Rumors that Anthropic might acquire robotics startup Physical Intelligence have hardened into reports of “actual” and “substantive” acquisition talks held in spring 2026, even though neither company has announced a deal as of mid‑July 2026. Those talks, first detailed by The Information and relayed by TechCrunch, suggested engagement beyond casual partnership exploration, placing the discussions within a broader push by frontier AI labs to secure footholds in robotics and embodied intelligence. Beyond these talks, both Anthropic and OpenAI have pursued an aggressive slate of AI acquisitions in 2026, with Anthropic reportedly closing four deals and OpenAI at least seventeen, highlighting a competitive race for sector leadership. Yet company newsrooms and investor disclosures remain silent on any transaction, leaving the status of the relationship unresolved.
Serious spring talks left Anthropic’s Physical Intelligence ambitions stalled in a public, unconfirmed limbo
The public cycle began when tech blogger Robert Scoble posted on X that an unnamed, “well‑known” investor had told him Anthropic was buying Physical Intelligence and that the deal had not yet closed. The claim offered no price, structure, or timeline and rested entirely on this single investor account, but was quickly amplified by AI Twitter and robotics commentators who framed it as evidence of Anthropic’s sudden move into physical intelligence.
Subsequent coverage traced the rumor back to Scoble’s post while emphasizing that, at best, it captured only part of a more complex situation. Reports from The Information affirmed that Anthropic and Physical Intelligence had indeed engaged in acquisition discussions earlier in the year, describing them as serious rather than speculative, but stopped short of indicating that terms were agreed or a signing was imminent.
TechCrunch characterized the talks as “actual” and “substantive,” underscoring that they went beyond routine business development, even as it noted that no definitive deal had emerged. Runtimewire’s analysis highlighted how the strongest public claim—that Anthropic was actively pursuing an acquisition—depended on Scoble’s single unnamed source and lacked corroborating transaction details or documentation.
Inside Physical Intelligence, CEO Karol Hausman responded to the swelling speculation with a brief Slack message that used a looping GIF of a character from “The Office” shaking her head no, signaling rejection of reports that an acquisition was imminent or already completed. The message did not address the earlier talks described by The Information and offered no detail on strategic alternatives, fundraising plans, or partnership options.
It functioned as a narrow denial of transaction rumors rather than a transparent account of the company’s broader relationship with Anthropic and its investors. Physical Intelligence develops robot foundation models, including its π0.5 model, which has become a widely used tool in robotics research and offers a control layer complementary to large language models such as Anthropic’s Claude. This evolving landscape underscores the urgency for continuous safety evaluation of AI systems as they integrate into physical domains.
Industry observers argued that acquiring such a platform could give Anthropic a ready‑made “robot‑control layer” and accelerate efforts to move from purely digital assistants toward embodied systems capable of operating real‑world machines. The rumored deal was consequently framed as a shortcut into physical intelligence at a moment when robotics is becoming a central battleground for leading AI labs.
For now, the outcome remains uncertain.
