More than half of the Pentagon’s workforce is already using generative AI at work. That’s not a pilot program. That’s adoption. According to TechCrunch, the Department of Defense has added ChatGPT Mil and Grok for Government to GenAI.mil, its centralized secure AI portal that now reaches all 3 million civilian and military personnel.
GenAI.mil launched last year with Google Gemini and was built specifically to let DoD employees access commercial AI models without routing sensitive data through consumer channels. The platform is exempt from the data collection baked into standard consumer products. That distinction matters a lot at this scale.
The two new additions serve different purposes. ChatGPT Mil, built through OpenAI’s Government program, targets routine unclassified work: documents, admin tasks, logistics, planning, and policy. Grok for Government, running on SpaceX’s Starshield satellite network, is positioned in harder military terms. The DoD says it will help personnel “execute missions faster and with greater precision” across contexts from supply chain management to acquisition research.
But the bigger story here is who’s missing. Anthropic’s Claude is not on the platform. The Trump administration labeled Anthropic a supply-chain risk after the company refused to give the Pentagon unrestricted access to its models and pushed for safety guardrails. That dispute is now in court. So the Pentagon’s model roster reflects not just technical preference, but active political friction in the AI industry.
The DoD has also struck separate deals with AWS, Microsoft, Nvidia, and Reflection AI as part of its wider push to build AI capacity. The pace of these partnerships signals that military AI is no longer a future concern. It’s procurement happening right now.
