Round 2, Anthropic

TL;DR: A federal judge temporarily blocked the Pentagon’s blacklisting of Anthropic, calling it “classic illegal First Amendment retaliation.” The government’s own documents undercut its case, and Anthropic argued it has already suffered damages in the hundreds of millions because of this whole fiasco. But this preliminary injunction isn’t the last word—and who ultimately wins is still very much up in the air.

What happened: Yesterday, a federal judge in California blocked the government from enforcing its blacklist of Anthropic using a supply chain risk label—at least for now. Judge Rita Lin wrote in her order that “nothing in the governing statute supports the Orwellian notion that an American company may be branded a potential adversary and saboteur of the US for expressing disagreement with the government.”

Much of the hearing centered on a post Defense Secretary Pete Hegseth made on X declaring that no military “contractor, supplier, or partner” could conduct “any commercial activity” with Anthropic. It took a federal judge to clarify that no, the designation doesn’t apply to a company whose only military connection is selling the Pentagon toilet paper. The government’s own lawyer conceded during the oral argument that the post had “absolutely no legal effect at all.”

A quick recap: In January, Hegseth issued a memo demanding all military AI vendors grant the Pentagon “any lawful use” of their technology. But Anthropic had two red lines—no mass domestic surveillance, no fully autonomous weapons. The dispute went public in late February, and the Pentagon slapped Anthropic with a supply chain risk designation reserved for foreign adversaries in early March.

The damage report: Anthropic argues that the ambiguity the government created with all of this has already cost the company $180 million in collapsed deals. Over 100 enterprise customers expressed “deep fear, confusion, and doubt” following the supply chain risk designation. Anthropic CFO Krishna Rao estimated the actions could cost it up to billions of dollars in 2026 revenue—noting that spooked customers drive away investors, and without investor capital, Anthropic can’t buy the compute it needs to stay in the AI race.

Threat or asset?: Court filings revealed that the day after designating Anthropic an “unacceptable national security threat” (but before it notified the company), Defense Under Secretary Emil Michael emailed Anthropic CEO Dario Amodei about contract terms, saying, “I think we are very close here.” Lin called this exchange “exceedingly difficult to square” with the government's simultaneous characterization of Anthropic as “hostile” and posing an “intolerable” risk.

Lin also flagged that Hegseth had told Anthropic he’d either label it a supply chain risk—a grave threat to national security—or invoke the Defense Production Act to compel it as essential to national security. She said this contradiction showed a “complete lack of prior notice or process” from the Pentagon.

The bottom line: One legal document filed in support of Anthropic described the Pentagon’s designation as “attempted corporate murder.” The judge’s reply during the hearing: “I don’t know if it’s murder, but it looks like an attempt to cripple Anthropic.” Lin put her ruling on hold for seven days so the government can challenge it, and a final verdict could be months out. —WK