A federal judge ruled Thursday that the Pentagon acted illegally when it branded the AI company Anthropic a "supply chain risk" earlier this year, finding the designation amounted to unconstitutional retaliation for the company's public criticism of the Defense Department's approach to artificial intelligence, a decision first reported Thursday night.
In a 59-page order, U.S. District Judge Rita Lin wrote that the government's punitive measures against the maker of the Claude chatbot "were based on a desire to make a public example out of Anthropic for its 'arrogance' in criticizing the government, not based on any articulable basis to believe that Anthropic would actually sabotage its model." Lin found the Pentagon had neither constitutional nor statutory authority to impose sweeping penalties over a company's policy disagreements with the administration, and concluded the move violated the Fifth Amendment's due process protections as well as First Amendment limits on retaliating against protected speech.
A dispute rooted in military AI use
The fight dates to February, when President Trump and Defense Secretary Pete Hegseth publicly accused Anthropic of endangering national security and ordered federal agencies to stop using its technology, after negotiations over Pentagon access to Claude broke down. At issue was Anthropic's refusal to permit unrestricted military use of its Claude models; chief executive Dario Amodei has said the company would not waive safeguards it maintains against uses such as mass surveillance and fully autonomous weapons. Anthropic sued the following month, calling the supply-chain designation an "unlawful campaign of retaliation," and the case proceeded through hearings in July before Thursday's ruling.
The episode has been read in Washington and in the tech industry as a test of how the administration handles AI companies that push back on its terms, particularly after a rival, OpenAI, signed its own Pentagon contract shortly after the government took action against Anthropic. Anthropic's rising commercial profile has continued regardless of the standoff: the company disclosed an expanded enterprise partnership with Salesforce this week that Wall Street treated as a sign of Claude's growing footprint inside corporate software.
Reaction and next steps
Anthropic welcomed the decision, saying in a statement that it remained "focused on working productively with the government to harness AI for our national security so all Americans benefit from this technology." The White House did not immediately comment, and legal observers expect the administration to appeal; a separate case involving different Pentagon AI regulations is already pending before a federal appeals court in Washington, D.C.
The ruling does not force the Pentagon to buy or use Anthropic's products, but it strips away the formal "supply chain risk" label that had barred agencies from doing so, at least for now. Lawyers involved in the case said the decision could have broader implications for how much latitude the executive branch has to penalize contractors — in AI or any other sector — over public disagreements with administration policy, a question likely to recur as the government deepens its reliance on private AI developers for both civilian and military systems.
