Full Breakdown
Pentagon Secures AI Deployments on Classified Networks
5/2/2026, 5:21:56 AM
Core Deal: AI on Classified Networks
The Department of Defense announced agreements to place AI models on its Impact Level 6 (secret) and Impact Level 7 (top-secret) networks. Contracts cover SpaceX, OpenAI, Google, Nvidia, Reflection AI, Microsoft, Amazon Web Services, and later Oracle.
Companies and Anthropic Exclusion
Anthropic, only AI on classified systems, was excluded after the Pentagon labeled it a “supply-chain risk” for refusing to remove safety guardrails on autonomous weapons and surveillance.
Timeline and Scale
OpenAI signed a classified-network deal in February 2026, Google in April, and the list was released on May 1 2026 with Oracle added later. The contracts give AI access to networks used by over 1.3 million DoD personnel via GenAI.mil.
Intended Benefits
The Pentagon says AI will “streamline data synthesis, elevate situational understanding, and augment warfighter decision-making in complex operational environments,” to achieve decision superiority and logistics, targeting, and intelligence cycles.
Official Statements
Release said the tools are for “lawful operational use” and that the deals “accelerate the transformation toward establishing US military as an AI-first fighting force.” Hegseth warned Anthropic would not accept, likening to “Boeing giving us airplanes and telling us who we can shoot at.” Michael said partners needed to avoid single-supplier reliance.
Opposition and Legal Issues
Anthropic says use of Claude could enable autonomous weapons and surveillance, rejecting the “all lawful purposes” clause. Civil-rights groups raise privacy concerns; the NSA’s continued use of Anthropic’s Mythos model underscores oversight gaps.
Conflicting Reports
Sources differ on the firm count—most list seven, later updates add Oracle as an eighth. Pentagon has not disclosed schedules, payment terms, or oversight mechanisms for “lawful operational use,” nor integration plans for warfighting systems.
Verbatim Quotes
- “What we’ve learned since we started this effort at the Department of War is that it’s irresponsible to be reliant on any one partner, and we learned that that one partner didn’t really want to work with us in the way we wanted to work with them,” — Emil Michael, Pentagon CTO
- “Boeing giving us airplanes and telling us who we can shoot at.” — Pete Hegseth, Defense Secretary
- “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,” — Pentagon statement
- “There’s a phenomenon called automation bias, where people can be prone to assume that machines work better than they actually do,” — Helen Toner, interim executive director, Georgetown University’s Center for Security and Emerging Technology
Next Steps
Acting Principal Deputy Chief Digital and AI Officer Andrew Mapes said models will be deployed on classified networks within months. Ongoing litigation with Anthropic and congressional scrutiny of “lawful operational use” suggest further policy reviews before full rollout.
