Pentagon official sees little chance to revive Anthropic AI deal

Emil Michael, the under secretary of defense for research and engineering, said Monday that Anthropic’s lawsuit was an “expected reaction”.

A top Pentagon official sees little chance of resuming negotiations with Anthropic PBC over military use of its artificial intelligence tools following the company’s legal challenge of an unprecedented government move to declare the firm a supply-chain risk.

"US Department of War" and Anthropic logos are seen in this illustration. (REUTERS)

Emil Michael, the under secretary of defense for research and engineering, said Monday that Anthropic’s lawsuit was an “expected reaction” and that the company’s attempt to reverse the supply-chain declaration wouldn’t alter the Pentagon’s decision.

“I don’t think there’s a scenario where this gets resolved in that way,” Michael said in an interview with Bloomberg News.

Michael spoke hours after Anthropic sued to block the Pentagon and other agencies from declaring the firm a threat to the US supply chain and barring it from government contracts. In court filings earlier Monday, Anthropic claimed the moves had violated the company’s rights to free speech and due process under the US Constitution.

An Anthropic spokesperson had no immediate comment when asked Monday about Michael’s remarks.

A former Uber Technologies Inc. executive now overseeing a Pentagon effort to accelerate AI adoption, Michael had held weeks of tense negotiations with Anthropic Chief Executive Officer Dario Amodei over terms for using the firm’s AI tools. Talks broke down roughly two weeks ago, after the company demanded assurances that its AI wouldn’t be used for mass surveillance of Americans or autonomous weapons deployment.

That prompted the Pentagon to declare San Francisco-based Anthropic a supply-chain risk, a move normally reserved for companies from adversarial nations. Until recently, Anthropic had provided the only AI system that could operate in the Pentagon’s classified cloud, and its Claude Gov tool has become a favored option among defense personnel for its ease of use.

The decision puts at risk a $200 million contract for Anthropic to provide the Pentagon with classified AI tools, and it could preclude the firm from partnering with other companies on their military work. Defense Secretary Pete Hegseth and President Donald Trump have outlined a six-month period for the military and US agencies to shift from Anthropic to other AI providers.

In one of its court filings, Anthropic expressed concern that the government’s actions have already hit its work with other federal contractors. It said that one vendor it has partnered with on custom applications warned that it may “suspend work or even remove Claude from existing deployments.” Other contractors, Anthropic said, “are raising concerns, pausing collaborations, and considering terminating contracts.”

On Monday, Michael reiterated a key Pentagon complaint that Anthropic’s public positions in calling for safeguards on military use indicated that the company wanted to have input on the chain of command and potentially operational control.

The company has rejected that characterization and insisted that all military decisions rest with the Pentagon. Amodei has also said that Anthropic would like to keep working with the military but wants those two usage restrictions to be honored in any defense contract.

Asked whether there could be any return to negotiations, Michael was blunt: “The talks are over. We’re moving on.”