Physical Intelligence: Why a Stripe Veteran Thinks Robots Are Finally About to Get Smarter

Have you ever wondered why neural networks can already write code and compose sonnets, yet still can't properly fold a T-shirt or neatly unload a dishwasher? For years we have watched humanoid robots stumble awkwardly at trade shows, leaving audiences with a mix of pity and irony. The problem was never the hardware: servos and sensors reached the necessary level long ago. The trouble was the brains. Or rather, their absence. Every robot had to be programmed for one narrow task, and any step off script turned an expensive machine into a heap of scrap metal. But Lachy Groom, the man who helped turn Stripe into a financial giant, believes that dead end is now behind us.

The startup Physical Intelligence did not come out of nowhere. This is not just another attempt to build a "smart hand." It is an ambition to create what the industry calls foundation models, but for the physical world.

If GPT is a model that understands the structure of language, PI is building a model that understands the physics of movement. To do it, Groom has surrounded himself with true rock stars of academia: Sergey Levine of Berkeley, Chelsea Finn of Stanford, Karol Hausman of Google DeepMind. These people have spent decades trying to teach machines to learn from their own experience.

They used to lack computing power and data, but the success of large language models (LLMs) showed that if you feed an algorithm enough information, quantity inevitably turns into quality.

Why does this matter right now? We are at a turning point. For a long time, robotics lived in isolation from breakthroughs in AI.

Engineers built robots, programmers wrote control algorithms, and the two groups rarely understood each other. Lachy Groom and his team decided to tear down that wall. They approach robotics as a data-processing problem.

Their goal is to assemble the world's largest library of physical interactions. How does a hand grip a cup? How much force does it take to open a door?

How do you compensate for a slippery surface? All these small things we do without thinking are fiendishly complex equations for a robot. PI plans to solve them once and for all by creating universal "software" that can be loaded into any mechanism.

Skeptics will of course bring up Elon Musk and his Optimus, or the folks at Figure AI. But there is a fundamental difference. While competitors try to build the perfect body, Physical Intelligence focuses on the intelligence that controls that body. Groom is playing the long game: he understands that the market for "robot brains" is potentially much larger than the market for the robots themselves. It is like an operating system: whoever captures this standard will control the entire automation industry in the next decade. And judging by how readily Silicon Valley investors are writing checks to this startup, they believe Groom's Stripe-honed grip can turn theoretical science into commercial triumph.

What does all this mean for us? Most likely, we will see the end of the era of specialized robots. We won't need separate machines for the warehouse, the kitchen or the hospital.

A class of devices will appear that learns on the fly, simply by watching what people do or by receiving instructions in natural language. It sounds like science fiction, but the foundation is being laid today. Groom is not just building a company; he is trying to prove that the physical world is as pliable for algorithms as the world of numbers and text.

If he is right, in five years we will look back on today's "dumb" robots with the same smile with which we now look at rotary phones.

The bottom line: Physical Intelligence is betting that robotics lacked not engineers but scalable intelligence. Can PI's "brain" become the Windows of the robot world?