Anthropic blocks Trump's military AI blacklist

What happens when an AI lab refuses to let the military use its models for autonomous warfare? If you are Anthropic, the government tries to ban you from the country’s supply chain.

Anthropic just notched a major legal victory against the Trump administration. On Thursday, U.S. District Judge Rita F. Lin issued a preliminary injunction halting the government’s recent moves to blacklist the AI giant.

Drawing lines against the Pentagon

The standoff began last month over Anthropic’s acceptable use policy. When negotiating military adoption of its software, Anthropic insisted on hard limits: its AI could not be used to power fully autonomous weapons systems or conduct mass surveillance.

The Defense Department pushed back, and the dispute escalated fast. The Trump administration slapped Anthropic with a “supply-chain risk” tag—a designation typically reserved for foreign adversaries like Huawei—and President Trump ordered all federal agencies to cut ties with the company.

The White House publicly branded Anthropic as a radical-left, woke company that was sabotaging national security. Anthropic CEO Dario Amodei fired back, calling the actions retaliatory and punitive, before taking the Pentagon and Defense Secretary Pete Hegseth to federal court.

The judge calls it out

Judge Lin did not mince words during the proceedings, noting that the administration's actions looked like an attempt to cripple Anthropic.

In her ruling, Judge Lin found that the government trampled on Anthropic’s free-speech protections by punishing the company for its safety stance. She ordered the administration to immediately rescind the risk label and back off the federal ban.

The court also put the government on a tight leash, demanding a compliance report by April 6 detailing how officials are following the order.

Following the decision, Anthropic expressed gratitude for the swift ruling and said it was pleased the court found it likely to succeed on the merits. The company reiterated that it still wants to work productively with the government on safe, reliable AI.

Why it matters

This ruling gives enterprise customers and developers who rely on Claude a much-needed breather. When the White House issued its directive, federal contractors scrambled to figure out whether keeping Anthropic’s API keys could get them blacklisted from government work. That immediate threat is now paused.

More broadly, this is a watershed moment for AI governance. Anthropic proved that an AI company can stand by its core safety red lines without folding to executive pressure. If the government could weaponize national security designations just because a vendor refuses to build autonomous weapons, every commercial AI safety policy in Silicon Valley would become meaningless overnight.

In our view, weaponizing a foreign-actor blacklist against a domestic AI leader was a reckless overreach, and the judge saw right through it. But this injunction is only a temporary shield. The legal fight is far from over, and navigating relationships with a hostile administration will be Anthropic's toughest political test yet.