Anthropic-Physical Intelligence Rumor Fuels AI Speculation

AI companies have been busy buying startups this year. Large acquisitions have become so common that many barely make headlines anymore.

Both Anthropic and OpenAI have expanded through acquisitions. They have bought developer tools, AI service companies, and product-testing startups. The goal is simple. Turn powerful AI models into profitable business products and move faster than competitors.

That is why a new rumor caught so much attention over the weekend.

Reports claimed Anthropic was planning to acquire robotics startup Physical Intelligence. The story spread quickly across AI Twitter. It continued circulating even after Physical Intelligence denied the claim.

Part of the excitement came from the company involved.

Physical Intelligence is one of the best-known robotics startups in the industry. It was co-founded by investor and entrepreneur Lachy Groom, whose influence in Silicon Valley has grown rapidly.

The company has already raised more than $1 billion. Earlier this year, reports suggested it was also seeking another $1 billion in funding at an $11 billion valuation.

Its AI model, called π0.5, has also become one of the most widely used foundation models in robotics research.

As it turns out, the rumor was not completely false.

According to The Information, Anthropic and Physical Intelligence did hold acquisition discussions earlier this year. The talks reportedly took place during the spring.

The original rumor was shared by tech blogger Robert Scoble on X. While some details may have been inaccurate, it appears there was at least some truth behind the speculation.

Physical Intelligence did respond to the reports, although the denial was brief.

According to The Information, CEO Karol Hausman addressed employees through Slack. His message reportedly included a GIF from The Office showing a character shaking her head “no.” That served as the company’s response to the acquisition rumors.

Lachy Groom did not publicly comment. He also did not respond to requests for comment. Anthropic has completed four known acquisitions this year. OpenAI has been even more active.

Since 2023, OpenAI has acquired at least 17 companies. Both firms are also preparing to enter the public markets.

Anthropic confidentially filed for an initial public offering on June 1. OpenAI reportedly followed with its own confidential IPO filing one week later. If both offerings move forward, they could become two of the largest stock market debuts in U.S. history.

The rumors also raised another question.

Why would Anthropic suddenly be interested in robotics?

Many researchers believe physical-world experience is important for building more advanced AI systems. Training on internet data alone may not be enough. Robots can interact with the real world and collect information that text cannot provide.

OpenAI has already explored that path.

Several years ago, the company developed a robotic hand capable of solving a Rubik’s Cube. It later shut down its robotics division in 2021. Co-founder Wojciech Zaremba later said the project was missing important pieces needed for achieving superintelligence.

The company changed course in 2024.

OpenAI quietly restarted its robotics research by building a humanoid robotics lab in San Francisco. In May, CEO Sam Altman officially confirmed the effort. He announced that OpenAI Robotics was hiring new talent.

Altman said the company plans to focus first on robots for infrastructure work. Its long-term vision is much broader. He believes personal robots could eventually become part of everyday life.

Anthropic has taken a different approach.

Instead of building robots, the company has focused on AI safety research. Its internal teams regularly test how advanced AI models perform in real-world situations.

One example is Project Fetch, introduced last November.

Researchers studied whether Claude could help people with little robotics experience program a robot dog.

Anthropic later expanded the project in June. The company said its newer AI model completed the same robotics tasks about 20 times faster than the best human-and-Claude team from the previous year.

Buying an experienced robotics company could save Anthropic years of development work.

However, the situation may not be so simple.

Physical Intelligence was founded about two years ago in San Francisco. Its founders include Lachy Groom, former Google researchers, and professors from Stanford and UC Berkeley.

Its list of investors also overlaps with OpenAI.

Major backers include Khosla Ventures, Thrive Capital, and Founders Fund. All three have also invested in OpenAI. More importantly, OpenAI itself is an investor in Physical Intelligence.

That makes OpenAI more than just a competitor. It is also a shareholder.

Because of that investment, some observers wonder whether OpenAI negotiated special rights. Strategic investors sometimes receive access to company information or even the first opportunity to purchase a business before it is sold to a rival.

If Physical Intelligence ever becomes available for acquisition, OpenAI could already have an advantage.

The company has an existing relationship with the startup. It is also investing heavily in robotics and competing directly with Anthropic.