Reuters analysis finds Grok barely used in U.S. government AI deployments, appearing in only three of 400+ documented cases compared to OpenAI's 230+

A Reuters review of federal government AI vendor records found Grok or xAI appeared in only three examples out of more than 400 documented cases, each time for basic uses like document drafting or social media management alongside competitors like Microsoft and OpenAI. By contrast, OpenAI's models appeared in more than 230 examples, while Google and Anthropic each appeared dozens of times.

Unnamed Pentagon sources told Reuters that Grok "is just not the best model out there," with staffers preferring Gemini or Claude instead. Public leaderboards show Anthropic, Google, and OpenAI dominate the top ranks, while Grok rarely cracks the top 10 outside occasional image or video categories.

SpaceX's IPO filing positions AI and Grok specifically as central to the company's investor pitch, claiming to have identified a $28.5 trillion opportunity primarily from enterprise AI. However, Musk has also admitted that xAI used OpenAI's models to help train and improve Grok through a process called distillation, meaning Grok cannot outperform the rival system it was trained on.