The Pentagon has officially brought two of the biggest commercial AI models into military operations: OpenAI’s ChatGPT Mil and xAI’s Grok for Government (distributed via SpaceX’s Starshield AI). Both have joined GenAI.mil, the U.S. Department of Defense’s centralized AI portal, which has already featured Google Gemini since launching last year.
GenAI.mil is designed to give Department of Defense personnel access to top commercial AI models without exposing sensitive government data to standard consumer channels. The crucial factor: the military versions are exempt from data collection that consumer products can rarely avoid.
According to the Department of Defense, the portal has already attracted more than 1.7 million users out of its total 3 million personnel. Developed under the OpenAI for Government initiative, ChatGPT Mil will offer an experience familiar to users of commercial ChatGPT, centered on chats, files, projects, and custom GPTs. The tool is positioned for unclassified administrative, logistics, planning, and policy work, with additional features slated to roll out over time.
Meanwhile, Grok is framed by the Pentagon in much broader, more military-oriented terms: helping service members “execute missions faster and with greater precision across a wide range of operational contexts,” from market research analysis for procurement personnel to supply chain management for logistics units.
Notably absent is Claude (Anthropic). Following a dispute with the administration, Anthropic was designated a supply chain risk after refusing to grant unrestricted access and insisting on maintaining certain safety guardrails. The company is currently suing the administration over the decision. In addition, the Pentagon has signed agreements with a host of other companies—including AWS, Microsoft, Nvidia, and Reflection AI—to bolster its AI capabilities, including deployments on classified networks.
The official integration of premier commercial AI models into military operations marks a turning point in both revenue and reach for OpenAI and xAI. It also demonstrates that the defense AI landscape is increasingly diverging along political lines, where a company’s presence on military portals hinges not solely on model quality, but also on its degree of cooperation with the government.