August 28, 2026, 16:21 GMT | Comment
It is “undisputed” that the US government unlawfully retaliated against
Anthropic by designating it a supply chain risk, US District Judge Rita Lin found Thursday, putting a period on her review of actions that marked a turning point in President Donald Trump’s second term, toward Big Tech interventions. But the dispute is far from resolved, with related litigation still pending and possible appeals.
It is “undisputed” that the US government unlawfully retaliated against Anthropic by designating it a supply chain risk, US District Judge Rita Lin found Thursday, putting a period on her review of actions that marked a turning point in President Donald Trump’s second term, toward Big Tech interventions. But the dispute is far from resolved, with related litigation still pending and possible appeals.
Ruling late Thursday on cross motions for summary judgement in the US District Court of the
Northern District of California, Lin found that it was “illegal and baseless” for the Department of Defense, which calls itself the Department of War, to impose a risk designation on Anthropic earlier this year (see
here).
The DoD acted on March 4, after Anthropic insisted that its federal contract for classified military projects must include redlines against fully autonomous lethal weapons and mass surveillance of American citizens (see
here).
Anthropic, developer of the Claude large language models, subsequently challenged the designation in two lawsuits, disputing the government’s authorities under Section 3252 of the US Code in California, and under Section 4713 of the US Code in the
US Court of Appeals for the DC Circuit (see
here and
here).
“Neither the Constitution nor the federal statute invoked by defendants allows them to impose sweeping penalties based principally on Anthropic’s critique of the administration’s views,” Lin wrote Thursday. Their actions violated the First Amendment, she said, and were arbitrary and capricious in violation of Section 3252.
In a statement, Anthropic celebrated the order. “We welcome the court’s ruling that this supply chain risk designation was unlawful,” a spokesperson said. “We remain focused on working productively with the government to harness AI for our national security so all Americans benefit from this technology.”
The White House and Pentagon did not immediately respond to requests for comment about whether they plan to appeal. But a pending decision from the DC Circuit, which declined to grant Anthropic interim relief this spring, muddies the legal landscape.
Lin has been critical of DoD’s actions for months, decrying in her preliminary injunction order the “Orwellian notion that an American company may be branded a potential adversary… for expressing disagreement with the government” (see
here). But the DC Circuit found in April that a stay would inappropriately constrain the DoD during wartime (see
here).
The circuit panel heard arguments in May, scrutinizing both sides, though US Circuit Judge Karen Henderson described the DoD’s actions as “spectacular overreach” in one pointed exchange (see
here).
Observers have noted that if the courts ultimately reach different conclusions on the merits, the public will have to grapple with conflicting designations under parallel authorities. The parties may seek clarification from higher courts (see
here).
Paul Lekas, head of policy and government affairs at the Software Information Industry Association, said Friday that Lin’s order provides detailed analysis for the DC Circuit to consider.
“There’s some pretty strong statements in the opinion about the government not making a showing sufficient to establish what is essentially a presumption in favor of the government on matters of national security,” Lekas told MLex. “I think that’s probably a key place where we’re going to see attention in the DC Circuit.”
In the months since DoD first blacklisted Anthropic, Trump’s White House has imposed other intermittent restrictions on the company and its competitors, responding to the cyber security risks posed by increasingly capable models, such as Anthropic’s frontline Mythos and Fable.
Anthropic temporarily withdrew those models in June after the White House imposed, then lifted, export controls, citing security concerns (see
here). And
OpenAI temporarily limited the rollout of its GPT‑5.6 models that month, saying it did so at the government's request (see
here).
More recently, the Trump administration has instituted a voluntary process for frontier developers to share their models with the government pre-release, the details of which have not been published. Members of Congress are calling for more transparent and consistent review processes, via legislation (see
here).
Lin’s Thursday order references the Trump administration’s ongoing dealings with Anthropic, saying they undermined its position in the case before her.
“Even now, the government is discussing collaboration with Anthropic on its new model, Mythos, in an array of sensitive contexts,” she said. “None of that is consistent with a genuine fear that Anthropic is a saboteur who would poison its software to harm national security.”
Regardless of where the courts ultimately land on Anthropic’s risk designation, the White House is signaling that it can’t afford to ice out a leading company, even as it searches for its preferred level of oversight of increasingly capable models.
Please email editors@mlex.com to contact the editorial staff regarding this story, or to submit the names of lawyers and advisers.
Tags
Sections:
Artificial Intelligence
Industries:
Computing & Information Technology
Geographies:
North America, United States