Physical Intelligence Is in Talks to Raise $1 Billion at $20 Billion — With No Revenue at All

Physical Intelligence, the San Francisco startup building what it hopes becomes the general-purpose brain for robots, is in talks to raise at least $1 billion at a valuation of roughly $20 billion, Bloomberg reported on October 27, citing people familiar with the matter. Existing backers Avenir and Lux Capital are in talks to lead the round. The company did not respond to requests for comment, and the terms are not final.

If it closes near those numbers, the round would nearly double the $11.5 billion valuation the company reached earlier in 2026, and mark an eightfold climb from the $2.4 billion valuation at which it raised $400 million in November 2024. The investor roster behind that ascent includes OpenAI, Jeff Bezos, Thrive Capital, and Alphabet's independent growth fund CapitalG.

What makes the number unusual is what the company does not have: revenue. Bloomberg notes the company has no disclosed commercial revenue. Physical Intelligence does not sell robots. It trains robot foundation models — the pi family — that take camera images and language instructions and output actions across different robot arms and mobile platforms. It has released weights for pi0 and pi0.5 openly under an Apache 2.0 license, with base checkpoints pretrained on more than 10,000 hours of robot interaction data, and its demo reel includes folding laundry, bussing tables, and sorting trash from dishes without task-specific programming.

The commercialization channel opened earlier this month. On October 13, the company launched the Pi API beta, which lets outside developers upload their own robot data, run fine-tuning on compute the company manages, and control their robots through a hosted inference gateway. Access runs through a waitlist and pricing has not been announced. Early partners include Actor, whose autonomous wheel loader scoops and loads material in a real commercial deployment with no operator in the cab, and Almond, whose dual-arm Axol robot learned to deburr its own assembly parts after four days of iteration on the API. The company itself flags the limits of cloud-hosted robot control, including network connectivity challenges at some deployment sites.

The valuation lands in a sector awash with capital. Robotics startups have raised nearly $48 billion this year, and competitors are scaling fast: Skild AI reached a $14 billion valuation and passed $100 million in annual recurring revenue in September. Physical Intelligence is betting the more expensive position — that one model generalizes across robot bodies the way language models generalize across tasks — is worth more than early revenue.

The bet has a logic. Data, not hardware, is the bottleneck in robotics, and a company that aggregates demonstration data across industries builds a moat no single robot maker can replicate. The comparison investors keep reaching for is early OpenAI: no product, enormous ambition, and a valuation that only makes sense if the general model wins.

But a $20 billion price on a pre-revenue company is a pure convexity trade, and the round is still being negotiated. If the talks close, Physical Intelligence would rank among the most valuable private AI companies in the world without having sold anything — a wager that the physical world, not the text box, is where the next platform fortune gets made.