Regulation
AI Minute Newsroom
2026-08-28
A judge has thrown out the Pentagon's 'supply chain risk' label on Anthropic — invoking national security, she wrote, is not a blank cheque to punish critics
On Thursday 27 August 2026 US District Judge Rita Lin, of the Northern District of California, set aside the Defense Department's designation of Anthropic as a supply chain risk and issued a permanent injunction barring the government from giving effect to it or to the directive ordering federal agencies to stop using Claude. Lin found the government had retaliated against Anthropic for criticising the administration's contracting position in public, in violation of the First Amendment, and had stripped the company of protected liberty interests without adequate notice or a meaningful chance to respond, in violation of the Fifth Amendment's due process clause. She wrote that 'the empty invocation of national security is not a blank check to punish and retaliate against government critics'. The fight began over a roughly $200 million deployment of Claude on classified systems: Anthropic refused to drop two contractual restrictions — no fully autonomous lethal weapons, no domestic mass surveillance of Americans — talks collapsed in late February, and Defense Secretary Pete Hegseth applied the label. The same judge had blocked the measures temporarily on 26 March, saying then that nothing in the governing statute supports 'the Orwellian notion that an American company may be branded a potential adversary'. Anthropic said it was pleased the court had found the designation unlawful; the government can still appeal.
Why it mattersEvery frontier lab writes usage limits into its government contracts, and until yesterday the open question was whether refusing to drop them could get a company branded a security risk and cut off from the entire federal market. A court has now said no — and said it on free-speech grounds, not procurement grounds. That makes the ruling portable: it is about what happens when an administration uses a national-security label to answer criticism, which is a lever that reaches well past this one company.
✓ Verified · 4 sources
▶ Related video: Anthropic Sues the Pentagon Over Ban on Claude AI Tech | Vantage with Palki Sharma
Read in the app — free, in 9 languages
Related stories
The labs whose models broke out and hacked people have now co-signed a letter asking the world to hurry up and defend itself
2026-08-28Australia's charts now require a human to have written the song and sung the lead — the rule starts Friday
2026-08-27Meta will pay up to $18 billion over how its recommendation systems held teenagers — and a third of it only lands if YouTube and TikTok match
2026-08-27Bill Gates wants a class of jobs that AI is not allowed to take — childcare and jury duty to start, funded by a tax on tokens
2026-08-26The FTC's proposal: if a machine set your price by reading your data, the shop has to tell you which data it read
2026-08-26