Pentagon Taps Nvidia, Microsoft And AWS To Bring AI To Classified Networks

The U.S. Department of Defense has signed new agreements with Nvidia, Microsoft, Amazon Web Services and Reflection AI that will let the Pentagon deploy their AI hardware and models directly on its classified networks for what it describes as “lawful operational use.”

The move follows earlier arrangements with Google, SpaceX and OpenAI, and is part of a broader push to make the U.S. military what the department calls an “AI-first fighting force.” In a statement, the Pentagon said these deals are meant to strengthen service members’ ability to maintain “decision superiority across all domains of warfare.”

The latest contracts land as the Pentagon is actively widening its base of AI suppliers following a high-profile clash with Anthropic over how its models can be used. According to the Defense Department, it is “accelerating” efforts to diversify AI vendors and avoid depending too heavily on any single provider.

The dispute centres on usage terms. The Pentagon sought unrestricted use of Anthropic’s AI tools, while the company insisted on limits designed to keep its technology from being used for domestic mass surveillance and autonomous weapons. The two sides are currently in court over the disagreement.

In March, Anthropic won an injunction blocking the Pentagon from labelling the company a “supply-chain risk,” a designation that would have had significant consequences for the AI firm. That legal fight is ongoing.

Against that backdrop, the Defense Department is emphasizing a multi-vendor approach. “The Department will continue to build an architecture that prevents AI vendor lock-in and ensures long-term flexibility for the Joint Force,” its statement said. It added that access to “a diverse suite of AI capabilities from across the resilient American technology stack” is intended to give military personnel the tools to act “with confidence” while defending against threats.

Under the new agreements, the Pentagon says it will deploy the companies’ AI hardware and models on its Impact Level 6 (IL6) and Impact Level 7 (IL7) environments. These designations cover some of the government’s highest security categories for data and systems tied to national security.

IL6 and IL7 environments are subject to strict protections, including physical security, tight access controls and auditing requirements. By placing AI systems in these zones, the Defense Department aims to:

• Streamline data synthesis
• Improve situational awareness
• Augment warfighter decision-making

While the department frames the new deals as focused on operational use in classified contexts, it is already running a separate, large-scale AI service for its workforce. The Pentagon said more than 1.3 million Defense Department personnel have used its GenAI.mil platform, a secure enterprise environment for generative AI.

GenAI.mil provides access to large language models and other AI tools within government-approved cloud setups. The Pentagon describes it as primarily geared toward unclassified work such as research, drafting documents and analysing data, offering a way for staff to use generative AI inside stricter security and compliance boundaries than public tools typically provide.

By pairing that broad, unclassified platform with newly authorized AI capabilities on IL6 and IL7 networks, the Defense Department is positioning AI as an integral layer across both its day-to-day knowledge work and its most sensitive operations.