Debate Ignited on AI Twitter Over Controversy of Anthropic-Physical Intelligence

This year has witnessed a notable spike in AI acquisitions, making them nearly routine. Anthropic and OpenAI have been at the forefront, acquiring developer tools, AI service firms, and startups that focus on product testing to boost their model capabilities and expand their enterprise reach. In this context, the recent speculation surrounding Anthropic’s possible acquisition of the robotics startup Physical Intelligence has gained considerable attention. The rumor circulated widely, even after a denial from the CEO of Physical Intelligence.

The intrigue certainly lies in the involved parties. Physical Intelligence is not your ordinary robotics firm. Co-founded by Lachy Groom, a rising Silicon Valley innovator, it has amassed over $1 billion in funding (and was reportedly in talks for another $1 billion at an $11 billion valuation this spring); its π0.5 model is recognized as one of the most prominent robotic intelligences in academic circles.

Interestingly, the acquisition rumor had some basis. According to The Information, Anthropic and Physical Intelligence did indeed have discussions regarding a potential acquisition this spring. Although tech blogger Robert Scoble’s post on X may have inaccurately characterized the events, he was correct in stating that talks occurred.

It’s interesting to note that Physical Intelligence’s response to the speculation wasn’t vehemently dismissive. CEO Karol Hausman informed employees that the rumors were false through a Slack message featuring a gif of a character from “The Office” shaking her head, as reported by The Information.

For his part, Groom did not respond to TechCrunch’s request for comment sent on Monday evening.

This year, Anthropic has officially completed four acquisitions, while OpenAI has been even more proactive, securing at least 17 companies since 2023. Both firms are also preparing for an initial public offering. Anthropic confidentially filed for its IPO on June 1, followed by OpenAI a week later, potentially indicating two of the most significant stock debuts in U.S. history are on the horizon.

So, what’s driving interest in robotics now? The most logical explanation is that grasping the physical world could be crucial for the advancement of superintelligent systems—something that can’t be solely derived from internet text.

OpenAI’s experience in this realm is particularly revealing. Initially, it created a robotic hand capable of solving a Rubik’s Cube but disbanded its entire robotics division in 2021, with co-founder Wojciech Zaremba later indicating a lack of necessary components for achieving true superintelligence. In 2024, the team reformed, quietly launching a humanoid robotics lab in San Francisco. CEO Sam Altman publicly announced hiring for “OpenAI Robotics” in late May, focusing on infrastructure robots in the short term while aiming for a personal robot for everyone long-term.

Conversely, Anthropic hasn’t established anything akin to OpenAI’s hardware lab. Instead, it has published several research papers through its internal group, evaluating frontier capabilities for safety. This includes Project Fetch from last November, where the team assessed Claude’s ability to assist non-experts in programming a robot dog, revealing in June that a newer model performed the tasks around 20 times faster than the best human-plus-Claude team from the previous year.

Acquiring a team with robotics expertise would enable Anthropic to bypass years of development. However, a potential complication arises: Physical Intelligence was founded in San Francisco around two years ago by Groom and former Google researchers, along with professors from Stanford and Berkeley, and its early investors mirror those of OpenAI, including Khosla Ventures and Thrive Capital. The Founders Fund—also a significant investor in OpenAI—was reportedly involved in Physical Intelligence’s latest funding round earlier this year.

Significantly, OpenAI is an investor in Physical Intelligence, raising the stakes; it’s not just an external observer but a stakeholder in a company that its main competitor may have been considering acquiring.

This raises questions about whether OpenAI’s initial investment included any information rights or a right of first refusal regarding a sale to a competitor—precautionary measures that major strategic investors might negotiate for such scenarios.

This creates the possibility that if Physical Intelligence is genuinely on the market, OpenAI—already a shareholder, well-connected to Groom, and eager to surpass Anthropic in the robotics sector—might have a stronger claim than Anthropic.

We reached out to OpenAI earlier today with these inquiries but did not receive a response.