The Anthropic-Physical Intelligence rumor roiling AI Twitter

AI acquisitions have gotten so routine this year that most of them barely make news anymore. Anthropic and OpenAI have both been on shopping sprees, scooping up developer tools and enterprise-services startups to turn model horsepower into revenue faster than the other guy. Which is exactly why a weekend rumor about Anthropic acquiring Physical Intelligence, the robotics startup, blew up the way it did, spreading fast even after a denial from the company's CEO.

The denial itself was less than convincing. Per The Information, Physical Intelligence CEO Karol Hausman told employees the reports were false via a Slack message featuring a GIF of a character from The Office shaking her head. Meanwhile, The Information also reported that Anthropic and Physical Intelligence did in fact hold acquisition talks this spring, meaning blogger Robert Scoble, whose X post kicked off the frenzy, may have botched the details while still being onto something real. Physical Intelligence isn't a nobody, either — co-founded by investor Lachy Groom, backed by more than a billion dollars, reportedly in talks this spring for another billion at an $11 billion valuation, and its π0.5 model is widely used in robotics research.

The bigger question is why any of this matters to a company like Anthropic, which has zero hardware history to speak of. The answer probably traces back to a belief that's spreading across the frontier labs: text alone can't get you to superintelligence, and physical-world understanding is a missing ingredient. OpenAI learned this the hard way, building a robotic hand that solved a Rubik's Cube back in the day, then shutting its robotics team down entirely in 2021 — co-founder Wojciech Zaremba later admitted the approach lacked what real superintelligence would require. That team quietly came back in 2024, and Sam Altman made