Court overturns Pentagon decision to label AI developer Claude as security threat
A US federal court has ruled that the Trump administration illegally added AI developer Claude - the company Anthropic - to a list of supply chain risks and restricted cooperation with it. The court concluded that the government effectively punished the company for its public stance on acceptable military use of artificial intelligence.
The ruling was made by Judge Rita Lin of the US District Court for the Northern District of California. She struck down the Pentagon's designation of Anthropic as a "supply chain risk" and barred federal agencies from complying with orders to stop using the company's products. Thus, the temporary block on these measures, imposed by the court back in March, has now become permanent.
In the 59-page decision, the court noted that citing national security alone does not give the government the right to punish companies for criticizing the government. The judge concluded that the decision regarding Anthropic was related not to proven risk of sabotage or interference with US military systems, but primarily to the company's conflict with the Pentagon over the rules for using Claude.
Why Anthropic fell out with the Pentagon
The conflict arose in early 2026 during negotiations over the use of Claude by the US military. The Pentagon insisted on the ability to use the model for any lawful military purposes, while Anthropic refused to remove two fundamental restrictions.
The company refused to allow the use of Claude for mass domestic surveillance of Americans and in fully autonomous weapons systems that could select and strike targets without human involvement. Anthropic explained that current AI models are not yet reliable enough for fully autonomous weapons, and mass domestic surveillance threatens fundamental rights.
After negotiations broke down, US Secretary of Defense Pete Hegseth ordered that Anthropic be designated a supply chain risk. This mechanism was created primarily to protect US military systems from potential interference or sabotage, and had not previously been applied publicly to an American company.
Anthropic from the outset maintained that this decision had no proper legal basis. The company emphasized that it supports the use of AI for US national security and has been working with the US military since 2024, but it is not ready to abandon the two stated restrictions.
Court sees punishment for criticizing the government
Key for the court was that the US government itself later moved away from its initial arguments about technical risks from Anthropic. As follows from the court ruling, the company has no hidden access to its models after they are deployed in secure systems, and its technology in this regard does not pose a greater risk than other closed AI models.
Meanwhile, administration officials publicly criticized Anthropic for its stance. President Donald Trump called the company "radical left" and accused it of trying to dictate to the military how to use technology. The court viewed the circumstances of the case as confirming that the government wanted to make an example of punishing the company for disagreeing with the administration's position.
For Anthropic, the ruling also has significant commercial implications. The company had warned that being on the "blacklist" could cost it billions of dollars due to loss of government and related contracts, as well as reputational damage among corporate clients.
Anthropic welcomed the court's decision. However, the legal battle with the US government is not fully over: a separate case regarding another Pentagon decision on the company's status is ongoing in Washington, so the government may still seek to maintain some restrictions.
