Anthropic chief seeks last-minute Pentagon deal to keep AI in military supply chain

Anthropic CEO Dario Amodei is pushing a compromise with the Pentagon after a heated dispute that left the AI company at risk of being blacklisted by the US government.

According to the Financial Times, Amodei has engaged in urgent negotiations with officials, including Emil Michael, under-secretary of defense for research and engineering, to reach an agreement governing military access to Anthropic’s AI models.

A successful outcome would allow the Pentagon to continue deploying the company’s technology and would avert a threatened designation as a supply chain risk that would effectively sever Anthropic from defense contracts and force military contractors to cut ties with the San Francisco-based AI firm.

Following a US operation to capture Venezuelan leader Nicolás Maduro in January, reports surfaced that Anthropic employees discovered through Palantir logs that Claude was used during the operation.

The application raises questions about compliance with Anthropic’s Acceptable Use Policy.

Combined with the company’s reluctance to allow its AI to be used for fully autonomous weapons and mass surveillance, this led to a dramatic breakdown in negotiations with the Pentagon.

The department is seeking broader permission for the AI to be used for any “lawful” purpose, which Anthropic fears could enable surveillance uses it opposes.

After Amodei rejected the government’s ultimatum, President Donald Trump ordered federal agencies to stop using Anthropic’s technology, and Defense Secretary Pete Hegseth designated the firm a national security supply chain risk.

Amodei accused the Pentagon and OpenAI of misrepresenting the issue. He also suggested that Anthropic was being sidelined partly because it has not praised Trump as enthusiastically as its rivals.

Anthropic, alongside OpenAI, Google, and xAI, landed a Pentagon deal worth up to $200 million to advance agentic AI for military use. Losing that foothold would represent a major setback for a company that has positioned itself as a leader in AI safety.