Unsealed emails show Anthropic-Pentagon clash over AI weapons redlines before supply-chain risk label

Court documents unsealed July 2, 2026 reveal that Anthropic Chief Executive Dario Amodei clashed with a senior Pentagon official over the company’s AI-safety redlines in the days before the Defense Department branded Anthropic a supply-chain risk, a label previously reserved for foreign adversaries.

The filings, part of Anthropic’s lawsuit against the US Department of Defense, add specifics to a dispute the company described publicly in March. In email exchanges from earlier this year, Amodei reiterated Anthropic’s refusal to let its Claude models be used for fully autonomous weapons systems or mass domestic surveillance of Americans. Emil Michael, the under secretary of defense for research and engineering, called those guardrails “just not workable” and warned there was “one more chance to align on core principles.”

When Amodei said the Pentagon appeared to want to “completely remove our redlines,” Defense Secretary Pete Hegseth designated Anthropic a supply-chain risk the next day; Anthropic disclosed the designation on March 5. US District Judge Rita Lin later blocked enforcement of that designation, finding the government’s actions likely retaliatory.

The emails are one side’s evidence in active litigation, and the Pentagon disputes the retaliation finding. What they document is the friction point: the dispute traces back to Anthropic raising concerns over whether Claude was used in the January 2026 US operation that captured former Venezuelan leader Nicolas Maduro.

The case tests whether an AI developer can hold contractual limits on military use without losing federal business. A ruling on the merits, or a settlement, will set the precedent.