Musk’s Grok Chatbot Falls Behind as ChatGPT, Claude, and Gemini Surge Ahead

Elon Musk’s AI chatbot, Grok, is finding it hard to stay competitive as new data shows its popularity and business use are falling behind rivals like ChatGPT and Claude. Despite a loud start, the "anti-woke" chatbot is seeing fewer downloads and a lack of paying customers.

Downloads Drop After Image Controversy

Earlier this year, Grok saw a massive surge in interest, peaking at over 20 million downloads in January. However, much of that growth was tied to a controversial feature that allowed users to create explicit, non-consensual deepfake images. The tool was widely criticised by lawmakers and child safety groups, especially after reports surfaced of it being used on images of celebrities and minors.

Following a wave of international bans and restrictions, Grok’s parent company, xAI, was forced to limit these features. Since then, the hype has faded. According to data from AppMagic, downloads plummeted to just 8.3 million in April - a drop of nearly 60% from its January peak.

Failing to Win Over Paying Users

While millions use free AI tools, the real test is whether people are willing to pay for them. For Grok, the numbers are stagnant. A massive survey of 260,000 Americans by Recon Analytics found that only 0.174% of people paid for Grok in the second quarter of 2026. In comparison, OpenAI’s ChatGPT is dominating the market, with more than 6% of respondents saying they pay for a subscription.

Grok is also struggling to get inside office buildings. Most companies are looking for AI to help with coding or writing emails, and they aren't choosing Musk’s tool. A study by Enterprise Technology Research showed that only 7% of companies are actively using Grok. Meanwhile, Google’s Gemini and Anthropic’s Claude are seeing much faster growth in the workplace.

A Surprising Deal with a Former “Enemy”

Perhaps the biggest sign that Grok is lagging is a new deal involving Musk’s other company, SpaceX. Despite Musk previously calling the AI firm Anthropic "evil," SpaceX has reportedly signed a massive deal to rent out its computing power to them. Anthropic will now use the "Colossus 1" data center in Tennessee-Musk's own hardware-to train its own AI models. While this deal will make billions for SpaceX, it suggests that Musk's infrastructure might be more successful at powering his competitors than his own chatbot. Grok for now remains integrated with the social media platform X, but without a major shift in strategy or business adoption, it risks becoming a niche product in a market increasingly controlled by OpenAI, Google, and Anthropic.

Read More: 'Amateur City': Nadella Blasts OpenAI Board as Microsoft Takes Center Stage in Musk Trial