Speculation Mounts Over Anthropic’s Potential Acquisition of Physical Intelligence

The AI sector has seen a surge in acquisitions lately, with major players like Anthropic and OpenAI aggressively acquiring developer tools and AI service startups to bolster their enterprise offerings. This trend recently sparked intense speculation regarding a potential acquisition of the robotics startup Physical Intelligence by Anthropic—a rumor that gained significant traction despite denials from Physical Intelligence’s leadership.

Physical Intelligence is a high-profile entity in the robotics space, co-founded by notable investor-operator Lachy Groom. The company has reportedly raised over $1 billion, with recent discussions hinting at an $11 billion valuation. Furthermore, its π0.5 model has become a prominent foundational model for robotics research.

Reports from The Information suggest the rumors were rooted in reality, noting that Anthropic and Physical Intelligence engaged in acquisition discussions earlier this spring. While tech commentator Robert Scoble may have lacked specific details, his reports correctly identified that meaningful negotiations had occurred.

Physical Intelligence has attempted to downplay the rumors; reports indicate that CEO Karol Hausman used a “The Office” GIF to inform employees via Slack that the acquisition reports were inaccurate. Lachy Groom has not commented on these reports.

While Anthropic has completed four acquisitions this year, OpenAI has been even more prolific, acquiring at least 17 companies since 2023. Both companies are also moving toward public markets, with both Anthropic and OpenAI having confidentially filed for IPOs this June—a move that could signal two of the largest stock market debuts in history.

The interest in robotics suggests that mastering physical-world interactions may be essential for achieving true superintelligence, a capability that cannot be gained through text-based training alone.

OpenAI’s history with robotics highlights this importance. After shuttering its initial robotics group in 2021, the company re-entered the field in 2024 by establishing a dedicated robotics lab in San Francisco. CEO Sam Altman has expressed a long-term vision of integrating robotics into daily life, starting with infrastructure-focused automation.

Anthropic, conversely, has focused on research regarding safety and frontier capabilities rather than hardware. Through initiatives like Project Fetch, Anthropic has demonstrated how its Claude models can assist in programming robotic systems, showing significant speed improvements in task completion over previous iterations.

Acquiring an established robotics team could allow Anthropic to bypass years of foundational research. However, complexities arise from Physical Intelligence’s investor base. The company was founded by prominent Silicon Valley figures and has backing from major investors like Khosla Ventures and Thrive Capital—the same group that supports OpenAI. Additionally, Founders Fund, a key OpenAI investor, was reportedly involved in Physical Intelligence’s recent funding.

Because OpenAI is already an investor in Physical Intelligence, the company is positioned not just as a bystander, but as a stakeholder in a firm its primary rival is interested in. This raises strategic questions regarding whether OpenAI’s investment includes rights of first refusal or other protective provisions that might block a sale to a competitor.

If Physical Intelligence is indeed a potential acquisition target, OpenAI’s existing ties to the company and its leadership may give it a stronger strategic claim than Anthropic. OpenAI has not responded to inquiries regarding these potential rights or interests.

Source link