A federal judge in San Francisco has temporarily halted two government actions against Anthropic: the Pentagon’s move to label the company a “supply chain risk” and President Donald Trump’s order directing federal agencies to stop using its technology.
U.S. District Judge Rita F. Lin said the measures appeared aimed at punishing the company and could cripple its business.
The ruling pauses a major Trump administration move against a U.S. artificial intelligence company while a broader legal fight continues.
The conflict, which grew out of a contract fight that became public in February, escalated after Anthropic CEO Dario Amodei said Claude should not be used for autonomous weapons or to surveil Americans. That position put the company at odds with the Pentagon over how the military could use the tool.
Download the SAN app today to stay up-to-date with Unbiased. Straight Facts™.
Point phone camera here
What the judge said about the government’s actions
In her order, Lin wrote that the government’s measures appeared “designed to punish Anthropic,” not protect national security, according to NPR and The Associated Press. She said the designation was likely unlawful and called it “arbitrary and capricious.”
Lin also said the law does not support branding an American company a potential adversary or saboteur simply because it disagreed with the government.
Anthropic argued the designation was retaliatory and could damage its business by costing the company customers and revenue. The Pentagon argued that Anthropic had become untrustworthy and that the military should decide on the lawful uses of the tools it buys.
What happens next
The order leaves the restrictions on hold while the broader lawsuit moves forward. Lin delayed her order for one week and said it does not require the Pentagon to use Anthropic’s products.
A separate, narrower Anthropic case remains pending in a federal appeals court in Washington, according to the AP.
