Pentagon deploys military versions of ChatGPT and Grok to 3 million defense personnel

The U.S. Department of Defense has rolled out military-specific builds of two leading generative AI systems—OpenAI's ChatGPT and xAI's Grok—expanding access across its 3 million uniformed and civilian workforce. Both tools are now available through GenAI.mil, a centralized, security-hardened platform that was established to let Pentagon employees use cutting-edge AI models while keeping sensitive government data off commercial networks.

Expanded access through a secure platform#

GenAI.mil, which debuted last year with Google Gemini, serves as a firewall between Department of Defense personnel and the data-collection practices endemic to consumer AI products. By running these models through a dedicated government infrastructure, military staff gain the functionality of widely-used generative AI tools without funneling classified or sensitive information through standard consumer channels.

The platform has gained significant traction since launch. Over 1.7 million of the department's total 3 million eligible users have already registered on GenAI.mil, demonstrating strong institutional appetite for AI-assisted workflows.

Two different mandates#

ChatGPT Mil mirrors the commercial product's interface and capabilities—chat, file handling, projects, and custom GPT configurations. The Defense Department intends to use it for routine, unclassified administrative work including logistics, planning, policy drafting, and document-heavy tasks. New features are expected to roll out over time.

By contrast, Grok for Government (developed under SpaceX's Starshield AI program, which pairs the model with secure satellite infrastructure) is positioned in explicitly operational terms. According to the Pentagon, Grok will enable faster mission execution, improve knowledge sharing across teams, and support a range of military functions—from supply chain logistics to market research for acquisition professionals.

The Anthropic absence and industry partnerships#

The addition of both models notably excludes Anthropic's Claude, a leading generative AI system. The omission underscores lingering tensions between the Pentagon and the AI safety-focused company. The Trump administration designated Anthropic a supply-chain risk after it refused to grant the military unrestricted access to its models, insisting instead on safety guardrails. Anthropic is currently challenging that designation in court.