US Secretary of Defense Hegseth announced on January 12, at SpaceX headquarters in southern Texas, that Grok, from Elon Musk's AI company xAI, will be integrated into Department of Defense networks, including classified systems. He stated that Grok would be operational "this month." On February 24, Axios reported the signing of the contract, citing confirmation from Department of Defense officials.

Until now, Anthropic's Claude had been the only available model for the military's highest-security systems (used for intelligence analysis, weapons development, and battlefield operations). The Department of Defense had been pressuring Anthropic amid conflicts over safeguards, increasing the need for alternative models. Secretary Hegseth is scheduled to meet with Anthropic CEO Dario Amodei at the Department of Defense on February 25, and officials expect the atmosphere to be tense.

In non-classified systems, Grok, Google's Gemini, and OpenAI's ChatGPT are already available, and both Google and OpenAI are in discussions regarding migration to classified environments. According to a New York Times report, Google is "close" to an agreement regarding the classified use of Gemini, while OpenAI is said to be "further off."

The announcement of Grok's integration came a few days after Grok faced international criticism for generating sexual deepfake images without user consent. Ofcom, the UK's online safety regulator, has launched a formal investigation into X, while Malaysia and Indonesia have blocked the platform entirely.

Sources: