Physical Intelligence Nears $1B Raise at $11B+ Valuation to Build General-Purpose Robot AI

San Francisco just keeps serving these quiet earthquakes, and Physical Intelligence is the latest tremor you feel before you understand it. Founded in 2024, led by CEO Karol Hausman, the company is now in talks to raise around $1B, with a valuation north of $11B. 4 months ago that number looked very different. Now it is doubling like it found a cheat code and decided not to tell anyone.

Karol Hausman is not pitching sci-fi. He is engineering it into something that folds your laundry and stacks your boxes without asking for applause. The π0 model is doing real-world work, not demo theater. Folding clothes, assembling boxes, handling the kind of tasks most software politely avoids. Turns out intelligence gets a lot more interesting when it has to touch something, when the edge cases are not bugs but the entire environment.

The capital circling this round reads like a table where nobody bluffs. Founders Fund and Lightspeed Venture Partners are in the mix, with Thrive Capital, Lux Capital, Jeff Bezos, NVIDIA, Index Ventures, and T. Rowe Price all leaning in again. When that many heavyweights stay close to the table, it is usually because they have already seen the cards, and more importantly, they understand how the game changes when software leaves the screen and starts moving atoms.

Here is the part people miss while chasing the headline number. This is not just another AI company scaling tokens and APIs. This is about teaching machines to deal with gravity, friction, and the chaos of the physical world. Software had it easy. Clean inputs, predictable outputs. Physical Intelligence is stepping into the mess where things slip, break, and refuse to behave, and still finding a way to make it look routine.

That valuation jump is not just momentum, it is signal. Enterprises are no longer asking if robots can adapt. They are asking how fast and how broadly. A general-purpose model that works across different machines and environments starts to look less like a feature and more like infrastructure. And infrastructure does not ask for attention. It becomes unavoidable.

There is also a lesson here for anyone building. The market is rewarding teams that collapse complexity into something usable. π0 is not trying to be everything at once, it is trying to be useful across enough real scenarios that it earns its place. That is how you move from interesting to inevitable, from demo to dependency. Congratulations to Karol Hausman and the Physical Intelligence team. And to the investors lining up again, because when the physical world becomes programmable, the upside is not just scale, it is gravity itself becoming part of the product.