The White House is weighing new guidance that could ease restrictions on Anthropic, potentially allowing federal agencies broader access to its latest artificial intelligence models despite an ongoing national security dispute, according to sources familiar with the discussions, Axios reported.
The draft policy, still under development by the White House, would give agencies a pathway to work around a controversial "supply chain risk" designation previously imposed by the Pentagon — a move that effectively sidelined the company from broader government adoption.
The effort marks a shift for the Trump administration, which had earlier characterized Anthropic as a significant security concern. One source described the internal push as an attempt to "save face and bring them back in."
At the center of the debate is Anthropic’s newest model, Mythos, which officials say has demonstrated advanced capabilities — including the potential to automate cyberattacks — while offering defensive applications.
Federal agencies are increasingly seeking access to the system as legal and policy disputes remain unresolved.
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 described as a productive discussion on future cooperation.
The White House is also convening industry leaders this week to help shape potential executive action, including "table read" sessions reviewing draft guidance that could scale back earlier restrictions issued by the Office of Management and Budget.
"The White House continues to proactively engage across government and industry to protect our country and the American people, including by working with frontier AI labs," a spokesperson told Axios, adding that any formal policy announcement would come directly from the president.
The policy deliberations come as tensions persist between the company and the U.S. Department of Defense, which imposed the supply chain designation after Anthropic refused to sign an agreement allowing its models to be used for "all lawful purposes," including potential applications in mass surveillance or autonomous weapons development.
Anthropic instead insisted on restrictions barring those uses — a stance the Pentagon argued made the company an unreliable partner. The dispute has since spilled into court, further complicating cooperation.
Despite the conflict, some federal entities — including the National Security Agency — are already using Anthropic's models, including Mythos, under existing agreements.
However, those arrangements rely on older terms of service that limit functionality and prevent access to the latest updates, according to sources familiar with the matter.
Other major AI firms, including OpenAI and Google, have signed agreements permitting the Pentagon to use their models under the "all lawful purposes" standard, though both companies maintain they enforce similar ethical guardrails.
Sources told Axios that while some Pentagon officials remain opposed to easing restrictions on Anthropic, others across the government view the standoff as counterproductive and are seeking a path forward.
It remains unclear whether the White House's proposed steps would resolve the Pentagon dispute or allow other agencies to bypass it.
Even if the supply chain designation is eventually lifted, officials say the underlying disagreement over acceptable uses of advanced AI systems is likely to persist.
Brian Freeman ✉
Brian Freeman, a Newsmax writer based in Israel, has more than three decades writing and editing about culture and politics for newspapers, online and television.
© 2026 Newsmax. All rights reserved.