D.C. Appeals Court Lets Pentagon Blacklist of Anthropic Stand For Now

The U.S. Court of Appeals for the D.C. Circuit refused on Wednesday to block the Pentagon’s decision to blacklist San Francisco–based AI lab Anthropic as a national‑security supply‑chain risk while litigation proceeds, even as a separate federal court in San Francisco has already ordered the Trump administration to remove similar labels. Anthropic sued in both courts last month, alleging the administration is unlawfully retaliating because the company has tried to restrict how its Claude chatbot can be used in fully autonomous weapons and domestic surveillance, while the White House has attacked Anthropic as a liberal company trying to dictate military policy. U.S. District Judge Rita Lin in San Francisco previously held that the administration overstepped by branding Anthropic a supply‑chain risk unqualified to work with defense contractors and directed the government to lift those stigmatizing designations, which it has begun to do according to new filings. The D.C. Circuit acknowledged Anthropic would likely suffer some irreparable harm but said the financial impact was not yet clear enough to justify its own emergency order, and set a fuller hearing for May 19. Tech‑industry groups warn the dueling rulings and shifting labels are creating serious uncertainty for U.S. AI firms vying with OpenAI and Google for military and government work, and raising concern that "national security" designations can be turned into a political cudgel against companies that resist certain Pentagon uses of their systems.