Anthropic plans legal challenge against Pentagon supply-chain designation

Dario Amodei said on Thursday that Anthropic intends to challenge the Department of Defence’s decision to classify the AI company as a supply-chain risk in court, describing the designation as “legally unsound.”

The statement arrived just hours after the DOD formally designated Anthropic a supply-chain risk, following a dispute that had stretched for weeks over how much authority the military should have over AI systems. A supply-chain risk designation can prevent a company from doing business with the Pentagon and its contractors. Amodei had drawn a clear boundary that Anthropic’s AI would not be used for mass surveillance of Americans or for fully autonomous weapons. At the same time, the Pentagon held the view that it should have unrestricted access for “all lawful purposes.”

In his statement, Amodei said that the overwhelming majority of Anthropic’s customers would not be affected by the supply-chain risk designation.

“With respect to our customers, it plainly applies only to the use of Claude by customers as a direct part of contracts with the Department of War, not all use of Claude by customers who have such contracts,” he said.

Offering an early indication of the case Anthropic is likely to make in court, Amodei argued that the Department’s letter designating the company as a supply-chain risk is narrow in what it actually covers.

“It exists to protect the government rather than to punish a supplier; in fact, the law requires the Secretary of War to use the least restrictive means necessary to accomplish the goal of protecting the supply chain,” Amodei said. “Even for Department of War contractors, the supply chain risk designation doesn’t (and can’t) limit uses of Claude or business relationships with Anthropic if those are unrelated to their specific Department of War contracts.”

Amodei also reiterated that Anthropic had been engaged in constructive talks with the DOD over the past several days, discussions that some observers believe may have been derailed when an internal memo he sent to staff was leaked. In that memo, Amodei described rival OpenAI’s work with the Department of Defence as “safety theatre.”

OpenAI has since signed an agreement to work with the DOD in place of Anthropic, a decision that has triggered backlash among some OpenAI employees.

In his Thursday statement, Amodei apologised for the leak, saying the company had not intentionally shared the memo or instructed anyone else to do so. “It is not in our interest to escalate the situation,” he said.

Amodei said the memo had been written within “a few hours” of a sequence of announcements, including a Truth Social post from the president stating that Anthropic would be removed from federal systems, followed by Defence Secretary Pete Hegseth’s supply-chain risk designation, and then the Pentagon’s announcement of its deal with OpenAI. He apologised for the tone of the memo, describing it as written on “a difficult day for the company,” and said it did not reflect his “careful or considered views.” He added that because it was written six days ago, it now represents an “out-of-date assessment.”

He concluded by saying that Anthropic’s highest priority is making sure American soldiers and national security personnel continue to have access to critical tools during ongoing major combat operations. Anthropic is currently supporting some U.S. operations in Iran, and Amodei said the company would continue supplying its models to the DOD at “nominal cost” for “as long as necessary to make that transition.”

Anthropic could challenge the designation in federal court, most likely in Washington. Still, the law underlying the decision may make it difficult to contest because it limits some of the normal avenues companies use to challenge government procurement actions and gives the Pentagon broad authority over matters related to national security.

As Dean Ball — a former Trump-era White House adviser on AI who has criticised Hegseth’s handling of Anthropic — put it: “Courts are pretty reluctant to second-guess the government on what is and is not a national security issue … There’s a very high bar that one needs to clear to do that. But it’s not impossible.”