Pentagon expands its AI arsenal with deals for Nvidia, Microsoft and AWS

The Pentagon is making another major push to integrate artificial intelligence across its operations. The Defense Department announced new deals with Nvidia, Microsoft, Amazon Web Services, and Reflection AI that will allow their AI technology and models to run on classified military networks for “lawful operational use.”

These partnerships add to existing agreements with Google, SpaceX, and OpenAI, showing how aggressively the military is pursuing AI capabilities. The Pentagon said the deals will “accelerate the transformation toward establishing the United States military as an AI-first fighting force” and strengthen warfighters’ decision-making abilities across all combat domains.

The timing reflects the military’s urgent need to diversify its AI suppliers after a messy legal battle with Anthropic over usage restrictions. The Pentagon wanted unrestricted access to Anthropic’s AI tools, while the company demanded guardrails to prevent their technology from being used for domestic mass surveillance and autonomous weapons. That dispute is still playing out in court, with Anthropic winning an injunction in March against being labeled a “supply-chain risk.”

The new partnerships matter because they show the Pentagon is serious about avoiding vendor lock-in while building what it calls a “resilient American technology stack.” Military leaders have learned from decades of being dependent on single contractors for critical systems. Having multiple AI providers means better negotiating power and backup options if any relationship goes sour.

The companies’ AI systems will run on Impact Level 6 and Impact Level 7 environments – high-security classifications for data critical to national security. These systems need physical protection, strict access controls, and regular audits. The AI will help with:

• Streamlining data synthesis from multiple sources
• Improving situational awareness in complex environments
• Supporting faster and better decision-making by military personnel

The Pentagon already has more than 1.3 million personnel using GenAI.mil, its secure platform for generative AI tools. This system provides access to large language models and other AI capabilities within government-approved cloud environments. It currently handles non-classified work like research, document drafting, and data analysis.

This expansion into classified networks represents a significant escalation. Military AI applications on classified systems could include real-time threat analysis, mission planning, and intelligence processing – capabilities that could provide major advantages in modern warfare where information speed often determines outcomes.

The broader context is a global AI arms race where China and other adversaries are also rushing to militarize artificial intelligence. The Pentagon’s vendor diversification strategy aims to tap into America’s tech industry strengths while maintaining the flexibility to adapt as AI technology rapidly evolves.