Anthropic and Trump Administration Hold High-Level Talks Despite Pentagon Dispute

Pentagon designated Anthropic a supply-chain risk after military AI negotiations failed

The label — typically reserved for foreign adversaries such as Chinese telecom firms — could block federal procurement of Anthropic’s technology across agencies. Negotiations broke down because Anthropic sought to restrict use of its models for fully autonomous weapons systems and mass domestic surveillance.

Anthropic is challenging the Pentagon’s supply-chain risk designation in court

Rather than accept the label, Anthropic filed a legal challenge, signaling it views the designation as unjust and potentially precedent-setting for how AI companies with ethical constraints are treated by the U.S. government.

CEO Amodei met White House’s Bessent and Wiles; both sides issued positive statements

The meeting — reported by Axios — included Treasury Secretary Scott Bessent and Chief of Staff Susie Wiles. Both issued statements covering shared priorities: cybersecurity, U.S. AI competitiveness, and AI safety. Jack Clark called the Pentagon dispute ‘narrow’ and separate from the company’s broader willingness to work with government.

Every federal agency except DoD reportedly wants access to Anthropic’s technology

An administration source told Axios the Pentagon’s hostility is isolated within government. Bessent and Fed Chair Jerome Powell also reportedly encouraged major bank executives to test Anthropic’s Mythos model, indicating strong advocates in economic leadership.

OpenAI announced a military deal after Anthropic declined — drawing consumer backlash

After Anthropic held firm on autonomous weapons safeguards, OpenAI quickly moved to secure its own military deal. That decision generated consumer backlash, illustrating the reputational trade-offs AI labs face when navigating government-military relationships.