The rumour of an Anthropic Physical Intelligence acquisition spread across AI Twitter last weekend faster than either company could contain it, and for good reason: even after Physical Intelligence’s chief executive pushed back, reporting from The Information confirmed the two companies had genuinely held acquisition talks this spring.
The denial itself became part of the story. According to The Information, CEO Karol Hausman told employees via Slack that the reports were untrue, accompanying the message with a gif of a character from The Office shaking her head no. Lachy Groom, Physical Intelligence’s co-founder, did not respond to a request for comment. That is a long way from a firm, formal rebuttal.
What Physical Intelligence Actually Is
Physical Intelligence is not a speculative startup. Founded in San Francisco roughly two years ago by Groom, former Google researchers, and professors from Stanford and Berkeley, the company had about 80 employees as of January 2026 and had raised just over $1 billion in total by that point.
Its funding history charts a steep trajectory. A seed round in March 2024 raised $70 million at a $400 million post-money valuation, according to American Bazaar Online. By late 2024, the company had raised a further $400 million at a $2.4 billion valuation, backed by Jeff Bezos, OpenAI, Thrive Capital, Lux Capital, Khosla Ventures, and Sequoia Capital. Then came a $600 million round led by Alphabet’s CapitalG, lifting the valuation to $5.6 billion, with Index Ventures, T. Rowe Price, and others joining the cap table.
Now, according to Bloomberg, the company is in talks to raise a further $1 billion at a valuation exceeding $11 billion, with Lightspeed Venture Partners in discussions to invest alongside returning backers Founders Fund, Thrive Capital, and Lux Capital.
At the centre of its technology is the pi-0.5 model: a 3 billion parameter vision-language-action system that competes with offerings such as NVIDIA’s GR00T for real-world robotic navigation. Co-founder Sergey Levine, speaking to TechCrunch, framed the company’s ambition plainly: ‘Think of it like ChatGPT, but for robots.’ Groom has been equally unguarded about capital appetite, telling reporters: ‘There’s no limit to how much money we can really put to work.’
Why an Anthropic Physical Intelligence Acquisition Would Make Strategic Sense
Anthropic has not built anything resembling a hardware lab. What it has published is a series of research efforts through its internal group testing frontier model capabilities for safety. Project Fetch, run last November, examined how well Claude could help non-experts programme a robot dog. A second phase, completed in June, found a newer model completed equivalent tasks roughly 20 times faster than the best human-plus-Claude team from the prior exercise. Buying an established team would let Anthropic skip years of that work.
The broader pull toward robotics is not Anthropic’s invention. OpenAI built a robotic hand that solved a Rubik’s Cube, shut the entire robotics group in 2021, then quietly rebuilt a humanoid robotics lab in San Francisco before CEO Sam Altman made it official in late May, describing near-term infrastructure robots and a personal robot as a long-term goal. Co-founder Wojciech Zaremba later said the earlier approach was missing pieces needed for real superintelligence, which tells you something about how seriously the company takes the category now.
The robotics sector as a whole raised €38.6 billion in venture capital in 2025, representing approximately 9% of total global venture investment, according to aggregated data cited alongside the Physical Intelligence funding discussions. The capital is flowing in one direction.
The complication for any Anthropic deal sits directly on Physical Intelligence’s cap table. OpenAI is already an investor in Physical Intelligence. Founders Fund, also a major OpenAI backer, is reportedly involved in the current funding round. Strategic investors at this level routinely negotiate information rights and rights of first refusal over sales to competitors. Whether OpenAI holds either of those provisions is unknown; the company did not respond to questions on the matter. But the investor overlap means that if Physical Intelligence is genuinely in play, OpenAI may have structural levers that Anthropic does not.
Both Anthropic and OpenAI are also heading toward public markets. Anthropic confidentially filed its S-1 with the Securities and Exchange Commission (SEC) on 1 June, days after announcing a $65 billion fundraising round; Wired reports the company’s current valuation stands at $965 billion, though the company has not set an IPO price or target raise. OpenAI filed the following week. Anthropic has made four known acquisitions this year; OpenAI has completed at least 17 since 2023.
If Physical Intelligence does return to the table, the more pressing question may not be whether Anthropic wants it, but whether OpenAI would allow it to be sold to anyone else.
