Physical Intelligence raises $1B, one of the largest late-stage computer vision rounds ever
What's the deal? Physical Intelligence, a robotics startup building general-purpose AI for robots, raised $1 billion in a late-stage round backed by OpenAI Startup FundDealroom has a profile for this one. Try Dealroom →, Sequoia CapitalDealroom has a profile for this one. Try Dealroom →, and Thrive Capital. The raise brings the company's total funding above $2 billion.
By the numbers: The round ranks among the largest late-stage venture deals ever recorded in US computer vision — in the top 1% by size across 154 comparable rounds.
What's the endgame? Founded in 2024, Physical Intelligence is developing AI that lets robots apply learned skills to new physical tasks. In April, the company said its latest model could fold laundry without task-specific training data.
Why now? The capital is fuelling rapid physical expansion. Physical Intelligence has agreed to sublease 80,000 square feet from Airbnb at 850 Brannan Street in San Francisco, adding to a 60,000-square-foot lease nearby and a 233,500-square-foot space in Mountain View. Its plans for the new offices have not been disclosed.
The bigger picture: Venture investors poured more than $12 billion into robotics startups last year, as so-called physical AI companies move from research toward commercial applications. Physical Intelligence is among a cluster of well-funded firms scaling up office footprints across San Francisco.
That demand is reshaping the local market. Citywide office vacancy fell to 28% in the third quarter, down from a 36.9% peak in late 2024, according to preliminary CBREDealroom has a profile for this one. Try Dealroom → data. The city is on track for its strongest annual net absorption on record, at 4.3 million square feet so far this year.
The signal: Money is now chasing robots that work in the real world, not just software that runs on screens. Physical Intelligence's billion-dollar raise and its aggressive land grab show how quickly physical AI is moving from the lab toward deployment.
Read more: therealdeal.com
Image credit: uwnews
