The Double Ruling Complicating the Anthropic Case: Judges Divided

A federal appellate court in the United States has rejected the request for a stay submitted by Anthropic against the decision of the Department of Defense (DoD) to place it on the blacklist of suppliers deemed a national security risk. The ruling, issued by a federal court in Washington D.C., denies Anthropic the possibility of temporarily suspending the DoD's measure, which in early March classified the company as a "supply chain risk." According to the Department, the use of technologies developed by the company - including the Claude model - would pose a potential threat to national security.

In justifying the rejection, the judges highlighted a balancing of interests in favor of the government: on one hand, a limited economic damage for a single private company, and on the other hand, the management of critical technologies in the military context during an active conflict.

This ruling comes alongside a previous decision, issued by a federal court in San Francisco, which had partially granted Anthropic's requests. In that case, the judge had imposed a preliminary injunction preventing the administration led by Donald Trump from banning the use of the Claude model by federal agencies.

The result is a hybrid situation: Anthropic remains excluded from contracts with the DoD, while it can continue to collaborate with other government agencies. Defense contractors, however, cannot use Claude in projects related to the Pentagon.

Anthropic has contested the Department's decision, calling it unconstitutional retaliation, arbitrary, and not in accordance with legal procedures. The company has also argued that the measure limits its freedom of speech.

The court recognized a potential irreparable harm to the company but considered that such impact is primarily economic and not related to a real compression of free speech. On the opposing side, the Department of Justice - led temporarily by Todd Blanche - described the decision as a victory for military readiness, reiterating that operational control over strategic technologies belongs solely to government authorities.

At the heart of the dispute is the lack of agreement between Anthropic and the Pentagon regarding the use of AI models. The DoD reportedly requested extensive access to the technologies for any legal purpose, while the company sought guarantees excluding uses in autonomous weapons or mass surveillance systems. The breakdown of negotiations - which occurred after a $200 million contract was signed the previous July - quickly led to legal confrontation. Particularly noteworthy is the fact that Anthropic is the first U.S. company to receive the designation of "supply chain risk," historically reserved for foreign entities deemed hostile.

The conflicting judicial decisions suggest an unstable regulatory landscape, with potential repercussions for the entire AI ecosystem, especially for companies involved in government and military projects.