A federal judge in San Francisco granted Anthropic a preliminary injunction, halting the Trump administration’s effort to blacklist the AI firm and ban federal use of its Claude models.
US District Judge Rita Lin found Anthropic likely to succeed on claims that the government’s actions constituted illegal First Amendment retaliation-punishing the company for publicly resisting military demands to remove safety safeguards.
The order blocks enforcement of President Trump’s directive and the Pentagon’s designation of Anthropic as a national security supply chain risk. The ruling is stayed for seven days to allow appeal.
Anthropic had secured a $200 million Pentagon contract and deployed Claude models across classified defense networks before refusing to permit use in fully autonomous weapons or mass surveillance of Americans.
The administration’s dual legal strategies-military blacklisting and civilian procurement bans-forced Anthropic to file separate lawsuits. A related case over civilian contracting remains active in Washington.
This marks the first known public blacklisting of a U.S. tech company over ethical AI policy.