Unconfirmed speculation emerged over the weekend that Anthropic may be exploring an acquisition of Physical Intelligence, a robotics startup focused on training AI models for physical tasks1. Neither company has commented on the rumors.
The speculation, which circulated widely on social media, has sparked debate about the next phase of AI industry consolidation. Physical Intelligence has been working on what it describes as "foundation models for robotics" — AI systems designed to learn to manipulate objects, navigate spaces, and perform complex physical tasks. The startup has raised funding from top-tier Silicon Valley venture capital firms, according to The Tech Buzz.
Industry analysts cited by The Tech Buzz suggest a deal for Physical Intelligence could be valued anywhere from several hundred million to over a billion dollars. Anthropic CEO Dario Amodei has been described as particularly interested in how AI systems interact with the physical world.
The rumor arrives amid a broader pattern of acquisition activity by foundation model companies. Both Anthropic and OpenAI have been acquiring talent, technology, and specialized AI companies throughout 2026, according to sources cited by The Tech Buzz. The report characterizes the foundation model companies as racing to build comprehensive AI ecosystems that extend beyond text and image generation into the physical world. OpenAI has separately been investing in robotics partnerships and acquisitions throughout 2026.
ANALYSIS The rumor is unconfirmed and sourced entirely from social media speculation and a single outlet's reporting, with no on-the-record statements from either Anthropic or Physical Intelligence. Readers should treat the acquisition claim with appropriate caution until either party provides confirmation or denial.
If substantiated, a move by Anthropic into robotics would represent a significant expansion beyond its core large language model business into embodied AI — a domain where Physical Intelligence's foundation-model approach to robotic manipulation could complement Anthropic's existing capabilities in language and reasoning.