Judge Rules Pentagon Illegally Punished AI Firm Anthropic in Retaliation for Criticizing the Government

SAN FRANCISCO — A federal judge late Thursday permanently barred the Trump administration from enforcing a set of measures aimed at cutting artificial intelligence company Anthropic off from the federal government, ruling that officials had unconstitutionally punished the firm for protected speech.

U.S. District Judge Rita Lin of the Northern District of California issued the 59-page ruling capping a monthslong legal fight between Anthropic and the Pentagon over the company's AI model, Claude, which is used across parts of the U.S. military. Lin found that the Defense Department's actions against the company violated both the First Amendment and the Fifth Amendment's due process protections.

The dispute traces back to February, when the Pentagon and Anthropic clashed over how far the military could go in using Claude. Anthropic had sought to keep in place internal safeguards preventing its model from being used for mass surveillance or fully autonomous weapons systems, while military officials argued they should be free to use the technology for any lawful purpose. When the two sides failed to reach an agreement, President Trump ordered federal agencies to stop using Claude, and Defense Secretary Pete Hegseth designated Anthropic a supply chain risk, a rarely used label that effectively sought to block private defense contractors from using the company's technology for military-related work.

In her ruling, Lin wrote that the government's actions "constituted unlawful retaliation in violation of the First Amendment" and found that Anthropic "was denied the pre-deprivation process required under the Fifth Amendment." She was sharply critical of the administration's stated justification for its actions, writing that the government's contemporaneous statements and conduct showed the moves 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 also pointed to a contradiction in the government's own conduct, noting that other federal agencies continued working with and meeting with Anthropic even after the Pentagon's supply chain risk designation took effect. "None of that is consistent with a genuine fear that Anthropic is a saboteur who would poison its software to harm national security," she wrote. She further wrote that "the empty invocation of national security is not a blank check to punish and retaliate against government critics."

The judge was careful to note that her order does not require the government to do business with Anthropic or prevent it from choosing a different AI vendor, "so long as those actions are consistent with applicable regulations, statutes, and constitutional provisions."

Lin's ruling followed a hearing in late July at which she had already signaled skepticism toward the government's position, telling attorneys the administration's stance was "really troubling" to her and appeared "at odds" with the First Amendment. At that same hearing, Justice Department lawyers argued that AI models are "so staggeringly enormous and opaque" that the Pentagon could not evaluate Claude the same way it would assess a piece of physical hardware.

Anthropic sued the Pentagon in March, calling the supply chain risk designation part of an "unlawful campaign of retaliation" tied to its refusal to allow unrestricted military use of its technology. Earlier in the litigation, Lin had already temporarily blocked the Pentagon from enforcing the supply chain risk label and had separately blocked enforcement of Trump's directive ordering federal agencies to stop using Claude.

An Anthropic spokesperson welcomed Thursday's ruling in a statement, saying the company remains "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 respond to a request for comment on the ruling.

The Pentagon's designation of Anthropic marked the first time the U.S. government had publicly applied the supply chain risk label to an American company under the statute, which was originally intended to protect military systems from foreign sabotage. Lin also noted an apparent contradiction between the blacklisting and an earlier threat by Hegseth to invoke the Defense Production Act against Anthropic, a move that Lin said implied the administration itself viewed the company as "essential to national security rather than a threat to it."

The government is expected to appeal Thursday's decision. A second, narrower lawsuit filed by Anthropic over a different rule the Pentagon has used to try to justify labeling the company a supply chain risk remains pending before the federal appeals court in Washington, D.C. It remains unclear whether the administration will wait for a ruling in that case before challenging Lin's decision.

Thursday's ruling represents a significant legal victory for Anthropic as the company continues to expand its commercial and government business. The decision could reopen paths for federal agencies and defense contractors to resume using Claude, potentially restoring business opportunities that had been cut off since the designation took effect earlier this year.