Has Anthropic Gained the Upper Hand in Pentagon Lawsuit?

A federal judge in San Francisco dealt a stinging rebuke to the Trump administration on Thursday, blocking its efforts to bar Anthropic from government contracts and ruling that the White House and Pentagon appeared to have punished the AI company for speaking out – a move the court called unconstitutional.

Judge Rita Lin granted Anthropic's request for a preliminary injunction, halting two sweeping directives: a presidential order demanding federal agencies "immediately cease" all use of Anthropic's technology and the Pentagon's designation of the company as a "supply chain risk". This national security label is  historically reserved for firms based in adversarial nations like China and Russia.

"Punishing Anthropic for bringing public scrutiny to the government's contracting position is classic illegal First Amendment retaliation," wrote Judge Lin in her ruling. 

The order prevents the Trump administration from implementing or enforcing both the presidential directive and the Pentagon's designation while the underlying lawsuit proceeds. A final verdict is likely still months away.

From contract dispute to constitutional crisis

The case has its origins in a US$200m contract Anthropic signed with the Department of Defence in July 2025. As the two sides began negotiating the deployment of Anthropic's Claude models on the Pentagon's GenAI.mil platform that autumn, talks broke down over a single, consequential clause.

The Pentagon sought broad permission to use Anthropic's tools for "any lawful use". Anthropic, led by CEO Dario Amodei, refused, insisting on explicit assurances that Claude would not be used to power fully-autonomous weapons systems or to conduct mass domestic surveillance of American citizens. The company declined to sign without those guardrails in place.

What followed was extraordinary, even by Washington's standards. In late February, Secretary of Defense Pete Hegseth declared Anthropic a supply chain risk, requiring defence contractors including Amazon, Microsoft and Palantir to certify they were not using Claude in their military work.

Days later, President Trump posted on Truth Social ordering all federal agencies to phase out Anthropic tools within six months.

"WE will decide the fate of our Country – NOT some out-of-control, Radical Left AI company," he wrote. 

Anthropic, which had never previously been described in such terms by a government partner, filed suit earlier this month, arguing that its business had suffered material harm and that the government had violated its constitutional right to free expression.

The judge's findings

In her order, Judge Lin was notably pointed about the administration's stated justifications. She observed that Trump and Hegseth had publicly labelled Anthropic "woke" and described its staff as "left-wing nut jobs" – language that had little bearing on any genuine national security concern.

"Nothing in the governing statute supports the Orwellian notion that an American company may be branded a potential adversary and saboteur of the US for expressing disagreement with the government," she wrote.

The Pentagon had argued that its concerns were legitimate – that Anthropic's refusal to accept new contract terms created a genuine security risk, but Judge Lin was unconvinced: "If this were merely a contracting impasse, DoW would presumably have just stopped using Claude."

Instead, she said the challenged actions "far exceed the scope of what could reasonably address such a national security interest". 

At a hearing two days before the ruling, Judge Lin pressed government lawyers on precisely why the supply chain risk designation had been applied. Her language in the written order was sharper still.

What comes next?

Because the administration relied on two distinct legal provisions to justify its actions, Anthropic has had to pursue them in separate courts. A second lawsuit challenging the formal designation has been filed in the U.S. Court of Appeals in Washington.

In a statement following Thursday's ruling, Anthropic said it was "grateful to the court for moving swiftly".

It added: "While this case was necessary to protect Anthropic, our customers and our partners, our focus remains on working productively with the government to ensure all Americans benefit from safe, reliable AI."

The court ruling this week serves as victory for Anthopic, but only an initial one. Anthropic's Claude models may continue to be used across government and by contractors for now, but the deeper question of whether Washington overstepped its authority in a politically-motivated dispute with one of the biggest players in AI is still to be answered.