FOLLOW US

The Verdict, Distilled

A California federal judge has ruled that the Pentagon's "supply chain risk" label against AI firm Anthropic was not grounded in genuine security concerns but in retribution for the company's public pushback against the administration. The 59-page opinion found violations of both the First and Fifth Amendments and orders federal agencies to stand down from enforcing the directive. The ruling marks the most significant legal setback yet for the White House's escalating campaign against tech companies that resist government demands.

By Samuel López | USA Herald

CALIFORNIA - A federal judge has just handed the Trump administration one of its most stinging legal defeats of the year — and the target wasn't a political rival or an advocacy group. It was an artificial intelligence company that simply refused to say yes to everything the Pentagon asked for.

U.S. District Judge Rita F. Lin of the Northern District of California ruled Thursday that the Department of Defense's designation of Anthropic PBC as a "supply chain risk to national security" had nothing to do with an actual threat assessment. Instead, according to the court, it was punishment — dressed up in the language of national security — for the company's decision to publicly resist the administration's contract demands.

The line that will likely define this case for years came directly from the bench. "The empty invocation of national security is not a blank check to punish and retaliate against government critics," Lin wrote in her decision. It's the kind of sentence that doesn't just decide a case — it reframes the entire dispute as a cautionary tale about the limits of executive power when it collides with the First Amendment.

“It is DECLARED that the Challenged Actions violate the First Amendment because they are unlawful retaliation against Anthropic for constitutionally protected expressive activities.” - Judge Rita F. Lin, from the Order of Final Relief - Filed 8/27/26

To understand why a federal judge would use language this pointed, you have to go back to late February. That's when the Department of Defense, under Secretary Pete Hegseth, pressed Anthropic to strip away safeguards built into its Claude AI models — restrictions the company had put in place specifically to prevent its technology from being used for two purposes: mass, dragnet-style surveillance of American citizens, and the development of fully autonomous weapons systems capable of using lethal force without a human in the loop.

Anthropic's leadership, led by CEO Dario Amodei, held the line. The company would not budge on those two guardrails, contract or no contract.

The administration's response was swift and public. Trump directed federal agencies to stop using Claude and other Anthropic products, and the Pentagon went a step further, branding the company a "supply chain risk" — a designation with real teeth, since it effectively barred the department from doing business with any contractor that also does business with Anthropic. Agencies were given a six-month runway to migrate away from the company's tools.

Anthropic didn't fold. It sued, arguing the designation was retaliatory, unconstitutional, and legally baseless. In March, Judge Lin agreed enough to issue a preliminary injunction on First Amendment grounds, calling the government's underlying theory "Orwellian." Thursday's ruling is the next chapter — and, for now, a decisive one.

Judge Lin's order found that the Pentagon's actions likely violated the First Amendment's protection against retaliation for protected speech, and the Fifth Amendment's guarantee of due process. The court also concluded the designation exceeded the government's authority under the Administrative Procedure Act — in plain terms, that the Pentagon didn't just act unfairly, it acted unlawfully.

Perhaps most damaging to the government's position was the court's observation about the Pentagon's own conduct. The opinion noted that defense officials continued pursuing business with Anthropic even after slapping it with a national-security risk label — behavior fundamentally at odds with treating the company as an actual threat. That contradiction became a key thread in the court's reasoning that the "risk" designation was pretextual.

The order requires the government to withdraw the directives issued against Anthropic and bars certain federal agencies from continuing to enforce Trump's original order to abandon the company's products.

But don't expect this fight to end here. Under Secretary of War Emil Michael pushed back hard on social media, characterizing the ruling as riddled with factual errors and accusing the court of encroaching on the President's constitutional role as Commander in Chief during what he described as a period of active conflict. Administration officials have signaled they still consider Anthropic's risk designation to be in effect while an appeal moves forward — meaning the practical fallout from Thursday's ruling remains, for now, contested territory.

Anthropic, for its part, struck a notably restrained tone in victory. A company spokesperson said it welcomed the court's finding that the designation was unlawful, while emphasizing that Anthropic remains interested in working constructively with the government on national security applications of its technology — just not on the administration's original terms.

For the moment, the ruling stands as a significant check on executive authority — and a reminder that even in an era of rapid AI deployment and national-security urgency, courts are still willing to ask whether the government's stated justification actually matches its conduct.

About the Author

Samuel López is a Senior Legal Analyst and investigative journalist for USA Herald, where he covers high-stakes litigation at the intersection of technology, national security, and constitutional law. His reporting has followed the Anthropic-Pentagon dispute since its earliest filings, offering readers context and analysis that goes beyond the headlines.

Stay Ahead of the Story

Subscribe to the USA Herald newsletter for exclusive investigative coverage delivered straight to your inbox.

👉 Follow us on X: https://x.com/RealUSAHerald