Pentagon approves eight firms’ AI for classified networks

The Department of Defense on Friday signed agreements with eight technology firms — SpaceX, OpenAI, Google, Nvidia, Reflection, Microsoft, Oracle and Amazon Web Services — certifying their AI systems to run on classified networks at Impact Levels 6 and 7. The arrangements allow those systems to operate on infrastructure that enforces strict access controls, network isolation and security-clearance requirements.

Impact Level 6 covers secret-level information. Impact Level 7 applies to intelligence systems that handle highly restricted national security data. Under the certified regime, AI tools must run on isolated, classified networks, with controlled connections to unclassified systems and management by cleared personnel.

The agreements expand the department’s internal AI platform, GenAI.mil, launched in December with Google’s Gemini model. The Defense Department reported that more than 1.3 million personnel used the platform in its first five months to generate tens of millions of prompts and deploy hundreds of thousands of AI agents. The department described GenAI.mil as a multi-provider environment intended to support data analysis, situational awareness and decision-making across military operations.

The approvals follow a series of recent AI contracts. In March 2025 the Pentagon hired Scale AI to build the Thunderforge planning system. The department has also made arrangements to incorporate ChatGPT and rival models such as Grok. The announcement did not disclose contract values.

In its 2026 budget request, the Defense Department sought $961.6 billion in total funding, including $33.7 billion for science and technology and autonomous systems. The department noted the new agreements build on existing federal investments in cloud computing, data infrastructure and AI.

The Pentagon said, ‘These agreements accelerate the transformation toward establishing the United States military as an AI-first fighting force and will strengthen our warfighters’ ability to maintain decision superiority across all domains of warfare.’

Tim Barrett, an Amazon Web Services spokesperson, said the company plans to expand support for U.S. military operations as the Defense Department proceeds with classified AI deployments. He added that AWS has been supporting the military for more than a decade and looks forward to continuing to help modernize defense capabilities.

The approvals come amid federal reviews of security and supply-chain risks tied to some AI providers. Officials are considering guidance that could restore broader federal access to a firm currently in dispute with the Defense Department over model use. The National Security Agency has begun deploying that firm’s newer model on classified networks while negotiations continue.

The Defense Department said the certified operating framework relies on existing federal cloud and data investments and is intended to use multiple vendors rather than a single provider.