The filing came one week after U.S. District Judge Rita F. Lin granted Anthropic a preliminary injunction, calling the government's actions "broad punitive measures" that could "cripple Anthropic" and describing the Pentagon's application of a military designation against an American company as likely "arbitrary, capricious" and in violation of the law.

Defense Undersecretary Emil Michael responded to Lin's ruling with a single line.

"This ruling is a disgrace and restricts Secretary Hegseth's operational flexibility."

• Emil Michael, Undersecretary of Defense for Research and Engineering

The Ninth Circuit has set an April 30 deadline for the Justice Department to file its arguments for why Lin's order should be overturned.

How the Standoff Got Here

The dispute traces back to a contract negotiation that broke down on February 27 after the Pentagon refused to accept two conditions Anthropic requires of all government customers: that Claude won't be used in autonomous weapons systems and won't be rolled out for domestic mass surveillance.

The Contract That Collapsed

Anthropic had been operating under a $200 million contract with the Pentagon. When renewal talks began, Defense Secretary Pete Hegseth's team demanded that any new agreement include no such restrictions - they wanted unrestricted access to Claude for "all lawful purposes." Anthropic refused. Hegseth signed a supply-chain risk designation under Section 3252 of Title 10, a rare military authority previously applied only to foreign adversaries. Trump then posted on Truth Social directing "EVERY Federal Agency in the United States Government to IMMEDIATELY CEASE all use of Anthropic's technology."

A Label Built for Foreign Adversaries

The Section 3252 mechanism has a specific origin: it exists to cut the military off from suppliers linked to countries like China and Russia. The statute had never been used against a U.S. company before. Lin's 43-page ruling described applying it to Anthropic as "likely both contrary to law and arbitrary and capricious," noting that "nothing in the governing statute supports the Orwellian notion that an American company may be branded a potential adversary and saboteur of the U.S. for expressing disagreement with the government."

Anthropic filed suit in California federal court on March 9, with a parallel case in D.C. Lin's March 26 injunction blocked both the supply-chain designation and the Trump ban, giving Anthropic an early win - though the court was clear that a preliminary injunction is not a final verdict.

What the Appeal Could Change

The critical variable is whether a Ninth Circuit panel agrees to issue a stay - a temporary suspension of Lin's order while the appeal proceeds. The Trump administration routinely requests stays after adverse rulings, and the current federal judiciary has been receptive to national security arguments. A three-judge panel would decide whether the appeal is likely to succeed before ruling on any stay.

For Anthropic

The company is commercially exposed in ways Lin's injunction currently masks. Defense contractors were told after the February designation that they'd need to certify zero exposure to Anthropic products to keep their Pentagon work. A number had already begun removing Claude from their pipelines before Lin's ruling paused the designation. A successful DOJ appeal reverses that protection.

For Federal Agencies

Several agencies had begun transitioning off Claude and moving to OpenAI's models or xAI's Grok before Lin blocked the ban. If the Ninth Circuit grants a stay, those transitions restart without warning.

For the Broader Industry

The case attracted an unusual coalition of friends-of-the-court filings: Microsoft, multiple tech trade groups, retired military leaders, and - in a turn nobody predicted - Catholic theologians. That breadth signals the industry sees this case as setting precedent on whether the government can use national security tools to punish companies for their AI safety policies.

Anthropic CEO Dario Amodei has held to two non-negotiable conditions across the dispute: no Claude deployment in fully autonomous weapons, and no domestic mass surveillance. Source: commons.wikimedia.org

What Happens Next

The Justice Department has until April 30 to file its brief at the Ninth Circuit. Anthropic then gets time to respond before a panel decides whether to grant a stay. If the panel stays the injunction, the federal ban resumes right away. If it declines, the case proceeds to a full appeal of Lin's ruling - a process that could take many months before any final resolution.

The D.C. parallel case adds another front. Anthropic filed in both jurisdictions to hedge against different outcomes at the district level, and the two cases may produce conflicting decisions before either reaches an appellate court.

On April 2, the same day the DOJ filed its appeal, California Governor Gavin Newsom signed an executive order giving California the authority to independently review any federal supply-chain risk designation before deciding whether to do business with the named company. It's a direct counter to the federal action against Anthropic, and a signal that state-level AI procurement has become a separate battleground in the same fight.

The James R. Browning U.S. Court of Appeals Building in San Francisco will be the next arena. The Ninth Circuit's decision on whether to grant a stay - before any ruling on the merits - may be the most consequential near-term decision in the case. Source: commons.wikimedia.org

The core legal question Lin left open is whether a Section 3252 designation can lawfully be used against an American company that refuses to waive its AI safety policies on ethical grounds. The Ninth Circuit will be the first appellate court to weigh in. Whether it does so before or after granting a stay may determine whether federal agencies spend the rest of the spring with or without Claude.

Sources:

• ABC News: Trump Administration Appeals Ruling Blocked Pentagon Action Against Anthropic
• Broadband Breakfast: Trump Administration Appeals Ruling That Blocked Pentagon Action Against Anthropic
• TechCrunch: Anthropic Sues Defense Department Over Supply-Chain Risk Designation
• DefenseScoop: Pentagon CTO Urges Anthropic to Cross the Rubicon on Military AI
• Calmatters: California Will Decide How Risky Its AI Startups Are, Newsom Says