Pentagon appears to be ‘punishing’ Anthropic in violation of free speech, judge says

Unlock the White House Watch newsletter for free

Your guide to what Trump’s second term means for Washington, business and the world

A US judge has said the Pentagon appears to have tried to punish Anthropic for going public with its contract dispute over the military use of AI, in violation of free speech protections.

Judge Rita Lin, who is overseeing Anthropic’s legal challenge to the defence department’s move to brand the start-up a supply chain risk, said its actions looked like “an attempt to cripple” AI lab.

“It looks like [the defence department] is punishing Anthropic for trying to bring public scrutiny to this contracting dispute, which would, of course, be a violation of the First Amendment,” Lin said in a hearing on Tuesday.

Lin’s comments suggest she is sceptical of the government’s arguments as she considers whether to impose an injunction against the Pentagon. She is expected to issue a decision as soon as this week.

She said the department’s actions were “troubling” and “don’t really seem to be tailored to the stated national security concern”.

The dispute centres on the use of Anthropic’s AI model Claude by the military. It has already been deployed in classified operations, including in the war against Iran and in the capture of Venezuelan leader Nicolás Maduro.

But negotiations over its contract with the Pentagon broke down after Anthropic said it refused to have its technology used for lethal autonomous weapons and mass domestic surveillance.

The Trump administration responded by labelling Anthropic a supply chain risk, a national security measure normally used for foreign companies based in adversary nations such as Russia or China.

President Donald Trump told US agencies to stop using the start-up, while defence secretary Pete Hegseth posted on X that all military contractors must end commercial partnerships with Anthropic.

After questions from Lin about the social media posts, a Pentagon lawyer told the court that Hegseth’s post did not constitute a legal action and should not be interpreted broadly.

Military contractors could still use Anthropic for work unrelated to the defence department, he added.

Anthropic argued that the fracas has led to almost a month of “profound uncertainty” for its commercial partners as well as “irreparable and mounting” damages.

Recommended

A broad application of the ban might have cut Anthropic off from vital data centre infrastructure provided by companies such as Amazon and Microsoft and cut deeply into its revenues.

Anthropic estimated that even a narrow view of the order could still mean hundreds of millions of dollars in annual revenue were at risk.

The company has claimed that the designation violates the First Amendment of the US constitution and is retaliation for refusing to comply with the department’s demands to loosen safeguards on its technology.