On August 31, 2026, the Department of Defense added OpenAI's ChatGPT Mil and xAI's Grok for Government to GenAI.mil, the Pentagon's centralized portal for generative AI tools. The move gives up to 3 million military and civilian personnel access to custom versions of the commercial models inside a secure environment that handles controlled unclassified information.
GenAI.mil launched on December 9, 2025 with Google's Gemini as its first model. The platform has since onboarded more than 1.7 million unique users, according to the department. Both new tools are built to keep sensitive government data out of ordinary consumer channels and exempt the military versions from the data collection typical of the public products.
What's new
- ChatGPT Mil — Developed through OpenAI's OpenAI for Government program. The department says it will focus on chat, files, projects, and custom GPTs, supporting document-heavy unclassified work such as administrative tasks, logistics, planning, and policy. Additional features are slated to arrive over time.
- Grok for Government — Delivered via SpaceX's Starshield AI, the secure satellite network built on Starlink technology. The Pentagon framed its capabilities in operational terms: "immediate productivity gains, stronger knowledge continuity and more secure and efficient collaboration" for warfighters, with use cases ranging from market research analysis for acquisition professionals to supply-chain management for logisticians.
Work on adapting ChatGPT for controlled unclassified information began in February 2026. OpenAI officials had indicated a deployment target of early July, but the final release slipped to late August. A defense official described the preparation as "arduous," requiring separate compute environments, user tracking, data compilation, and security reviews by both the Defense Information Systems Agency and the National Security Agency.
Why it matters
The expansion signals the Pentagon's strategy of offering personnel a choice of frontier models for different tasks rather than standardizing on a single provider. A defense official said Gemini is preferred for search and scraping, while ChatGPT suits text-based work, and the department wants users to pick the tool that fits the job. The additions also underscore the ongoing absence of Anthropic's Claude, which the Trump administration labeled a supply-chain risk after the company declined to provide unrestricted access and insisted on safety guardrails — a designation Anthropic is contesting in court.
Our take
By fielding multiple commercial models on a single accredited portal, the Pentagon is treating generative AI as interchangeable infrastructure rather than a strategic bet on one lab. That approach reduces vendor lock-in but shifts the burden to the department to maintain parallel security stacks, compute allocations, and model-update cadences — a sustainment challenge that will grow with each new integration.
