ChatGPT just fell below 50% market share: building and getting found in a multi-model world

For the first time in three and a half years, ChatGPT is no longer the default AI assistant. Sensor Tower's June 2026 data puts ChatGPT at 46.4% of the consumer assistant market — below half for the first time since launch — with Google's Gemini at 27.7% and Anthropic's Claude at 10.3%. The headline number matters less than what it signals: people have stopped treating 'the AI' and 'ChatGPT' as synonyms, and they're actively choosing different tools for different jobs.

The shift is qualitative, not just quantitative. Claude users convert to paid plans at 13% — the highest of any assistant — which tells you the market is segmenting by intent rather than collapsing toward a single winner. Casual questions, coding, long-document work, and research are increasingly routed to whichever model is best at that task. The era of one assistant doing everything for everyone is ending, and the era of deliberate model comparison is beginning.

For anyone trying to get found, this breaks a comfortable assumption. If your visibility strategy was 'show up in ChatGPT', you were optimizing for a single retrieval system that now serves under half the audience. Gemini draws on Google's index, Claude has its own retrieval behavior, and Perplexity is different again. Generative Engine Optimization is no longer a one-engine game — the content and structure that get you cited have to work across several systems with different citation logic.

For anyone building with AI, multi-model is no longer a nice-to-have architecture; it's the realistic one. Tying your product to a single provider's API made sense when one model dominated. With three credible engines — and open-weights closing the gap behind them — provider lock-in is now a genuine business risk. The teams in the best position kept a thin abstraction layer and an eval suite, so swapping or routing between models is an afternoon's work, not a rewrite.

The practical move this quarter is to stop thinking in terms of 'the AI' and start thinking in terms of a portfolio. Test your most important buyer questions across ChatGPT, Gemini, Claude, and Perplexity and see who cites you where. Architect new features so the model is a swappable component, not a foundation. The market just told you it's plural — build and market like you believe it.