US Court Rejects Anthropic’s Request for Removal From Security Threat List

Just a year ago, Anthropic was the darling of the U.S. military. The company was the first to deploy its models across the Department of Defense's classified networks and signed a contract worth $200 million. But then came a dispute with the Pentagon, and Anthropic found itself blacklisted. A federal appeals court in Washington, D.C., rejected Anthropic's request to temporarily suspend the supply-chain risk designation imposed on the company by the Pentagon. The three-judge panel ruled that the balance of interests favored the government, not the technology company.

The dispute with the Pentagon and the path to court

The dispute stems from a simple but fundamental disagreement. The Pentagon wanted complete freedom from Anthropic to use the Claude AI model "for all lawful purposes." Anthropic refused and set two red lines: Claude must not be used for mass surveillance of U.S. citizens and must not power fully autonomous lethal weapons.

Defense Secretary Pete Hegseth regarded this as an act of rebellion and, in late February 2026, designated Anthropic a national security supply-chain risk on the social network X. President Trump subsequently ordered all federal agencies to immediately stop using Anthropic's technology, with agencies such as the Department of Defense receiving a six-month transition period.

Anthropic argued that this was pure retaliation for its stance on artificial intelligence safety. The company filed two lawsuits in two different courts because the Pentagon had applied two separate legal designations under different laws.

While the Washington appeals court denied the request for a suspension, a federal court in California took the opposite approach. Judge Rita Lin issued a preliminary injunction on March 26 preventing the government from enforcing the ban on the use of Claude outside the Pentagon. The result is two conflicting rulings in effect at the same time:

• Anthropic may not enter into new contracts with the Department of Defense
• Pentagon contractors must certify that they do not use Claude when working for the military
• Other government agencies, however, may continue to use Anthropic
• The Pentagon itself will continue using Anthropic's products for another six months

Consequences of the court's decision for Anthropic

The three-judge appeals court panel was not unsympathetic. It acknowledged that Anthropic "will most likely suffer some irreparable harm," but characterized its interests as primarily financial. The company claimed that the designation could cause it to lose billions of dollars. The judges also rejected the argument that freedom of speech had been restricted. They wrote that Anthropic had not demonstrated that its speech had actually been restricted during the dispute.

On the other hand, the court granted Anthropic one important concession: expedited review. Oral arguments are scheduled for May 19. "We are grateful that the court recognized the need for a swift resolution, and we are confident that the courts will ultimately confirm that these designations are unlawful," an Anthropic spokesperson said.

Acting Attorney General Todd Blanche hailed it as a victory. He wrote on X: "Military authority and operational control belong to the commander in chief and the Department of War, not a technology company."

The dispute took on a personal dimension after the media published an internal letter from Anthropic CEO Dario Amodei. In the 1,600-word text, Amodei accused the Pentagon of punishing the company because it "did not pay dictatorial tribute to Trump," while OpenAI and its president, Greg Brockman, had contributed to Trump's campaign.

Amodei later apologized for the "tone" of the letter, but its political message was clear. And OpenAI? It did not hesitate. As soon as the Pentagon parted ways with Anthropic, Sam Altman's company immediately offered the government replacement AI services.

The irony of the whole affair is that Anthropic was the one trying to address safety concerns responsibly. But the appeals court judges gave precedence to another argument: the U.S. military is engaged in an active military conflict with Iran, and severing ties with a key supplier of AI technology at such a time poses an undue risk.

"On one side is the relatively limited financial risk to a single private company. On the other is judicial interference in how and through whom the Department of War obtains vital AI technology during an active military conflict," the judges wrote in their decision.

The supply-chain risk designation has historically been reserved for foreign adversaries such as China's Huawei. Anthropic is the first American company to receive it.

Sources: foxnews.com, thehill.com, reuters.com and cnbc.com