THE RULING
● Who: U.S. District Judge Rita Lin, Northern District of California, in a 59-page order issued Thursday.
● What: The supply-chain risk designation violated the First Amendment and Fifth Amendment due process. Government ordered to rescind all directives.
● The line being quoted: the empty invocation of national security is not a blank check to punish and retaliate against government critics.
● Not over: a second Anthropic case remains pending in Washington D.C., where an appeals court ruled against the company in April.
How this started
Anthropic was the only major AI company working with the Pentagon on classified systems. Negotiations broke down over contract language: the Defense Department wanted terms permitting Claude to be used for any lawful purpose, and Anthropic would not agree.
The two sticking points were domestic surveillance and autonomous weapons. Anthropic's position is that AI models are not reliable enough to be trusted with autonomous weapons, and that domestic surveillance violates rights. The Pentagon's position is that private companies should not be able to constrain military action.
Both are coherent positions and the disagreement is genuine. What followed is what the court ruled on.
After talks collapsed, Defense Secretary Pete Hegseth designated Anthropic a national security supply-chain risk — a label normally applied to companies that might expose military systems to infiltration or sabotage by foreign adversaries. It was the first time a U.S. company had been publicly designated under that statute.
What the judge found
| Finding |
Detail |
| First Amendment |
The actions constituted unlawful retaliation for criticising the government |
| Fifth Amendment |
Liberty interests stripped without adequate notice or a meaningful chance to respond |
| The designation itself |
Described as arbitrary and capricious, and as illegal and baseless |
| The government's evidence |
Submissions defending the decision were characterised as slim |
| The motive |
A desire to make a public example of Anthropic for its arrogance in criticising the government |
| Remedy |
Government ordered to rescind all directives issued against the company |
THE CONTRADICTION THAT DECIDED IT
Lin pointed to an inconsistency in the government's own conduct. Hegseth had earlier threatened to invoke the Defense Production Act against Anthropic — a mechanism used to compel production from suppliers deemed essential.
You cannot simultaneously argue a company is so essential to national security that you might compel it to supply you, and that it is a security threat to be cut off. One of those had to be untrue.
The judge was explicit that this is not a ruling about vendor choice. The Defense Department remains free to select whichever AI vendor it wants, and the suits would not require the Pentagon to restart work with Anthropic. What the court found unlawful was the mechanism used — a designation reserved for foreign sabotage risk, applied to a domestic company over a contract disagreement.
It is not settled
A second Anthropic case is pending before the federal appeals court in Washington D.C., over a separate designation that could exclude the company from civilian government contracts.
That venue has already gone the other way. In April the D.C. Circuit rejected Anthropic's request to block the Department of War from blacklisting its technology, reasoning that the balance favoured the government — a contained financial risk to one private company against judicial management of how the department secures AI capability during an active military conflict.
So two federal courts have now reached opposite conclusions on closely related questions. That is the kind of split that tends to travel upward.
Why it matters beyond one company
Strip out the parties and the question is whether a supplier can decline specific uses of its product without being labelled a threat for doing so.
| If you are... |
Why this matters |
| An AI vendor with usage restrictions |
This is the first ruling testing whether declining a use case can be treated as a security risk |
| Buying AI for government work |
Any lawful purpose clauses are now visibly contested territory in procurement |
| Building on Claude |
Nothing changes commercially. Anthropic reported its first operating profit in Q2 and the blacklisting did not slow it |
| Watching the IPO timeline |
Removes a disclosure item, though the D.C. case remains outstanding |
Anthropic executives have said the designation could cost billions in lost business and reputational harm. The company's stated aim in litigating was to return to the position it held before the designation, not to compel the Pentagon to buy from it.
FAQ
What did the judge rule?
U.S. District Judge Rita Lin found that the Pentagon's designation of Anthropic as a national security supply-chain risk violated the First Amendment through unlawful retaliation and the Fifth Amendment through inadequate due process, and ordered the government to rescind all directives against the company.
Why was Anthropic blacklisted?
After contract negotiations broke down over the Defense Department seeking language permitting use of Claude for any lawful purpose. Anthropic declined over domestic surveillance and autonomous weapons, and was subsequently designated a supply-chain risk.
What is a supply-chain risk designation?
A label under a procurement statute aimed at companies that could expose military systems to infiltration or sabotage by foreign adversaries. This was the first time it was publicly applied to a U.S. company.
Is the dispute over?
No. A second Anthropic case is pending before the federal appeals court in Washington D.C. over a separate designation, and that court ruled against Anthropic on a related request in April.
Does the Pentagon have to use Claude now?
No. The judge was explicit that the Defense Department can select whichever AI vendor it prefers. What was struck down was the designation and the directives, not the procurement choice.
Does this affect Claude availability?
No. Commercial and consumer access is unaffected and always was. This concerns federal contracting.