The White House is developing guidance that could let federal agencies sidestep Anthropic’s supply-chain risk designation and adopt new AI models, including Mythos, Axios reported. Reuters said it could not immediately verify the report. Axios reported that officials are considering a draft executive action. The measure could give the Trump administration a way to resolve its dispute with Anthropic, according to two sources cited by Axios.
Anthropic declined to comment on the report. The White House also did not immediately respond. The development follows a dispute between Anthropic and the Pentagon earlier this year. The department labeled the Claude-maker a supply-chain risk after the company declined to remove guardrails on its AI systems.
Those guardrails restrict the use of Anthropic’s AI for autonomous weapons and domestic surveillance. The dispute limited the company’s standing with parts of the federal government.
The draft guidance could allow agencies to work around the designation without formally removing it. That route would create a possible offramp while leaving the Pentagon’s concerns unresolved.
The issue now centers on Anthropic’s newest model, Mythos. Trump’s comments and the reported guidance came weeks after Anthropic introduced the system as its most advanced AI model.
Experts have said Mythos may hold a strong ability to identify cybersecurity weaknesses and develop ways to exploit them. That capability has placed the model at the center of federal AI discussions.
At the same time, the tool could attract interest from agencies focused on cybersecurity. The reported guidance would help those agencies onboard new models despite the existing risk flag.
President Donald Trump signaled a warmer tone toward Anthropic last week. He said the company was “shaping up” after CEO Dario Amodei met with White House officials.
When CNBC asked whether a Pentagon deal could emerge, Trump said it was possible. He added that the administration wants “the smartest people.”
Read More: CLARITY Act Talks Advance After White House, Senators Reach Tentative Deal
Axios reported that key Pentagon players remain firm on the issue with Anthropic. Still, other stakeholders see the dispute as counterproductive and want a way out. That split leaves the talks uncertain. The report said both sides could still return to difficult negotiations, even if the draft guidance moves forward.
The situation also fits into a wider federal debate over advanced AI models. Agencies and lawmakers are weighing how to use powerful systems while managing national security concerns.
Recent discussions have focused on cybersecurity, infrastructure, and the risks linked to next-generation AI tools. Those concerns have grown as models gain stronger technical abilities. For now, the draft guidance remains under development. Its final form could determine whether agencies gain a clearer path to Anthropic’s models or face another round of talks.
The White House’s reported draft guidance could give federal agencies a path to use Anthropic’s Mythos despite the Pentagon’s supply-chain risk flag. The move reflects ongoing tensions between AI access, safety guardrails, and national security concerns. The outcome will indicate how Washington balances advanced AI use with oversight.