A federal judge in Northern California ruled that the Trump administration illegally blacklisted Anthropic, delivering a major setback to Washington's effort to punish the AI company for setting safety boundaries on its models.
The ruling finds that the Pentagon's designation of Anthropic as a "supply chain risk" was arbitrary, retaliatory, and violated the company's First and Fifth Amendment rights. The judge's 59-page opinion said internal government records showed the move was motivated by frustration over Anthropic's public refusal to remove red lines on autonomous weapons and mass surveillance, rather than any genuine security concern.
The decision immediately vacates the blacklisting order, which had barred Anthropic from federal contracts and instructed Pentagon contractors to stop working with the company. While the administration can appeal, the ruling preserves the company's ability to compete for government AI work for now, at a moment when frontier AI labs are becoming central to national security strategy.
The case carries broader implications for how far Washington can go in coercing AI companies over safety, ethics, or speech. Rather than settling the question, the ruling turns the dispute into a longer-running legal and political fight over who gets to set the rules for the most powerful AI systems.
Image source: i.ibb.co