On this week’s Bricks, Bucks & Bytes rundown, Foundamental’s Patric Hellermann put a number and a date on the future of physical automation. The West spends roughly $300 billion a year on industrial automation by his estimate, and 75% of it has to be deployed through projects – configured, wired, programmed, and commissioned by contractors. Hardware and software are solved, he argues. Deployment capacity is the bottleneck, and almost nobody is funding it. His prediction: the “Palantir moment,” when hardware investors realize this, arrives within 15 months.
No guest this week, which usually means the sharpest takes. Martin asked Patric Hellermann for his thesis on physical AI, and what came back was less a talking point and more a set of core convictions the Foundamental general partner has clearly been sitting on for a while.
He opened with a quiz. What does the West spend each year on industrial automation, excluding China? Martin guessed $200 billion for the whole world. Dustin DeVan reached for a trillion. Patric’s answer: about $300 billion for the West alone, a figure he says comes from years of triangulation because nobody publishes it directly. Treat it as his informed estimate rather than an audited number – he was upfront about that on air.
What he built on top of that number is the interesting part, because it cuts against where most robotics capital is flowing right now.
Automation Is a Contractor Business
Three quarters of the spend goes through projects, and that changes who matters
Here is the split Patric laid out. Of that $300 billion, only about a quarter is ship, plug and play automation. Think pool cleaning robots, lawn mowers, floor cleaners – kit that arrives in a container, gets wheeled off a flatbed, and works after minimal setup.
The other 75%, call it $230 billion, is project deployment. Someone has to configure the systems, run simulations, do the logic and motion programming, add the cages, pull the wiring. That work looks a lot more like a construction trade than a product sale, which is exactly his point.
It is a contractor business, actually. So if we want to automate the West, we have to go through that.Patric Hellermann, General Partner, Foundamental
For anyone running an electrical, mechanical, or industrial services firm, this reframing matters. In Patric’s read, the firms that deploy automation are the constraint on the entire market, and constraints tend to have pricing power.
The Hardware Race Is Already Over
Asia won components and subsystems, and the installation data backs him up
Patric’s second conviction is blunter: the hardware game is won, and Asia won it. Actuators, motors, harmonic drives, and the subsystems built from them – arms, cobots, CNC machines. Outside of high-tech niches like grippers and optical equipment, or a genuine sovereignty argument, he does not understand why Western venture capital keeps pouring into these systems.
The installation numbers give his claim some spine. According to the International Federation of Robotics, 542,076 industrial robots were installed globally in 2024, more than double the total a decade ago. China alone accounted for 54% of deployments, and for the first time Chinese manufacturers outsold foreign suppliers in their home market. Patric quoted the figure from memory on air, about 540,000 a year with China taking more than half, and he was almost exactly right.
Dustin pushed the point further with a line that will annoy a lot of cap tables: “Humanoids are gonna be the last robotics that we actually deploy.” That squares with what we found in our breakdown of the 2026 construction robotics report, where narrow, high-utilization machines are winning real deployments while general-purpose platforms stay parked in the corner.
The automation shift, decoded weekly
Deals, deployments, and honest takes on the technology reshaping construction and industry.
Join 3000+ ReadersThe Palantir Moment
When the hardware money realizes it needs forward deployment
So if hardware is unlocked and software is unlocked, what stops the West from moving its automation spend from $300 billion toward the $1 trillion a year Patric believes is needed to sustain production and sovereignty? His answer: nothing technical. The missing layer is capitalizing and servicing the contractors who deploy this stuff. Financing so they can run more projects in parallel. Insurance products built for automation work. Support infrastructure that lets a mid-sized integrator take on triple the volume.
That gap is what sets up his prediction. “We’re 15 months away from the Palantir moment in physical automation,” he said, defining it as the point where everyone who put money into hardware realizes hardware was never the bottleneck, and goes looking for whoever can forward-deploy physical automation “in rural Alabama and rural Germany.” Palantir spent years being mocked for its forward-deployed engineer model before the market decided that deployment capability was the product. Patric expects the same reversal here.
The WIIFM is different depending on your seat. If you run an industrial or electrical contracting business, the deployment capability you already have is about to become the scarce asset in a much bigger market. If you’re a ConTech or robotics founder, the open lane is the capital, servicing, and insurance rails for the people doing the installs, rather than yet another arm or humanoid. Deals like All3’s full-stack robotics bet show investors circling the space, but by Patric’s logic most of them are still aiming at the wrong layer.
| Layer | State of play | Hellermann’s verdict |
|---|---|---|
| Hardware and components | Asia dominates supply of actuators, motors, arms, and subsystems | Already won, stop funding it |
| Software and controls | Mature, capable, widely available | Unlocked, not a constraint |
| Project deployment | Fragmented contractor and integrator market doing 75% of the work | The real bottleneck |
| Capital and servicing for deployers | Barely any dedicated financing, insurance, or scaling support | The opportunity nobody is seeing |
It’s Patric Hellermann’s term for the point when investors who funded robotics hardware realize hardware was never the bottleneck, and start hunting for companies with deployment capability instead. The analogy is Palantir, whose forward-deployed engineering model was dismissed for years before the market repriced it as the core asset. He puts the moment 15 months out.
It’s Hellermann’s own triangulated estimate of annual Western spend on industrial automation, which he acknowledged on air can’t be sourced from a single published dataset. The adjacent hard number is the IFR’s count of 542,076 industrial robots installed globally in 2024, with China taking 54% of deployments. (Source)
Because deploying automation is contractor work: configuration, wiring, cages, commissioning. If Western automation spend grows from $300 billion toward $1 trillion a year, the firms that install and service it capture a large share of that growth. Contractors with automation deployment capability would move from the supporting cast to the constraint on the whole market, which is a strong negotiating position.
Not in a good way. Dustin DeVan’s view on the episode was that humanoids will be the last robots actually deployed at scale, and industry research points the same direction: specialized, narrow machines are winning real jobsite and factory deployments while general-purpose platforms remain experimental. (Source)
Related Articles
Construction Robotics in 2026: 4 Workflows Where Robots Are Actually Working
Uber’s Founder Just Bought an Autonomous Mining Company
All3 Bags $25m to Roboticise Everything Above the Foundation
