A federal judge in California has stopped the Defense Department from enforcing a national security supply-chain risk designation against Anthropic. The ruling keeps the Claude developer in contention with the Pentagon over how artificial intelligence should be used in military settings and how far federal contracting restrictions can reach.
U.S. District Judge Rita Lin found that the Pentagon acted outside the law when it directed companies working with the U.S. military to cut ties with Anthropic. The judge said the government could not invoke national security as an unlimited justification to penalize critics of official policy.
The dispute centers on a February decision by Defense Secretary Pete Hegseth to classify Anthropic as a supply-chain risk. That designation is generally applied to companies tied to countries viewed as threats to the United States and is meant to address risks such as foreign infiltration or sabotage of military systems.
Anthropic’s lawsuit, filed March 9 in California federal court, argues that the Pentagon’s action was unlawful and unsupported by evidence. The company says the designation conflicted with earlier favorable assessments of Claude by the military.
The company also alleges constitutional violations. Anthropic claimed that the sanction was a punishment for the company’s stated views on AI safety, which is an infringement of the right to free speech under the First Amendment. Anthropic also invoked the Fifth Amendment, arguing that the company was not allowed to challenge the designation before it was imposed, violating the right to due process.
Anthropic has said its AI systems are not dependable enough for safe deployment in autonomous weapons. The company also opposes domestic surveillance on civil-rights grounds.
The Pentagon has maintained that private technology companies should not be able to limit military operations. It has also said the designation was driven by Anthropic’s refusal to accept revised contract terms, rather than by the company’s public position on AI safety. The department argued that the company’s stance raised concerns about how its technology might be used.
Anthropic welcomed the court’s decision, saying it remained committed to working with the government on national security uses of AI that benefit Americans.
During an earlier stage of the litigation, Judge Lin said the government’s approach appeared aimed at damaging Anthropic’s business and discouraging public discussion about the military use of AI.
Anthropic is also pursuing a separate case in Washington, D.C., challenging another Pentagon supply-chain risk designation. That action could affect the company’s ability to compete for civilian government contracts.
