OpenAI Reaches Agreement with U.S. Department of War to Deploy AI on Classified Networks

Agreement to Deploy AI on Classified Networks

OpenAI CEO Sam Altman announced on February 28, 2026 via X: "Tonight, we reached an agreement with the Department of War to deploy our models in their classified network." The announcement came hours after the Trump administration banned rival Anthropic from federal systems, making it a pivotal moment in AI's relationship with U.S. defense.

Core AI Safety Principles

Altman emphasized that two of OpenAI's most important safety principles are embedded in the agreement: a prohibition on domestic mass surveillance and the requirement for human responsibility in the use of force, including for autonomous weapon systems. The Department of War agreed to these principles and reflected them in law, policy, and the contract itself.

Technical Safeguards and Deployment

OpenAI will build technical safeguards to ensure models behave as intended and will deploy Field Deployment Engineers (FDEs) to support model safety on the ground. All deployment will occur exclusively on cloud networks.

Call for Industry-Wide Standards

Altman urged the Department of War to offer these same terms to all AI companies: "We are asking the DoW to offer these same terms to all AI companies, which in our opinion we think everyone should be willing to accept." This agreement sets a notable precedent for how AI companies can maintain safety principles while partnering with government defense agencies.