Anthropic and Trump administration ease tensions as IPO nears

Anthropic’s tense standoff with the Trump administration has started to ease just as the company moves toward Wall Street, a shift that could shape how investors judge regulatory risk for the San Francisco AI giant and its rivals.

The company confidentially submitted a draft S-1 to the SEC on June 1 for a proposed initial public offering, one day after it disclosed a $65 billion Series H funding round at a $965 billion post-money valuation. Those numbers put Anthropic among the most closely watched AI companies in the market, where policy risk can move as fast as product launches.

The dispute with Washington began earlier this year after Anthropic refused to allow the U.S. military to use its models for domestic surveillance and fully autonomous weapons systems. In March, the Pentagon designated Anthropic a supply-chain risk, the first U.S. company to receive that label, a move that bars tens of thousands of contractors from using Anthropic’s AI in military work. Anthropic said it received a letter on March 4 confirming the designation and said it would challenge the action in court. The Pentagon is still vigorously defending its position, leaving the core legal fight unresolved.

Even so, the temperature has dropped. Relations improved after Dario Amodei visited the White House in mid-April, and officials later invited him to a planned May 21 signing of an executive order on artificial intelligence before that event was canceled. Anthropic later said it looked forward to collaborating with the White House on implementing the order. The company has also held discussions with National Cyber Director Sean Cairncross about its Mythos model and ways to protect critical infrastructure from AI-enabled cyberattacks.

That conversation matters because Anthropic launched Claude Mythos Preview on April 7 and has pitched it as a general-purpose model with cybersecurity applications. Through Project Glasswing, Anthropic says access to Claude Mythos Preview helps identify and mitigate software vulnerabilities, underscoring how frontier models are becoming central to cyber defense even as they raise fears about misuse. POLITICO reported on April 28 that Cairncross was expected to chair a White House meeting with tech and cyber firms on Mythos-related concerns.

President Trump’s June 2 executive order on advancing AI innovation and security sharpened that picture. It explicitly emphasizes strengthening cybersecurity and protecting critical infrastructure, signaling that the White House wants advanced models available for defense while still keeping pressure on companies it has already penalized. For Anthropic, that means the path to an IPO is not just about growth and valuation. It is also a test of how much regulatory risk public investors are willing to price in, and whether rivals can expect the same mix of punishment, negotiation and selective cooperation from Washington.