predictable, human-readable motion. Dense 3D perception also unlocks behaviors including stair navigation, curb and elevation detection, uneven terrain traversal, and dynamic obstacle avoidance in shared spaces.

Read more here about it!

LimX Dynamics robot enabled with RealSense technology

Dexterity’s Foresight world model tackles truck loading
→ 4D box packing agent reasons across 3 spatial dimensions + time, making placement decisions in <400ms
→ What 100M+ autonomous production actions reveal about real-time physical manipulation vs observational models

ABB Robotics partners with NVIDIA for industrial Physical AI
→ RobotStudio HyperReality delivers 99% simulation-to-reality correlation, cuts deployment costs 40%, accelerates time-to-market 50%
→ Why exporting fully parameterized stations as USD files into Omniverse matters for synthetic data training

Yann LeCun’s AMI Labs raises $1.03B at $3.5B valuation
→ World models learning from reality vs language, starting with fundamental research not quick product releases
→ Why Nabla partnership signals focus on real-world applications where LLM hallucinations have life-threatening repercussions

Rhoda AI raises $450M Series A at $1.7B valuation
→ FutureVision platform learns from hundreds of millions of internet videos to predict physical world behavior
→ What anticipating actions dozens of times per second demonstrates about handling industrial unpredictability

Dexterity delivers a big leap for truck loading! 🚛

Dexterity has announced a major leap forward in its Physical AI stack, anchored by Foresight, a state-of-the-art world model and a 4D box packing agent. These advancements help solve some of the most physically demanding and hardest-to-staff tasks, such as truck loading.

Foresight is Dexterity’s physics-consistent world model, a real-time, transactable representation of the physical environment that enables robots to perceive, reason, and act. It represents a new class of world model, built not for observation, but for physical manipulation at production scale. In autonomous truck loading, Foresight powers Dexterity’s dual-armed superhumanoid robot, Mech, with a 4D box packing agent that reasons across three spatial dimensions plus time, determining where to place each package onto an evolving wall of freight.

This is a combinatorial problem far more complex than the game of Go, with near-infinite input variation, up to 400 potential placements per box, and multiple walls packed simultaneously. Foresight makes each placement decision in under 400 milliseconds, jointly optimizing density, stability, reachability, and dual-arm parallelism, while predicting how each placement affects the integrity of the entire truck. To date, Foresight has been trained with experience from over 100 million autonomous actions in production.

Built on Foresight, Dexterity’s agentic framework coordinates perception, decision, and motion agents that operate asynchronously to automate truck loading, package sortation, and other applications. The architecture is interpretable and safety-first, giving operators visibility into why the system makes each decision

Read more here!

Foresight, a world model developed by Dexterity

ABB Robotics partners with NVIDIA to deliver an industrial-grade Physical AI at scale! 👾

ABB Robotics and NVIDIA have announced a breakthrough partnership that brings industrial-grade physical AI to the factory floor. By integrating NVIDIA Omniverse libraries directly into its RobotStudio programming and simulation suite, ABB Robotics will deliver physically accurate simulation capabilities, dramatically cutting engineering time, reducing deployment costs by up to 40%, and accelerating time to market by as much as 50%.

The new product, RobotStudio HyperReality, will be available in the second half of 2026 and is already drawing strong interest from ABB’s global customer base. Early pilots include Foxconn, the world’s largest electronics manufacturer, and Workr, a U.S.-based robotic workforce company bringing advanced automation to small and medium-size manufacturers. RobotStudio HyperReality exports a fully parameterized robot station: robots, sensors, lighting, kinematics, and parts, as a USD file into NVIDIA Omniverse. There, ABB Robotics’ virtual controller runs the same firmware as the physical robot, enabling 99% correlation between simulation and real-world behavior.

Foxconn is piloting the technology in consumer electronics assembly, where delicate metal components and frequent product variations make automation challenging. Using HyperReality, Foxconn trains robots virtually with synthetic data, achieving high accuracy when deployed on the production line while reducing setup time and eliminating costly physical testing. Workr is integrating its WorkrCore physical AI platform with ABB industrial robots trained with synthetic data to deploy advanced automation to SMBs, planning to demonstrate AI-powered robotic systems at NVIDIA GTC 2026 that can onboard new parts in minutes without programming expertise.

Read more here about it!

Yann LeCun's AMI Labs raises $1.03 billion to build world models! 🌍

AMI Labs, the new venture cofounded by Turing Prize winner Yann LeCun after he left Meta, has raised $1.03 billion at a $3.5 billion pre-money valuation. AMI is working on world models, or AI that learns from reality, not just from language. The round was co-led by Cathay Innovation, Greycroft, Hiro Capital, HV Capital, and Bezos Expeditions, with participation from NVIDIA, Samsung, Sea, Temasek, Toyota Ventures, and angels including Tim and Rosemary Berners-Lee, Jim Breyer, Mark Cuban, Xavier Niel, and Eric Schmidt.

CEO Alexandre LeBrun predicts that “world models” will be the next buzzword, with every company calling itself a world model to raise funding within six months. However, he believes AMI Labs is fundamentally different, with a goal to understand the real world. The startup’s first partner will be Nabla, the digital health startup where LeBrun serves as chairman. As Nabla’s CEO, LeBrun reached the same conclusion as LeCun on the limitations of large language models where hallucinations could have life-threatening repercussions.

AMI Labs is a very ambitious project that starts with fundamental research, not a typical applied AI startup that can release a product in three months. It could take years for world models to go from theory to commercial applications. The French AI lab was reportedly seeking just €500 million last December but ended up raising some €890 million, likely thanks to its team. In addition to LeCun as chairman and LeBrun’s entrepreneurial track record, the team includes Meta’s VP for Europe Laurent Solly as COO, and researchers Saining Xie as chief science officer, Pascale Fung as chief research & innovation officer, and Michael Rabbat as VP of world models.

The funding will give AMI Labs runway to bankroll its two main cost centers: compute and talent. LeBrun will prioritize quality over quantity to build the team in four key locations: Paris (headquarters), New York (where LeCun teaches at NYU), Montreal (where Rabbat is based), and Singapore (to recruit AI talent and be close to future clients in Asia). Although AMI Labs has no plans to generate revenue for now, it plans to engage with prospective customers early on, putting models in real-world situations with real data and real evaluations.

Read more here about it!

Robotics maps #6 (Shenzhen 🇨🇳)

Fantastic graphic about humanoid ecosystem in Shenzhen.

Rhoda AI raises $450M at $1.7B valuation, unveils robot intelligence platform! 💾

Rhoda AI has raised $450 million in a Series A funding round that values the company at $1.7 billion and unveiled FutureVision, a robot intelligence system it says can handle the unpredictability of industrial environments. The round drew backing from Khosla Ventures, Temasek, Mayfield, Premji Invest, and Capricorn Investment Group, among others.

Rhoda’s FutureVision platform works by first studying hundreds of millions of internet videos to learn how objects move and how the physical world behaves. It then uses that knowledge to constantly anticipate what is about to happen around it and translate those predictions into physical movements, a cycle it repeats dozens of times per second. The approach targets a longstanding problem in robotics: most machines perform well in controlled, predictable environments but struggle when something unexpected happens.

The company, which emerged from stealth on Tuesday, says its platform is designed to integrate with a wide range of robotic hardware, allowing manufacturers and logistics operators to deploy intelligent robots without rebuilding existing systems. Rhoda expects to eventually license FutureVision to companies running robotic hardware and software platforms.

Breakthroughs in AI models that help robots understand language, interpret visual information, and predict how the physical world behaves, combined with growing investment from major tech and robotics companies, are driving robotics adoption.

Read more here about it!

Rhoda AI team with their robots

BMW deploys humanoid robots in Europe for first time
→ Hexagon’s AEON tested at Leipzig plant for high-voltage battery assembly and EV production
→ Why April 2026 broader test and summer pilot signal shift from lab to stable shift operation

RLWRLD raises $26M, total seed hits $42M
→ Physical AI foundation models trained on real factory/warehouse data vs clean lab settings
→ What multimodal data from unpredictable live operations reveals about simulation limitations

Wayve closes $1.2B Series D at $8.6B valuation
→ End-to-end AI Driver runs entirely onboard, zero-shot deployment in 500+ cities across 70 countries
→ Why licensing to automakers vs vertical integration matters for capital-light scaling

Unitree unveils As2 quadruped robot
→ 5 m/s top speed, 4+ hour runtime unloaded, 20 km range with 648Wh battery
→ What 90 N·m joint torque and 5 N·m/kg torque-to-weight ratio demonstrate for payload strength

BMW deploys humanoid robots in Europe for the first time! 🇪🇺

BMW is placing humanoid robots on the factory line at its Leipzig plant, marking the first time the company has deployed them at a European facility. A small fleet of Hexagon Robotics humanoids is being tested on assembly lines and in high-voltage battery production for electric vehicles, an area where employees currently wear cumbersome protective gear. A broader test deployment is planned for April 2026, with a full-scale pilot launching in summer 2026.

Hexagon’s humanoid, AEON (unveiled in June 2025), features a human-like body that can accommodate various hand, gripper, and scanning attachments, and moves dynamically on wheels. The Leipzig pilot began in December 2025, focusing on high-voltage battery assembly and component manufacturing where precision, safety, and ergonomics are critical. BMW explicitly positions humanoid robotics as a complement to existing automation, not a replacement for workers, aiming to “relieve employees and further improve working conditions” in repetitive or physically taxing roles.

BMW’s production chief Milan Nedeljković, set to become CEO in May, emphasized that the symbiosis of engineering expertise and artificial intelligence opens up entirely new possibilities in production. BMW calls this “Physical AI”—intelligent systems that perceive, reason, and act within complex, real-world manufacturing environments. According to Michael Ströbel, BMW’s head of process management, humanoids could eventually allow BMW to bring work in-house that currently goes to suppliers.

The Leipzig pilot builds on lessons from Spartanburg, South Carolina, where Figure AI’s Figure 02 robot worked daily ten-hour shifts for ten months in 2025, supporting production of over 30,000 BMW X3s. The robot handled precise removal and positioning of sheet metal parts for welding, moving over 90,000 components, logging roughly 1,250 operating hours, and covering approximately 1.2 million steps. The transition from lab to production floor happened faster than expected, with motion sequences transferring smoothly into stable shift operation.

Read more here!

BMW deployed first humanoid robots in Europe!

Physical AI startup RLWRLD raises $26M! 🤖

RLWRLD, a physical AI startup developing robot foundation models for industrial environments, has raised $26 million in Seed 2 funding. This brings the Seoul-based company’s total seed capital to $42 million, following its $14.8 million Seed 1 round in April 2025. The round was led by global venture firm Headline Asia and Z Ventures (the corporate venture capital arm of Z Holdings and Yahoo Japan–LINE), with participation from strategic investors including CJ Logistics, Kakao Investment, Lotte Ventures, Hanwha Asset Management, and Mirae Asset–Emart Investment Fund.

Unlike robotics systems trained primarily in clean, controlled lab settings, RLWRLD’s models are trained using high-precision, multimodal data collected directly inside real factories, warehouses, and service environments. The company exposes its models to the unpredictability of live operations, crowded spaces, variable lighting, human interaction, and imperfect conditions, that are difficult to replicate in simulations. RLWRLD’s scalable, multi-sensor system captures complex demonstrations from diverse industrial settings, generating rich data for robust, adaptable models.

RLWRLD is executing proof-of-concept and robotics transformation projects with enterprise partners across logistics, retail, and service environments in Japan and South Korea. Several initiatives with major companies, including CJ Logistics and Lotte, have progressed from memoranda of understanding (MOUs) into joint industrial validation phases. The company is actively working with multiple strategic investors on joint projects that include using real-world operational data to train and validate its models.

Read more here about it!

Wayve raises $1.2B with plans to bring robotaxis to London! 🚖

Wayve Technologies has closed a Series D round of $1.2 billion, bringing its post-money valuation to $8.6 billion. The London-based company plans to use the funding to accelerate commercialization of its end-to-end AI platform for autonomous driving.

Founded in 2017, Wayve’s AI Driver doesn’t rely on deterministic, map-based systems and runs entirely on onboard vehicle compute. The unified AI platform spans SAE Level 2 “hands-off” operations through Level 4 “eyes-off” driving across vehicles, brands, and markets. The company’s AV2.0 foundation model replaces the “sense-plan-act” architecture with a single neural network, trained on data from over 70 countries and a wide range of vehicles for easier generalization to new markets.

Wayve claims to be the first and only AV developer to drive zero-shot in more than 500 cities across Europe, North America, and Japan, meaning without city-specific fine-tuning before deployment. The company licenses its software directly to automakers, providing tools to customize driving models for specific vehicles and brands. By partnering with manufacturers and mobility platforms rather than vertically integrating, Wayve asserts its approach enables autonomous vehicle technology to scale globally with lower capital intensity.

This year, Wayve plans to conduct robotaxi trials with Uber (a past and current investor) in London. Under the partnership, Wayve will deploy its AI Driver in L4-capable vehicles from participating automakers, while Uber will own and operate the fleet. In 2027, consumers will be able to buy AI Driver-equipped passenger vehicles, starting with L2+ capability that allows the vehicle to steer, navigate, and respond to traffic under driver supervision.

Read more here about it!

Robotics maps #5 (Odense 🇩🇰)

Odense is one of Europe’s strongest robotics ecosystems, especially for collaborative and industrial robots. At the center is Syddansk Universitet (University of Southern Denmark/SDU), which has a leading robotics research department and works closely with startups. Many founders and engineers come directly from SDU labs, making it easy to turn research into real companies.

A major boost came from the success and acquisitions of Universal Robots and Mobile Industrial Robots. Their global growth and later acquisitions created serial founders and angel investors, capital recycled back into new startups, and a strong global reputation for Odense robotics. Startup support initiatives like Odense Robotics help with global commercialization and partnerships, allowing more robotics startups to be incubated locally with support from experienced mentors who have already built and scaled global robot companies.

Key Odense-based companies include Universal Robots (collaborative robotic arms, acquired by Teradyne for ~$285M), Mobile Industrial Robots/MiR (autonomous mobile robots for internal logistics, acquired by Teradyne for ~$148M), and Teradyne Robotics (Teradyne’s robotics division). OnRobot provides plug-and-play end-of-arm tooling for collaborative robots (~$55M+ raised), while ROICO develops robotic solutions for industrial painting applications.

Other notable players include Spin Robotics (collaborative robot torque tools for automated screwdriving), Enabled Robotics (mobile manipulator robots for flexible logistics automation), Nord Modules (modular top modules for AMRs to expand logistics functionality), and Smooth Robotics (automated welding solutions for cobots). Essential Robotics builds autonomous healthcare solutions, XiniX AI develops AI software and vision for robotics, ARIS Robotics creates advanced robotic systems for industrial and research applications, RobSub builds underwater robotic systems and ROVs for marine applications, and Nordbo Robotics develops force-torque sensor technology and adaptive control software for robots.

Unitree Robotics revealed their new robot dog! 🐕

Chinese robotics firm Unitree Robotics has unveiled the As2, a high-performance quadruped robot built for speed, payload strength, and advanced autonomous capabilities. Weighing about 18 kg (40 pounds) with its battery, the robot delivers up to 90 N·m of maximum joint torque and can reach running speeds above 5 meters per second (approximately 11 mph).

The As2 is powered by a 648Wh, 15,000mAh lithium battery that allows more than four hours of runtime when unloaded, covering distances beyond 20 km (12.4 miles). Even when carrying a 33 lbs load, it maintains more than 2.5 hours of operation with a walking range exceeding 13 km (8.08 miles). Designed for real-world environments, the robot carries an IP54 rating for dust and rain resistance and can operate in temperatures from -20°C (-4°F) to 50°C (122°F).

The quadruped demonstrates impressive mobility metrics. It can climb 25 cm stairs, handle slopes up to 40 degrees, and mount vertical platforms as high as 50 cm. Maximum continuous walking load reaches around 15 kg (33 pounds), while its standing payload capacity goes up to 65 kg (143 pounds) in the EDU variant. Speed varies by model: the As2 AIR runs between 0 and 3.0 m/s, while the PRO and EDU versions reach 3.7 m/s, with peak speeds approaching 5 m/s under no-load conditions.

Mechanically, the As2 features 12 degrees of freedom driven by low-inertia, high-speed inner rotor PMSM motors with industrial-grade crossed roller bearings for high precision and load capacity. The torque-to-weight ratio stands at approximately 5 N·m/kg, enabling powerful yet agile motion. The robot integrates dual encoders for each joint, local air cooling, supply voltage from 36V to 50.4V, plus Wi-Fi 6, Bluetooth 5.2, speaker, microphone, front light, and HD camera across all versions.

Read more here about it!

Newest addition to Unitree Robotics family

Apptronik raises $935M total at $5.3B valuation
→ Series A extended from $350M to $935M as valuation triples in a year
→ Why Google DeepMind, GXO, and Mercedes partnership signals embodied AI beyond fixed instructions

Weave Robotics ships $7,999 laundry-folding robot
→ Isaac 0 pivots from mobile to stationary desk-mounted system with “ship-now” mentality
→ What remote teleoperation for edge cases reveals about training pipeline strategy: 5-10 second human corrections improve autonomous models weekly

AutoPallet’s upside-down robots solve density
→ YC-backed startup demos ceiling-mounted magnetic robots for palletizing/depalletizing
→ Why freestanding modular superstructure matters: dense automation in existing warehouses without redesigning material flow

Allonic raises €6M pre-Seed for robotic body platform
→ 3D Tissue Braiding produces integrated compliant structures in single automated process vs assembling hundreds of parts
→ What bio-inspired manufacturing demonstrates: tendons, joints, and soft tissues formed together continuously

Apptronik has now raised $935M at a $5B+ valuation! 💰

Apptronik building humanoid robots for Google DeepMind and others, has re-opened its Series A to raise a total of $935 million. While the company didn’t disclose the valuation, TechCrunch learned separately that its post-money valuation is now about $5.3 billion, roughly triple the initial Series A valuation of around $1.75 billion.

Apptronik previously announced a $350 million Series A a year ago, then expanded it to $415 million due to strong demand. Now it has raised another $520 million from earlier investors Google, Mercedes-Benz, and B Capital, alongside new investors. The company says investors have paid progressively more for shares in each subsequent extension. The startup maintains it’s still in early development stages and wasn’t actively seeking funding, but rather dealing with inbound interest.

Part of the excitement is Apptronik’s partnership with Google DeepMind, GXO, and Mercedes-Benz to deliver embodied AI, robots capable of perceiving their environment and taking physical action based on reasoning rather than just following fixed instructions. The robot is being built for tasks like unloading trailers, picking warehouse inventory, and tending machinery.

Despite the early-stage funding label, Apptronik’s humanoid work dates back to 2013 when members of the Human Centered Robotics Lab from the University of Texas competed in the NASA-DARPA Robotics Challenge with a robot called Valkyrie. The company was formally founded in 2016, and NASA has maintained a partnership with Apptronik as it developed its own humanoid robot, Apollo.

Read more here!

Apptronik founders with their robot named Apollo

Weave Robotics launches a $7,999 laundry-folding robot 👕

San Francisco-based Weave Robotics has announced Isaac 0, a stationary laundry-folding robot priced at $7,999. Founded in 2024 by Apple and Carnegie Mellon alumni Kaan Doğrusöz and Evan Wineland, Weave had previously announced a mobile, telescoping robot named Isaac capable of autonomous tidying and laundry folding. The company is now prioritizing a “ship-now” mentality with this desk-mounted system that ditches the wheeled base for immediate domestic utility.

Isaac 0 is a stationary torso with a neck (4 degrees of freedom) and dual arms (6 degrees of freedom each), capable of folding an average load of laundry in 30 to 90 minutes. The robot handles t-shirts, hoodies, pants, towels, and undergarments, though it cannot yet handle large blankets or turn garments right-side out. The platform features a carbon fiber structure to minimize weight, a telescoping torso, and an 18” x 19” footprint.

Weave is offering two ownership paths for Bay Area residents: a one-time payment of $7,999 or a $450 monthly subscription. A critical component is “Remote Op” teleoperation, while the robot runs autonomously for common items, if it encounters a difficult sweater or unknown textile, a Weave specialist can “sub in” remotely for a 5-to-10-second correction. This data is then fed back into the training pipeline to improve the robot’s autonomous models weekly.

Despite the pivot to a stationary form factor, the mobile version remains part of the roadmap. To maintain loyalty among early adopters, customers who purchase Isaac 0 will receive preferred pricing when they transition to the wheel-based Isaac, which is designed to combat domestic “entropy” by autonomously tidying surfaces and returning objects to their proper places.

Read more here about it!

Weave Robotics Isaac 0 folding clothes

How ‘upside-down’ robots solve palletizing density! 🔄

AutoPallet Robotics made the first public demonstration of its novel palletizing/depalletizing system at Manifest 2026. The Y Combinator-backed startup (founded in 2024) is solving warehouse density challenges with small autonomous mobile robots that drive upside down, magnetically affixed to a steel plate positioned over the target workspace. The robots lower a vacuum gripper to a pallet or conveyor to acquire a box, then lift it up and carry it to the target drop location.

The freestanding modular superstructure bolts into the warehouse floor, allowing operators to drop high-density automation into existing buildings without redesigning their entire material flow. Co-founder Eric Miller explained the company started from the question: “What would a purpose-built, modern robotic architecture for case handling look like if you didn’t assume a robot arm on a pedestal?” By putting small, efficient robots on the ceiling, AutoPallet gives operators dense, flexible sortation and palletizing where traditional automation solutions don’t fit.

AutoPallet robots can receive mixed streams of diverse cases, sort them across many pallet positions, and directly build dense pallets in the same zone. Because all complexity lives in the ceiling of the workcell, pallets and conveyors can be packed tightly under the panels, delivering floorspace density impossible to achieve with traditional loop sorters, tilt-trays, or arm-based cells.

The robots are completely self-contained, battery-powered units. The mobile base has an onboard battery for drive propulsion, and the gripper has its own battery to power the vacuum pump. Each robot communicates with others through a wireless mesh network to share instructions and avoid collisions.

Read more here about it!

Robotics maps #4 (London 🇬🇧)

London has built a strong robotics ecosystem thanks to excellent engineers and researchers from universities like Imperial College London and UCL, both well known for robotics, AI, and engineering. Many robotics founders and early employees come directly from these institutions. The city is also home to Google DeepMind, one of the world’s leading AI labs, whose work on robot learning, control, and general AI has pushed forward how robots learn and adapt in the real world.

London has one of Europe’s strongest investor ecosystems. As a major global finance hub, it’s easier to find venture capital, corporate investors, and early customers, especially for robotics companies in logistics, healthcare, and automation. The city is very international and business-friendly, making it easy to hire talent from around the world, set up a company, and sell globally.

Key London-based robotics companies include Humanoid (general-purpose AI-driven humanoid robots), Automata Tech (easy-to-deploy robotic automation for SMBs), and Shadow Robot (advanced dexterous robotic hands for research and industrial automation). Paddington Robotics designs small autonomous robots for retail and service environments, while KAIKAKU builds robotics and AI solutions to optimize warehouse and logistics processes.

Other notable players include Automated Architecture/AUAR (spatial computing and robotic systems for construction and design), Recycleye (AI-powered robotics for waste processing, ~$20M+ raised), and Neuracore (AI-based perception and planning software, $10M+ raised). Apian Health develops autonomous robotic systems for medication dispensing and hospital logistics, Slamcore offers high-performance SLAM navigation software (~$6M+ raised), and Moley Robotics builds fully automated robotic kitchen systems (~$30M+ raised). Dexory develops autonomous warehouse robots with AI software for inventory scanning ($80M Series B and $165M Series C), while Extend Robotics builds tele-operation technology for robot fleet orchestration.

Allonic raises €6M for robotic body manufacturing platform! 🦾

Budapest-based robotics startup Allonic has raised €6 million ($7.2 million) in pre-Seed funding to industrialize a new way of producing complex, dexterous robotic bodies through its manufacturing platform. The round was led by Visionaries Club, with participation from Day One Capital, Prototype, SDAC Ventures, and TinyVC, plus over a dozen angel investors including OpenAI and Hugging Face, as well as research institutions like ETH Zurich and Northwestern University.

Founded in 2021, Allonic is pioneering 3D Tissue Braiding, a manufacturing platform for producing integrated, compliant, bio-inspired robotic bodies through a single, fully automated production process. Instead of assembling hundreds of individual parts, Allonic’s tendons, joints, and load-bearing soft tissues are formed together in one continuous process. A robotic finger, for example, can be built from a small number of bone-like elements held together by hundreds of fine fibers that wrap around and anchor into the structure, similar to connective tissue in the human body.

Allonic’s platform combines proprietary hardware and software that allows users to configure high-level robotic designs and translate them automatically into production code, similar to slicing in 3D printing. Multiple materials can be integrated into a single structure, enabling elastics, wiring, and sensing elements to be embedded directly during production. Advanced robotic components or complete limbs can be manufactured directly from digital designs without managing complex supply chains or distributed assembly processes.

Co-founder and CEO Benedek Tasi explains that while attention focuses on intelligence and software, hardware still holds many of the hardest problems. The trade-offs between durability and softness, dexterity and strength have been dictated by manufacturing limits.

Read more here about it!

Carbon Robotics introduced a large model for plants

RobCo raises $100M Series C, valuation tops $500M
→ German automation startup accelerates US expansion in San Francisco and Austin → Why full-stack platform approach (hardware + software) matters more than point solutions

Bedrock Robotics’ $270M Series B for operator-less excavators
→ Total funding exceeds $350M as company moves toward fully autonomous deployments in 2026
→ What retrofit kits reveal about scaling strategy: leveraging existing fleet infrastructure vs selling new machines

World’s fastest humanoid hits 10 m/s
→ MirrorMe’s Bolt claims speed record with human-like proportions and optimized power system
→ China’s push into robotic athletics gaining momentum since first World Humanoid Robot Games

Carbon Robotics launches first Large Plant Model
→ 150M labeled plants enabling laser weeding adaptation in minutes vs weeks
→ What compounding data flywheel from global LaserWeeder fleet demonstrates about real-world learning loops

LimX Dynamics closes $200M Series B
→ TRON 2’s modular architecture reconfigures limbs between arms and legs without hardware customization
→ Why agentic OS (COSA) combining reasoning with whole-body control matters more than fixed scripts

RobCo raises $100M to expand growing US presence! 🇺🇸

German automation startup RobCo has raised $100 million in a Series C round to advance its physical AI roadmap, expand enterprise deployments, and accelerate its US market presence. The round was led by Lightspeed Venture Partners and Lingotto Innovation (an arm of Exor), with participation from Sequoia Capital, Greenfield Partners, Kindred Capital, Leitmotif, and the Friedkin Group. Bloomberg reported the funding takes the company’s valuation above $500 million.

Founded in 2020 by three researchers at the Technical University of Munich, RobCo develops AI-powered flexible robotic hardware kits for SME manufacturing applications like machine loading, unloading, and palletizing. The Munich-based company builds hardware and software as a single full-stack platform, with robots that can acquire task-specific skills through demonstration and self-learning rather than manual programming, enabling faster deployment and easier adoption.

RobCo expanded into the US in 2025 and currently operates in San Francisco and Austin. The company says the US has become a strategic priority and a major growth market, with manufacturers rapidly adopting automation in response to labor constraints and reshoring initiatives. Co-founder and CEO Roman Hölzl stated the funding will help Robco become the dominant AI robotics company for manufacturing in the US and Europe.

RobCo previously raised $42.5 million in 2024 in a round also led by Lightspeed. Alexander Schmitt from Lightspeed said the firm is excited to double down and co-lead the $100 million round, noting that Robco has continued to raise the standard for what modern robotics can look like in real-world production.

Read more here!

RobCo founders with their modular bots in Munich

Bedrock Robotics’ $270M Series B to scale the operator-less excavators! 🚜

Bedrock Robotics has raised $270 million in Series B funding, bringing its total funding to over $350 million. The round was led by CapitalG and the Valor Atreides AI Fund, with participation from Xora, 8VC, Eclipse, Emergence Capital, Perry Creek Capital, NVentures (NVIDIA’s venture capital arm), Tishman Speyer, MIT, Georgian, Incharge Capital, C4 Ventures, and others.

The San Francisco-based company emerged from stealth in July 2025 with $80 million in Series A to build retrofit kits for excavators. Co-founder and CEO Boris Sofman, who previously worked at Waymo, told The Robot Report the company is working towards its first fully operator-out, completely operatorless deployments later this year. This marks a major milestone where autonomy capabilities are mature, safety systems are ready, and the company understands the job interfaces needed.

The investment will accelerate Bedrock’s efforts to transform how general contractors build, from deploying individual autonomous machines to orchestrating fully connected fleets. Much of the funding will support data collection, model development, testing, deployment, and expansion of capabilities and volume. The company plans to grow its team and expand its geographical footprint, eventually becoming a giant ecosystem provider by leveraging the existing infrastructure of companies that already buy and manage these machines.

The market opportunity is substantial. McKinsey estimates a cumulative $106 trillion investment is required by 2040 to address global infrastructure needs, while the construction industry needs to attract an estimated 349,000 net new workers in 2026 to meet demand. Bedrock recently expanded its leadership team with Vincent Gonguet as head of evaluation (previously led AI safety for Meta’s Llama models) and John Chu as head of people (former Waymo role overseeing 400% headcount growth).

Read more here about it!

Bedrock Robotics is retrofitting construction machines to be autonomous

World’s fastest humanoid robot runs 10 meters per second! ⚡

Chinese robotics firm MirrorMe Technology has unveiled Bolt, claiming it as the fastest-running humanoid robot to date with a peak speed of 10 meters per second during real-world testing. The company says Bolt is the first humanoid robot of its size to reach such velocities outside the lab, marking a new benchmark in bipedal robotics.

Standing 175 centimeters (5.7 feet) tall and weighing 75 kilograms (165 pounds), Bolt reflects MirrorMe’s vision of the humanoid robot’s “ideal form,” combining human-like proportions with performance-focused engineering. The robot is built with newly designed joints and a fully optimized power system intended to replicate natural human movement while sustaining extreme speeds. In promotional footage, Bolt took shorter strides than founder Wang Hongtao but compensated with a much faster cadence.

MirrorMe’s goal extends beyond speed records. The company aims to develop “super-species” robots capable of approaching or surpassing human athletic abilities through advanced motion perception and physical performance. Last year, its Black Panther II ran 100 meters in 13.17 seconds on Chinese TV, reportedly outperforming Boston Dynamics in speed tests.

Formally established in May 2024 and headquartered in Shanghai with a core team from Zhejiang University, MirrorMe’s product lineup includes Baobao, Apollo, and Black Panther II, primarily designed for scientific research. The company has treated speed as a long-term priority since 2016, pushing the boundaries of robotic athletic performance. China’s growing focus on robotic athletics was highlighted last August when Beijing hosted the first World Humanoid Robot Games.

Read more here about it!

Robotics maps #3 (Paris 🇫🇷)

Paris is emerging as one of Europe’s strongest robotics hubs, mixing world-class tech talent, big startup energy, and strong public support for deep tech. The city has excellent research and engineers from places like Inria (a major robotics and AI research institute) and top schools around Paris/Saclay. This matters because robotics startups often start from research and need smart people who can build real systems.

Paris is one of Europe’s biggest startup hubs thanks to STATION F, one of the world’s largest startup campuses. It gives founders access to mentors, partners, and investors, making it easier to meet the right people fast. France is unusually strong in government-backed deep-tech funding through programs like French Tech 2030 and support from Bpifrance, providing crucial backing for robotics companies that need time and money for prototypes, hardware, and testing.

Key Paris-based robotics companies include Genesis AI (physical-AI lab building generalist robots and foundation models, launched with $105M), Stanley Robotics (robotic valet parking systems), and UMA (general-purpose mobile/humanoid robots for industrial tasks). LeRobot from Hugging Face offers an open-source robotics stack with datasets and models to lower barriers to robot learning, with Pollen Robotics in their portfolio.

Other notable players include Enchanted Tools (friendly humanoid robots like Mirokaï for service environments, ~$17M seed), Fuzzy Logic Robotics (quick robot programming platform, €2.5M seed), and Inbolt (real-time 3D vision guidance for industrial arms, ~$16-17M Series A). Heex Technologies builds smart data software for edge AI (€6M seed), while Wandercraft develops robotic exoskeletons and humanoid R&D ($75M total funding). SoftBank Robotics EMEA HQ commercializes service robots like Pepper and NAO, Galam Robotics builds modular warehouse automation (€10M Series A), Exwayz develops GPS-free 3D localization for autonomous robots (€1M), and phospho (YC W24) builds tools to connect, control, and train AI models on robots.

Carbon Robotics launches the world’s first-ever Large Plant Model! 🌱

Carbon Robotics has announced a major breakthrough with the launch of the world’s first Large Plant Model (LPM), the most advanced AI model for plant detection and identification. Trained on the largest, most diverse, and fastest growing agricultural dataset ever built with 150 million labeled plants, the LPM enables farmers to start laser weeding any field or crop in minutes.

The LPM serves as the foundation for Carbon AI, the decision-making brain operating across all Carbon Robotics products. As the company’s global LaserWeeder fleet operates in fields daily, the system continuously ingests real-world data in a compounding data flywheel effect, strengthening the model exponentially and improving performance for all LaserWeeders worldwide. Trained on diverse crops, weeds, soil types, climates, and growth stages, the LPM creates an unprecedented foundation for plant recognition and decision making.

A new feature called Plant Profiles demonstrates how this AI translates into practical value. It enables farmers to quickly tailor the foundational LPM to their unique crops, weeds, and field conditions. Operators can select just two to three images in the iPad Operator App to add to their Plant Profile and the system immediately adapts, optimizing LaserWeeder behavior for their specific fields. Unlike other AI-based systems that require substantial farmer input and take weeks or months to create new models, Plant Profiles enables adaptation in minutes.

Read more here about it!

Carbon Robotics introduced a large model for plants

LimX Dynamics picks up $200M for expansion! 💸

LimX Dynamics has completed a $200 million Series B financing round to accelerate research and development and market expansion with new and existing shareholders, plus ecosystem partners. The Shenzhen-based company, founded in 2022, is developing full-size humanoid robots and embodied AI with a mission of bringing artificial general intelligence into the real world, focusing on robot hardware design and manufacturing, motion-control foundation models, and a physical-world-native agentic OS.

LimX said its TRON 2 and LimX COSA systems have attracted industry attention in recent months. Thanks to its modular architecture, TRON 2 can be reconfigured to use its limbs as manipulator arms or legs for walking. Unveiled in December 2025, this unified design allows developers and industry users to adapt the robot to different applications without customizing hardware for individual tasks, serving as a reusable, scalable, and continuously evolving robotic foundational platform.

LimX COSA is an agentic operating system developed from the ground up for robots to function in the physical world. The company says it combines high-level reasoning and decision-making with whole-body motion control. With COSA, the Oli humanoid no longer relies on fixed scripts. Instead, the OS orchestrates models, skills, and hardware as a unified system, enabling robots to autonomously execute complex tasks by interpreting instructions, perceiving surroundings, planning actions, and adjusting behaviors in real time to handle dynamic environments.

Leading Chinese and international investors joined the Series B round, including Stone Venture, JD, Oriental Fortune Capital, and CoStone Capital. Shangqi Capital (backed by SAIC Motor), NIO Capital, Future Capital, and other existing shareholders also participated in the round.

Here's more about this robot!

LimX Dynamics humanoid picking up tennis balls

1X’s world model lets NEO learn from video.
→ AI update enables learning tasks from internet-scale video without prior examples. → Why inverse dynamics models translating visualizations into movements matter more than teleoperation datasets.

Microsoft enters robotics with Rho-alpha
→ First robotics model from Phi series adding tactile sensing and continuous learning from human feedback.
→ What NVIDIA Isaac Sim on Azure reveals about synthetic data strategy for overcoming real-world scarcity.

Zipline raises $600M at $7.6B valuation
→ U.S. expansion to Houston, Phoenix, and four more states in 2026.
→ Why 2M deliveries and 15% weekly growth signal operational proof over hype.

Unbox Robotics closes $28M Series B
→ Swarm-intelligence software with modular 3D sortation hardware scaling across India, Europe, and U.S.
→ Why minimal fixed infrastructure and dynamic coordination address throughput bottlenecks.

Humanoid completes Siemens POC deployment
→ HMND 01 Alpha hits 60 totes/hour, 90%+ success rate, 8+ hour uptime in real production environment.
→ What two-week on-site validation at Erlangen factory demonstrates beyond lab conditions.

1X launches world model enabling NEO to learn tasks by watching videos! 📹

1X has announced its latest 1X World Model, an AI update that enables the NEO humanoid robot to turn any request into an AI capability on demand using a video model grounded in real-world physics. After years of development, NEO can now learn from internet-scale video and apply that knowledge directly to the physical world, even performing tasks with objects and environments it has never encountered before.

With this update, users can give NEO a simple voice or text prompt, and the robot uses what it’s looking at to generate visualizations of future actions. A built-in inverse dynamics model then translates these into precise movements to complete the request. Demonstrations show NEO handling completely novel tasks like operating a toilet seat, opening a sliding door, ironing a shirt, and brushing a human’s hair without any prior examples in its dataset, highlighting the transfer of broad human knowledge through the World Model.

Where traditional AI models for humanoid robots have relied on data collected by human operators, the 1X World Model enables NEO to collect its own data and autonomously master new capabilities. The model also benefits from improvements in video models, since it uses a video model at its core. This addresses a major bottleneck where improvement in AI capabilities for humanoids has been limited by the speed at which robot data can be collected.

NEO is designed for household use and available through 1X’s early access program for $20,000, which includes priority delivery in 2026. A subscription model will also be available for $499/month.

Watch their launch video here:

NEO from 1X cleaning up at home

Microsoft presents Rho-alpha, its first robotics model! 🤖

Microsoft Research has announced Rho-alpha (ρα), its first robotics model derived from the Phi series of vision-language models. The model translates natural language commands into control signals for robotic systems performing bimanual manipulation tasks, expanding beyond typical vision-language-action models by adding tactile sensing and working toward continuous learning from human feedback during deployment.

Rho-alpha is designed to make physical systems more adaptable to dynamic situations and human preferences. The model achieves tactile-aware behaviors through co-training on trajectories from physical demonstrations and simulated tasks, combined with web-scale visual question answering data. Microsoft uses NVIDIA Isaac Sim on Azure to generate physically accurate synthetic datasets, helping overcome the scarcity of diverse real-world robotics data, especially for tactile feedback and less common sensing modalities.

The model is currently under evaluation on dual-arm setups and humanoid robots. Recent demonstrations show Rho-alpha controlling a tactile sensor-equipped dual-UR5e-arm setup to perform tasks like plug insertion and toolbox packing. When the robot struggles, human operators can provide real-time corrective guidance using intuitive teleoperation devices like a 3D mouse, and Microsoft is working on enabling Rho-alpha to learn from this feedback during operation.

Microsoft is inviting organizations to express interest in the Rho-alpha Research Early Access Program, with broader availability planned via Microsoft Foundry later.

Another top company joining the robotics race.

Read more here about it!

Robots with the new Microsoft's model Rho

Zipline’s expansion with $600M in new funding! 🚁

Zipline has raised $600 million in fresh funding at a $7.6 billion valuation to fuel its U.S. expansion, with plans to launch operations in Houston and Phoenix early this year and expand to at least four states throughout 2026. The round includes participation from Fidelity Management & Research Company, Baillie Gifford, Valor Equity Partners, and Tiger Global.

Founded in 2014, Zipline developed its own drone delivery ecosystem, including logistics software, launch and landing systems, and aircraft. The company started commercially in 2016 delivering blood in Rwanda and now operates in five African countries, several U.S. cities, and Japan. Last year, Zipline launched a home delivery service in the U.S. using its Platform 2 drones, which carry up to eight pounds and travel within a 10-mile radius. Its larger Platform 1 drones handle long-range deliveries covering 120 miles round trip for enterprise and government clients.

The company’s delivery numbers show strong momentum. After completing 1 million drone deliveries in 2024, Zipline surpassed 2 million deliveries this week. U.S. deliveries have grown about 15% week-over-week for the last seven months. The home delivery service started in Pea Ridge, Arkansas, and Dallas-Fort Worth with Walmart and more than a dozen restaurant brands, with plans to launch in Seattle. Commercial partners include Panera, Chipotle, Crumbl, Blaze Pizza, Wendy’s, and Little Caesars.

Co-founder and CEO Keller Cliffton calls 2026 the company’s breakout year, stating that when deliveries are faster, cleaner, safer, and cheaper, demand grows exponentially. Zipline faces competition from Flytrex, DroneUp, Amazon Prime Air, and Wing, the Alphabet-owned subsidiary that has partnered with Walmart and plans to expand to another 150 stores through 2027.

Read more here about it!

Zipline's aircraft in action

Robotics maps #1 (Zurich 🇨🇭)

Zurich has built one of the world’s strongest robotics ecosystems, with everything startups need close by and well connected. The city draws top engineering talent from ETH Zürich, one of the best robotics and AI universities globally, with many successful startups coming directly from ETH research. The presence of Disney Research and RAI Institute keeps the region at the frontier of physical AI.

Switzerland is home to global companies in robotics, manufacturing, and automation like ABB Robotics, which often work with startups as partners or early customers. Zurich offers strong access to funding, especially for deep-tech and robotics, with investors comfortable with long development cycles and complex hardware products. The city’s stability, safety, and central European location make it easier to attract international talent and scale globally.

Key Zurich-based robotics companies include ANYbotics (autonomous four-legged robots for industrial inspection, $150M+ raised), Gravis Robotics (AI autonomy for construction machines, ~$27M raised), and Verity (autonomous warehouse inventory drones, ~$60M+ raised). mimic develops AI-powered robotic picking for logistics, while Bota Systems makes high-precision force-torque sensors for robot tactile awareness.

Other notable players include Duatic (human-scale robotic arms), RIVR (wheeled-legged robots for last-mile delivery, $22M seed), Flexion Robotics (humanoid robot AI “brain”, $50M raised), and FORGIS (AI edge software for industrial machines, and Sevensense, which built AI-powered 3D vision navigation tech, was acquired by ABB Robotics in January 2024.

That’s quite an ecosystem, right? 😮‍💨

Unbox Robotics secures $28M in Series B! 📦

Unbox Robotics, a supply chain robotics technology company, has raised $28 million (INR 243 crore) in Series B funding led by ICICI Venture and Redstart Labs (Infoedge), with co-investment from F-Prime, 3one4 Capital, Navam Capital, Force Ventures, and other existing investors. The round, comprising primary and secondary capital, will strengthen leadership and engineering teams, accelerate new product development, and expand market presence across India and select international markets.

Founded with the vision of transforming intralogistics through intelligent automation, Unbox Robotics designs and deploys modular robotic systems that enable faster, scalable, and more efficient order fulfillment for enterprises. The company’s platform combines proprietary swarm-intelligence software with modular 3D robotic sortation hardware, allowing large fleets of robots to coordinate dynamically and scale throughput with minimal fixed infrastructure.

Unbox Robotics has built a growing customer base across Europe, the United States, and India, supporting global e-commerce, retail, and third-party logistics operators. The company also created meaningful liquidity for employees through its ESOP program as part of the transaction.

Read more here about it!

Co-founders of Unbox Robotics

Humanoid and Siemens’ industrial deployment! 🏭

Humanoid and Siemens have completed a proof of concept demonstrating the HMND 01 wheeled Alpha robot in industrial logistics at a Siemens facility. The POC was structured in two phases: in-house development with a “physical twin” for testing and optimization, followed by a two-week on-site deployment at the Siemens Electronics Factory in Erlangen, Germany, where the robot operated in a real-world production environment.

The HMND 01 Alpha conducted a tote-to-conveyor de-stacking task, autonomously picking totes from a storage stack, transporting them to a conveyor, and placing them at the designated pickup point for human operators. The robot repeated this sequence until the stack was fully empty, demonstrating how robots can handle repetitive logistics tasks. The system met all target metrics, including 60 tote moves per hour, operation with two different tote sizes, continuous autonomous task execution for over 30 minutes, and uptime exceeding 8 hours.

Both the overall pick-and-place success rate and the autonomous pick-and-place success rate exceeded 90%. Humanoid and Siemens said this POC is a first step toward long-term strategic collaboration, with plans to expand the scope, add use cases, and potentially progress toward a broader rollout deploying more robots across Siemens’ facilities based on the robot’s specific skill set.

Humanoid founder and CEO Artem Sokolov emphasized the company’s commercial focus on creating robots that deliver measurable value in real-world settings. This week, Humanoid also partnered with Schaeffler Technologies, agreeing to bring hundreds of robots into Schaeffler’s production facilities over the next five years.

Here's more about this robot!

Skild AI hits $14B valuation:
→ $1.4B Series C led by SoftBank tripling valuation in 7 months.
→ The race to make robots adaptable without lengthy retraining.

Schaeffler plans to deploy hundreds of humanoids:
→ Partnership with Humanoid for wheeled platform across global facilities.
→ Why joint actuator supply deals signal industrial readiness over prototype hype.

X Square Robot raises $140M in funding:
→ Foundation models for physical world with real-robot RL at scale.
→ What autonomous food delivery demos reveal about causal inference and zero-shot generalization.

Mobileye acquires Mentee Robotics for $900M:
→ Intel subsidiary pivots automotive vision tech to humanoids.
→ Why $24.5B automotive pipeline funding “costly enterprise” matters.

NEURA unveils next-gen humanoids at CES:
→ 4NE1, 4NE1 Mini, and quadruped all integrated with Neuraverse OS.
→ Why shared learning platforms across robot networks could be the real differentiator.

Robotics software maker Skild AI hits $14B valuation! 💰

Skild AI, which builds foundation models for robots, has raised a $1.4 billion Series C round that values the company at more than $14 billion. The round was led by SoftBank, with participation from NVIDIA, Macquarie Group, 1789 Capital, and others. This represents more than a tripling of its valuation in just seven months, jumping from $4.5 billion last summer.

Founded in 2023, Skild AI develops general-purpose robotic software and foundation models that can be retrofitted to a variety of different robots and tasks without requiring extensive additional training. The models are designed to learn from watching humans perform tasks, addressing one of the biggest hurdles in robot adoption: the sheer amount of training required for robots to learn each new task.

CEO Deepak Pathak confirmed the company has now raised more than $2 billion to date. The funding comes amid a major push into learn-as-you-go robotic software alongside rising hype around humanoids. Being able to learn and adapt on the fly would clear the path for broader robotic adoption in both personal and industrial use cases.

Skild AI joins other startups like Field AI and 1X pursuing easily adapted robotic software. The race is on to build AI systems that can make robots more flexible, intelligent, and ready for real-world deployment without lengthy task-specific programming.

Read more about it here!

Schaeffler will deploy hundreds of humanoid robots! 🤯

Schaeffler has announced a strategic partnership with Humanoid (London-based humanoid maker) to deploy hundreds of humanoid robots across its global production facilities over the next five years.

Under the agreement, Schaeffler will become Humanoid’s preferred supplier for joint actuators used in its wheeled platform. The companies will explore joint development of next-generation actuators, with Schaeffler’s planetary gear actuator combining a two-stage planetary gearbox, electric motor, encoder, and controller in a compact unit for precise movement and continuous operation. Initial deployments are scheduled to begin with beta-stage robots in 2026 and 2027, focusing on validating technical integration, operational performance, and compliance with system safety and security requirements.

Following validation, Humanoid plans to offer its products under robot-as-a-service (RaaS) models for the gamma phase, with both RaaS and CapEx options available for later stages. Schaeffler CEO Klaus Rosenfeld stated the company wants to play a key role in the growing humanoid robotics market by leveraging its manufacturing excellence and industrialization expertise.

Read more here about it!

Humanoid HMND-01 robot working at Schaeffler facility

X Square Robot secures $140M in funding! 🔥

X Square Robot has closed its Series A++ funding round, raising RMB 1 billion (about $140 million) to advance its foundation models for the physical world. The Shenzhen-based company, founded in 2023, is building general-purpose, end-to-end embodied AI that combines vision-language-action models with world models to help robots handle real-world tasks in unstructured environments.

The company’s WALL-A system uses world models to predict actions and causal inference to understand feedback, boosting zero-shot generalization for mobile manipulation. Large-scale, real-robot reinforcement learning lets the foundation model learn through physical interaction, allowing robots to refine their skills autonomously. X Square also released WALL-OSS in September 2025, an open-source version designed to democratize embodied intelligence and speed up community innovation.

X Square recently put its tech to the test with autonomous food delivery using its Quanta X1 wheeled bimanual robot. Powered by WALL-A, the robot successfully navigated indoor and outdoor tasks, handling strong winds, deformed packaging, and visual occlusions. Like a human, it used causal inference to “fill in the blanks” when objects were partially hidden and self-corrected without human help when facing operational issues.

The company claims to be the first in China to scale up real-world data resources, developing advanced data-capture tools including teleoperation, exoskeletons, and the Universal Manipulation Interface.

Read more here about it!

Quanta X1 wheeled bimanual robot from X Square Robot

Robotics maps #1 (Munich 🥨)

I decided to start creating these maps that show the most dense robotics areas in the world.

Starting with Bavarian capital, expect more maps like this in coming weeks! :)

Munich has become one of Europe’s top robotics hubs, thanks to strong engineering talent, industry connections, and deep tech funding. Technical University of Munich trains top robotics and AI engineers, while companies like BMW and Siemens provide partnerships and early customers. Student groups like RoboTUM show the city’s growing interest in robotics.

Munich’s robotics scene covers everything from factory automation to warehouse logistics and outdoor cleaning. Key players include robominds (robots that learn from human demos), Franka Robotics (force-sensitive robotic arms, acquired by Agile Robots for ~€33 million), Agile Robots (raised over $270-380 million for intelligent automation), and RobCo (modular automation with €39 million Series B). Magazino, now owned by Jungheinrich, builds AI-driven warehouse robots, while Filics raised €13.5 million for its autonomous pallet movers.

Startups are tackling specific problems: Angsa Robotics makes autonomous outdoor cleaning robots that detect and pick up trash, sewts automates textile handling with €7 million Series A, and Circus Group builds robotic systems for automated food production. Olive Robotics develops AI-enabled sensor hardware, while Intrinsic makes industrial robots easier to program and use.

Major players like Universal Robots and Exotec also have Munich offices. With access to venture capital, accelerators, and government funding for deep tech, Munich’s robotics ecosystem keeps growing strong.

Mobileye acquires startup Mentee Robotics for $900M! 💸

Mobileye, the Intel subsidiary known for its automotive computer vision chips and advanced driver assistance systems, has announced a $900 million acquisition of Mentee Robotics, marking what co-founder and president Amnon Shashua calls “Mobileye 3.0.” The deal includes approximately $612 million in cash and up to 26.2 million shares of common stock, and is expected to close in the first quarter of 2026.

Mentee Robotics, co-founded by Shashua in 2022, will continue as an independent unit within Mobileye. Mobileye aims to apply its vehicle navigation technology to humanoid robots, leveraging its $24.5 billion automotive revenue pipeline over the next eight years to fund this costly enterprise.

Shashua recused himself from the Mobileye board’s consideration and approval of the deal due to his role as chairman, co-founder, and significant shareholder of Mentee. The transaction, approved by both the Mobileye board and Intel, is expected to modestly increase Mobileye’s operating expenses in 2026 by a low-single-digit percentage.

The acquisition comes as humanoid robotics sees a surge in interest and investment across the industry. Mobileye announced the deal during CES in Las Vegas, just one day after revealing that a “top 10 automaker” will purchase 9 million of its EyeQ6H-based Surround ADAS systems, bringing total estimated future deliveries to over 19 million units.

Read more here about it!

NEURA Robotics unveils next-gen humanoids! 🦿

NEURA Robotics showcased during CES 2026, its next-generation humanoid 4NE1 designed in collaboration with Studio F.A. Porsche, alongside two brand-new robots: the 4NE1 Mini and NEURA Quadruped Robot.

All robots are fully integrated with the Neuraverse, NEURA’s operating system that connects robots to a shared platform where learning and experience can be pooled and reused across the entire network.

The 4NE1 Mini stands 52 inches tall and delivers cognitive intelligence in a compact, education-friendly form factor with natural language interaction and computer vision. The NEURA Quadruped Robot tackles challenging terrain with less than a meter in height, a 48.5 lb payload, and 360° environment vision for industrial and exploratory missions. These products position NEURA as the first Western supplier to bring such robots into series production, competing directly with existing Asian manufacturers.

CEO David Reger has hosted their presentations at NEURA’s booth alongside guests including Spencer Huang from NVIDIA and David Kehr from Schaeffler. All NEURA robots are powered by NVIDIA Isaac GR00T and simulated using NVIDIA Isaac Lab.

Here's more about this robot!

the new generation of 4NE-1 humanoid robots from NEURA Robotics

🤖 Atlas Goes Industrial
→ Boston Dynamics shows the new Atlas at CES 2026
→ Why this is less about athletic demos and more about factory readiness
→ What the partnership with Google DeepMind changes, and what it doesn’t

🏡 Home Robots, Take Two
→ SwitchBot introduces a general-purpose household robot
→ A more realistic take on home robotics: platforms + ecosystems, not single gadgets

🧠 NVIDIA’s Physical AI Stack
→ New Cosmos models and NVIDIA’s latest GR00T release
→ Why simulation, synthetic data, and reasoning models are becoming the real bottleneck

🎾 Tennis playing humanoid
→ A humanoid rallying tennis balls on stage
→ What dynamic vision and motion planning actually demonstrate — beyond the spectacle

🇮🇹 Italian humanoid partners up with AMD
→ Generative Bionics unveils its first humanoid concept
→ Why full-body tactile sensing is emerging as a serious industrial requirement

Boston Dynamics shows the new Atlas during CES! 🆕

Boston Dynamics has shown its new Atlas humanoid, which is now moving out of the lab and into real factory use. The robot appeared onstage at CES 2026 and is already being prepared for deployment at Hyundai factories, starting with a site in Savannah, Georgia.

This version of Atlas is built for industrial work. It has 56 joints, human-scale hands with touch sensing, and can lift up to 110 pounds. It’s designed to do repetitive tasks like handling parts and moving materials. Boston Dynamics says that while Atlas already moves well, turning it into a usable product means making sure it can work safely around people, not just perform athletic motions.

After showing the robot, Boston Dynamics announced a partnership with Google DeepMind. The goal is to use DeepMind’s AI models to improve how Atlas understands its surroundings and reacts to people. Instead of relying on fixed, pre-programmed behaviors, the companies want Atlas to learn from experience and adapt to new situations.

Atlas will be the first robot used for this joint work. Hyundai is supporting the effort with new training facilities and by feeding real factory data back into the system.

Read more about it here!

Another robot to take care of our households 🏡

At CES 2026, SwitchBot is presenting Smart Home 2.0, a home ecosystem built around AI-driven robotics rather than isolated smart devices.

The main reveal is onero H1, a general-purpose household robot designed to handle chores that are still difficult to automate. Instead of focusing on one task, onero H1 is built as a multi-task platform that can adapt to different home scenarios and work alongside SwitchBot’s existing devices.

The robot has 22 degrees of freedom and runs an on-device vision-language-action model. It combines vision, depth sensing, and tactile feedback to better understand objects and physical contact.

Alongside robotics, SwitchBot also introduced the Lock Vision Series, a smart deadbolt that uses 3D structured-light facial recognition. The system projects over 2,000 infrared points to build a 3D face map, enabling fast unlocking while resisting spoofing attempts. All biometric data is stored locally rather than in the cloud.

Read more here about it!

New Cosmos and GR00T models from NVIDIA 🧠

At CES 2026, NVIDIA unveiled several new AI models focused on physical AI — meaning AI that understands, plans, and acts in the real world rather than just text or static images.

These models are intended to help developers build and test robots more quickly, simulate realistic environments, and give machines stronger reasoning skills.

Cosmos Reason 2

• A new reasoning vision-language model (VLM) that helps robots and AI systems see, understand, and interact with the physical world.

• It goes beyond simple detection — it can use common sense and physics knowledge to plan and make multi-step decisions.

• It leads benchmarks for physical reasoning accuracy and supports long-context understanding (important for robots tracking moving objects over time).

Cosmos Transfer 2.5 & Cosmos Predict 2.5

• These models generate large-scale synthetic video and simulated environments that mirror real-world conditions.

• Synthetic data is used to train and test robot perception and control policies in simulation before deploying in the real world.

• Transfer 2.5 focuses on translating simulated scenes into realistic-looking versions, while Predict 2.5 focuses on predicting how environments evolve over time.

Isaac GR00T N1.6

• A vision-language-action (VLA) model designed specifically for humanoid and complex robots.

• It uses Cosmos Reason for the “brain” that ties perception and reasoning together, enabling full-body control — from walking to object interaction.

• Partner companies are already using GR00T to simulate, train, and validate robot behaviors before deployment.

Read more here about it!

Meme of the week 🤖

A humanoid robot playing tennis… here we go. 🎾

Vision means anticipating what comes next.

To start the new year, a human and a robot exchange tennis shots.

Walker S2 tracks the ball, plans its movement, and returns it with controlled timing and force. The rally shows perception, prediction, and whole-body balance working together in real time.

What a great dynamic vision, and motion planning in humanoid robots developed by UBTECH Robotics.

Generative Bionics presents its first humanoid 🇮🇹

Italian robotics company Generative Bionics has unveiled GENE.01, its first humanoid robot concept, at CES 2026. The reveal follows the company’s recent $81 million funding round and marks a shift from long-term research into early industrial demonstrations. Generative Bionics says its first commercial humanoid is planned for Q4 2026.

A central feature of GENE.01 is its full-body tactile sensing. Touch and force sensors are distributed across the robot’s surface and treated as a primary input, not an add-on. This allows the robot to detect contact, pressure, and small interactions in real time, which is critical for safe operation around people and objects in industrial settings.

The robot uses an onboard computing stack based on AMD CPUs, GPUs, and FPGAs, enabling fast local processing without relying on external systems. Generative Bionics describes its Physical AI approach around three pillars: using the body itself as part of computation, combining touch with vision and force sensing, and building on an open platform that can evolve over time.

GENE.01 is intended to be the first in a series of humanoids optimized for different industrial tasks, while sharing a common physical and behavioral design. Over 2026, the company plans to run demonstrations as it moves toward real-world deployment.

Here's more about this robot!

💰 Galbot’s $300M round
→ A bet on mobile manipulators over full humanoids
→ Why 24/7 warehouse deployments matter more than flashy hardware

🏭 New Atlas from BD to be presented during CES
→ Hyundai Motor Group shares a new AI robotics strategy ahead of CES 2026
→ The new Boston Dynamics Atlas debut at CES 2026

📉 iRobot’s bankruptcy
→ How price competition and tariffs exposed a fragile business model
→ What this says about the limits of consumer robotics

🍱 Chef Robotics’ latest system
→ What 80M+ meals reveal about reliability as a competitive moat
→ Why incremental engineering beats moonshot designs in factories

🧹 LG’s two-armed home robot
→ A serious hardware signal, but still short on proof
→ What to watch for at CES

Galbot raises $300M to expand robot deployments 💰

Galbot has raised over $300 million in its latest funding round, bringing total funding to $800 million and valuing the company at $3 billion. The Beijing-based startup says the capital will be used to scale deployments, continue technical development, and expand internationally.

Founded in 2023, Galbot focuses on mobile manipulators rather than full humanoids. Its approach combines large-scale synthetic data for pre-training with real-world data for post-training, allowing faster iteration of embodied AI models and smoother transfer to physical robots.

The company’s main product is the G1, a mobile manipulator with a wheeled base, torso, and two arms. It’s designed for tasks like inventory handling, replenishment, delivery, and packaging. Galbot says G1 robots have been running 24/7 in warehouse environments for over a year, and are also being tested in hospitals, where they assist with tasks in patient rooms, pharmacies, and navigation.

Galbot reports deployments and pilot projects with industrial partners including CATL, Bosch, Toyota, and Hyundai, and claims orders for thousands of units. Beyond enterprise use, the company has launched Galbot Store, an autonomous retail concept powered by G1 robots, now operating in more than 30 cities in China.

Read more about it here!

New Atlas from Boston Dynamics to be presented at CES 2026 🚨

Hyundai Motor Group says it will present its AI robotics strategy at CES 2026. The announcement will be made during Hyundai’s Media Day on January 5, 2026.

A key moment will be the first public stage debut of the new Atlas humanoid from Boston Dynamics.

Beyond Atlas, Hyundai plans to show how it is scaling robotics using its Software-Defined Factory (SDF) approach. The company says this lets it manage the full robotics lifecycle, from hardware and AI development to training, deployment, and operations, all within its broader group ecosystem.

From January 6–9, Hyundai will run a public exhibition at CES featuring live demonstrations of Atlas, Spot, and MobED, along with interactive scenarios showing how its robots could operate in factories, logistics facilities, and everyday environments.

Read more here about it!

iRobot files for bankruptcy amid competition and tariffs 📉

iRobot, the company behind the Roomba robot vacuum, has filed for bankruptcy protection in the U.S. after years of mounting pressure from cheaper competitors and rising manufacturing costs.

The company is using a pre-packaged Chapter 11 process, under which its main manufacturing partner, Picea Robotics, will take ownership. iRobot said the filing is not expected to disrupt its products, apps, or customer support.

iRobot has struggled to compete with lower-cost Chinese rivals, forcing it to cut prices while continuing to invest heavily in new technology. On top of that, U.S. tariffs hit hard: import duties on goods from Vietnam, where most Roombas for the U.S. market are made, added about $23 million in costs this year alone.

The company’s decline has been sharp. Valued at $3.6 billion in 2021 during the pandemic boom, iRobot is now worth roughly $140 million. Its shares dropped more than 13% following the bankruptcy news.

Founded in 1990 by MIT researchers, iRobot originally worked on defense and space projects before launching the Roomba in 2002. Despite its troubles, the company still holds about 42% of the U.S. robot vacuum market and 65% in Japan.

A planned $1.7 billion acquisition by Amazon collapsed last year after EU regulators raised competition concerns, removing what might have been iRobot’s clearest exit.

Read more here about it!

Meme of the week 🤖

Chef Robotics introduces its most capable meal-assembly robot 🍱

Chef Robotics has launched Chef+, a new version of its AI-powered meal assembly robot designed for large food manufacturing facilities. The company says the system is built on lessons from more than 80 million meals produced in real customer environments and focuses on reliability, efficiency, and food safety.

Chef+ is designed to solve common production bottlenecks. It doubles ingredient capacity, meaning fewer refills and less downtime, while keeping roughly the same footprint as a human worker.

Reliability was another focus. Chef+ routes wiring through sealed channels, uses industrial IP cameras that perform better in cold environments, and includes moisture management for pneumatic systems. These changes aim to increase uptime and reduce maintenance issues seen in high-volume facilities.

Food safety and usability were also improved. The open-frame design makes surfaces easier to clean, ingredient pans slide in quickly with a locking mechanism, and the touchscreen works reliably with gloves. Self-leveling feet and integrated handles make it easier to reposition on the floor.

On the performance side, Chef+ adds more compute power and a three-camera vision system to track trays and conveyor speed in real time, allowing precise placement even when line conditions change.

Chef Robotics also introduced a new “pat-down” gripper that automatically flattens meals before sealing, replacing a repetitive manual task. Using vibration and vision tracking, the system improves tray consistency, reduces spillage, and lowers food waste.

Chef+ is already running at customer sites and is now available across the U.S., Canada, and the U.K.

More about it here >>

LG teases a two-armed home robot 🧹

LG has previewed a new home robot called CLOiD, which it plans to show in full at CES next month. Unlike the wheeled home companion LG revealed last year, CLOiD has two articulated arms and five individually actuated fingers on each hand, suggesting a focus on physical tasks rather than simple monitoring or assistance.

LG says each arm has seven degrees of freedom, giving the robot a more human-like range of motion for handling objects. From the limited images released so far, CLOiD appears capable of grasping household items like towels, though LG hasn’t demonstrated complete tasks yet.

The robot’s “head” houses onboard compute along with a display, speaker, camera, and sensors for navigation and voice interaction. LG also says CLOiD uses its so-called Affectionate Intelligence, which the company frames as AI designed to better understand user intent and respond naturally. What that means in practice remains unclear.

Details on autonomy, payload, speed, battery life, or real-world reliability haven’t been shared. CES should clarify whether CLOiD is a near-term consumer robot or an early platform preview.

Here's more about this robot!

Humanoid robotics startup 1X Technologies has announced a strategic partnership with EQT, the global private equity firm and investor in 1X through EQT Ventures. The goal: deploy up to 10,000 of 1X’s NEO humanoid robots across EQT’s portfolio companies between 2026 and 2030.

While individual companies will decide if and how they adopt the robots, the partnership gives EQT’s portfolio early access to 1X’s production capacity and integration support. The focus is on real-world environments where robots work alongside people, including logistics, warehousing, manufacturing, facility operations, and healthcare.

1X initially positioned NEO as a household helper, capable of handling everyday chores like tidying and laundry. The company says consumer shipments will begin in 2026, but this new partnership marks a clear shift toward commercial deployment. NEO is designed to operate safely around humans, a requirement for both homes and workplaces.

The move puts 1X in direct competition with companies like Agility Robotics, Figure, and Boston Dynamics, all racing to place humanoids into paying customer sites. Agility’s Digit and Figure’s humanoids have already begun early commercial deployments, setting the bar for real-world performance.

EQT says the partnership is about addressing labor shortages and safety challenges, not replacing workers. With €267 billion in assets under management and hundreds of portfolio companies worldwide, EQT offers a large testbed for scaling humanoid robots beyond pilots and into everyday operations.

If successful, this collaboration could become one of the largest real-world rollouts of humanoid robots to date.

Read more about it here!

Generative Bionics raises $81M to build humanoids! 🇮🇹

Italian humanoid startup Generative Bionics has raised €70 million ($81.2M) to move its robots from research labs into real-world deployment. The funding will support product development, physical AI training, and construction of the company’s first production facility. Its first industrial contracts are expected to be announced in early 2026, with a full humanoid reveal planned for CES in Las Vegas.

Generative Bionics spun out of the Italian Institute of Technology (IIT) in 2024, building on more than 20 years of humanoid research. IIT has granted the company exclusive licenses to key technologies developed across projects like iCub, ergoCub (industrial assistance), and iRonCub (a flying humanoid).

The company’s humanoids are based on three core pillars:

• Dense tactile and force sensing for safe physical interaction

• A Physical AI architecture that allows robots to learn directly from real environments

• Advanced AI methods designed to handle extreme conditions, including heat and outdoor settings

The funding round was led by CDP Venture Capital’s AI fund, with participation from AMD Ventures, Eni Next, Duferco, RoboIT, and Tether.

With deep academic roots and fresh capital, Generative Bionics is positioning itself as one of Europe’s strongest bets in human-centric humanoid robotics.

Read more here about it!

Meme of the week 🤖

Luxonis launches a standalone edge AI vision platform 👁️

Luxonis has introduced OAK 4, a new generation of its spatial AI cameras that brings 52 TOPS of on-device AI compute, roughly 40× more than the previous generation. Unlike many vision systems, OAK 4 runs fully standalone, with no external computer or cloud inference required.

Each OAK 4 device is built around Qualcomm’s QCS8550 processor, combining GPU, ISP, and AI acceleration while running Yocto Linux. This allows multiple computer-vision models, including stereo depth, object detection, and tracking — to run at high frame rates directly on the camera, reducing latency, bandwidth usage, and server costs.

The hardware integrates a rich sensor stack depending on the model: stereo depth cameras, a 48 MP RGB sensor, IR projection and illumination, a 9-axis IMU, microphone, and wide field-of-view optics. Luxonis also added dynamic stereo calibration to maintain depth accuracy as lighting and temperature change. All components are housed in an IP67-rated industrial enclosure designed to handle shock, vibration, and harsh environments.

OAK 4 targets robotics, automation, logistics, smart cities, and ADAS applications. Pricing starts at $749, with devices available now and early-access models planned for early 2026.

More about it here >>

A robot that builds furniture from voice commands! 🪑

Researchers at MIT have developed a “speech-to-reality” system that lets a robot build physical objects simply by listening to a spoken request. Tell the robot, “I want a simple stool,” and it can design and assemble one in under five minutes.

The setup uses a robotic arm mounted on a table and a chain of AI models. First, speech recognition and a large language model interpret the request. Then 3D generative AI creates a digital design, which is broken down into modular building blocks. Planning software figures out a feasible assembly order, and the robot executes it in the real world.

So far, the team has used the system to build stools, chairs, shelves, small tables, and even decorative objects like a dog statue. Unlike 3D printing — which can take hours or days, this approach assembles objects quickly from reusable components.

The project connects research from natural language processing, generative design, and robotic assembly, and grew out of work at MIT’s Center for Bits and Atoms. The goal is to make design and fabrication accessible to people without CAD or robotics experience.

Next, the researchers plan to improve the strength of the structures and explore scaling the approach using multiple robots.

Here's more about this robot!

Humanoid, a new robotics company from the UK, has introduced the HMND 01 Alpha, a bipedal robot that reached stable walking only two days after final assembly. The team built the entire prototype in just five months, showing how fast humanoid development is accelerating.

The speed came from a simulation-heavy workflow. Using NVIDIA Isaac Sim and Isaac Lab, the robot was trained on 52.5 million seconds of virtual locomotion — work that would normally take well over a year. After that, Alpha needed only 3.2 million seconds of on-robot training before taking its first real steps.

Alpha is 179 cm tall with 29 joints, can handle 350 N pushes, and carries up to 15 kg with both arms. It supports either five-finger hands or simple grippers. For perception, it uses six RGB cameras, depth sensors, and a microphone array. An NVIDIA Jetson Orin AGX and Intel i9 power onboard compute, and a swappable battery gives roughly three hours of runtime.

The robot can walk, turn, run, squat, hop, and recover from force in any direction. It can also work with other robots, manipulate objects, and communicate using screens, audio, and LEDs.

Humanoid is targeting industrial work first, manufacturing and logistics, with an eye toward home assistance in the future. Its modular design means hands, arms, or external coverings can be upgraded without redesigning the whole machine.

Alpha builds on Humanoid’s earlier wheeled platform and is positioned as a practical, adaptable robot built for real work, not just showcase demos.

Read more about it here!

Nebius Physical AI Awards 🤖📍

The Nebius Robotics & Physical AI Summit and Awards takes place on December 9 in Mountain View, bringing together many of the operators, researchers, and investors pushing robotics from demos to real deployment.

Speakers include:

• Jonathan Hurst (Agility Robotics)

• Robert Nishihara (Anyscale)

• Adrian Macneil (Foxglove)

• Charles Wong (Bifrost AI)

• Kevin Black (Physical Intelligence)

• Amit Goel (NVIDIA)

Investors from Shanda Grab Ventures, Radical Ventures, N47, Khosla Ventures, BMW Ventures, and Accel will also be in the room.

Topics span autonomy architectures beyond “LLMs + actuators,” the rise of simulation and synthetic data, and where leading VCs are placing bets across the Physical AI stack.

For robotics builders and investors, it’s one of the most relevant gatherings of the year. Attendance is invite-only — apply and mention “Lukas” on your registration.

Sign up here!

SoftBank and Yaskawa team up to develop “physical AI” 🧠

SoftBank and Yaskawa Electric have announced a new partnership to build “physical AI” systems aimed at office and service environments, spaces that traditionally challenge robots because of clutter, people, and constantly changing tasks.

SoftBank defines physical AI as robots that can interpret sensor and camera data, make decisions with AI models, and then execute flexible physical actions. Instead of single-purpose machines, the goal is to create robots that can switch between many tasks and handle environments where unpredictability is the norm.

Japan’s shrinking workforce is one driver. Offices, hospitals, schools, and retail spaces all need automation, but they require robots that can adapt quickly rather than follow rigid scripts. That’s where the two companies say they can combine strengths.

Yaskawa is contributing a mobile manipulator based on its MOTOMAN NEXT arm, known for safety and precise motion control. SoftBank is providing AI-RAN, its communications system that blends AI with radio access networks, plus edge computing and building management integration to give robots real-time environmental awareness.

In their first joint use case, the companies built a virtual office environment to test a robot that tracks facility data, manages supplies, and coordinates tasks. SoftBank’s vision-language model (VLM) converts sensor data into high-level instructions, while Yaskawa’s vision-language-action (VLA) model turns those instructions into specific robot motions.

The companies say this approach could enable “multi-skilled” robots, a single machine capable of handling many roles inside a building. They plan to expand the collaboration to broader office and social robotics applications.

Read more here about it!

Meme of the week 🤖

Chinese humanoids raises $140M Series A! 💰

ROBOTERA, a fast-growing Chinese robotics company, has raised nearly RMB 1 billion (~$140M) in Series A+ funding and signed a new strategic partnership with the United Nations Industrial Development Organization (UNIDO).

Founded in 2023, ROBOTERA is building general-purpose robots that learn and adapt through real-world interaction. Its product lines include bipedal humanoids, wheeled service robots, and dexterous robotic hands. The company says its systems are already deployed at nine of the world’s top 10 tech companies.

ROBOTERA’s humanoid platform has shown strong performance in varied conditions, including autonomous walking in snow. It also set jumping records at the 2025 World Humanoid Robot Games with a 1.47 m long jump and 95.64 cm high jump. Its XHAND 1 dexterous hand, designed for reinforcement learning research, can manipulate over 100 tools and is widely used in labs worldwide.

The agreement signals growing global interest in embodied intelligence and positions ROBOTERA as a key player in next-generation robotics.

More about it here >>

Zipline with $150M to scale medical drone delivery! 🚁

The U.S. Department of State is partnering with Zipline, committing up to $150 million to help the drone delivery company expand its medical logistics network across Africa. The funding uses a pay-for-performance model, the U.S. releases money only when African governments sign expansion contracts and commit to long-term usage.

Zipline said the partnership could triple its coverage, growing from 5,000 to 15,000 hospitals and health facilities, and give up to 130 million people near-instant access to blood, vaccines, and essential medications. African countries will pay up to $400 million in utilization fees as part of the expansion.

Zipline’s impact is already significant: in regions where deliveries once took 13 days, drones now complete them in under 30 minutes. Facilities gain reliable access to stocked “remote pharmacies,” helping eliminate stockouts and reduce waste from spoiled supplies.

Each new Zipline hub is staffed entirely by local employees, creating skilled jobs and strengthening health infrastructure. Nigeria’s health minister highlighted improvements in treatment rates and healthcare access across the country’s three existing Zipline-served states.

Since launching in 2016, Zipline has completed 1.8 million autonomous deliveries across four continents, all without safety incidents. With this new U.S. backing, its logistics network is set to become one of the largest and most impactful autonomous delivery systems in the world.

Here's more about this robot!

Agile Robots has introduced Agile ONE, its first humanoid robot designed specifically for industrial environments. The Munich-based company, known for its dexterous robotic hands and force-sensitive arms, is positioning the system as a practical tool for tasks like material handling, machine tending, pick-and-place, and fine manipulation.

Hands include fingertip and force-torque sensors in every joint. This gives the robot enough precision for delicate tasks while still handling heavier operations, a combination many current humanoids struggle to balance.

The robot runs on AgileCore, the company’s software platform that ties together strategy, motion planning, and low-level control. Its AI has been trained on one of Europe’s largest datasets of real industrial workflows, along with simulated and human-collected data, helping the robot better generalize to real factory conditions.

Agile ONE is also designed with human-robot interaction in mind. It uses proximity sensors, expressive “eyes,” and a chest display to communicate status, and its bright, approachable design is meant to fit safely alongside workers and other robots.

Full production begins in early 2026 at a new facility in Bavaria, with all hardware built in-house. Agile Robots has grown quickly since its founding in 2018, doubling revenue each year and reaching €200 million.

Read more about it here!

SOSV Robotics Matchup (cool robotics event)

The SOSV Robotics Matchup is a weeklong, fully virtual event that connects robotics founders with investors actively looking to fund robotics. It focuses solely on curated 20-minute 1:1 meetings, making it a streamlined alternative to costly in-person events.

Who It’s For

• Founders building robotics hardware or software with real commercial traction.

• Investors seeking qualified robotics startups and clear visibility into who’s active in the sector today.

How It Works

Participants specify their investment or startup criteria—stage, geography, verticals, robotics category, and check size. Both sides review profiles before accepting meetings, ensuring only high-intent conversations.

SOSV also provides a sortable list of all investors and startups, giving founders a rare, up-to-date view of active robotics funds.

In short, the Robotics Matchup is a focused, effective, and accessible way for founders and investors to meet and close deals.

Sign up here!

Jeff Bezos steps back with Project Prometheus 🤖

Jeff Bezos is reportedly returning to hands-on leadership for the first time since stepping down as Amazon’s CEO in 2021. According to The New York Times, the Amazon founder is backing a new AI venture called Project Prometheus, which has already raised a staggering $6.2 billion and positioned Bezos as co-CEO.

Bezos will share the role with Vik Bajaj, a veteran of Google’s life sciences division and co-founder of Verily and Foresite Labs. Bajaj recently left Foresite to help launch Prometheus, which aims to build AI systems for engineering and manufacturing — spanning aerospace, automotive, computers, and other sectors of the so-called “physical economy.”

While details remain largely under wraps, the startup’s direction reportedly resembles the work of Periodic Labs: simulating the physical world to accelerate scientific and industrial problem-solving. Prometheus has already assembled a team of nearly 100 researchers, including talent from Meta, OpenAI, and Google DeepMind, signaling ambitions well beyond a typical stealth startup. With billions in funding and heavyweight leadership, Project Prometheus is shaping up to be one of the most closely watched AI launches of the year.

Read more here about it!

Meme of the week 🤖

Flexion Robotics raises $50M to build autonomy stack 🤖

Flexion Robotics has closed a $50 million Series A to accelerate development of its reinforcement-learning and sim-to-real platform designed to power humanoid robots across multiple morphologies and tasks. The Zurich-based startup, founded in 2024, believes generative AI and LLMs can finally break robotics free from brittle, task-specific pipelines built on scripted behaviors.

Flexion’s autonomy stack spans three layers. At the top sits a command layer driven by language models that handle common-sense reasoning, break down natural-language requests into subtasks, and ground them in the robot’s environment. Below that, a motion layer uses a vision-language-action (VLA) model trained largely on synthetic data, then fine-tuned on real-world edge cases. At the bottom, a control layer provides low-latency, transformer-based whole-body control with a modular skill library that can be recombined into new behaviors on the fly.

The company’s latest system, Flexion Reflect v0, pushes toward generalizable autonomy. LLM and VLM agents plan and schedule tasks, while a general motion generator produces short-horizon, collision-aware trajectories for everything from grasping to whole-body navigation. A reinforcement-learning tracker then executes commands robustly across varying terrains and control spaces.

More about it here >>

RAI Institute teaches robots to play baseball ⚾️

Researchers at the RAI Institute in Cambridge, Massachusetts, have released a new demo showing two robots that can throw, catch, and even hit baseballs with impressive accuracy. Instead of following preprogrammed motions, the robots adjust in real time to unpredictable ball trajectories.

Founded by former Boston Dynamics engineer Marc Raibert, the RAI Institute focuses on robots that can move and think quickly. Each robot uses flexible joints and adaptive control systems that absorb impacts rather than stiffening or breaking on contact, allowing smooth throws and safe catches.

High-speed sensors track the ball the instant it leaves a hand, while prediction models estimate where it will land. Motors then position the glove or bat with precise timing. The robots can throw at 70 mph, catch pitches up to 41 mph, and hit balls thrown at 30 mph from about 7 meters away.

Their hardware is purpose-built for speed: lightweight carbon-fiber arms, soft joints, fast electric actuators, and low-mounted batteries for stability during swings. A fusion of camera data and motion models drives the robot’s split-second decisions.

While this looks like a fun sports demo, it showcases something more important, robots handling fast, unpredictable objects safely and accurately. Skills like these are directly relevant to warehouse automation, disaster response, sports training, prosthetics, and even orbital robotics.

Today it’s baseball. Tomorrow it’s any task where reflexes, timing, and real-time decision-making matter.

Here's more about this robot!