Pentagon Signs Contracts with Nvidia, Microsoft, and AWS to Deploy Classified AI

Pentagon's Three Major Contracts Bring Classified Networks into the AI Era

The U.S. Department of Defense (DOD) recently announced the signing of significant cooperation agreements with Nvidia, Microsoft, and Amazon Web Services (AWS), planning to comprehensively deploy artificial intelligence technologies across its classified network systems. This series of deals marks a critical step in the U.S. military's push toward AI militarization, while also reflecting a major strategic shift in the Pentagon's approach to AI vendors.

Three Tech Giants, Each with a Distinct Role

The three companies bring unique strengths to the AI domain. Nvidia, the world's leading GPU chipmaker, will provide the core computing infrastructure needed to power AI model operations for the DOD. Microsoft, leveraging its extensive experience with the Azure Government cloud platform, will handle cloud computing and AI service integration within classified environments. AWS, drawing on its deep expertise in government and intelligence sectors, will take on the deployment and operations of large-scale AI workloads.

The central goal of this three-way collaboration is to enable AI capabilities to truly penetrate the DOD's classified networks. For years, progress in introducing cutting-edge AI technologies to classified networks has been slow due to extremely stringent security compliance requirements. The signing of these agreements signals that the Pentagon is breaking through this bottleneck, accelerating the application of large language models, computer vision, and intelligent analytics to critical scenarios such as intelligence processing, operational planning, and logistics support.

Anthropic Dispute Sparks Vendor Diversification Strategy

Notably, a key driver behind these deals is the Pentagon's earlier dispute with Anthropic over AI model usage terms. Anthropic's usage policies for its AI model Claude reportedly included clauses restricting military applications, which directly conflicted with the DOD's requirements.

This dispute prompted the Pentagon to fundamentally reassess its AI procurement strategy. The risks of over-reliance on a single or limited number of AI vendors had become abundantly clear — any shift in a vendor's usage terms, ethical stance, or business strategy could introduce significant uncertainty into the DOD's AI deployment plans. Consequently, the DOD has "doubled down" on its AI vendor diversification strategy, establishing partnerships with multiple tech giants simultaneously to spread risk.

The U.S. Military AI Race Heats Up

The contract signings also reflect the broader acceleration of the AI race within the U.S. defense sector. In recent years, the Pentagon has steadily increased its AI investments, from the establishment of the Chief Digital and Artificial Intelligence Office (CDAO) in 2023 to a series of AI pilot programs, as the U.S. military systematically builds its AI warfighting capabilities.

At the same time, these deals have reignited ethical debates surrounding military AI applications. Anthropic's earlier stance represented the cautious attitude of some Silicon Valley companies toward AI militarization, while the active participation of Nvidia, Microsoft, and AWS demonstrates that, when faced with national security imperatives and substantial contracts, more tech companies are choosing to partner with the DOD.

From a geopolitical perspective, the AI competition between the U.S. and China is rapidly expanding from the civilian sector into the military domain. The Pentagon's move is not merely a technological upgrade but a strategically vital effort to ensure the United States maintains its lead in military AI.

Outlook: A Classified AI Ecosystem Takes Shape

As these contracts move into execution, the DOD's classified AI ecosystem will take shape at an accelerated pace. In the short term, Nvidia's computing power, combined with the cloud platform capabilities of Microsoft and AWS, will significantly enhance AI processing performance across classified networks. In the long term, the vendor diversification strategy will help the Pentagon build a more resilient and competitive AI supply chain.

For the global AI industry, the military sector is becoming an increasingly important market that cannot be ignored. The vendor strategy adjustments triggered by the Anthropic incident also serve as a reminder to every AI company: in the unique arena of defense, the willingness to cooperate matters just as much as technical capability.