Anthropic’s dispute with the US government began in February after the AI company objected to the use of its technology for mass surveillance and fully autonomous lethal weapons.
President Donald Trump subsequently ordered federal agencies to stop using Anthropic’s products, while Defence Secretary Pete Hegseth designated the company a “supply chain risk” to national security.
In a 59-page ruling, US District Judge Rita Lin rejected the government’s justification, saying national security could not be used to punish companies for criticising government policies.
The judge barred the agencies named in the case from enforcing Trump’s order and overturned Hegseth’s designation of Anthropic as a supply chain risk. The ruling, which took effect immediately, makes permanent a temporary suspension of the sanctions imposed in March. The government can appeal the decision.
Anthropic welcomed the ruling, saying it remained committed to working with the US government on the safe use of AI for national security purposes.
The dispute followed Anthropic’s refusal to permit the US military to use its Claude AI models for fully autonomous lethal weapons or mass surveillance of Americans.
Trump had previously criticised the company on social media, calling it a “radical left, woke company”.
Judge Lin said the sanctions appeared to have been motivated by a desire to publicly punish Anthropic for challenging the government rather than by a genuine national security threat.
Claude had reportedly been used in several classified and sensitive military systems. Reports said the technology was also used during a US operation to capture Venezuelan leader Nicolas Maduro in January.
The Pentagon’s supply chain designation, issued in March, terminated Anthropic’s military contracts and prevented other defence contractors from using its technology.
The court warned that the sanctions could have broader consequences for government contractors and discourage companies from challenging government policies.
However, Anthropic has not secured a complete victory. A separate Pentagon sanction based on federal procurement rules remains in effect while another court considers the matter.
The White House has also criticised Anthropic CEO Dario Amodei over his repeated warnings about the potential risks of artificial intelligence, including its impact on employment.
