Anthropic CEO Revives Pentagon AI Contract Talks

The air in the Pentagon’s negotiating rooms has rarely been this charged, or this fragile. Dario Amodei, the CEO of Anthropic, is back at the table, not with a triumphant pitch but a salvage operation. His company’s relationship with the U.S.

military imploded on Friday after weeks of bitter public feuding, driven by Anthropic’s refusal to hand the Department of Defense unrestricted keys to its AI models. The fallout was swift: rivals like OpenAI wasted no time rushing into the gap, eager to claim the defense work Anthropic had left on the table. Now, Amodei is reportedly meeting with under-secretary Emil Michael to broker a new contract, hoping to keep Claude in the Pentagon’s hands.

But the deeper story is more personal, and more political. Amodei blames the collapse squarely on what he sees as a loyalty test: “We haven’t donated to, or given dictator-style praise to Trump,” he said. In an era where access to power can hinge on allegiance, Anthropic is betting that technical capability still outweighs political alignment.

It’s a last-ditch attempt to prevent being branded a “supply chain risk” and frozen out of defense work entirely.