Defence Secretary Pete Hegseth has summoned Anthropic CEO Dario Amodei to the Pentagon on Tuesday morning for talks focused on the U.S. military’s use of Anthropic’s AI model Claude, according to Axios.
The meeting is taking place amid escalating tensions between the Pentagon and Anthropic, with the Defence Department reportedly weighing whether to designate the AI company as a “supply chain risk”—a label typically applied to foreign adversaries. The threat stems from Anthropic’s refusal to permit the Department of Defence to use Claude for two specific purposes: mass surveillance of Americans and the development of weapons that can fire without human involvement.
Anthropic signed a $200 million contract with the Department of Defence last summer. Since then, the relationship has grown more strained, particularly after reports said Claude was used during the January 3 special operations raid that resulted in the capture of Venezuelan president Nicolás Maduro. This episode brought the disagreements into the open and intensified scrutiny over how the model is being deployed in real-world operations.
A source cited by Axios described the Pentagon’s posture toward Anthropic as ultimatum-driven, suggesting Hegseth is delivering a blunt message to Amodei: cooperate with the Defence Department’s desired use cases or risk being pushed out entirely. Whether that hard line is a negotiating tactic remains unclear, since replacing Anthropic inside sensitive defence workflows would be difficult and disruptive.
Even so, the consequences being discussed are significant. A “supply chain risk” designation would effectively void Anthropic’s existing Defence Department contract. It could compel other Pentagon contractors and partners to stop using Claude as well, cutting it out of defence-related ecosystems beyond direct government procurement.
