Federal agencies logged their AI tool usage last year, and Grok barely showed up. That's the finding of a new Reuters report, which combed through government records and found xAI's chatbot - Elon Musk's self-described "truth-seeking" AI - almost entirely absent from official federal use. The Verge covered the Reuters findings and paired them with a blunt conclusion: the chatbot is neither especially good nor especially popular.
For the average person using AI tools at work, this probably matches lived experience. Grok launched in 2023 with a lot of noise and a platform advantage that almost no other AI company has - mandatory distribution to hundreds of millions of X users. Despite that, it hasn't become the default tool anyone actually reaches for when they need to write, research, code, or think. The federal records are just the government's version of that same pattern.
The Platform Advantage That Didn't Convert
Musk's theory was that owning the platform meant owning the distribution. Grok gets surfaced to X users constantly - built into the interface, promoted in posts, attached to Premium subscriptions. By raw visibility, it should have a massive user base. The problem is visibility and daily active use are different things. You can put a chatbot in front of millions of people and still not get them to trust it for their actual work.
ChatGPT didn't win its position through a social media platform. It won by being the tool people genuinely found useful, told their colleagues about, and kept coming back to. Claude built credibility by being notably better at following instructions and avoiding hallucinations (confident but wrong outputs) than its competitors. Grok's positioning - as a based, anti-woke alternative that tells it like it is - appeals to a specific audience but doesn't solve any real workflow problem.
What the Federal Records Actually Signal
Government procurement and usage data is often a lagging indicator. Agencies move slowly, have security requirements that limit which tools they can use, and tend to standardize on what their IT departments have already approved. ChatGPT and Claude have made inroads through enterprise agreements and government-specific products. Grok hasn't had the same focus there.
But the federal absence isn't the main story - it's just the most concrete data point in a longer pattern. Grok's benchmark performance has generally been competitive on paper but unremarkable in practice. Its integration with X data gives it an edge for real-time social media queries, but that's a narrow use case. Most people using AI tools daily - for writing, coding, customer support, content creation - aren't choosing Grok.
xAI has the resources to keep iterating. Musk's companies have pulled off unlikely reversals before. But building a genuinely useful AI assistant requires trust, reliability, and a reason to switch that's bigger than political alignment. Right now, Grok doesn't have a compelling answer to the question every new user asks: why this instead of what I'm already using?