Pentagon taps Nvidia, Microsoft, AWS for classified AI

The Pentagon is pushing artificial intelligence deeper into its most sensitive systems, signing deals with Nvidia, Microsoft, and AWS to deploy AI on classified networks.

The move signals more than a technical upgrade. It shows the Department of Defense wants broader control over how it buys and uses AI after a high-profile dispute with Anthropic over the terms governing its models. Reports indicate the department has doubled down on a diversification strategy, spreading its bets across major vendors instead of leaning too heavily on any single provider.

The Pentagon’s latest AI push is also a procurement story: more vendors, more leverage, and fewer choke points.

That matters because classified environments impose tougher demands than ordinary enterprise systems. Bringing AI into those networks requires not just powerful models and chips, but trusted cloud infrastructure, security controls, and operational support that can function under strict government rules. Nvidia, Microsoft, and AWS each occupy a different piece of that stack, giving the Pentagon a mix of computing power, software, and cloud reach.

The decision also reflects a bigger tension running through the AI market. Governments want cutting-edge tools, but they do not want to hand strategic dependence to one company or accept restrictive terms that could limit how those tools operate in mission-critical settings. Sources suggest the Defense Department now sees supplier diversity as a security feature, not just a contracting preference.

What comes next will shape how quickly AI moves from pilots to real-world defense operations. The central question is no longer whether the Pentagon will use AI in classified environments, but how it will manage risk, oversight, and vendor power as that effort scales. If this strategy holds, the department could set a template for how other agencies buy AI when the stakes extend far beyond efficiency.