TL;DR:
- Judge Rita Lin ruled Thursday: The Pentagon's supply chain risk designation of Anthropic is blocked. The Trump administration's ban on federal agencies using Claude is halted.[1][2]
- The ruling: A 43-page opinion calling the government's actions "Orwellian" and "classic illegal First Amendment retaliation"[5]
- The finding: The Pentagon labeled Anthropic a threat because of its "hostile manner through the press," punishing speech, not security risk[6]
- What's next: Lin delayed implementation one week to allow government appeal. The government could take this to the Ninth Circuit.[4]
- The bottom line: A federal judge just said the government can't punish companies for publicly refusing to build surveillance tools[3]
The Ruling
U.S. District Judge Rita Lin didn't mince words.
"Nothing in the governing statute supports the Orwellian notion that an American company may be branded a potential adversary and saboteur of the U.S. for expressing disagreement with the government."[5]
That's from her 43-page ruling issued Thursday in San Francisco federal court. Lin granted Anthropic a preliminary injunction, blocking the Pentagon's supply chain risk designation and halting President Trump's directive banning federal agencies from using Anthropic's technology.[1][2]
The ruling came two days after a March 24 hearing where Lin pressed government lawyers to explain how a domestic policy dispute justified using a statute designed for foreign adversaries.[4]
First Amendment Retaliation
The judge found what Anthropic argued all along: the Pentagon punished the company for speaking up.
"Punishing Anthropic for bringing public scrutiny to the government's contracting position is classic illegal First Amendment retaliation," Lin wrote.[1][3]
The evidence? The Pentagon's own records. Lin cited internal documents showing the Defense Department designated Anthropic as a supply chain risk because of its "hostile manner through the press."[6]
That's the Pentagon admitting the designation came from bad press coverage, not actual security concerns.
When Anthropic CEO Dario Amodei said publicly that Claude wouldn't be used for mass domestic surveillance or fully autonomous weapons, he was exercising protected speech. The government's response was to label his company a national security threat alongside Huawei and ZTE.[2]
What This Means
The preliminary injunction freezes the status quo while the full case proceeds. Anthropic is no longer designated a supply chain risk. Federal agencies can resume using Claude, at least for now.
But Lin was clear about the limits. She delayed implementation for one week to let the government appeal. The administration could ask the Ninth Circuit to stay her order while they fight this out.
Still, Lin's ruling creates significant obstacles. She found Anthropic is "likely to succeed on the merits" of its claims, legal speak for "you're probably going to win this case." That makes it harder for the government to get her injunction reversed on appeal.
The Backstory
This started in February when the Pentagon terminated a $200 million contract with Anthropic over the company's AI safety policies.
The Pentagon wanted "unfettered access" to Claude across all lawful purposes. Anthropic wanted assurances its technology wouldn't be used for mass domestic surveillance or autonomous weapons systems without human oversight.
When negotiations broke down, President Trump issued an executive order on February 28 banning all federal agencies from using Anthropic's products. On March 3, the Pentagon invoked the supply chain risk statute, a tool designed to address foreign adversaries compromising U.S. systems.
Anthropic sued on March 9 in two courts. Microsoft filed an amicus brief warning this set a "dangerous precedent." Thirty-seven researchers from OpenAI and Google, plus 22 retired military officers, filed briefs backing Anthropic.
The Precedent
This is the first time the supply chain risk statute has been used against an American company over a policy dispute. The only previous use was against Switzerland-based Acronis AG, involving actual foreign ownership concerns.
Lin's ruling establishes that the statute can't be weaponized against domestic companies for their public statements. That matters for every tech company watching this case.
If the government can label you a national security threat for saying you won't build surveillance tools, every company with a government contract has to choose: keep quiet about your ethics, or risk everything.
Lin said no. The Constitution protects companies that speak up, even when the Pentagon doesn't like what they're saying.
What's Next
The government has a week to appeal before Lin's order takes effect. The case will continue regardless: this was just a preliminary injunction, not a final ruling.
Senator Elizabeth Warren's investigation into the Pentagon's Anthropic designation and OpenAI's new contract is ongoing. She's demanding answers about what safeguards were actually negotiated.
OpenAI, meanwhile, has a Pentagon contract with language that critics say includes loopholes for the exact surveillance Anthropic refused. The EFF called it "weasel words."
The bigger question: will Congress step in? Amodei himself said it's "Congress's job" to address AI surveillance risks. Right now, these battles are being fought in courtrooms and contract negotiations, not legislative chambers.
Sources
- CNBC: Anthropic wins preliminary injunction in DOD fight as judge cites 'First Amendment retaliation'
- CNN: Judge blocks Pentagon's effort to 'punish' Anthropic by labeling it a supply chain risk
- NPR: Judge temporarily blocks Trump administration's Anthropic ban
- Axios: Judge temporarily blocks Pentagon's ban on Anthropic
- Fortune: Federal judge temporarily halts Pentagon's 'Orwellian' ban on Anthropic's AI technology
- Washington Post: Judge blocks Pentagon order branding Anthropic a national security risk
Published: March 27, 2026