Inside The European Factory Where Robots Are Building Robots

Chen’s two viral loops to make humanoid robots real

Ask most humanoid CEOs why they think they’ll win and you’ll hear about data, or models, or capital. Chen’s answer is structural, and it’s different from what I’ve heard from robot company CEOs in the U.S. and China.

Large language models, he says, continually improve because the growth engine is understood, mapped out and fundamentally simple. It’s easier, because it’s entirely digital, and it rests on a foundation of internet-scale text. “You just create the whole channel — model, data, compute and outcome,” he told me. “All you need to do is basically pump in money. So if you’re pumping money, then the real results will come out. And this flywheel is working really fine.”

Physical AI is not that simple.

“You really have to create the model, the body, physical part, sensory system, whatever, and then application. And only when application and the valuable applications are massively deployed, then this will come back, drive capital, drive the compute, and drive the body and drive the application again.”

In other words, you’re dealing with the real world, the physical world, and only by doing real and valuable work in this real physical world can you drive economic value, which you then inject back into the physical AI growth flywheel.

Which leads to his thesis.

“For me, the wheel is two levels, two layers, let’s say. Layer one is technical. Layer two is also business,” Chen said. “If you cannot close that, it doesn’t matter how much capital you have.”

That’s an interesting comment in September 2026, as the amount of capital flowing into robotics and physical AI continues to grow. Humanoid robot companies raised $2.7 billion in August alone, most of it in China. In the United States, Figure has committed $3.5 billion for AI compute to drive its physical AI. In Germany Neura Robotics, an hour up the autobahn from Agile Robots in Metzingen, raised up to $1.4 billion in June, one of the largest humanoid robot rounds on record.

But Chen says Agile Robots has raised roughly $1.5 billion over seven years, is backed by SoftBank, and isn’t worried.

“Competition is always good. I’m trained in a way to compete,” he said. He’ll concede the Americans have an edge: “US companies, they have certain … it’s very unique, unfair sort of advantage. That’s something I have to admit.” But he won’t concede the win based on cash alone. “I don’t think any business will be sustainable by only pumping in money.”

Nobody’s factory will have 10,000 humanoids

Here’s where Agile Robots diverges most sharply from the pure-play humanoid startups: the humanoid isn’t the product. In fact, any individual robot is not the product. Rather, the factory is the product.

Specifically, the factory of the future.

“The goal is to solve the problem, and the second thing is to solve it with the least amount of automation as possible to give them the best ROI,” Rory Sexton, Agile Robots’ executive director, told me during a day at the company’s Munich facilities. “Sometimes you just need a smart robot, not a very complex robot.”

Sexton spent 20 years at Apple. He joined the day the iPod launched, and by the time he left he was responsible for 120 plants with 2.5 million workers. He came out of retirement to take on the challenge of reinventing how we make stuff. And his view of where humanoids actually sit is refreshingly unromantic.

“Humanoids, I think for many years are not going to operate independently,” he said. “They’ll be part of a set of solutions that go in.”

That’s why Agile Robots builds the whole stack: robotic arms, dextrous hands, autonomous mobile cargo robots, conveyors, vision, the AgileCore software layer that uses AI to help companies plan and deliver full-suite automation solutions, and a portfolio of acquired integrators including idealworks, BÄR Automation, Franka Robotics, India’s XNG, and thyssenkrupp’s former automation engineering arm, now rebranded Krause Automation, which added roughly 650 people and ten sites across Europe and North America.

“You’ve got to look at the problem and then go: what are the parts of physical AI and normal automation?” Sexton said. “You’ve got to merge those together to bring one solution.”

The factory, therefore, is the machine. Components within it can include humanoid robots and many other kinds of robots.

Practically, that also solves an integration problem the standalone humanoid companies in large part still have to answer. Agile Robots’ automation suite is already wired into customers’ manufacturing execution systems. “Putting a semi-humanoid on top of it keeps it integrated,” Sexton said. “All the route planning, the task planning, it’s just another part of the system.”

What the factory floor actually says

I was able to tour the factory floors. Not all the workers building robots are robots, but at least two lines that build the arms and the torsos are entirely automated. So is testing.

Agile Robots will ship thousands of robot systems out of its German operations this year, Sexton told me, and believes it could scale to 8,000 to 9,000 a year, hitting full capacity next year. There’s a fair bit of work to do to hit those ambitious estimates: current throughput runs about 10 units a day, at three to four days per robot. Some customer orders come in increments of hundreds; others are smaller. Hardware upgrades on new robots are fast: there’s a roughly two-month upgrade cycle.

Humans still do the most challenging parts of the work: hands and final assembly are done by human technicians.

The company doesn’t release pricing. But Sexton offered the yardstick customers use: two shifts of laborers in China cost about $80,000 a year, and industrial buyers typically want a two-year payback. Do that math and the price band for a viable humanoid becomes fairly obvious, assuming it is capable enough and fast enough to do the job.

Interestingly in a time of increased tariffs and protectionism, especially in the United States, manufacturing is going to be regional. Agile Robots will build robots in Asia and America as well as Germany. But the German supply chain for critical components remains, in his words, extra important. “A lot of people want to get into this supply chain,” he said.

Also: the company is not locking it down to any one supplier, nor trying to build everything in-house. Agile Robots plans multiple battery suppliers and multiple motor-actuator suppliers.

And some of those plants won’t have lights on. “We don’t need people to build arms or torsos,” Sexton said. “We will have some dark factories.”

Demand for automation is peaking right now, he added. Agile has a Franka Robotics plant an hour from Munich that’s selling, in his words, like hotcakes. To add capacity he needs another 40 to 50 trained workers, and that will take many months, making robots and automation a tempting solution.

“You don’t need higher-order math to see that the demand is there,” Sexton says.

The solution to limited workforces: robots as a service

Ask Sexton what humanoids and other robots uniquely unlock and he lands on flexibility. And a robots-as-a-service delivery model.

“You could do a leasing model,” he said. “I’ve got Christmas coming. I want another 50 of them, please, and you take them back come December.”

That only works, of course, if the machines are genuinely general: if the physical AI they embody is smart enough to do many tasks, and if the body carrying that brain is capable enough as well. A fleet leased out for holiday fulfillment for a logistics facility in Q4 has to do something else in Q1 and Q2.

“You need flexible robots that can do multiple things,” Sexton said. “The more flexible the robot, the more useful it is. And hence the humanoid discussion.”

He thinks the endpoint for the factory of the future is a category shift: “Manufacturing as a service, or service as a service, is coming.”

But there’s a reason why Agile Robots’ currently-shipping humanoid robots are wheeled, not legged.

“There is no regulatory environment in the West, Europe and North America, for legged, bipedal humanoids working with humans,” he said. “There’s no standards for this yet.” The company’s wheeled platforms are already certified. And, interestingly, they’re very capable at stairs, despite having no legs, the company told me.

The United Nations approach: hybrid vigor

Both executives kept returning to the same idea, from different directions: that one of the company’s key structural advantages is that it isn’t from anywhere in particular.

“In terms of the company makeup, we’re sort of like the United Nations,” Sexton said. Agile Robots is German, Indian, European and Chinese in roughly equal measure, with more than 60 nationalities represented.

“We don’t have a German culture, per se. We have an Agile culture, and that’s very important. The Chinese are unbelievably fast at supply chain and the art of the impossible. Germany is very, very good at design and deep mechatronic knowledge. On the AI side, Germany is very good, but clearly California and America, that’s very strong. If you’re going to build a successful company in this, you’ve got to take it all.”

Chen frames it as Europe’s second chance.

He’s blunt that the continent lost round one in AI: Europe and Germany, he said, got lost in the last foundation model competition. But he thinks the physical round is different. “Europe stands a chance for the physical AI, because it is not only software. It is heavily, heavily software and hardware in combination.”

He also thinks the regulatory and tariff walls going up around this technology are entirely predictable, and temporary. Agile Robots didn’t go multinational to hedge geopolitical risk, he said; it was built that way from day one.

But he’s watched this movie. “If you look at the history of the last 100 years, this is happening and happening again and again and again for every new technology.” China once made Volkswagen take a joint venture partner. Everyone protects, then everyone exports. “This is never-ending. It’s just history.”

Which brings it back to Asimov, and to the other writer Chen invoked during our conversation: the pioneering French science fiction author Jules Verne, who described the submarine roughly a century before anyone built one. There’s a line from Verne posted at the Paris airport that Chen quotes as his operating principle.

“Whatever mind can imagine, mind will realize it.”

Eight billion personal robots is a lot to imagine. A personal robot for everyone on the planet is also challenging, not least due to cost and socio-economic differences.

Agile Robots’ bet is that you don’t get there by imagining harder.

You get there by peeling protective film off smartphones – yes, the company demoed a robot that does just that — one production line at a time, and adding tasks and capability and potential, until Chen’s twin loops close, and the technology and the economics all just start to make sense.