A federal judge ruled the night of August 27 that the Pentagon acted illegally when it designated Anthropic a "supply chain risk to national security" -- finding the label was retaliation for the company's public criticism of the Defense Department's plans for AI on the battlefield, not a genuine security judgment. U.S. District Judge Rita Lin's 59-page order, reported the next morning, sided with Anthropic on summary judgment: the full merits of the case, not just a preliminary finding.
The dispute traces back to a July 2025 contract that made Claude the first frontier AI model approved for use on classified US government networks. Sometime after that, the Pentagon asked Anthropic to waive the restrictions it had built into that deal -- rules against using Claude for mass domestic surveillance or fully autonomous weapons systems. Anthropic refused. Defense Secretary Pete Hegseth designated the company a supply chain risk on February 27, 2026, a category ordinarily reserved for suppliers suspected of foreign sabotage or compromised hardware, not a US lab declining a request on safety grounds.
Two separate Anthropic-Pentagon fights, not one
| N.D. Cal. -- this ruling Anthropic PBC v. U.S. Dept. of War | D.C. Circuit -- the other case still open | |
|---|---|---|
| What's being challenged | The Feb. 27 designation itself, on First and Fifth Amendment grounds | A different rule the Pentagon is using to try to reach the same result |
| Status as of Aug. 28 | Decided for Anthropic on summary judgment | Still pending |
| What happens next | Government hasn't said whether it will appeal this ruling | Awaiting a ruling |
Anthropic sued March 9, 2026, alleging the designation violated its First Amendment right to free speech by retaliating against its public safety stance, and its Fifth Amendment right to due process by giving it no chance to dispute the label before it took effect. Lin questioned the Pentagon's "troubling" conduct in a preliminary ruling that March, and the Trump administration appealed that preliminary finding on April 2 -- a separate, narrower case over a different rule that remains pending in the D.C. Circuit today. A hearing on the full merits followed July 30; Lin's summary-judgment order, finding almost entirely for Anthropic, came out roughly four weeks later.
Anthropic's response was measured rather than triumphant. "We welcome the court's ruling that this supply chain risk designation was unlawful," a spokesperson told reporters. "We remain focused on working productively with the government to harness AI for our national security so all Americans benefit from this technology." The Pentagon did not immediately respond to requests for comment on the ruling.
The stakes were never just reputational. Anthropic executives have said the designation could have cost the company billions of dollars in lost business -- a blacklisted status with a federal government that is itself one of the largest single buyers of enterprise AI, across intelligence, defense and civilian agencies. (This newsroom previously reported that Anthropic's own March 9 court declaration cited revenue "exceeding $5 billion to date," separate from its roughly $19 billion reported annualized run-rate around the same period -- the same filing this lawsuit produced.)
- Regains eligibility for the federal contracts the designation blocked -- unless a court stays the ruling pending appeal.
- Gain a precedent that refusing a Pentagon request on safety grounds isn't grounds for blacklisting -- but a government appeal could still narrow or reverse it.
- A federal judge's written finding that the designation was retaliatory, 'illegal and baseless,' not a good-faith security judgment.
- Left unresolved by this ruling, which is about retaliation, not about whether Claude should be usable for those purposes.
What the ruling doesn't do is settle the argument that started all of this. Judge Lin found the Pentagon's *process* unlawful -- that it punished Anthropic for saying no, rather than making a genuine security case for saying yes. It says nothing about whether the Pentagon's underlying request, waiving Claude's restrictions on mass surveillance and autonomous weapons, was itself reasonable. That fight, and the parallel D.C. Circuit case over a different rule aimed at the same outcome, is still open. For an administration that has already appealed one loss in this case, the likeliest next step is another one.
- A federal judge ruled August 27 that the Pentagon's Anthropic 'supply chain risk' designation was illegal retaliation.
- Judge Rita Lin's 59-page order found it aimed to punish Anthropic's public 'arrogance,' not a real security risk.
- The designation followed Anthropic's refusal to let Claude be used for mass surveillance or autonomous weapons.
- Anthropic says the designation could have cost it billions in lost federal business; it welcomed the ruling.
- Caveat: a separate, narrower case is still pending, and the Pentagon hasn't said whether it will appeal this ruling too.