The Ministry of National Defense (War Department) signed a classified-use agreement with major artificial intelligence (AI) companies. Anthropic, which has barred the use of its AI for surveillance inside the United States or for autonomous lethal weapons, was excluded from the agreement.
On the 1st (local time), the Ministry of National Defense said it signed agreements with seven corporations: SpaceX, OpenAI, Google, Nvidia, Rebellion, Microsoft, and Amazon Web Services.
The Ministry of National Defense said it would bring these corporations' advanced AI technologies onto the ministry's classified networks for lawful operational use, adding that the agreement will accelerate the process of transforming the U.S. military into an AI-first fighting force and strengthen the ability to maintain decision-making superiority for our warfighters across all domains of the battlefield.
It also said it will continue to build an architecture that prevents vendor lock-in and ensures long-term flexibility for the joint force.
While many major AI corporations were included in the agreement, Anthropic was left out. Analysts said the company's policy of limiting the military uses of its AI likely played a role. Earlier, Anthropic clashed with the Ministry of National Defense after saying its models should not be used for large-scale surveillance targeting Americans or for fully autonomous weapons. According to the New York Times (NYT), all corporations that signed the agreement that day agreed to let the ministry use their technologies for "all lawful purposes."
The NYT said ministry officials expect the agreement to increase pressure on Anthropic to withdraw its existing stance. After releasing its new AI model, "Mythos," last month, Anthropic has continued talks with the U.S. government even amid litigation with the Ministry of National Defense.
Some government agencies are already reported to have obtained access to Mythos and are using it. President Donald Trump, before departing for an event in Florida that day, briefly told reporters that he had a "good meeting" when asked about a meeting between Anthropic and Director Wiles.
Emil Michael, the Ministry of National Defense's chief technology officer (CTO) and Vice Minister for research and engineering, appeared on CNBC that day and said, "Because of Anthropic's supply chain risk issues, we gave each office under the ministry a six-month grace period to transition to other systems," but added, "However, the Mythos issue being handled across the government, not just at the ministry level, is a separate national security matter."
He added, "Because that model has specialized capabilities to find and block cyber vulnerabilities, we need to use it to further harden our networks."