ChatGPT's market share just fell below 50% for the first time

22 June 2026 · 4 minute read

Three and a half years after launch, ChatGPT has dropped below 50% global AI assistant market share for the first time. According to data tracked across analytics firms, ChatGPT held 46.4% of the market in May 2026. It launched in November 2022 and effectively owned the category. By December 2024 it still commanded 65.3%. By December 2025 that had slipped to 52.8%. Now it is below half.

The redistribution is concrete. Gemini reached 27.7% by May 2026, up from 18.2% in December 2024. Claude hit 10.3%, up from just 3% six months earlier. Grok holds 3.3%, DeepSeek 3.2%, Perplexity 2.8%, Meta AI 2.5%, Microsoft Copilot 1.6%.

What actually changed

ChatGPT is not shrinking in absolute terms. It has 1.1 billion monthly active users. Gemini has 662 million, Claude has 245 million. The total market is growing faster than ChatGPT is adding users, which is how you lose share while still growing.

Three things drove the shift. Gemini's multimodal capabilities improved significantly in early 2026, particularly for image and document tasks. Google pushed Gemini into Google Workspace, which has 1.8 billion users worldwide; that is a distribution channel no standalone AI product can easily match. And Claude's subscriber conversion rate hit 13%, the highest of the main assistants, suggesting the people who try it are finding enough value to pay.

There was also a trust event. OpenAI's US Department of Defense contract announcement in early 2026 prompted a segment of privacy-focused users to look elsewhere. Some went to Claude. Some to Gemini. Few came back.

Why this matters beyond the headline number

OpenAI is preparing an IPO at a valuation that sources have cited at up to $1 trillion. Falling below 50% market share in your flagship product while still the market leader is not a crisis. But it is a data point that investors will scrutinise.

The narrative a trillion-dollar valuation needs is one of dominance. Falling share tells a different story: the AI assistant market is becoming a multi-player competition, not a winner-takes-all outcome. That tends to compress the multiples investors will pay.

The counterargument is that ChatGPT still has twice the share of its nearest competitor, is growing in absolute terms, and has GPT-5.5, Daybreak, and new voice features in its pipeline. That argument may well hold. Whether it holds through six months of IPO roadshow questioning is a separate matter.