Capturing this week’s zeitgeist

AI coding tools let industrial engineers build custom software and connect messy plant data in hours rather than months. Instead of waiting on IT, teams can quickly spin up tailored dashboards to monitor live machinery or draft code for tricky shop-floor workflows that usually take forever to configure.

Replacing core systems of record like your ERP or MES is another story /X/. These platforms hold decades of fragile business rules, custom logic, and historical data. When AI-generated code makes even a tiny mistake, it can freeze a production line or mess up account balances across the entire company. For now, AI is fantastic for building quick tools around your existing systems but probably keep them away from pulling the plug on your core foundation unless you’ve mastered them like a process technology.

Subscribe now

Assembly Line

This week’s industrial breakthroughs and frontier technologies of the built world.

We Made a Knob for SmarterEveryDay to Print Millions

Slant 3D counters SmarterEveryDay’s claim /YouTube/ by arguing that while individual 3D printers take longer per part than injection molding, mass-production “print farms” achieve equal or higher overall throughput through massive parallelization. By converting the high capital cost of a single molding machine into hundreds of lower-cost printers running concurrently, manufacturers eliminate expensive tooling, avoid single-point failure risks, and gain the flexibility to iterate part designs instantly via software.

Loading Hondas in the Heartland: How Honda’s Indiana Auto Plant Ships its vehicles

Automotive finished-goods logistics relies on highly synchronized multi-modal transport, where factory workers precision-park new vehicles, shuttle crews, and use specialized connecting ramps inside enclosed rail auto-racks to rapidly roll vehicles across entire multi-car trains. While relocating assembly, such as Honda keeping the Civic in Greensburg, Indiana, rather than Mexico/Reuters/, allows automakers to bypass heavy import tariffs, it concentrates immense pressure on domestic rail infrastructure and local carrier yards to manage high outbound volume efficiently.

• Shipping Ford Trucks Across the Country: A Look at How Autoracks are Loaded /YouTube/

Several major automakers implement autonomous end-of-line logistics, utilizing two distinct architectural approaches:

On-Vehicle Autonomous Driving (Tesla BYD, NIO, XPeng)

Tesla: Employs its onboard Full Self-Driving (FSD) software to have newly built vehicles drive themselves straight off the assembly line at the Fremont Factory /X/ and navigate over a mile to their assigned staging lanes and loading docks without human drivers.

Infrastructure-Guided Automation (BMW, Mercedes-Benz, General Motors)

BMW deploys an “Automated Driving In-Plant” (AFW) system across plants like Dingolfing and Leipzig /BMW/. Instead of relying on the car’s built-in sensors, the factory uses a stationary off-vehicle LiDAR infrastructure and cloud movement planners (developed with Seoul Robotics and Embotech) to remotely pilot 5 Series, 7 Series, and MINI Countryman models over 1 km through test tracks directly into logistics yards.

Mercedes-Benz & Audi use automated valet parking (AVP) and infrastructure-driven plant routing systems in plants like Factory 56 (Sindelfingen) /Audi/. External cameras and LiDAR networks track vehicle location and issue remote steering, braking, and throttle commands over local networks to move vehicles between production steps and staging bays.

BYD & Chinese EV makers leverage integrated smart-factory architecture to autonomously route EV units off assembly lines into yard staging. Utilizing factory-wide digital twin tracking alongside onboard camera/radar suites, vehicles self-navigate to charging pads, quality inspection lines, and shipping bays.

Inside a high-tech lab designing food for a hotter world

The video highlights high-throughput plant phenotyping using automated greenhouses equipped with 3D scanners, thermal sensors, and AI computer vision to continuously analyze crop performance under simulated climate stress. By mapping non-destructive traits like photosynthetic efficiency, root growth, and water retention to their specific plant genomes, researchers can dramatically accelerate traditional cross-breeding without resorting to genetic modification. The overall trajectory points toward creating climate-resilient, disease-resistant crops capable of maintaining food yields in a warmer, more volatile global environment.

Have you tried the Exponential Industry Arcade yet? Safety comes first in Hazard Sweep…

New Product Introduction

Highlighting new and innovative facilities, processes, products, and services

Terraform Industries Hydrogen Electrolyzer Breakthrough

Last week the c team at Terraform Industries pulled off yet another first, the sustained production of >99.9% pure H2 from our vertically integrated, California-manufactured electrolyzer stack while it was coupled directly to a solar array at our Muroc desert test site.

Most commercial electrolyzers need carefully conditioned power from expensive battery-meditated backup systems. Ours runs directly off the sun. Clouds pass, the day turns to night, and we maintain purity from a stack whose Bill of Materials cost is well below $100/kW.

Terraform’s single-minded focus on capex reduction has allowed us to convert sunlight to hydrogen in an unprecedented demo at a cost below $2/kg. If that wasn’t enough, we have a crystal clear plan of steady execution to push that cost below $1/kg in the coming years.

Read more from Founder Casey Handmer on X

Oklo’s Groves Reactor Achieves First Criticality in Under a Year

Oklo Inc., an advanced nuclear technology company, announced that its Groves Isotope Test Reactor reached first criticality after achieving a controlled, self-sustaining nuclear chain reaction at low power. The milestone comes less than a year after groundbreaking and follows U.S. Department of Energy (DOE) authorization through the DOE Reactor Pilot Program (RPP).

The DOE’s RPP created a pathway that allowed engineering, construction, commissioning, and operational preparation to advance alongside DOE’s safety review and authorization. Groves demonstrates that the domestic nuclear industry can once again move from design through construction, authorization, and startup on timelines that are measured in months rather than years when developers, suppliers, and regulators work together on an integrated deployment approach.

Read more at Oklo

Business Transactions

This week’s top funding events, acquisitions, and partnerships across industrial value chains.

Hadrian Raises $1.37B Series D to Build Highly Automated Factories to Accelerate America’s Industrial Renewal

Hadrian, the advanced manufacturing company building highly automated factories for America, announced $1.37 billion in new equity financing to accelerate the buildout of the domestic manufacturing capacity needed to deliver America’s most critical defense, aerospace, and industrial systems.

The financing will support new factories, expanded R&D, and additional production capabilities as Hadrian builds the industrial infrastructure required to deliver complete mission-critical systems at speed and scale. At a time when demand for trusted U.S. manufacturing capacity continues to outpace supply, Hadrian is expanding the AI-powered production network needed to strengthen deterrence, accelerate delivery, and restore America’s ability to build.

Read more at PR Newswire

Valor Atomics Announces a $1B Series B Led By Sequoia

Valar Atomics is excited to announce the closing of a $1 billion Series B financing led by Sequoia Capital. As part of this new partnership, Valar welcomes Sequoia partner Shaun Maguire to Valar’s board of directors. Sequoia was joined in this round by Apandion, Atreides Management, Conviction, Dream Ventures, HOF Capital, Point72, Riot Ventures, Snowpoint Ventures, Valor Equity Partners, and other both new and longtime investors. Finally, in addition to the $1B of equity, Valar is excited to announce the closing of a $200 million credit facility led by Erebor Bank, as administrative agent, and J.P. Morgan, alongside Crescent Cove and Hercules Capital.

The preliminary stage of Valar’s mission has been a success. Ward 250’s operability has been proven. More importantly, the speed and efficiency of a vertically integrated, hardware focused approach stands for all to see. So, what does the new Series B financing mean for Valar? It allows us to make the next mission critical evolution: Valar will transition from demonstrating the operability of an integrated reactor system to producing fleets of them en masse. This has been Valar’s mission since its inception. A project can build one reactor, but a fleet requires manufacturing. This series B financing is the fuel for this mission. It will allow us to fully execute a vertically integrated, hardware first approach; one that encompasses everything from reactor deployment and long-term operations to fuel production. Valar will not rely on outside suppliers for the fuel needed to deploy and operate. We will construct it ourselves in labs placed alongside the very reactors that same fuel will run. These reactors will bring unprecedented energy abundance that matches the scale of their production.

Read more at Valar Atomics

• Base Power Announces $1B Series D and Launches Base Core, First-of-its-Kind Home Battery Built in the United States /Business Wire/. Go inside the factory with Sourcery /YouTube/.

• Ore Energy Raises $43 Million /Newswire/ to scale its iron-air battery technology. Ore’s batteries, designed to store renewable electricity for up to 100 hours, can solve one of the biggest barriers to the energy transition: affordable, long-duration energy storage.

Mariana Minerals Raises $310 Million Series B Led by Khosla Ventures to Accelerate Critical Minerals Production

Mariana Minerals, a software-first, vertically integrated critical minerals company, today announced it has raised $310 million in Series B financing. The round was led by Khosla Ventures, with continued support from Andreessen Horowitz (a16z) and Breakthrough Energy Ventures, and new participation from Greenoaks, Halo Fund, Pax Ventures, StepStone Group, BHP Ventures, Washington Harbour Partners, Greycroft, General Innovation Capital Partners, Mitsubishi Corporation, In-Q-Tel (IQT), and Earthshot Ventures, with a number of additional strategic capital partners joining the round. This round brings the company’s total parent and project capital raised to date to approximately $400 million.

Mariana engineers, builds, and operates mines and refineries using its proprietary MarianaOS platform, which integrates capital project execution (CapitalProjectOS), refinery operations (PlantOS), and mine management (MineOS) into a single AI and machine learning stack. MarianaOS compresses traditional 5-to-10 year project execution timelines by roughly half and reduces commissioning timelines well below industry benchmarks.

Mariana’s Series B financing unlocks continued buildout of the company’s active project portfolio, including its flagship sites Copper One in Southeastern Utah and Lithium One in East Texas. Copper One, acquired by Mariana in late 2025, restarted mining operations under autonomous orchestration within 4 months and is ramping toward a production target of 50,000 metric tons of refined copper per year. Lithium One, which broke ground in Q4 2025, is the world’s first GWh-scale facility for producing lithium from oil & gas produced water, with commercial production targeted for the first half of 2027.

Read more at PR Newswire

• LoopX Innovation Inc., a Physical AI company building machine intelligence for uncontrolled mining environments, announced it has closed a $4.05M Seed financing led by BDC Seed Venture Fund and Orion Industrial Ventures /Business Wire/

Shelfmark Raises $3.5 Million to Automate Manufacturing’s Most Underserved Production Lines and Make them Smarter Every Run

Shelfmark, a Physical AI company providing production intelligence for continuous-flow manufacturing, announced a $3.5 million seed round to bring greater autonomy to one of manufacturing’s most overlooked environments: the fast-moving lines that produce materials on reels, rolls and continuous webs. Led by Armory Square Ventures, with participation from Grand Ventures, Hyde Park Angels, Argon Ventures and Cultivation Capital, the round brings Shelfmark’s total capital raised to approximately $5 million.

At the heart of Shelfmark’s Physical AI engine is a perception system combining in-line industrial cameras, spatial sensing and proprietary deep-learning vision models that inspect and interpret high-speed continuous materials in real time under dynamic plant conditions. Shelfmark reports 99.5% defect-detection accuracy in customer deployments that have halved inspection labor costs and generated returns of up to 7x compared with manual inspection.

Read more at GlobeNewswire

• NavVis Raises USD 85M Series D Round to Supply the Data Foundation for Physical AI /Business Wire/. NavVis combines survey-grade reality capture 10x faster than traditional methods with NavVis IVION, a collaborative enterprise cloud platform that turns spatial data into shared, trusted, always-current spatial twins.

The Kimi K3 Frontier

First on July 16th Moonshot AI released Kimi K3, introducing a novel architecture and training recipe that delivers roughly 2.5 times the intelligence per unit of compute compared to its predecessor, Kimi K2 /Kimi/. The release triggered market volatility reminiscent of DeepSeek’s R1 launch in January 2025, which erased nearly $590 billion from Nvidia’s market cap in a single day.

Beyond raw model performance, Moonshot demonstrated Kimi K3 autonomously generating a functional 4 mm² semiconductor design in just 48 hours using the open-source Nangate 45nm cell library /Si2/. Every semiconductor design step from RTL synthesis down through placement, routing, and simulation was executed entirely within open-source flows, marking a significant milestone for AI agents tackling complex, long-horizon engineering tasks. Moonshot similarly showcased automated workflows across GPU compiler development, kernel optimization, digital asset creation, and scientific research.

Wall Street reacted swiftly as shares of EDA giants Cadence and Synopsys each fell approximately 9% in the following trading session as investors digested the implications for proprietary chip-design software /Yahoo! Finance/.

All content published on this newsletter is based on public information and independent research. Opinions are authors own and have been sanitized through AI engines which can make mistakes. This newsletter is not financial advice, and readers should always do their own research before investing in any security.

Situational Awareness

Concurrently, Leopold Aschenbrenner’s Situational Awareness LP (SALP) fund entered July with heavy, highly leveraged exposure to the AI infrastructure complex. As the market spooked following the Kimi K3 announcement, SALP’s portfolio value plummeted 67%, triggering severe margin calls amid thinning liquidity across its core positions as first reported July 30th /FT/.

To satisfy those demands, the fund liquidated the bulk of its public equity book to Citadel and unwound its remaining leverage /WSJ/. While year-to-date performance remains positive at roughly +80%, the drawdown brought the fund dangerously close to permanent capital loss, an episode Aschenbrenner candidly characterized as “unacceptable” in a recent letter to partners /BI/. The forced transfer of these concentrated positions onto Citadel’s balance sheet effectively absorbed a major source of sell pressure, coinciding with a partial rebound across the impacted equities.

Aschenbrenner assumed that because a technology is transformative, the capital that builds it will earn its cost.
There is no relationship between those two things. In fact, it’s more likely not to be true.

/Porter Stansberry on X/

The July sequence illustrates how rapidly a perceived shift in model efficiency can reprice both enabling hardware and the leveraged funds positioned around it. But while financial markets reacted aggressively to Kimi K3’s compute efficiency, has any constraint really changed?

Let’s dive in starting from the top of stack...

Read more

Capturing this week’s zeitgeist

The relocation of the 1937 Westinghouse Atom Smasher from Forest Hills to Cranberry Township ends a decade of decay for the world’s first industrial Van de Graaff particle accelerator. Early researchers at this site studied fundamental particle behaviors, laying technical groundwork for practical nuclear power, medical X-rays, and radar systems. Saving these physical remnants keeps historical engineering achievements visible to young inventors who are building new hardware. Seeing physical machinery from past technological leaps helps current developers solve the todays complex manufacturing challenges.

Try GeoSourced inside the Exponential Industry Arcade to trace historical discoveries, supply chains, and industrial culture. Can you pin global trade flows and locate significant technological heritage sites worldwide? 👇

Launch the Exponential Industry Arcade

Assembly Line

This week’s industrial breakthroughs and frontier technologies of the built world.

The $1/Hour Robot Is Coming: Industry Leaders from ANYbotics, NEO, Boston Dynamics, & Agility Robotics Explain What’s Next

Almost $1 Billion Later, the US Still Can’t Make a Medical Glove

Establishing domestic nitrile glove production requires rebuilding a dead chemical process from the ground up. Workers must synthesize nitrile butadiene rubber by heating and cooling butadiene and acrylonitrile inside massive specialized reactors to strip out impurities. America abandoned this complex polymer refining thirty years ago. Restarting the industry means constructing multi-story chemical plants with five-hundred-thousand-dollar pressure vessels before a single glove rolls off an assembly line. Because domestic plants lack these facilities, local factories buy foreign raw latex from China, South Korea, and Italy to stay open.

Read more at Bloomberg

Flashback to July 2023, “Why America’s Largest Tool Company Couldn’t Make a Wrench in America“ /WSJ/

Applied Intuition: A Billion Intelligent Machines

The simplest way to understand Applied Intuition is that it builds the brains for machines, and the tools other companies use to build those brains. If a manufacturer wants its tractor, truck, or mining vehicle to drive itself, it can buy the intelligence from Applied Intuition or use its platform to develop its own.

The discussion centers on why the most important companies of the next 25 years will all be physical AI companies, Dana, their new agentic platform for developing and deploying these systems, and how the company raised a billion dollars without spending any of it.

We led Standard Bots’ $200 million Series C because we believe it’s the company built to reinvent the robot arm for physical AI and to bring the industry back to America.

In this public thesis we will dive into:

• The birth of the robot

• A typical robot arm’s hardware and software architecture

• The technical limitations that hold back today’s arms

• How Standard Bots is reinventing the arm from first principles

Read more from RoboStrategy on X

New Product Introduction

Highlighting new and innovative facilities, processes, products, and services

Agentic Engineering Harness: Arche Smith

Design for Manufacturability is what really matters in the end, so Arche integrated hundreds of real catalog STEP parts directly into its Smith’s model reasoning. Design with real parts!

Read more from Arche on X

Applied Electrodynamics’ WaveSight Brings Radio Frequencies into Human Perception

What would the world look like if we could see beyond visible light, across the entire electromagnetic spectrum?

Read more from Founder & CEO Bilgehan Avser on X

Business Transactions

This week’s top funding events, acquisitions, and partnerships across industrial value chains.

🇺🇸☢️ Commonwealth Fusion Systems Raises Another $1 Billion

This capital raise is the single largest funding round among fusion energy companies worldwide since CFS announced its $1.8 billion Series B round in 2021. With this capital, and the $863 million the company raised last year, CFS has now raised a total of $4 billion. This $4 billion represents about 30 percent of the total capital raised by the fusion industry to-date, reinforcing CFS’ position as the world’s largest and leading fusion company.

CFS will use the funds raised to further accelerate its progress to commercialization. In parallel to completing the assembly of its SPARC fusion demonstration machine, CFS continues to move forward with development of the world’s first grid-scale fusion power plant, called ARC, at the company’s Fall Line Fusion Power Station in Chesterfield County, Virginia.

Read more at PR Newswire

• 🇺🇸🔋 Antora Closes $550 Million in Series C Funding to Meet Surging Energy Demand and Reindustrialize America /Business Wire/. The announcement comes on the heels of the company’s deployment of one of the largest battery storage projects in the world, a 5 gigawatt-hour system in South Dakota that advanced from initial construction to delivering energy in under 12 months. Antora is now scaling that proven approach, with a growing pipeline of signed agreements with hyperscalers and industrial leaders.

• 🇺🇸☢️ Antares Raises $470M Series C to Deploy Nuclear Microreactors for Critical Missions /Business Wire/ comes weeks after Antares took its Mark-0 reactor critical at Idaho National Laboratory. It was the first privately developed non-light-water reactor to achieve criticality in the United States in more than four decades.

• 🇬🇧☢️ Nuclear Turbines raises £15M for compact reactors it says could undercut fossil fuels /TNW/. The technical idea is a substitution. Nuclear Turbines replaces the steam turbine systems at the heart of a conventional plant with high-temperature turbine technology borrowed from jet engines and gas-fired power stations.

🇺🇸🛰️ K2 Space Raises $500M Series D at $6.8B Valuation to Scale Large, High-Power Satellites

K2 Space, the leading manufacturer of big, high-power satellites, announced a $500 million Series D funding round at a $6.8 billion valuation. Kleiner Perkins and ICONIQ led the round, with participation from CapitalG, Lightspeed, Altimeter, Spark Capital, Sands Capital, ARK Invest, T. Rowe Price Associates, and other existing investors.

“In addition to expanding our factory footprint and growing our accomplished workforce, we’ll also be scaling up the satellite itself,” said Neel Kunjur, K2 Space Co-Founder and CTO. “We’ll be introducing our 100kW Giga-class satellite in the second half of 2028, which will be the foundational platform for deploying massive amounts of compute on orbit.” In addition to building and deploying the Giga satellite, K2 will focus on delivering large constellation orders on schedule for its customers’ ambitious programs over the next 24 to 48 months. K2 Space’s next mission, Trinity, will launch multiple satellites in Q2 2027.

Read more at PR Newswire

🇺🇸💿📐🧠 ChipAgents Expands Series A Funding to $134 Million as Demand Grows for Agentic AI in Semiconductor Design

ChipAgents, the category leader in agentic AI platforms for semiconductor design, announced an additional $60 million in Series A2 financing, bringing the expanded round to $134 million just six months after its initial Series A close. The financing reflects continued investor confidence in the company’s commercial momentum, with new investor B Capital joining Bessemer Venture Partners, Micron, MediaTek, Ericsson, ScOp, and other existing investors.

The funding will support ChipAgents’ next phase of expansion as the company scales customer deployments, grows its engineering and go-to-market teams, and accelerates development of its AI-native semiconductor design platform. Since its founding in 2024, the company has raised more than $134 million.

Read more at Business Wire

• 💿📐🧠🤝 Synopsys Advances Agentic AI Chip Design with AMD and Microsoft /PR Newswire/. This milestone builds on the existing collaboration and previously announced agentic specification to RTL workflow, marking the first EDA applications available for evaluation on Microsoft Discovery. Customers can request evaluation access through Synopsys. AMD is actively evaluating the application of autonomous workflows to accelerate the development of their next-generation products.

• 💿📐🧠 Kimi K3: Open Frontier Intelligence /arXiv and HuggingFace/. The RTL code is available on GitHub.

🇺🇸📐🧠 Engineering AI startup, P-1 AI, Announces Its Series A Financing Led by NEA, Adding Jeff Immelt to the Company’s Board

P-1 AI, Inc., the company building Archie, an AI mechanical and electrical engineer for industrial teams, announced the initial closing of its Series A financing round led by New Enterprise Associates (NEA). The Series A, on the heels of the company’s $23 million seed round led by Radical Ventures in 2025, reflects strong product execution and commercial traction. “Interest in Archie from major industrial firms—employers of most of the world’s engineers—has been extraordinary and emphasizes the need to scale our product and deployment teams,” said Paul Eremenko, the company’s co-founder and CEO. “This capital will enable us to grow the team and 10X the compute for training custom models to meet customer needs.”

P-1 AI built Archie to maximize anthropomorphism so that it fits seamlessly into existing engineering teams and workflows. Archie today is at the level of a junior mechanical and electrical engineer, with quantitative intuition over the product design space and the ability to use complex engineering tools—the same tools his human teammates use. “Paul’s approach is distinctive: bringing advanced AI capability into the workflow of engineering and industrial teams in an engineer-native form factor. That matters,” said Immelt. “By fitting naturally into how engineers collaborate, P-1 AI can shorten cycle times, improve competitiveness, and deliver measurable results in mission-critical environments.”

Read more at GlobeNewswire

• 🇬🇧🤖🧠 Kinematic Trees raises £585,000 to scale robotics intelligence using principles of brain evolution /Kinematic Tree/. Funds will be used to build Kinematic Trees’ commercial, robotics and software engineering teams across London, Nottinghamshire and Sheffield in the UK, in line with product development, international deployment and major advancements in the robotics industry.

See the unseen connections driving industrial tech. Paid subscribers get Overcapacity exclusive industrial ecosystem ontologies that map how people, capital, and technology connect. Turn raw news into actionable strategic context.

Capturing this week’s zeitgeist

You can literally retrace all the forks in the tech tree and just look for “Here’s what we knew at the time”, and then you can insert what YOU know NOW, in 2026, and see if the other forks that weren’t pursued would bear fruit today.

You have to remember we arrived at many of those things by making decisions with incomplete information.

/Object Zero on X/

That’s the core thesis of Exponential Industry. By systematically retracing historical tech-tree forks and pairing them with modern materials, advanced manufacturing, and AI that augments the tacit knowledge of the original researchers and their networks, we can bypass the incremental dead-ends of today’s dominant technologies. History is littered with good ideas abandoned under incomplete information; the work of Exponential Industry is to surface the entrepreneurs and capital that can turn those abandoned industrial technology detours into high-probability exponential breakthroughs.

Subscribe now

Assembly Line

This week’s industrial breakthroughs and frontier technologies of the built world.

You can now 3D-print the core of an MRI scanner.

A machine that hospitals pay $1.1 million to $3.4 million for has been broken open. The OSI² ONE and its educational siblings deliver real images of heads and limbs for a fraction of the cost, using a permanent-magnet Halbach array, 3D-printed structures, and fully open designs.

This is not a toy or a simulation. Working systems already produce in-vivo images in Leiden, Utrecht, Berlin, and Uganda.

The magnet alone has 396 carefully oriented neodymium cubes in a cylindrical Halbach array with costs totaling about $1,370. A complete scanner lands between $28,500 and $68,000 depending on the console and coils you choose.

Read more at Brian Roemmele on X

Why NVIDIA & BlackRock Are Betting $2.3B on This Startup

A visit to PsiQuantum’s quantum computer factory in Milpitas, California - the last production chip fab to close its doors in Silicon Valley, now ground zero for the world’s largest quantum computer.

• BBC Tech Now’s Adrienne Murray has rare access to Microsoft’s Quantum Lab in Denmark to find out how the chips are being developed /YouTube/

Performant robotic manipulation with real-world reinforcement learning

✍️ Authors: Kun Lei, Huanyu Li, Dongjie Yu, et. al

Real-world robotic manipulation in homes and factories demands reliability, efficiency, and robustness that approach or surpass skilled human operators. We present a real-world reinforcement learning (RL) framework, RL-100, for achieving complete task success under a predefined evaluation protocol built on diffusion visuomotor policies. RL-100 unifies imitation and RL under a single clipped proximal policy optimization surrogate objective applied in the denoising process, yielding conservative, stable improvements across offline and online stages. To meet deployment latency, a lightweight consistency distillation compresses multistep diffusion into a one-step controller for high-frequency control. The framework is task, embodiment, and representation agnostic and supports both single-action and action-chunking control. We evaluated RL-100 on eight diverse real-robot tasks, from pushing and bowling to pouring, cloth folding, unscrewing, multistage juicing, and long-horizon box folding. Under our predefined protocol, RL-100 achieved 100% success in the evaluated trials (1000 of 1000 episodes), including up to 250 of 250 consecutive trials on one task. It matched or surpassed expert teleoperators in time to completion. Without retraining, a single policy attained ∼90% zero-shot success under environmental and dynamics shifts, adapted in a few-shot regime to substantial task variations (86.7%), and remained robust to human perturbations (about 96%). Our juicing robot served customers continuously for about 7 hours without failure when deployed zero-shot in a shopping mall. These results suggest a potential path to deployable robot learning by starting from human priors, aligning training objectives with human-grounded metrics, and reliably extending performance beyond human demonstrations.

Read more at Science Robotics

The Hidden American Manufacturing Advantage

WattFactory

The Variety Trap

The West’s manufacturing strategy is changing…

Read more

20 days ago · 6 likes · 1 comment · Glen Turley

New Product Introduction

Highlighting new and innovative facilities, processes, products, and services

CuspAI Launches ‘AI Materials Foundry’ a Global Network to Accelerate Breakthrough Discoveries

Software-led materials discovery requires four things: high quality training data at scale, to make AI predictions reliable; compute powerful enough to screen at molecular resolution across billions of candidates; synthesis infrastructure to move from digital design to physical reality; and domain expertise to interpret what the machine finds and know what to do with it.

CuspAI’s AI Materials Foundry assembles all four, for the first time, providing the production infrastructure for industrial materials discovery. Rather than relying on the physical constraints of a single, isolated laboratory, CuspAI’s ecosystem approach creates a compounding intelligence loop: breakthroughs achieved within the network have potential to accelerate discovery timelines across the entire global value chain.

Read more at Business Wire

• By uniting advanced semiconductor packaging materials with high-efficiency data center thermal management, Solstice Advanced Materials and Element Solutions seek to create a specialized end-to-end platform capturing high-margin demand across every critical hardware layer of the expanding AI supply chain /Solstice/.

RTX’s Pratt & Whitney Valox™ 1500 Engine Completes Key Design Milestone

Pratt & Whitney, an RTX business, has successfully completed a key design milestone for its Pratt & Whitney Valox™ 1500 engine, a new powerplant intended for semi-autonomous collaborative platforms. A comprehensive digital assessment confirms the design is on track to deliver an optimal mix of performance, affordability, and speed to market.

Leveraging proactive investments in digital engineering, Pratt & Whitney has significantly de-risked the development timeline and is now focused on maturing the design for follow-on ground testing. In late 2025, Pratt & Whitney received a contract award valued at more than $10 million from the U.S. Air Force to advance the engine design.

Read more at RTX

• Boeing Sees ‘Volatile’ Supply Base as it Boosts Jet Output /Bloomberg/

Business Transactions

This week’s top funding events, acquisitions, and partnerships across industrial value chains.

• TruArc Partners Closes on $1.2 Billion for TruArc Partners Fund V /Business Wire/ focused on middle-market investments in specialty manufacturing and business services

ATOMS is announcing a $1.7B equity investment with lead investor a16z

With the ATOMS launch a few months ago, there was a huge amount of interest after many years of ninja level stealth. My vision letter outlined Physical AI in industry, but it cruised at a fairly high altitude. I figured this moment would be a good time to connect the dots between our bold vision for the future with the actual business we are building /Travis Kalanick on X/, and bring to life what A16Z and other great investors see in it.

The impetus for this transformation is Industrial AI - Systems of software, sensors, robotics and AI used to automate the operations of entire industrial sectors. Industrial AI is getting ready to power the next Industrial Revolution. Tesla calls it Amazing Abundance. We call it the Age of Atoms.

Read more at Atoms

• SoftBank Weighs Deal for Robotics Startup Gravis /Bloomberg/. A deal could ultimately value Zurich-based Gravis at more than $500 million. Founded in 2022 as a spinout from ETH Zurich university, Gravis makes hardware and software that helps earthmoving equipment operate autonomously.

Sila Secures $300 Million in Private Funding to Ramp Gigascale Anode Manufacturing and Strengthen America’s Technology Sovereignty

Sila, an advanced battery technology company, today announced it has secured $300 million in a private equity round led by Atreides Management, and Sutter Hill Ventures. The round is joined by 8VC, Bessemer Venture Partners, Matrix Partners, funds and accounts advised by T. Rowe Price Associates, Inc, and other existing and new investors. The raise will accelerate U.S. production of Sila’s silicon-carbon (Si/C) anode technology, Titan Silicon® and support a planned Phase 2 expansion at Sila’s plant in Moses Lake, Washington.

Sila’s team pioneered the modern silicon-carbon (Si/C) anode, a breakthrough technology that delivers 20-40% higher energy density than traditional graphite, enabling batteries that are smaller, lighter, and capable of ultra-fast charging. By pairing American intellectual property with gigascale manufacturing in the United States, Sila is building the blueprint for supply chain security of critical technology.

Read more at Business Wire

• Sofab Inks Secures $6 Million Seed Round to Break the Reliability Barrier for Perovskite Solar /PR Newswire/

Arrakis Comes out of Stealth With $38M to Help Industrial Enterprises Compete in the AI Era

Arrakis, the AI deployment company helping industrial companies install AI agents for mission-critical operations, announced a $30 million Series A led by Blossom Capital, with participation from existing investor Accel, which also led the company’s $7.5 million seed round.

Arrakis combines elite forward-deployed AI engineers with a model-agnostic enterprise AI platform and applied AI research tuned to industrial operations, letting companies leverage their data to train, deploy, and scale custom AI agents in weeks, not quarters. The result is an AI transformation executed at warp speed, using the right model for the right task, and giving companies full control over performance, resilience and cost. Arrakis operates to the strictest data and security standards: customer data and IP remain theirs, and the company never uses them for third-party model training. Rather than replacing existing technology stacks, Arrakis deploys AI agents into existing systems, improving operational throughput without multi-year migrations. The company ties a significant proportion of its fees directly to customer outcomes, aligning its commercial model with the value delivered.

Read more at Business Wire

1872 Launches Autonomous Steel Fabrication Factory Model in Cincinnati and Closes $15 million Seed Funding Round

1872, a startup focused on AI-native manufacturing, announced its opening, introducing what will be an automated steel fabrication factory model in Cincinnati’s Camp Washington neighborhood. The company, founded by three former SpaceX engineers, is propelled by a $15 million seed funding round, one of the largest in Ohio history, made through private funds advised by The O.H.I.O. Fund, a private fund investment adviser seeking to generate strong returns for investors while creating economic growth in Ohio.

1872 introduces a unique AI-based factory automation solution to an industry that has struggled to innovate and has been unable to keep up with rising demand for steel products in the modular construction and infrastructure markets. The company’s proprietary system integrates AI software that understands the physical world with advanced automation from companies such as Ohio-based Path Robotics. This technology automates heavy fabrication tasks that have traditionally relied on manual labor, now in critically short supply. Early automations are in place at 1872, with the company expecting to achieve its full autonomous vision in 2027.

Read more at Business Wire

Looking for additional content? Paid subscribers get more stories delivered each week as part of Overcapacity!

Click to access ⤵️

Access Bonus Media

This weeks digest:

Read more

Capturing this week’s zeitgeist

As America continues to celebrate its 250th anniversary, Union Pacific’s Big Boy No. 4014, the world’s largest operating steam locomotive, continues its historic coast-to-coast tour, thundering across the nation as a living tribute to the engineering ingenuity that helped build and unite the country /NPR/.

Remaining public displays are scheduled /UP/ in St. Louis on July 19, Kansas City on July 21-22, and Hays, and Kansas on July 25. Come see the restored 133-foot, 1.2-million-pound marvel captivates crowds with its raw power, haunting whistle, and the sheer spectacle of steam-era might, drawing families and rail enthusiasts alike to witness a once-in-a-lifetime connection to America’s industrial heritage.

Not in America? Exponential Industry is developing a game to experience industrial heritage across the world. Stay tuned!

Heard on the floor

“It’s an AI problem. It’s not a data problem, and it’s not an energy problem,” he said. “If you’re a tier-one AI researcher, where are you going to work? I don’t think Shell’s on that list.”

Applied Computing Chief Executive Callum Adamson /TNW/

Yes, there will be winners in the robotic supply chain, but the big outsized winners will be those that build and sell the robots themselves: Tesla, Figure, 1X, Apptronik, etc.

And the biggest will have in-house data collection and in-house robotic AI models as well as significant vertical integration in motors, actuators and other key hardware.

/Phil Trubey on X/

Assembly Line

This week’s industrial breakthroughs and frontier technologies of the built world.

How China came to dominate the global EV factory boom

• A New Semi Every 8 Minutes: Inside Tesla’s Truck Factory /Core Memory/

• Decommissioning the original Model S & X assembly line in just 46 days /Tesla/

• The landscape of automotive companies and their robotics bets /Vedant Nair/

Inside the Self-Driving Manufacturing Lab

Georgia Tech’s Advanced Manufacturing Pilot Facility is using AI, robotics and autonomous experimentation to accelerate manufacturing development. Dr. Tom Kurfess explains how the self-driving lab is helping manufacturers test and de-risk new technologies before they reach production.

The Lab Mistake That Might Revolutionize Computing

✍️ Authors: Mario Lanza, Sebastian Pazos

The brain is roughly one million times as energy efficient at many of the comparable tasks we set for AI. To try to approach these efficiencies, a radically different way of computing called neuromorphic engineering is seeking to build electronic components and circuits that act more like the brain’s neurons and the synapses that connect them.

Working in our laboratory in 2024, one of my students was measuring a memory circuit that consisted of one transistor and one memristor—a type of nonvolatile memory device first fabricated in 2008. The student’s memristor circuit was built from two-dimensional material atop a silicon microchip containing MOSFETs. The MOSFETs were created in a commercial foundry using fabrication technology called the 180-nanometer node, which was cutting-edge in the year 2000.

One day the student forgot to connect the bulk terminal of the transistor. What he observed was a sudden increase in current with high nonlinearity that self-relaxed when the voltage was ramped down (a phenomenon called a hysteresis loop). This was a very promising neuronlike behavior!

Read more at IEEE Spectrum

• Neuromorphic principles meet reinforcement learning for virtual robotic control /Neuromorphic Computing and Engineering/

RoboTTT: Context Scaling for Robot Policies

✍️ Authors: Yunfan Jiang, Yevgen Chebotar, Ruijie Zheng, et. al

Recent robot foundation models operate with single-step or short-history visuomotor context. We introduce Test-Time-Training Robot Policies (RoboTTT), a robot model and training recipe that scale visuomotor context to 8K timesteps, three orders of magnitude beyond state-of-the-art policies, without growing inference latency. At this context length, we unlock new robot capabilities: one-shot in-context imitation from human video demonstrations, on-the-fly policy improvement, robustness to perturbations, and stronger performance on multi-stage, long-horizon tasks.

Read more at arXiv and NVIDIA Research

• WAM-TTT: Steering World-Action Models by Watching Human Play at Test Time /arXiv/

• Claude plays robotics /Anthropic/

Chipstrat

The Next Trillion-Dollar Chip Company

Two AI chip startups had huge financial outcomes in the past year. Groq was acquired for $20B, and Cerebras IPO’d to a $40B+ market cap. Huge, but not trillion-dollar huge. And in this environment, an inference-first system that runs frontier models and scales to gigawatts of compute has a real shot at a trillion-dollar market cap if it can scale, grow, and IPO…

Read more

a month ago · 27 likes · Austin Lyons

New Product Introduction

Highlighting new and innovative facilities, processes, products, and services

Kimi K3: Open Frontier Intelligence

As an early proof of concept, Kimi K3 designed a chip to serve a nano model built on its own architecture. In a single 48-hour autonomous run, K3 built, optimized, and verified the chip using open-source EDA tools on the Nangate 45nm library. Within 4 mm², the chip closes timing at 100 MHz and sustains over 8,700 tokens/s decode throughput in simulation, packing 1.46M standard cells, 0.277 MB of SRAM, and an INT4 MAC array with fused dequantization. A chip built by a model, for a model, reflects K3’s long-horizon agentic capabilities.

Read more at Kimi

Archer Announces Zee, AI Foundation Model Purpose-Built for Aviation, a Key Pillar of Its Physical AI Strategy

Archer created Zee, what it believes to be the world’s leading aviation-specific foundational model that delivers a unified aviation intelligence platform built on ADS-B, ATC communication, maps and charts, aircraft state, terrain and weather data. Zee is trained on real-world operational data, aggregated through Archer’s proprietary data pipeline and a global network of over 6,000 ADS-B receivers.

On a typical day, more than 45,000 flights move through American airspace. That flight activity generates a constant flow of data in the form of radio calls, navigation inputs and aircraft-state that pilots and air traffic controllers piece together in real time. Zee is built to understand these disparate data sources as a single system. The goal of Archer’s approach with this unified aviation intelligence platform is to deliver a lower latency, better performing model capable of running on-device requiring no connectivity, critical for use in a wide range of aviation environments from air taxis and UAVs to commercial airliners and air traffic management.

Read more at Archer

Business Transactions

This week’s top funding events, acquisitions, and partnerships across industrial value chains.

Walden Robotics Launches with $300 Million to Put General-Purpose Robots to Work Today

Walden Robotics, a full-stack Physical AI company building and deploying general-purpose robots that continuously learn and improve while performing real work, launched out of stealth with $300 million in funding. The round, which values the company at $1.1 billion, is co-led by Toyota (Toyota Motor Corp, Toyota Invention Partners, and Toyota Ventures) and Deviation Capital, with participation from NVIDIA, Boeing, AE Ventures, Samsung Ventures, Prologis Ventures, CoreWeave Ventures, and financial partners Calibrate Ventures, Colle Capital, Shine Capital, NextView Ventures, Squarepoint Capital, One Madison Group, KAS Venture Partners, and Menlo Ventures, among others.

Walden’s product builds on a decade of foundational AI and robotics research the Walden team helped pioneer, including Diffusion Policy and Large Behavior Models (LBMs), the frontier model class that powers Walden’s robots and lets them quickly learn new tasks and continuously improve through real-world practice. Walden launched out of Toyota Research Institute in January 2026 and began working with customers across multiple industries immediately. Since February, Walden’s general-purpose robots have been doing useful work in production at a Toyota plant in North America, moving from first pilot to real work in under two months.

Read more at Business Wire

• LimX Dynamics Closes US$200 Million Pre-IPO Round /LimX/

• microagi raises $55 million seed round to power Europe’s robotic reindustrialisation /microagi/

TerraFirma Raises $115M to Accelerate Construction on Earth and Beyond

TerraFirma, a tech-enabled, vertically integrated construction company focused on critical infrastructure, announced it has raised approximately $115 million, including a $100 million Series A led by Kleiner Perkins, with participation from Bain Capital Ventures, Glade Brook Capital Partners, BANNER VC, Saga Ventures, Trust Ventures, Definition, PEAK6, Magnetar Capital, and Ravelin Capital. Angel investors include founders, executives, and engineers from companies including SpaceX, Anduril, Base Power, Shinkei, and Hadrian.

TerraFirma is building the robotic infrastructure stack for modern construction. The company’s full-stack platform combines AI-enabled pre-construction software, a remote command-and-control center, and retrofitted semi-autonomous heavy machinery, including excavators, dozers, loaders, rollers, skid steers, and more, converted into robots that no longer require an operator in the cab. Instead, skilled operators orchestrate entire fleets from screens, using the same intuition and judgment they’ve built over years, now extended across multiple machines at once. The result is a system that can make each operator up to 300% more effective, accelerating project execution, reducing costs, and creating jobs that are safer and higher-paying than traditional equipment operation.

Read more at Business Wire

• Monumental raised $32 million to put more robots to work building the world /Monumental/. They run a fleet of more than 100 robots laying real brick on job sites across Europe today.

• Hyperion Robotics raises €6.4 million to bring physical AI to European infrastructure /EU Startups/ using large-scale additive manufacturing and localized robotic microfactories to replace traditional, labour-intensive concrete processes with a cleaner, faster alternative. It claims to use up to 75% less material and to reduce program times by up to 3x.

Senra Systems Announces $65 Million Series B, Plans for Third Manufacturing Facility

Senra Systems, a software-driven manufacturing company modernizing wire harness production, announced it has raised $65 million in a Series B funding round and plans to accelerate its expansion with a third factory location. The round brings the company’s total funding raised to over $112 million and was co-led by Lowercarbon Capital and Interlagos, with participation from General Catalyst, Sequoia Capital, Andreessen Horowitz, Founders Fund, Dylan Field, CIV, 8VC, The Friedkin Group, Jaws Estates Capital, Sozo Ventures and Alumni Ventures.

Senra manufactures complex wire harness systems used in aircraft, spacecraft, launch vehicles, satellites, defense systems and other advanced platforms. The company’s manufacturing model is powered by Amp, its proprietary software platform that integrates quoting, engineering, manufacturing, supply chain management and production workflows into a unified system. Unlike traditional wire harness manufacturing environments that rely heavily on fragmented manual processes, Senra’s integrated manufacturing approach enables faster onboarding, reduced production variability and accelerated delivery timelines.

Read more at PR Newswire

• Bezos’ Blue Origin Said to Seek Funds at $130 Billion Valuation /Bloomberg/

• Ravee Optics Secures $6 Million Seed Round as the Need to Move More Data in Space Explodes /Business Wire/

Lucid Weighs Going Private or Chapter 11 as Adviser Reports to Board

Restructuring adviser AlixPartners has been asked to deliver its findings to Lucid‘s board before its next meeting, two people familiar with the matter told EV. The strategic options under review include taking the Saudi-backed EV maker private or filing for Chapter 11 bankruptcy protection, one of the sources said.

Read more at EV

• AE Industrial Establishes Specialty Materials Platform with Investment in Powder Alloy Corporation /AE Industrial Partners/

Looking for additional content? Paid subscribers get more stories delivered each week as part of Overcapacity!

Click to access human curated reads ⤵️

Access Bonus Media

This weeks digest:

Read more

Capturing this week’s zeitgeist

Exponential Industry’s mission is to be the premier source and mixed-media aggregator for the business culture of high-tech manufacturing, energy, and supply chains. We are building a thriving community of founders, engineers, operators, and capitalists who are accelerating innovation in the physical world. Our tenets:

• Direct from the source - We select from primary sources across millions of company blogs, press releases, podcasts, deep-dive articles, social media posts, and videos directly from where they are originally published.
  (Otherwise mishaps like the WIRED exclusive on the 1X hands launch happen)

• Credit and attribution - We prominently display the authors and publication to give credit where credit is due.

• Straight to the point - Our excerpts seek to answer what and why. We highlight the technology, value proposition, and results so each reader is briefed on a wide range of cross-discipline topics and is encouraged to dive deeper if the content is enticing to explore further.

• Private and opt-in by default - Substack is used to deliver this publication directly to your inbox.

Please spread the word of Exponential Industry by sharing the publication:

Share Exponential Industry

Heard on the Floor

“This country is either going to start making things again or it’s going to circle the drain.”

Mike Rowe on the Shawn Ryan Show /YouTube/

Assembly Line

This week’s industrial breakthroughs and frontier technologies of the built world.

🗜️ How to Build a Full Body Ultrasound

This project is actually crazy, combining 40 ultrasound probes into one machine, an underwater elevator to slide you through them, and the computational expertise to build an image of your insides just based on the echoes of what bounced around you.

• Announcing Midjourney Medical /Midjourney/

• How The World’s Best Cricket Balls Are Made in Australia /YouTube/

🛰️ Space Can Have Big Impact on the Economy

Money is flooding into the space economy, from SpaceX’s record IPO to Blue Origin’s first outside fundraising effort. But while capital is abundant, rocket launches are not. Founders Fund Partner and Varda Space Chairman Delian Asparouhov explains why access to space could become the industry’s biggest bottleneck.

🔻 The Debt-Fueled Collapse of China’s Top Machine Tool Maker

Shenyang Machine Tool (SYMT), once the world’s top machine tool maker, collapsed in 2019 due to unsustainable debt-fueled expansion and reliance on unmastered foreign technology. Following a bankruptcy reorganization, the state-owned conglomerate General Technology Group (GENERTEC) took control of the Shenyang Machine Tool Group to salvage and restructure the business.

☀️🚰 New 3D Photothermal Structure Enables Efficient Solar Desalination and Crop Irrigation

✍️ Editor: Zhang Nannan

Researchers from the Institute of Process Engineering, Chinese Academy of Sciences, and Shenzhen University have developed a new three-dimensional (3D) photothermal structure that greatly improves solar evaporation efficiency.

The researchers used Hansen solubility parameter theory to tightly combine polyethylene terephthalate (PET) chains with HoMS /Advanced Materials/. Accelerated seawater aging tests found no detectable particle detachment from the material after 30 days of continuous exposure. In addition, no active free radicals were detected when the material was exposed to light, suggesting potentially good durability and reliability during long-term use.

The new technology also shows promising economic potential. The researchers estimate that after two years of operation, the cost of water produced by this technology would be lower than that of commercial bottled water. If it proves capable of stable long-term performance, the new material may offer a practical solution for sustainable freshwater production in regions facing water shortages.

Read more at Chinese Academy of Sciences

Industrial Sector Insights

BMW's Spartanburg will be the proof if a US export base can carry electrification

BMW Group has closed out a $1.7bn investment in South Carolina, marking it with the "Home of X" event at Plant Spartanburg on 30 June and the premiere of the fifth-generation X5…

Read more

a month ago · 1 like · Dan Mottram

New Product Introduction

Highlighting new and innovative facilities, processes, products, and services

⚡🛢️ Making a new gas turbine, how big should it be?

“There’s a sweet spot in the 40-50MW range where the turbine has most of the efficiency of a gigantic unit yet is also much more manufacturable and deployable. This is the sweet spot is in total cost per unit energy, inclusive of capital expenditure.

This is why Superpower sits at the ever magical 42MW.”

Read more at Blake Scholl on X

• neuralCAD-Edit: An Expert Benchmark for Multimodal-Instructed 3D CAD Model Editing /GitHub/

🫲🦾🧠 NEO’s Hands: An API to the Physical World

Our NEO hand is read-write. Developed from the ground up by our world-class engineering team, it runs quasi-direct-drive tendons via the 1X Tendon Drive at low gear ratios of approximately 5:1 to 15:1. All 25 degrees of freedom (22 fully actuated DoF in the fingers and palm, plus 3 at the wrist) are natively force-controlled and fully backdrivable. Push on a finger and it yields — and reports exactly how hard you pushed. Force flows out and information flows back through the same physical path. We call this force transparency, and it’s what turns a push into a measurement.

These 25-DoF hands are IP68 waterproof and food-safe, allowing NEO to wash its own hands like a human. Far from fragile, they deliver powerful, useful strength: Peak torques reach 3.5 Nm at the thumb CMC and 2.6 Nm at the finger MCP joints, with distal flexion forces up to 45N. The wrist delivers 17.75 Nm of torque. This enables strong whole-hand grasps, tool use, lifting and carrying, opening doors, pushing loaded carts, and precision pinch under load, all while retaining full dexterity. With ±0.2 mm positioning accuracy, they work (and perceive) in the small-object regime where most human labor actually happens. Those aren’t demo modes. They’re the natural output of enough independent, force-controlled DoF at human scale.

Read more at 1X.tech

• AnyHand: A Large-Scale Synthetic Dataset for RGB(-D) Hand Pose Estimation /GitHub/

• NVlabs/ProtoMotions is a GPU-accelerated simulation and learning framework for training physically simulated digital humans and humanoid robots /GitHub/ and transfers directly to the Unitree G1 humanoid robot zero-shot.

• Robotics precision-transmission supplier Laifual Drive listed in Hong Kong /Hong Kong Exchange/

Business Transactions

This week’s top funding events, acquisitions, and partnerships across industrial value chains.

🇺🇸♨️ Quaise Energy Raises $134 Million in Initial Close of Series B to Build World’s First Superhot Geothermal Power Plant

Quaise Energy, a leading developer of utility-scale superhot geothermal energy, announced the first close of its Series B financing, raising $134 million with additional equity and debt capital expected to close imminently. The round was led by Prelude Ventures, with strategic investments from JERA and Idemitsu, two of Japan’s largest energy companies. Nearly all existing investors, including Safar Partners, participated in the round.

The proceeds from Series B will fund Project Obsidian, the world’s first commercial superhot geothermal power plant, and the continued development and commercialization of Quaise’s millimeter wave drilling system towards depths in excess of 5 km. Project-level equity and debt financing is concurrently being raised as the company advances toward first revenues secured by yet-undisclosed commercial off-take partners.

Read more at Business Wire

• 🇺🇸🔋 EnergyX Secures $225 Million Strategic Investment from Eni to Advance the Black Giant™ Lithium Project in Chile /PR Newswire/

• 🇺🇸♨️ Hephae Energy Technology Raises $17.8 Million for Commercial Deployment of Superhot Drilling Tech to Advance Geothermal /EINPresswire/

🇧🇪🥔 Polysense raises $10.7M to shape the food factory of the future

Polysense has built AI-native quality control and process optimization technology for food manufacturers. The $10.7M round was led by Felix Capital, with participation from Fortino Ventures, Syndicate One and angel investors including 100IN. This funding goes toward expanding to more production lines worldwide, growing our team, and making deployment faster.

• 🇺🇸📸 Oxipital AI Closes Series A Following Major Commercial Expansion and Rapid Adoption of its V-CORTX AI Vision Platform /PR Newswire/

🇬🇧🎛️🧠 Luffy AI raises £8.1M to put self-tuning AI inside the world’s electric motors

Luffy AI, an Abingdon startup building what it calls neuroplastic AI for the real-time control of physical machines, has raised £8.1m in a Series A round. The round was led by BGF, which bills itself as the most active equity investor in the UK and Ireland, and the money is earmarked for turning pilots into commercial deployments. It is a large ambition for a modest round, and one that rests on whether a self-tuning motor can prove itself on a real factory floor rather than in simulation.

Read more at TNW

🇬🇧📺🏭 Pixel-Flo Raises £5.25 Million Seed Round to Unlock MicroLED Displays for the Mass Market

Pixel-Flo, a University of Sheffield spin-out addressing a critical bottleneck in MicroLED display manufacturing with its proprietary Continuous-Flo™ Mass Transfer technology, has raised £5.25 million in Seed funding. The round was led by Northern Gritstone with additional investment from SCVC, the Parkwalk Northern Universities Venture Fund, and German investment firm, HTGF.

Pixel-Flo is developing the cost reduction breakthrough critically needed for MicroLED display manufacturing, through its highly scalable, cost-efficient approach. MicroLED is the next wave of display technology, offering 2–5x higher brightness and 2–4x greater efficiency than traditional displays. However, volume adoption of microLED is currently limited by the fundamental scalability challenges of mechanical mass transfer techniques. By comparison, Pixel-Flo’s proprietary fluidic self-assembly technology offers extremely high continuous throughput, delivering the high-performance of MicroLED at radically reduced processing and material cost. This will allow MicroLED enabled products from smartwatches to TVs to be made available at mass market price points.

Read more at Pixel-Flo

Looking for additional content? Paid subscribers get more stories delivered each week as part of Overcapacity!

Agility, the U.S. company behind the Digit humanoid with established paid deployments in logistics environments such as GXO warehouses /GXO/, is positioning itself to become the first pure-play humanoid robotics firm listed on a major North American exchange under the ticker AGLT. The transaction, which includes significant PIPE investment and is expected to close later in 2026, provides public-market access and capital to accelerate scaling of its Robot-as-a-Service model.

Meanwhile, Unitree is one of the world’s highest-volume humanoid shippers, with thousands of units delivered primarily in Asia, has filed its IPO application in March 2026 seeking roughly RMB4.202 billion (~$610 million). The move reflects strong revenue growth and positions the Chinese leader to fund expanded manufacturing and AI development on public markets.

Together, these developments signal growing investor confidence and infrastructure readiness for the humanoid market. They demonstrate that companies with tangible commercial traction and production scale can attract substantial public capital, shifting the narrative from early-stage R&D and pilots toward a phase of broader commercialization, supply-chain maturation, and measurable economic potential.

Constraint Capital #7 analyzes humanoid bills of materials (BOM) to identify supply chain constraints and value opportunities.

All content published on this newsletter is based on public information and independent research. Opinions are authors own and have been sanitized through AI engines which can make mistakes. This newsletter is not financial advice, and readers should always do their own research before investing in any security.

Lessons from Memory

Before we dive into humanoids, let’s look back at lessons from components of AI Data Centers. Exponential Industry was quick to identify high-bandwidth memory as the top constraint in the AI data center buildout since issue #2 in February and every month since (March #3, April #4, May #5) skipping June. For those willing to take the risk, this has been potentially lucrative.

But why…

Read more

Click to access ⤵️

Access Bonus Media

This weeks digest:

Read more

Capturing this week’s zeitgeist

In the story of America, the drive to build complex machinery with one’s own hands and to seek true independence from constraints, whether of empire, circumstance, or limited resources, has always formed the living core of the national mythos we call the American Dream. It began with the ragtag forces who, bankrupt and outmatched, refused to accept defeat and instead improvised the decisive victory at Yorktown through raw determination, borrowed ingenuity, and unbreakable will, proving that a people without pedigree or treasury could still forge their own future. That same spirit echoes today in every garage, workshop, and dorm room where individuals turn scrap, code, and sheer grit into advanced industrial systems from high-speed aerospace prototypes to computer systems, reminding us that the Dream was never about inherited advantage but about the radical belief that anyone, through relentless effort and inventive resolve, can create the machinery of their own freedom and prosperity.

Share

Assembly Line

This week’s industrial breakthroughs and frontier technologies of the built world.

🇬🇧🪙 How a New Steel Alloy Is Invented!

The Warwick Manufacturing Group (WMG) is a world-leading research and education group, creating knowledge to deliver impact - for industry, people, and the environment.

🇺🇸🚗🏭 Ford’s AI Hiccups Lead Carmaker to Rehire ‘Gray Beard’ Engineers

Ford initially assumed its AI systems could completely take over vehicle quality control and engineering tasks. However, because experienced engineers left the company before their institutional knowledge could be properly captured, the AI lacked the deep “judgment data” it needed to succeed. Instead of catching design flaws, the automated tools actually amplified weak inputs.

To fix the technical gap, Ford brought back the 350 engineers to:

• Rebuild the data pipelines feeding the AI to ensure it was learning from proper veteran engineering judgment.

• Refine the automated tools and mentor junior staff.

• Overhaul software QA by creating a dedicated 40-person software quality assurance team and introducing more than 100,000 AI-powered automated tests to catch edge cases late in development.

For the first time in 16 years, Ford secured the #1 spot among mainstream brands in J.D. Power’s Initial Quality Study (IQS), scoring 152 problems per 100 vehicles. Specific segments like the F-150, Mustang, and Super Duty won best-in-segment for the second consecutive year.

Read more at Bloomberg

• 🇨🇳🚙🏭 Inside NIO’s AI factory, where cars build themselves /The Korea Herald/. In this 100 percent automated process, where 941 robots orchestrate every movement, NIO EVs roll off the line at a rate of “one car every minute.” To boost production efficiency, NIO adopts the “Island production method,” replacing conveyor belts with a fleet of autonomous rovers that transport vehicles and components between workstations.

• 🇩🇪 SAP Gives AI Product Oversight to CEO and COO in Reshuffle /Bloomberg/

🦾 Automate Show 2026 Recap

The conversation traces the industry’s shift away from early-stage humanoid hype toward the practical, real-world deployment of Physical AI and edge computing. It then transitions into how software orchestration, digital twins, and advanced kinematics are solving labor shortages and preserving vital manufacturing knowledge.

• Vention collaborates with FANUC and Universal Robots on software-defined automation /Robot Report/

🇺🇸🖨️3️⃣ Advanced 3D printing creates origami-inspired structures

Researchers at the U.S. Department of Energy’s Manufacturing Demonstration Facility at Oak Ridge National Laboratory have introduced an innovation in additive manufacturing by integrating origami-inspired 3D printing techniques with hybrid composites — materials made from mixed reinforcing components. By eliminating the need for molds to fabricate a part, this process produces lightweight, cost-efficient structures with faster build times and greater adaptability.

The method enables the production of complex geometries that traditional mold-based methods cannot economically achieve. It also allows the fabrication of objects larger than the printing machine itself, reducing capital costs and boosting production efficiency. In a test print, researchers found that eliminating molds reduced fabrication time by 95 percent and costs by 90 percent for printing a unique design compared to conventional mold-based composite manufacturing. ORNL has filed a patent, and the team is in the process of preparing the innovation for future licensing.

Read more at ORNL

• 🇮🇹🇺🇸 Data-driven smart wire arc additive manufacturing: a qualification-oriented cyber-physical system framework /The International Journal of Advanced Manufacturing Technology/

• 🇨🇳 An executable machine-tool-level digital twin for closed-loop CNC milling /Robotics and Computer-Integrated Manufacturing/

• 🇮🇹 Event-Driven Reinforcement Learning Enables Long-Horizon Control in Semiconductor Fabrication /arXiv/

Citrini Research

Humanoid Robots

Yes, robots…

Read more

a year ago · 446 likes · 48 comments · Citrini

New Product Introduction

Highlighting new and innovative facilities, processes, products, and services

🇺🇸☢️ Nuclear Reactor Powers Nvidia AI Chip in US First

Valar Atomics, a California-based nuclear startup, generated power from an advanced reactor to run an Nvidia AI chip. While just a trickle of electricity was produced, it’s the first time a next-gen reactor has done so in the US. On the heels of a demonstration of Valar’s Ward 250 reactor connecting to the Nvidia Blackwell chip at the company’s site in Utah, Valar Atomics CEO Isaiah Taylor joins Ed Ludlow on “Bloomberg Tech.”

🇨🇳💿 Hangzhou Garen claims world’s first 6-/8-inch gallium oxide wafer production line

Chinese semiconductor firm Hangzhou Garen Semiconductor has established the world’s first mass-production line compatible with both 6-inch and 8-inch homoepitaxial gallium oxide wafers. Most commercially available gallium oxide wafers are limited to 2 to 4 inches due to severe manufacturing and quality-scaling challenges.

Garen overcame these hurdles using two core technical innovations:

• Proprietary Cast-Growth Single-Crystal Technology: Produces ultra-thick gallium oxide crystals while drastically cutting the consumption of iridium, a costly precious metal. Alongside an ultra-thin substrate process, this increases substrate output by three to four times and slashes substrate costs by more than 80% per wafer.

• Optimized MOCVD Epitaxy Process: Delivers epitaxial layers thicker than 10 micrometers with a thickness variation below 1%. This high level of uniformity significantly improves manufacturing yields.

Read more at Interesting Engineering

• Gallium: The Inelastic Byproduct Powering AI Semiconductors (GaN, GaAs & InP) /eⁱ/

Business Transactions

This week’s top funding events, acquisitions, and partnerships across industrial value chains.

🇫🇷🇺🇸🤝 Schneider Electric announces agreement to acquire Cognite, unlocking a new level of strategic intelligence for Industrial AI

Schneider Electric, a global leader in energy management and automation, announces that it has entered into a definitive agreement to acquire 100% of Cognite Holding B.V. (“Cognite”), a leading provider of industrial data and AI software, in an all cash transaction valued at $3.1 billion.

Cognite’s scalable, open architecture combined with the capabilities of AVEVA CONNECT means that analytics and industrial AI can ingest industrial data across the asset lifecycle and throughout customers’ existing data ecosystems and investments. The combination extends CONNECT with enterprise-wide data contextualization capabilities and leading-edge agentic AI.

Read more at Cognite

🇨🇦 Stathera Announces US$55 Million Series B to Scale Silicon Timing and Accelerate Next-Generation AI Data Center Solutions

Stathera Inc., a fabless semiconductor company pioneering silicon timing solutions, announced the close of an oversubscribed US$55 million Series B financing. As the market consolidates around a dominant incumbent, the round positions Stathera to become the leading independent alternative for AI data center and hyperscale customers. The round was led by new investor Maverick Silicon with continued backing from existing investors Celesta Capital, BDC Capital, MediaTek Innovation Fund, TXC Corporation, and Ultratech Capital Partners. The financing brings total funding to US$75 million.

Stathera will use the proceeds to fund mass production of its GEN2 silicon timing portfolio while accelerating development of its next-generation platform for AI data centers and growing the company’s engineering and commercial teams. With this expansion, Stathera will establish a Silicon Valley office to engage leading AI, data center, and hyperscale customers.

Read more at PR Newswire

• 🇺🇸 Omen AI Raises $31M Series A to Bring Continuous Fluid Intelligence to the Machines Powering the AI Economy /PR Newswire/

• 🇫🇷 VSORA strengthens its funding to launch the commercial roll-out of Jotunn8, its AI inference processor, and execute its ambitious technology roadmap /GlobeNewswire/.

🇰🇷 CarbonSix Secures $40M Series A to Deploy Physical AI Across Global Manufacturing

CarbonSix, a pioneer in Physical AI for the manufacturing sector, announced today that it has raised $40 million (approx. KRW 60 billion) in a Series A funding round from a syndicate of premier U.S. and South Korean venture capital firms. The round was co-led by DSC Investment and LB Investment. New investors joining the round include IMM Investment, Korea Development Bank (KDB), SV Investment, Cortentia (U.S.), and ASQ (A Squared, U.S.). Demonstrating strong continued confidence in CarbonSix’s technology and scalability, all existing seed-round investors—Foothill Ventures, Storm Ventures, Zeitgeist Capital, Xquared and CarbonBlack Fund—fully participated with follow-on investments.

At the core of CarbonSix’s competitive edge is its proprietary “data flywheel” business model. Instead of dumping massive, generalized datasets into abstract foundation models, CarbonSix provides practical automation tools that factories can use immediately. As clients operate these tools, high-quality, task-specific data is naturally captured and fed back into the system. This creates a powerful compounding loop: tool usage drives data accumulation, which refines the AI models, ultimately delivering even smarter automation tools back to the factory floor.

Read more at PR Newswire

• 🇺🇸 Robot hand company, Proception, settles Tesla trade secret suit and announces $11M raise /TechCrunch/

• 🇺🇸 Luxonis Raises $14 Million Series A to Power the Perception Layer Behind Physical AI /Business Wire/

🇺🇸 Arcturus Emerges from Stealth with $8M in Seed Funding to ‘Rewire the World’ with Metals Infused with Carbon Nanomaterials

Arcturus, an advanced materials company that infuses carbon nanomaterials into metals to unlock step-change improvements in electrical, thermal, and mechanical performance, emerged from stealth with an $8 million seed round led by Initialized Capital, with participation from 1517, Breakthrough Energy Discovery, Toyota Ventures, and Wireframe Ventures.

Arcturus is developing a proprietary process to infuse carbon nanomaterials into metals, creating a new class of high-performance materials designed to improve electrical, thermal, and mechanical performance. The company is building these materials as a drop-in solution for critical components across electrical systems, with target applications including motor windings, bus bars, heat sinks, and, over time, transmission lines and broader grid infrastructure.

Read more at GlobeNewswire

• 🇺🇸 Altillion Secures $5 Million Seed Round to Extract Iodine and Critical Minerals from Oilfield Wastewater and Industrial Brines /GlobeNewswire/

Looking for additional content? Paid subscribers get more stories delivered each week as part of Overcapacity!

Click to access ⤵️

Access Bonus Media

This weeks digest:

Read more

Capturing this week’s zeitgeist

The rise of “Labs” is essential because large venture-scale industrial opportunities ignore traditional disciplinary boundaries and require teams to master disparate fields across thermodynamics, fluids, mechanical and electrical engineering, materials science, and semiconductor physics, often by learning them on the fly when no single set of experts exists. This model was exemplified by Elon Musk’s transition from PayPal to founding and scaling SpaceX and Tesla, and it is now enabled by applied Labs that create dedicated, low-friction environments for cross-discipline collaboration and rapid iteration, something traditional siloed organizations or pure SaaS teams cannot replicate when building full-stack physical systems. The approach continues to be validated by software-founded companies such as Boom Supersonic (led by ex-Amazon engineers now developing supersonic jets) and AI leaders like OpenAI and Anthropic that have moved into hardware (see below), where only boundary-spanning Labs deliver the enduring technical depth and commercial value that venture capital bets now require.

Share

Assembly Line

This week’s industrial breakthroughs and frontier technologies of the built world.

Beyond the solver: How Geometric Deep Learning is reshaping CAE

A finite element mesh is an irregular graph. A point cloud from a 3D scan is an unordered set. A CAD surface is a manifold. None of these live comfortably on a regular grid. If you try to flatten them into one, as early AI-for-engineering approaches did, you lose the very geometric relationships that make the physics meaningful. You’re essentially crumpling a map to fit it into a square box, then wondering why the roads don’t connect.

GDL extends deep learning to these non-Euclidean domains – graphs, meshes, point clouds, and manifolds – by building neural networks that respect the underlying geometric structure of the data. The theoretical backbone comes from the concept of symmetry: a well-designed geometric neural network should produce consistent predictions regardless of how you rotate, translate, or renumber the input. This property – known as equivariance – is not just mathematically beautiful but also physically essential.

Read more at Siemens

• Siemens CEO on Industrial AI’s Future /YouTube/

Washington plant creates jet fuel from water, air and electricity

Alaska Airlines and Microsoft join Twelve to mark commercial operation of AirPlant One — the first plant in the United States to produce E-Jet fuel from CO2, renewable electricity, and water at scale /Twelve/. AirPlant is a eManufacturing factory that produces E-Jet® SAF and E-Naphtha with power-to-liquid technology that works like industrial photosynthesis—transforming CO2, water, and renewable electricity into hydrocarbons.

• AquaPoro Raises $5M to Advance Technology that Generates Net New Water from Air /Business Wire/

ViserDex: Visual Sim-to-Real for Robust Dexterous In-hand Reorientation

Experiments show that the pose estimator trained with 3DGS data outperforms those trained using conventional rendering data in challenging visual environments. We validate the system on a physical multi-fingered hand equipped with an RGB camera, demonstrating robust reorientation of five diverse objects even under challenging lighting conditions. Our results highlight Gaussian splatting as a practical path for RGB-only dexterous manipulation.

Read more at GitHub

• KinetIQ Ascend: Toward 100% Reliable Manipulation and Superhuman Speed /Humanoid/

• Kyber Labs releases a capable, force-aware hand that can do useful work around people /Kyber Labs on X/

• A world running foundation models on a billion robots won’t be doing it locally. /Gianluca on X/

Praxis Currents

Moneyball for Physical AI

In 2002, the Oakland Athletics won 103 games despite maintaining the third-lowest payroll in Major League Baseball. This advantage emerged because the market for player assets was mispriced: legacy scouts favored subjective aesthetics, stolen bases, and batting averages, whereas forward-looking management mathematically isolated on-base percentage, the …

Read more

2 months ago · 10 likes · Animesh Garg

• What Makes a Robotics Business Economically Viable /JacklouisP on X/

New Product Introduction

Highlighting new and innovative facilities, processes, products, and services

OpenAI and Broadcom unveil LLM-optimized inference chip

OpenAI and Broadcom (NASDAQ: AVGO) unveiled Jalapeño, OpenAI’s first Intelligence Processor: an accelerator architected around OpenAI’s vision for the future of LLM inference, and the first AI accelerator in a multi-generation compute platform the companies are building together to make advanced AI faster, more reliable, and more accessible to more people.

Jalapeño was co-developed from initial design to manufacturing tape-out in just nine months, and the custom AI accelerator program represents what we believe to be the fastest ASIC development cycle ever achieved in high-performance advanced semiconductors. That speed reflects deep software-hardware co-development with OpenAI’s engineering teams, Broadcom’s silicon implementation expertise, and the use of OpenAI models to accelerate parts of the design and optimization process.

Read more at OpenAI

• “When it comes to co-design of chips, he thinks the real value of companies like $AVGO, Mediatek, and $MRVL lies in their $TSM allocation and memory allocation, which they got sooner than everyone else” /Richard Jarc on X/

• Micron and Anthropic Announce Strategic Agreement to Scale Next-Generation AI Infrastructure /Micron/. Micron and Anthropic will analyze how memory and storage subsystems perform across various workloads and interact across the full infrastructure stack. This effort is expected to drive advances in memory and storage performance, energy efficiency and enhanced token economics in Anthropic’s AI infrastructure.

IBM Debuts World’s First Sub-1 Nanometer Chip Technology

IBM’s new sub-1 nm chip packs nearly 100 billion transistors onto a chip the size of a fingernail, nearly twice the density of IBM’s 2 nm chip, unveiled in 2021. Enabled by a series of structural and material innovations, including IBM’s groundbreaking three-dimensional nanostack architecture, the technology demonstrates how continued gains in performance and efficiency remain possible even as chip features approach atomic dimensions.

To produce this chip, IBM researchers developed an entirely new transistor architecture, called “nanostack,” the industry’s first known three-dimensional, nanosheet-based design /IBM Research/. Nanostack represents a major advance beyond nanosheet technology, the industry’s current leading-edge architecture, invented by IBM. The nanostack design vertically stacks and staggers transistors, taking advantage of 3D sequential integration to pack more transistors onto a chip. The design also unlocks the use of different material combinations within each stacked layer, optimizing performance and power efficiency of each transistor independent of the other.

Read more at IBM

Introducing Un-0: Generating Images with Coupled Oscillators

The best AI models today are conventional deep networks with transformer backbones. However, there is also a long history of alternatives that seek energy efficiency by leveraging the dynamics of a physical system, such as the noisy, time-varying behavior of analog circuits that compute with analog voltage and current instead of conventional digitized numbers.

To exploit these alternative computing methods, the AI task needs to be mapped efficiently to the dynamics of the physical system. Un-0 validates that modern AI workloads can run more efficiently on physical substrates than on today’s hardware.

Why oscillators? In the brain, rhythmic activity and synchronization are pervasive, and have long been hypothesized to do computational work like binding distributed features into coherent percepts, gating communication between regions, and organizing the timing of spikes (Gray et al., 1989; Buzsáki, 2006; Fries, 2015). Coupled oscillators are among the simplest mathematical models of that kind of behavior, which makes them a natural primitive to study for neuro-inspired models of computation (Winfree, 1967; Kuramoto, 1975; Ermentrout, 1996; Ermentrout et al., 2010).

Read more at Unconventional AI

Business Transactions

This week’s top funding events, acquisitions, and partnerships across industrial value chains.

Qualcomm to Acquire Modular

As AI scales, efficiency, not capability, becomes a constraint. Performance-per-watt drives the cost of inference, and cost determines what scales. Meeting this demand requires more than hardware. Developers need software that connects system-level optimization with heterogeneous, disaggregated compute, turning silicon performance into reliable and efficient AI services across accelerators, environments, and use cases.

Modular provides an open, AI-native software stack that enables AI to run efficiently across hardware architectures. Built by engineers who helped create much of today’s AI infrastructure, Modular’s unified platform runs models with industry-leading performance across CPU, GPU, NPU, and custom ASIC architectures without re-writes for each accelerator. For developers and enterprises, that means building once, deploying across any environment with lower total cost of ownership. Modular is supported by an open, industry-friendly, vendor-neutral developer community committed to improving the portability and efficiency of AI infrastructure.

Read more at Qualcomm

• AMD Acquires MEXT to Advance Memory Optimization for Compute Infrastructure /AMD/. MEXT has developed innovative AI-powered predictive memory technology designed to make flash behave more like DRAM, helping expand usable memory capacity while maintaining performance and efficiency.

Nearfield Instruments Secures $380 Million Series D Funding in Largest Ever Deep-Tech Funding Round in The Netherlands

Nearfield Instruments, a leader in advanced semiconductor 3D metrology and process control, announced the successful closure of a $380 million Series D funding round. The transaction values the company at $1.6 billion, marking a major milestone in Nearfield’s development into a global leader in semiconductor metrology and inspection.

Nearfield’s innovative metrology and inspection solutions provide the accurate, reliable and high-throughput measurements needed to control advanced processes, improve yield and ensure manufacturability. By enabling critical measurements for High Numerical Aperture Extreme Ultraviolet (High-NA EUV), Gate-All-Around (GAA) and Complementary Field-Effect Transistor (CFET architectures), and hybrid-bonded 3D integration, Nearfield plays a central role in making the next generation of AI computing scalable, energy-efficient, manufacturable and reliable.

Read more at Business Wire

Agility Robotics to Go Public Through Merger with Churchill Capital Corp XI

Agility Robotics, Inc., a leading humanoid robotics and physical AI company, and Churchill Capital Corp XI (NASDAQ: CCXI) (“Churchill”), a publicly traded special purpose acquisition company, announced they have entered into a definitive business combination agreement. Upon closing of the Transaction, the combined company is expected to operate as Agility and be listed on a major North American exchange under the ticker symbol “AGLT.”

Agility’s mission is to build robot partners that augment the human workforce and lead the adoption of humanoids everywhere. Its flagship humanoid robot, Digit, is a general-purpose, human-centric robot Made for Work™ and currently operating in manufacturing, distribution, and logistics environments to fill chronic physical labor shortages. Through more than a decade of development, Agility has established itself as one of the only humanoid robotics companies with multiple years of operational experience in real customer environments. The Company is supported by leading strategic investors and partners across the AI, technology, VC, and industrial ecosystem, including DCVC, NVIDIA, Amazon, SoftBank Vision Fund 2, Schaeffler, Foxconn, Abico, and Playground Global.

Read more at Agility Robotics

• Apptronik’s Robots Can Strut, But Its CEO Isn’t Bragging Yet /Bloomberg/

• onsemi to Acquire Synaptics to Enable the Next Generation of Intelligent Systems for Physical AI /onsemi/

Looking for additional content? Paid subscribers get more stories delivered each week as part of Overcapacity!

Click to access ⤵️

Access Bonus Media

This weeks digest:

Read more

Capturing this week’s zeitgeist

The magic of the Automate Show is the feeling that you are stepping directly into the future, where the cutting edge of industrial tech becomes a tangible, hands-on reality. As thousands of engineers, founders, and industry leaders gather this week at McCormick Place, the show floor transforms into a high-energy hub of live demonstrations, breakthrough AI, and game-changing humanoid robotics. Whether you are watching groundbreaking ideas pitch live at the Startup Challenge or trading insights with peers between sessions, it is a unique space where the next big technology and the right connection come together to reshape your business.

A few interesting exhibitors I’ve had the pleasure of first connecting with at previous Automate shows:

• Copia Raises $26M in Funding to Unify OT Code Management, Backup, and Recovery /PR Newswire/ used to accelerate Copia’s mission to bring modern code management and recovery capabilities to the operational technology (OT) teams who are the first responders in the event of an industrial cyberattack or downtime incident.

• Podium Automation Raises $18M Series A to Modernize the Backbone of Industrial Automation /PR Newswire/

• CESMII Releases i3X™ 1.0: A Complete, Implementation-Ready Open API Specification for Manufacturing Interoperability /CESMII/

Share

Assembly Line

This week’s industrial breakthroughs and frontier technologies of the built world.

🇪🇺🏭🧠 AI for Industry Is Europe’s Hope to Salvage Manufacturing Edge

While Europe may have lost the consumer chatbot wars to the US and China, a critical new battleground is emerging on the factory floor. European manufacturers are racing to integrate industrial AI using real-time sensors, predictive maintenance software, and automated quality control to create highly efficient, autonomous “smart factories.” This transition is a matter of survival, with over $1 trillion in manufacturing value at risk of leaving Western Europe due to high costs and labor shortages. However, the clock is ticking: industry experts warn that Europe has a narrow two-to-three-year window to successfully deploy these advanced systems before falling irreversibly behind Asian competitors, who are moving rapidly to combine robotics and AI.

Read more at Bloomberg

• 🤝🏭📊 Siemens partners with Databricks and FFT to turn production data into scalable AI-driven insights /Siemens/. Siemens Industrial Edge and FFT DataBridge connect shopfloor and plant data directly to the Databricks Platform – no IoT middleware required

• ☁️📊💨 Ingesting the Milky Way: Petabyte-Scale with Zerobus Ingest /Databricks/. By ingesting 1PB within 24 hours, Databricks demonstrates Zerobus’s ability to maintain continuous throughput of 12 GB/s to a single table.

• 🇪🇺📑 Europe’s golden age is over. Here’s the playbook for the next one. /Azumuta/ Eight imperatives, six of the sharpest voices in European manufacturing, distilled into one document designed to be read in one sitting.

• 📑 MIT’s Initiative for New Manufacturing builds momentum /MIT News/

🖨️🔋 The Secret Revolution in Battery Technology: 3-D Printing

The core uniqueness of 3D-printed battery technology lies in its ability to transform energy storage from a rigid, predetermined component into a fluid structural element that can fill any arbitrary geometry or empty space within a device. Unlike traditional manufacturing, which relies on chemical innovations and standard cylindrical or pouch shapes, this additive manufacturing approach is entirely chemistry-agnostic—boosting existing lithium-ion, sodium-ion, and solid-state designs alike. Furthermore, it revolutionizes the production process itself by enabling a solvent-free dry manufacturing method that eliminates massive, energy-intensive drying ovens, drastically reducing production footprints while making complex, high-capacity material designs physically viable.

Read more at WSJ

• 🚙🔋 GM may ditch LFP batteries for future EVs /Reuters/ and instead is focused on developing a different battery chemistry, lithium manganese-rich, or ​LMR, which the company has said costs about the same as LFP to make in the ​U.S., but can store more energy for the same weight and size.

• 🤝🚙🔋 QuantumScape Announces Agreement with Honda on Solid-State Battery Technology /QuantumScape/

🦾🧠 An Overview of Modern AI Robotics from First Principles

There is a deceptively simple way to describe what physical AI is all about, a way in which anyone with a STEM background will intuitively understand. Like all other AI models, a model which controls a robot is also a function. It takes in observations (camera pixels, joint angles, the felt resistance of a gripper, etc) and it outputs actions, the next set of positions and torques for its motors. Everything else, all the complicated algorithms, training methodologies, and data scaling theories are in service of producing a good version of that function and embedding it into the weights of a neural network.

The way out is the way humans learn: practice, against a critic. This is reinforcement learning: letting the robot try, scoring the outcome, and reinforcing what worked. The catch is that you can’t supervise a robot at every moment or hand it an example for every possible situation, and in the real world (unlike a simulator) attempts are serial, slow, and require a human to reset the scene. RL in the physical world proves very difficult as a result. Self-improvement algorithms hence fix major bottlenecks in performance for robotics, especially with regards to generalization.

Read more at Interlatent

The Diligence Stack - By Creative Strategies

AI Server Demand Is Becoming Three Markets

Read more

2 months ago · 5 likes · 3 comments · Ben Bajarin

New Product Introduction

Highlighting new and innovative facilities, processes, products, and services

🧫 Launching Midjourney Medical & The Midjourney Scanner

The major computational task is figuring out how to change waves into images. Basically - as waves travel through the water and your body they change shape. The shape of these waves changes whenever there is a change in density or stiffness (i.e., going from water to skin to fat to muscle to bone). By looking at how the shapes of all the waves change, we reconstruct a detailed map or ‘image’ which basically lets us figure out what’s in there. All of these images come together to cover a 3D map of your body, down to a fraction of a millimeter, that looks a lot like today’s MRIs but at nearly a hundred times the speed.

The next 12 months are about refining our algorithms and hardware on a daily basis. We’ll be doing research trials to show off the raw capabilities of our system, moving towards a 2nd generation hardware design, and we’ll do build-out for our first “research spa” which will become the promethean site that enables mass-scale health scanning.

Read more at Midjourney

• 📏 Lumafield Introduces Neptune Performance and Triton Performance Connecting Metrology-Grade Measurement to Production-Line Inspection /Lumafield/

• 🧪 Energy-efficient codon optimization on thermodynamic hardware /arXiv and THRML/. The first concrete pharmaceutical application mapped to thermodynamic hardware with energy estimates grounded in prototype measurements. All code is released under an open-source license.

🦾🧠 Qwen-Robot Suite: A Foundation Model Suite for Physical World Intelligence

The Qwen family of foundation models already gives strong perception and reasoning about the physical world. But seeing is not acting: the gap between vision and language understanding and physical control remains the central bottleneck for embodied intelligence. The Qwen-Robot Suite bridges this gap with three foundation models — Qwen-RobotNav, Qwen-RobotManip, and Qwen-RobotWorld. Nav unifies five navigation task families through a controllable observation protocol. Manip turns heterogeneous robot data into a coherent canonical space, enabling cross-embodiment training at scale. World co-trains 20+ embodiments via a natural-language action interface under one world model. Together, they enable an agentic system where general intelligence translates directly into physical action.

Read more at Qwen

• 🦾🧠 SC3-Eval: Evaluating Robot Foundation Models via Self-Consistent Video Generation /arXiv/. Joint work between Physical Intelligence and the NVIDIA Cosmos team /X/.

• 🦾📊 XDOF Raises $70 Million To Build Infrastructure For Robot Foundation Models /Pulse 2.0/. XDOF has open-sourced ABC-130K, which it describes as the largest open-source teleoperation dataset /Hugging Face/.

Business Transactions

This week’s top funding events, acquisitions, and partnerships across industrial value chains.

• 🇨🇭♻️ GR3N raises €15.5M to build the world’s first microwave-powered PET recycling plant / TNW/. GR3N’s technology, called MADE (Microwave Assisted DEpolymerisation), can process all PET.

🇪🇸🦾 THEKER raises €73M to deploy AI-native factory robots that learn on the job

THEKER, the Barcelona-based AI robotics company, has raised €73 million ($85 million) in a Series A to scale its generalist factory robots across industrial production environments. The round was led by CRV, with participation from Samsung, LVMH, Cathay Innovation, 20VC, Henkel Ventures, Korelya, and Bright Pixel Capital. It marks Samsung’s first-ever investment in a Spanish company, LVMH’s first bet on the Spanish startup ecosystem, and CRV’s first investment in Spain.

The round comes less than a year after THEKER closed Spain’s largest-ever seed round at €18 million. The speed from seed to Series A reflects what the company calls real commercial deployment momentum, not just lab demos.

THEKER’s pitch is a new category of industrial robot: AI-native, generalist machines that adapt in real time to changing environments, mixed SKUs, irregular shapes, and operational variability without manual reprogramming. Unlike traditional industrial robots that are rigid and costly to reconfigure, THEKER says its systems deploy in days and continuously learn in production.

Read more at TNW

• 🇭🇰🦾 Humanoid robot maker EngineAI files for a Hong Kong IPO /TNW/

• 🇬🇧🦾 Fuse: We’re building robots /Alan Chang on X/. Energy is the bottleneck for AI; our goal is to unleash it, and robots are how we make sure it never becomes one again.

🇺🇸🪙 Foundation Alloy Industrializes New Metals Platform with $22M Series A

Foundation Alloy, the 21st-century metals company building a new way to engineer alloys, today announced $22 million in Series A financing to scale production on its MetalsFIRST™ platform—a fully integrated, solid-state metallurgy technology—to industrial volumes. The round was led by Voyager Ventures, with participation from Trust Ventures, Yamaha Motor Ventures, America’s Frontier Fund, Overlap Holdings, Material Impact, Engine Ventures, and El Cap.

Foundation Alloy’s MetalsFIRST approach is a fully integrated, solid-state platform encompassing composition design, mechanical alloying, shape forming, and sintering, that produces engineered alloys without ever entering the molten state. This process enables faster, simpler manufacturing cycles, capital-efficient industrial scale up, and access to alloy compositions and properties impossible to achieve with legacy melt processes. Foundation Alloy’s specialty and stainless steels are engineered for the most demanding applications while requiring 80%–90% fewer production steps, while its Molyclast® product line includes MC1200, a molybdenum alloy that delivers more than 3x the strength of commercial alternatives. This combination of superior performance and manufacturability unlocks new capabilities for demanding applications across the industrial space.

The funding transitions Foundation Alloy into an industrial supplier. This summer, the company is opening a new 36,000 sqft facility in Massachusetts, standing up an additional modular production cell with Re:Build Manufacturing in southern New Hampshire, and doubling headcount across production, engineering, and commercial operations.

Read more at PR Newswire

• 🇺🇸🪙 Cuprum Metals raises $19.4M Series A to accelerate Copper Extraction Technology /Cuprum/ using a water-based chemical leach, with the flexibility to operate using sea water, to achieve high recovery rates at significantly faster processing times.

• 🇺🇸🪙 NOX METALS Raises an $11.5M Seed Round to Reindustrialize America /NOX/

• 🇬🇧⚙️ Doncasters seeks to raise up to $747M in U.S. IPO /Yahoo/ which manufactures high-precision alloy components for demanding conditions.

🇺🇸♨️ Critical Energy Raises $22M to Mass-produce Modular Geothermal Turbines

Based in Los Angeles and founded in 2024 by Spencer Jackson, Critical Energy just closed a seed round led by Susa Ventures and Upfront Ventures. Other investors include MaC Venture Capital, Susquehanna Sustainable Investments, Humba Ventures, Scribble Ventures, Underground Ventures, and Silicon Valley Bank, which provided venture debt.

Critical Energy designs modular geothermal turbines. These are container-sized units built in-house that turn heat into electricity using a closed-loop fluid system and custom machinery. The company says it can reduce deployment times from years to weeks by building most of the system off-site. It has already built a pilot facility in Los Angeles and is moving toward larger commercial projects.

Read more at TFN

• 🇺🇸♨️ Mantle Energy, Inc (Mantle), has successfully closed a $5 million seed round led by 17Shoals /PR Newswire/ leveraging techniques to create heat in oil and gas formations in a way that facilitates both Enhanced Oil Recovery and Geothermal Energy to provide energy security and a lower carbon baseload power.

Looking for additional content? Paid subscribers get more stories delivered each week as part of Overcapacity!

Click to access ⤵️

Access Bonus Media

This weeks digest:

Read more

Capturing this week’s zeitgeist

When we tune into a massive sporting event, we’re usually watching the athletes, but behind the scenes, a different kind of elite performer is keeping the whole show from collapsing. YOU. This year’s modern sports spectacles are high-stakes marvels of heavy industrial engineering, where prepping a venue means pushing logistics and science to their absolute limits. Whether it’s turfgrass scientists treating the 2026 World Cup pitches like a living "layer cake" /UT/, growing stadium sod in Colorado and racing it across the country in refrigerated trucks to beat the clock /YouTube/, or structural engineers dropping a 600-ton, 92-foot-tall steel behemoth nicknamed "The Claw" directly onto the historic White House lawn for UFC Freedom 250 /ESPN/, the margin for error is zero. This week we explore the incredible technology changing agronomy, role of the robots and AI, and precision logistics. Enjoy!

Share

Assembly Line

This week’s industrial breakthroughs and frontier technologies of the built world.

Why robotic arms are now being integrated with CNC machines

✍️ Authors: Lou Ferrell

For years, automation in CNC environments meant a dedicated unit performing one task. Modern integrations look different. A robotic arm now loads a raw blank, transfers it between machines, inspects the finished part, and routes it downstream — all without human involvement.

The infrastructure being built today by FANUC, Universal Robots, KUKA, ABB and RoboDK is already functional. What remains an open question is how much further the role of the robot expands as artificial intelligence matures. A robot that loads a part today follows fixed logic. But one that monitors spindle load, detects tool wear in real time and decides autonomously whether a part passes inspection is an entirely different kind of participant in the manufacturing process.

Read more at The Robot Report

What Can SLS 3D Printing Really Do? | The SLS Mindset

Selective laser sintering (SLS) is one of the most capable AND one of the most overlooked additive manufacturing processes for polymer parts. In this episode of the SLS Mindset series, Nate Stevens of Lifestyle Additive and explore what makes SLS different: no supports, complex geometry without compromise, assembly consolidation into single parts, and production-ready material performance.

• Formlabs’ new 3D printer is poised to reshape manufacturing /Fast Company/. The Fuse X1 is designed to bring large-format SLS printing to more engineering teams and production lines. The new machine is a selective laser sintering (SLS) printer built for manufacturers, engineering teams, product developers, and 3D printing service bureaus. The printer can turn out production-quality parts in less than 24 hours. (It starts at $84,999, is available for order today, and is expected to begin shipping in the fourth quarter of 2026.)

BlueME: Robust Underwater Robot-to-Robot Communication Using Compact Magnetoelectric Antennas

✍️ Authors: Mehron Talebi, Sultan Mahmud, Adam Khalifa, Md Jahidul Islam

We present the design, development, and experimental validation of BlueME, a compact magnetoelectric (ME) antenna array system for underwater robot-to-robot communication. BlueME employs ME antennas operating at their natural mechanical resonance frequency to efficiently transmit and receive very-low and low-frequency (VLF/LF) electromagnetic signals underwater. We detail the system’s design, simulation, fabrication, and integration onto low-power embedded platforms, emphasizing portability and scalability. For performance evaluation, we deployed BlueME on mobile robot platforms in freshwater (lake) and saltwater (ocean) trials. Our tests demonstrate reliable signal transmission and detection using the BlueME antenna system at distances beyond 700 m while consuming less than 10 W of power. Ocean trials demonstrate that the system operates effectively under challenging conditions, such as turbidity, obstacles, and multipath interference, which are factors that typically degrade acoustic and optical methods. Our analysis also examines the impact of complete submersion on system performance and identifies key deployment considerations. This work represents the first practical underwater deployment of ME antennas outside the laboratory and implements the largest VLF/LF ME array system to date. BlueME demonstrates significant potential for marine robotics and automation in multirobot cooperative systems and remote sensor networks.

Read more at IEEE and UF

Commercializing cutting-edge machine vision is notoriously difficult. Whether it’s Maneva's Video-to-Action AI /YouTube/, Lumafield’s accessible industrial CT /Lumafield/, or Cognex’s high-speed vision systems /The Robot Report/, a brilliant product is only half the battle. Trista Li studied 40+ vision AI companies and their unique go-to-market strategies required to win in this space:

Deploy 95

Is Industrial AI GTM Built Different?

This newsletter covers go-to-market strategy for founders selling AI and hardware into manufacturing and industrial environments. If that’s you, or you’re investing in this space, you’re in the right place…

Read more

2 months ago · 6 likes · Trista Li

New Product Introduction

Highlighting new and innovative facilities, processes, products, and services

This Nvidia Challenger Says Its AI Chip Is 10x Faster Than A GPU

There’s a new challenger to Nvidia that says its chip can run AI inference at ten times the speed of a standalone GPU. The new chip from D-Matrix is called Corsair, and it’s now in volume production with commitments from hyperscalers, neoclouds and frontier AI labs. In an exclusive interview with CNBC, CEO Sid Shath explains how Corsair bypasses the DRAM shortage by relying on SRAM directly on the chip, and how that tight integration means Corsair can transfer data using five times less energy. It’s a novel approach to memory that’s led to huge gains for other chip startups in recent months. Cerebras’ blockbuster $95 billion IPO in May landed it among tech’s largest ever debuts, and Groq received $20 billion from Nvidia in the AI giant’s largest purchase ever in December. Now CNBC asks whether D-Matrix could be next.

• Inside The Chip Factory 1,000 Times Cleaner Than An Operating Room /YouTube/

• Rapidus Completes 150 Billion Yen Funding Round from Japan Government /PR Newswire/ to support its evolution from the R&D stage to full-scale 2nm logic semiconductor manufacturing by 2027.

• TensorWave Raises $350 Million Series B at $1.55B Valuation to Expand Global AMD-Powered AI Infrastructure /Business Wire/

• Granarium Technologies raises EUR 1 million to commercialise the world’s first renewable, affordable supercapacitors for grid stability and industrial reliability /PR Newswire/

Claude Fable 5 and Mechanical Design (V8 engine)

Claude Fable 5 and Claude Mythos 5 /Anthropic/. Fable 5 is state-of-the-art on nearly all tested benchmarks of AI capability, showing exceptional performance in software engineering, knowledge work, vision, scientific research, and many other areas.

• Statement on the US government directive to suspend access to Fable 5 and Mythos 5 /Anthropic/. Anthropic suspends new AI tools over US government security concerns /BBC/.

• Will AI-augmented CAD make manufacturability easier to integrate into design? /Brain Ratliff on X/

• Robotics will not have a clean Llama moment /The Robot Report/

We’re Raising the Curtain: Introducing Azumuta Labs

It’s a working space where we run real experiments, put prototypes in front of real operators and engineers, document what happens, and share it here before we have all the answers.

That means you’ll see results that surprised us. Approaches that didn’t work out. Ideas that are solid in a controlled environment but need more work before they’re production-ready. We’ll share all of it, including the failures, because the path to useful manufacturing AI is shorter when we stop repeating each other’s dead ends.

Read more at Azumuta

Business Transactions

This week’s top funding events, acquisitions, and partnerships across industrial value chains.

NEURA Robotics Announces Record Series C of up to $1.4 Billion to Accelerate the World’s Leading Physical AI Platform

NEURA Robotics, the pioneer in cognitive robotics and creator of the Neuraverse, announced a landmark Series C financing with a total round size of up to $1.4 billion to accelerate its mission of building the world’s leading Physical AI platform. The financing brings together global leaders across AI, robotics, compute, manufacturing and industrial infrastructure, including Tether, Qualcomm Technologies, Inc., Amazon, NVIDIA, imec.xpand, Bosch, Schaeffler, European Investment Bank, Lingotto Horizon, InterAlpen Partners and others.

NEURA is building a new category of AI infrastructure where cognitive robots continuously learn, collaborate and operate across real world environments through a shared intelligence ecosystem called the Neuraverse. Unlike traditional robotics companies focused on isolated machines or narrow industrial automation, NEURA combines robotics, AI, sensors, edge compute and large scale learning infrastructure into one unified platform architecture designed for global deployment.

Read more at Business Wire and The Robot Report

PhysicsX Announces $300M Series C to Accelerate Physics AI for Industrial Engineering

PhysicsX, the physics AI company for industrials, announced an oversubscribed $300 million Series C financing at a valuation of approximately $2.4 billion. The round is led by Temasek, with participation from new investors M&G Investments and Intrepid Growth Partners, alongside existing investors including Applied Materials, Atomico, General Catalyst, July Fund, NGP, NVIDIA, Radius, and Siemens. Temasek first invested in PhysicsX in 2025, and has played an instrumental role in supporting the company’s international expansion and growth.

This round will accelerate the company’s global growth, the expansion of its platform capabilities, and frontier research, including the development of larger, more powerful pre-trained physics AI models, known as Large Physics Models /arXiv/.

Read more at PhysicsX

• Flux Raises $5M to Give Engineering Leaders Ground-Truth Visibility in the Age of AI /PR Newswire/

Standard Bots Raises $200 Million Series C At $1 Billion Valuation To Scale AI-Native American Robotics

Standard Bots, the largest U.S. manufacturer of AI-native industrial robots, announced a $200 million Series C financing led by RoboStrategy and existing investor General Catalyst, valuing the company at $1 billion.

The funding will support the expansion of the company’s manufacturing footprint in Glen Cove, New York, where Standard Bots is increasing its facility to 70,000 square feet. The company said the expanded operation will strengthen its ability to design, assemble, and deploy American-made robots at scale.

Read more at PR Newswire and The Robot Report

• Mesoware Announces $1.5M to Build AI-Powered Robots for Manufacturing /Manufacturing Tomorrow/

Innovafeed raises $59m, scales production, cuts R&D headcount as insect ag hits “industrialization phase”

French insect ag firm Innovafeed has raised a €51 million ($59 million) round and unveiled plans to consolidate activities at its commercial-scale black soldier fly larvae production facility in Nesle, France.

The round, which takes its cumulative funding including grants to over $500 million, was backed by existing shareholders Creadev, QIA, Temasek, FFC, ABC Impact, and ADM along with banking partners, said Innovafeed, which was founded in 2016 by Aude Guo, Bastien Oggeri, and Clément Ray.

Read more AFN

• Leaf Raises $13M Series B to bring AI to Agribusiness /PR Newswire/

• SWARM Engineering Raises $10M Series A to Transform Operational Decisions with AI /Business Wire/

Maneva Raises $27M to Tackle Inefficiencies Across the $46T Manufacturing Industry

Maneva, the Video-to-Action AI company building the intelligence layer for modern manufacturing, announced $27 million in Series A funding, accelerating its mission to close the $18 trillion global manufacturing efficiency gap. The Series A round was led by U.S. Venture Partners (USVP), with participation from returning investors Bling Capital and Freestyle Capital, alongside a group of leading angel investors and funds. This fundraise brings Maneva’s total capital raised to $38.4M.

Maneva’s Video-to-Action AI turns existing factory cameras into real-time operational decisions. It begins with deep learning computer vision, giving factories the ability to see, understand, and act on what’s happening on the floor, then expands into agentic systems that connect, coordinate, and optimize the entire operation. It is the deep learning engine for modern manufacturing, and the foundation of everything Maneva builds.

Read more at Maneva

Looking for additional content? Paid subscribers get more stories delivered each week as part of Overcapacity!

Click to access ⤵️

Access Bonus Media

This weeks digest:
https://substack.exponentialindustry.com/p/glass-manufacturing-ai-pinball

Read more

Capturing this week’s zeitgeist

The enduring success of Stern Pinball offers valuable lessons in managing production complexity at the intersection of heritage craftsmanship and modern tech. While often perceived as a nostalgic pastime, manufacturing a pinball machine is an intensive engineering challenge that requires balancing thousands of unique parts within a highly constrained physical footprint. As an avid player myself, I closely followed Austin Weber’s coverage of Stern’s operational evolution and took away the following core lessons:

• Intricate Packaging & Multi-Disciplinary Design: Each machine represents a dense engineering and spatial challenge. Design teams must integrate thousands of distinct parts, from mechanical solenoids to digital HD displays, within a fixed footprint. This requires tight synchronization between mechanical, electrical, and software engineering during a long development cycle.

• The Physics of Quality Control: Because pinball is an entirely physical, physics-based experience, there is zero margin for error in geometry. A fraction of a degree in playfield tilt or a millimeter of deviation in a ramp’s placement completely alters the game’s flow, making precision physical placement a critical quality baseline.

• Labor Efficiency Over Automation: Stern demonstrates that for certain high-complexity, high-variety products, manual craftsmanship remains superior to robotics. With 18 to 24 labor hours required per unit, the assembly line relies on highly organized human cells to navigate the unique wiring harnesses and mechanical subassemblies specific to each title.

• Technological Convergence: A modern machine is a complex hybrid system. It must seamlessly balance legacy electromechanical feel (pop-bumpers and flippers) with modern digital infrastructure like IoT connectivity and advanced animatronics, forcing a design philosophy that respects tactile tradition while adopting modern software standards.

• Operational Flexibility: Because the company juggles multiple high-profile pop-culture licenses and various model tiers (Pro, Premium, Limited Edition), the factory floor must remain highly agile. The lesson here is the value of a flexible manufacturing layout that can pivot production focus rapidly to align with real-time market demand.

Share

Assembly Line

This week’s industrial breakthroughs and frontier technologies of the built world.

☀️ How One Man Powers a Global App from a Dutch Cow Barn

Just outside Amsterdam, a converted cow barn is quietly running a global payment app used by over a million people around the clock. To keep it online, Wietse Wind and his installer Danny built a massive 540 kWh LiFePO4 storage system with 150 solar panels, multiple Quattro-II and MultiPlus-II inverters, and full grid redundancy using Victron Energy equipment. This is the story of the largest residential ESS install either of them had ever built.

• 🐂 A North Dakota startup raised $10 million from ranchers and neighbours to build the operating system for beef cattle /TNW/

🎛️ State of Digital in Process Manufacturing

✍️ Author: Dimitris Paraskevopoulos

Prior IoT Analytics research, specifically on IoT use cases, has found that process manufacturers typically lag behind discrete manufacturers in adopting digital technologies. This is largely due to long asset replacement cycles (more than 50 years in some cases), among other reasons. Oil & gas, though, is often regarded as ahead of other process industries due to its remote nature, with established SCADA and remote connectivity architectures.

Smart sensors top the list of most deployed technologies for process manufacturers. Companies are also adopting bolt-on wireless sensors, such as US-based industrial AI company Augury’s bolt-on Halo sensors, which monitor machine parameters such as vibration, temperature, and magnetic field data and transmit insights to the cloud for AI-driven analysis.

Read more at IoT Analytics

• 📑🎛️ Industrial AI: The New Wave of Innovation Hitting Heavy Industry /Energize Capital/

🫙 AI for the Glass Industry

Georg Katzlinger explores the surprising complexity of the flat glass industry and the unique challenges it faces including how AI solutions are revolutionizing the way custom glass orders are processed, from deciphering handwritten requests to integrating with legacy ERP systems. Georg shares his journey from engineering to AI entrepreneurship, revealing why deterministic, human-in-the-loop automation is essential for manufacturers.

💿 A new way to build chips: Sequentially stacking silicon to extend Moore’s law

✍️ Author: Michael O’Boyle

The most efficient approach, known as monolithic three-dimensional integration, builds each layer directly on top of the previous one to maximize interlayer connectivity. However, achieving this has been a longstanding technical challenge. Preparing high-quality silicon and fabricating high-performance devices normally require processes operating at 1,000 degrees Celsius, hot enough to destroy the metal wiring. For upper layers beyond the first, the temperature constraint, or “thermal budget,” is strictly set to 400 degrees.

Current commercial three-dimensional chips are made by fabricating semiconductor devices on separate wafer substrates first and then bonding those wafers or dies together. While this approach has enabled successful products such as high-bandwidth memory and 3D V-Cache chips, it comes with substantial limitations. The alignment between layers is necessarily coarse, and the micron-scale vertical connections between layers called through-silicon via or TSV are relatively large and sparse.

Read more at University of Illinois and Nature

• 💿🤝 NVIDIA and TSMC Bring AI Into Fabs to Advance Semiconductor Design and Manufacturing /NVIDIA/

🚀 Tom Mueller: SpaceX’s #1 Employee Built the Merlin Engine. Now He’s Building Impulse Space

• 🚀 Impulse Space Raises $500M Series D to Build In-Space Mobility Infrastructure for the Space Economy /GlobeNewswire/

New Product Introduction

Highlighting new and innovative facilities, processes, products, and services

Siemens powers the next phase of industrial AI with Intelligence Center X

Siemens announced Intelligence Center X, new industrial AI orchestration software designed to help organizations turn industrial AI from isolated experimentation into scalable, real world business impact through a hybrid workforce where people and AI agents work together with shared context, workflows and lifecycle intelligence.

By connecting data, models and workflows on a single governed foundation, Intelligence Center X enables companies to deploy AI-driven applications and agents faster, with full traceability and control. Intelligence Center X combines the Mendix™ low-code platform with Siemens’ Graph Studio and AI Studio software from the Rapidminer® portfolio to create the enterprise context, business-specific lifecycle intelligence, and orchestrated agents and intelligent applications businesses needed to turn AI ambition into real value gains. Across industries, companies have invested in AI but are struggling to scale beyond pilots because data is fragmented, governance is inconsistent, and AI insight does not connect to real workflows. Intelligence Center X is designed to close that gap as the production-ready system that orchestrates people and AI agents together, on top of what enterprises already own, with full auditability and policy controls.

Read more at Siemens

• 🫙 Vivix Vidros Planos achieves 4x faster quality resolution by scaling agentic AI with Mendix and Snowflake /Mendix/

• 🤝 Siemens and HighByte partner on industrial data operations to scale Industrial AI /Siemens/

🖨️🛥️ Voltage Vessels and CEAD developed a demonstration platform using LFAM, which is informing Voltage Vessels’ basalt fiber-reinforced PETG material developments

✍️ Author: Grace Stubbins

Voltage Vessels (Maui, Hawaii, U.S.), a startup founded by Sam Young, has submitted a 6-meter 3D printed rigid hull inflatable boat (RHIB) prototype for U.S. maritime defense evaluation. The hull’s technical execution was led by CEAD (Delft, Netherlands), using its large-format additive manufacturing (LFAM) system — an industrial-scale printer used by multiple defense contractors for hull-scale composites production — and HDPro thermoplastic composite (TPC) materials.

Separate third-party testing commissioned by Voltage Vessels measured the materials’ tensile strength along the print direction at roughly 108 MPa, versus approximately 49 MPa for the HDPro composite used as a benchmark. Eclipse X9 also retained more than 90% of its strength after long-term saltwater immersion, with water absorption below 0.4%. Additional testing performed at the University of Maine’s Advanced Structures and Composites Center (UMaine ASCC) further validated the material’s mechanical performance under project ID UM-TC-23-1008.

Read more at CompositesWorld

Business Transactions

This week’s top funding events, acquisitions, and partnerships across industrial value chains.

• 🤝 Fanuc, Google advance industrial robotics as part of recent AI deals /Manufacturing Dive/

🇩🇪🦾 SoftBank in Early Talks to Back $800 Million Agile Robots Round

Agile Robots is in discussions to raise about $800 million in new funding from backers including SoftBank Group Corp., as investors seek out companies deploying artificial intelligence in real-world settings. SoftBank is in talks to contribute more than $300 million, according to people familiar with the matter, who asked not to be identified as the information is private. The company previously led a funding round in Agile Robots.

Read more at Bloomberg

• 🇺🇸🦾 🧠 Generalist raises $400M to scale its general-purpose AI models /The Robot Report/

• 🇨🇦🦾 🧠 Mecka AI raises $60 million to train robots with human data sourced from body sensors and iPhones /Fortune/

🇮🇱 DriveNets secures $410 million in latest funding round, AMD joins as investor

DriveNets, an Israeli networking software firm, said it has raised $410 million in its latest ​funding round led by Bessemer Venture Partners ‌and Atreides Management, taking its total capital-raise so far to $1 billion. New investors AMD and Red Dot Capital joined ​the round, while existing investors Pitango and ​D1 Capital Partners also participated. The company would ⁠use the proceeds to meet the surging ​demand for large-scale AI infrastructure, CEO Ido Susan ​said.

DriveNets allows telecommunications operators and data centers to build and manage networks using standard, off-the-shelf hardware, rather ​than costly ​proprietary systems. ⁠Its technology supports high-speed connectivity and AI workloads.

Read more at Reuters

• Dragos Acquires Phosphorus to Bring OT-Native Cybersecurity to the Full xOT Environment /PR Newswire/

🇺🇸⛏️🧠 Terra AI raises $20M to Accelerate Mineral and Reservoir Exploration to Meet Global Critical Mineral and Energy Demands

Terra AI, the artificial intelligence platform helping solve subsurface uncertainty for mineral and energy development, announced it raised $20 million Series A led by Khosla Ventures and strategic investment from BHP Ventures.

Terra AI’s patented technology integrates nearly every type of exploration data to generate millions of geological models, enabling teams to evaluate a comprehensive range of potential subsurface scenarios. This allows project teams to identify the best drilling and geophysical approaches to reduce key uncertainty and rigorously assess project economics earlier than ever before. By going beyond geological visualization and delivering risk-quantified decision insights, Terra AI helps explorers on virtually every important decision across their portfolio from determining which assets to acquire in a portfolio to accelerating and de-risking resource development efforts. The outcome: resource pipelines are developed faster and more efficiently to help the world meet its resource needs.

Read more at PR Newswire

• ⛏️ CopperTech Metals Announces Public Filing of Registration Statement for Proposed Initial Public Offering /Business Wire/. CopperTech’s flagship asset, the Konkola Copper Mines, is a high-grade copper and cobalt producer strategically located in Zambia’s Copperbelt Province.

• ⛏️ Red Metals Raises $10M to Reshore American Copper Refining and Build a New Supply Chain for Finished Copper Products /Business Wire/ including a $70 million facility in Charleston, South Carolina.

🇺🇸👓 Phosio Raises $4 Million Seed Round to Bring Affordable AR/AI Glasses to Mass Market

Phosio announced the closing of a $4 million seed financing round to accelerate commercialization of its thin-film materials platform for next-generation AR/AI glasses. The round was led by MESH, with participation from Sierra Ventures, TEL Venture Capital, Silicon Catalyst Venture, Willamette Valley Capital, and Silicon Catalyst Angels.

As AI shifts from phones to ambient interfaces, the industry is racing toward lightweight, all-day wearable glasses. But optics and display costs remain the biggest barrier to mass adoption. Today’s AR/AI glasses rely on complex, costly display systems that are bulky, power-hungry, and difficult to manufacture at scale. Phosio’s proprietary thin-film platform integrates high-performance display functionality directly into lightweight, everyday eyewear, eliminating the cost, bulk, and power demands that have slowed mainstream adoption.

Read more at Business Wire

• 🇺🇸🪪 Alitheon Secures $8 Million Series A1 to Scale “Biometrics for Things” Across Global Supply Chains /PR Newswire/

Looking for additional content? Paid subscribers get more stories delivered each week as part of Overcapacity!

The three most important technology shifts and product development strategies Gavin highlights are:

The Disaggregation of Prefill and Inference and its Impact on Hardware Lifespans

One of the most critical structural shifts in how AI models run is the architectural decoupling of the “prefill” phase from the “decode” phase during inference [49:46].

• The Shift: Baker explains that prefill—where the model processes and understands the prompt or input context—is fundamentally a memory capacity-bound problem [51:17]. Conversely, decode—the subsequent generation of output tokens—is a memory bandwidth-constrained process [51:25].

• The Competitive Strategy: This canvas allows chip designers to make highly targeted, optimized trade-offs [51:32]. Startups (such as Groq or Cerebras) can build hyper-specialized chips optimized specifically for one part of this loop.

• The Financial Impact: Crucially, Baker points out that companies can now stick specialized custom chips (like Cerebras wafer-scale hardware or Groq LPUs) in front of existing, older hardware architectures like Nvidia Hoppers or Amperes [49:46]. This approach uses the legacy GPUs for the prefill stage and custom silicon for decode, extending the useful economic lifespan of data center GPUs from 1–2 years to potentially 10–15 years [49:28].

Ongoing question for investors: How long before a GPU depreciates? /CNBC/

The Transition to Custom ASICs and the “Terafab” Talent Strategy

As standard GPU architectures face physical limits and intense market constraints, product strategy is shifting toward hyper-specialized ASICs (Application-Specific Integrated Circuits) and alternative foundry setups [43:30].

• The Shift: Major players are leaning away from general GPUs to develop custom internal accelerators (e.g., Amazon’s Trainium 3) [43:30, 43:46]. To win, startups can’t just try to make a “better GPU”; they have to choose unique trade-offs within the “iron triangle” of hardware design (optimizing tightly for speed, memory, or cost) and do something fundamentally difficult to replicate [44:14, 47:58].

• The Competitive Strategy: Baker highlights a major domestic foundry initiative (a joint venture involving SpaceX/Tesla, dubbed the “Terafab,” partnering with Intel) to establish bleeding-edge semiconductor manufacturing capacity in the U.S. [27:48].

• The Financial Impact: Rather than operating like traditional corporate entities, Baker states that hardware leadership requires an aggressive, human-centric recruiting approach to compete with Taiwan Semiconductor (TSMC) [28:45, 29:59]. This includes recruiting elite global engineering talent and building highly accommodating cultural enclaves specifically tailored to entice the absolute best hardware minds to move and work domestically [30:15, 30:28].

All content published on this newsletter is based on public information and independent research. Opinions are authors own and have been sanitized through AI engines which can make mistakes. This newsletter is not financial advice, and readers should always do their own research before investing in any security.

Which leads to the last of the major maneuvers and the focus on this article…

The Move Toward “Orbital Compute” (Racks in Space)

Read more