The White House is preparing draft guidance that could allow federal agencies to move past earlier restrictions on Anthropic’s artificial intelligence systems and begin adopting newer models, including its most advanced system, Mythos, according to people familiar with the discussions.

The proposed changes would mark a significant adjustment in federal policy after months of tension between the company and U.S. defense officials over security designations and usage terms.

The draft executive action under review would be part of a broader federal AI framework and could give agencies more discretion to determine whether Anthropic’s systems meet procurement and security requirements, according to Axios. One source cited by the outlet described the effort as a practical attempt to “save face and bring them back in,” reflecting growing pressure within parts of government to restore cooperation with major AI developers.

White House officials have already begun convening technology companies and government stakeholders this week to shape the potential order, including discussions focused on deployment standards for frontier models like Mythos. Those sessions reportedly include structured “table reads” of draft language that could soften or effectively reverse earlier Office of Management and Budget guidance discouraging the use of Anthropic systems in federal agencies.

A White House spokesperson said the administration continues to engage across government and industry on artificial intelligence policy and national security priorities, adding that any formal announcement would come directly from the president, the outlet reported. Anthropic declined to comment.

The policy reconsideration comes amid a broader dispute that has stretched across multiple agencies, particularly the Department of Defense, which previously issued a supply chain risk designation against Anthropic. The designation limited how the company’s AI systems could be deployed in federal environments following disagreements over military usage conditions and safety constraints.

The dispute stems from Anthropic’s refusal to allow unrestricted military use of its Claude model, particularly for mass domestic surveillance or fully autonomous weapons development. Defense officials have argued that such restrictions reduce operational flexibility, while Anthropic has maintained that strict usage limits are essential for safety and governance. The Pentagon has continued to use older versions of Claude under legacy terms while negotiations remain unresolved, according to The Wall Street Journal.

Despite the ongoing disagreement, Anthropic systems remain partially embedded in federal workflows. The National Security Agency has already integrated the company’s Mythos model into certain analytical systems, according to Reuters, which has reported on the agency’s increasing reliance on commercial AI for cybersecurity and intelligence support functions.

The White House push also follows a series of high-level meetings between Anthropic leadership and senior U.S. officials. White House Chief of Staff Susie Wiles and Treasury Secretary Scott Bessent recently met with Anthropic CEO Dario Amodei in what both sides described as an introductory but productive discussion on potential collaboration pathways, according to Axios.

The Anthropic dispute has also played out alongside broader competition among leading AI companies for federal contracts. OpenAI and Google have both secured agreements allowing their models to be used under “all lawful purposes” frameworks in classified environments, according to Reuters, though those arrangements include internal safeguards intended to limit misuse while enabling broader deployment in defense systems.

Defense and intelligence agencies have been divided over how quickly advanced AI should be integrated into operational systems, according to The Washington Post, especially as capabilities in cyber operations and autonomous decision-support tools continue to evolve.