DOJ Defends Pentagon's Supply-Chain Risk Label on Anthropic
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Justice Department attorneys filed a response in San Francisco federal court arguing the Pentagon's supply-chain risk designation of Anthropic does not violate the company's First Amendment rights, writing that "The First Amendment is not a license to unilaterally impose contract terms on the government"
- Judge Rita Lin scheduled a hearing for next Tuesday to decide whether to grant Anthropic's request to resume business with the Pentagon while the lawsuit proceeds
- Anthropic could lose "up to billions of dollars in expected revenue this year" if the supply-chain risk designation holds, the source reports
- The government argued DoD became concerned that Anthropic staff might "sabotage, maliciously introduce unwanted function, or otherwise subvert" national security systems, particularly if Anthropic's corporate "red lines" were crossed
- Claude is currently the only AI model cleared for use on the Defense Department's classified systems, but DoD is working to deploy alternatives from Google, OpenAI, and xAI "in the next few months"
- Anthropic believes its models should not be used to facilitate broad surveillance of Americans and are not yet reliable enough to power fully autonomous weapons
- Court briefs supporting Anthropic were filed by Microsoft, a federal employee labor union, AI researchers, and former military leaders — none were filed in support of the government
Why it matters: A Tuesday hearing in San Francisco will determine whether Anthropic can keep working with the Pentagon during litigation, with billions in expected revenue on the line. The Defense Department is simultaneously racing to replace Claude — currently the only AI cleared for its classified systems — with alternatives from Google, OpenAI, and xAI within months, meaning Anthropic could find itself locked out of defense work even if it ultimately prevails in court.
