Judge rules Pentagon's Anthropic blacklist unlawful

In March, the Pentagon labeled Anthropic a "supply chain risk," cutting off military use of its Claude chatbot. On Thursday, a federal judge said that move was illegal. Judge Rita Lin ruled the label was retaliation for Anthropic publicly criticizing the administration's policies. Nine federal agencies had joined the ban. The judge called it unconstitutional under the First Amendment.

This wasn't a minor paperwork fight. Being labeled a supply chain risk can block a company from federal contracts and scare off other customers. Anthropic sued in March. It argued officials punished the company for warning that AI systems could be misused, not for any real security flaw. The Justice Department disagreed. It told the court that lifting the label could confuse the military. That confusion could hit at the worst time, during real operations. Judge Lin sided with Anthropic on every point. She ruled the label had no factual basis. It existed, she said, only to silence a critic.

Keep reading — what to watch

Will the Pentagon appeal, or accept the ruling and move on? The case exposed a bigger tension. The Trump administration wants AI companies working closely with the government. But some of those same companies keep warning publicly about AI risks. Anthropic's CEO has been especially vocal about that danger. This ruling suggests that criticizing the government doesn't strip a company of legal protection, even during a security dispute. Other AI companies facing their own government pressure will be watching closely.