The White House is drafting new guidance that could ease restrictions on Anthropic, signalling a reversal from its earlier stance which labelled the AI giant as a supply chain risk. According to a report by Axios, sources familiar with the matter said that the administration is exploring ways to enable the federal agencies to onboard Anthropic’s most advanced model, Mythos, despite ongoing disputes with the Pentagon. This rethink in the stance of White House is followed by US President Donald Trump’s recent comments that Anthropic was ‘shaping up’ in the eyes of the administration. Officials are now considering executive action that would provide agencies flexibility to work with the company, even as the Pentagon remains entrenched in its legal battle over Anthropic’s exclusion. One source described the effort as a way to “save face and being them back in”.
The meeting of White House and Anthropic
Earlier this month, White House Chief of Staff Susie Wiles and Treasury Secretary Scott Bessent met with Anthropic CEO Dario Amodei in what both sides called a productive introductory meeting. The administration is also convening companies across sectors this week to discuss best practices for deploying Mythos, including “table reads” of draft guidance that could soften the Office of Management and Budget’s directive against using Anthropic.In a statement, the White House said it continues to engage with government and industry to protect the country and its people, adding: “Any policy announcement will come directly from the President and anything else is pure speculation.” Anthropic declined to comment.
Pentagon and White House were once aligned in blacklisting Anthropic
The Pentagon and White House were once aligned in blacklisting Anthropic, denouncing the company as unreliable. But Mythos, which has demonstrated both the ability to automate cyberattacks and defend against them, has sparked renewed interest across agencies. The NSA is already using Mythos, while the Pentagon remains barred from accessing the latest updates due to contractual disputes.Anthropic has refused to sign agreements allowing its models to be used for “all lawful purposes,” drawing red lines against mass domestic surveillance and fully autonomous weapons. The Pentagon argued this refusal showed the company was not a reliable partner, leading to its unprecedented supply chain risk designation.
What’s next
It remains unclear whether the White House’s potential executive action would resolve the Pentagon’s fight or simply make it easier for other agencies to work with Anthropic. Multiple sources told Axios that while Pentagon officials remain firm, other stakeholders see the feud as counterproductive and want an offramp.Even if restrictions are lifted, contentious negotiations could resume. Meanwhile, rivals OpenAI and Google have already signed agreements with the Pentagon under the “all lawful purposes” standard, positioning themselves as more flexible partners in classified settings.
