Anthropic wins court battle over Pentagon risk label

U.S. District Judge Rita Lin in California ruled that the Trump administration's designation of Anthropic as a supply-chain risk was illegal. The judge stated that Defense Secretary Pete Hegseth's labeling of the Claude developer as a national security threat amounted to what Lin called "unlawful retaliation" in violation of the First Amendment. Lin also found that the decision was arbitrary and capricious, and that the government denied Anthropic due process under the Fifth Amendment.

The conflict began when Anthropic established strict safety guardrails to prevent the Pentagon from using its Claude models for mass surveillance of American citizens and fully autonomous weapons. In response, Hegseth and President Donald Trump labeled the company a supply-chain risk and ordered all federal agencies to cease working with them. However, Judge Lin noted that the government's actions were contradictory, pointing out that Hegseth had previously proposed applying the Defense Production Act to Anthropic and that the Department of Defense continued to pursue a contract with the company while collaborating on its new cybersecurity model, Mythos.

Anthropic initiated this legal battle by filing two complaints against the Department of Defense in March, with one lawsuit in California and another still ongoing in Washington, D.C. Judge Lin emphasized that the government cannot use national security as a blank check to punish critics, especially since Anthropic lacks any backdoor access to its technology once delivered to the military. For AI practitioners and enterprise developers, this ruling establishes a critical legal precedent. It suggests that AI vendors can enforce ethical guardrails and safety terms in government contracts without fearing arbitrary, retaliatory blacklisting from federal agencies.