For the first time since ChatGPT's launch, OpenAI holds less than half of the global AI assistant market: 46.4%. That's still enormous, and its 1.1 billion monthly users dwarf everyone else in absolute terms. But the symbolism is strong: the era of "AI is ChatGPT" is over.
For three years, talking about consumer AI meant talking about ChatGPT. OpenAI's product was so dominant that its name had become synonymous with AI, just as Frigidaire became synonymous with the refrigerator. The 2026 State of AI report published by Sensor Tower (a firm specialising in app market analysis) now marks a turning point: this absolute dominance is crumbling.
The numbers behind the shift
Here's how the global AI assistant market breaks down according to the report.
| Assistant | Market share | Trend |
|---|---|---|
| ChatGPT (OpenAI) | 46.4% | Declining, below 50% for the first time |
| Gemini (Google) | 27.7% | Rising sharply |
| Claude (Anthropic) | 10.3% | Rising, leader in monetisation |
| Grok + Perplexity | Under 5% combined | Far behind |
These figures need to be read carefully. ChatGPT remains number one, and by a wide margin in absolute terms: 1.1 billion monthly active users. Losing market majority doesn't mean losing the market. But when you go from "crushing everything" to "a dominant player among others", the narrative changes.
The metric that should worry OpenAI
The most interesting figure isn't the raw market share, it's the paid conversion rate. At Claude, 13% of users pay for a subscription, the highest rate in the sector. In other words, Anthropic has fewer users, but those users are far more willing to pull out their credit cards.
That's a major strategic difference. A free user costs money (they consume compute without generating any revenue). A paying user funds the company. Claude has positioned itself on the professional and demanding segment, where people pay for quality, while OpenAI rules over the consumer volume, which is harder to monetise.
The report notes a telling detail. When OpenAI signed a contract with the US Department of Defense in February 2026, there was a measurable spike in ChatGPT uninstalls. Translation: users are starting to choose their AI based on values and trust, not just performance. The most powerful model no longer wins automatically. The one you trust matters just as much.
What this really reveals
The shift in market share tells a deeper story than a simple corporate battle. It shows that users are maturing. In the early years, people used ChatGPT by default, because it was the only name they knew. Today, they compare. They test Gemini for search, Claude for writing or coding, ChatGPT for general use.
This is exactly what happens when a market normalises. At the start of a technology, one player sweeps everything up. Then competition sets in, usage becomes specialised, and users turn into informed consumers rather than followers. What we're seeing in 2026 looks like the end of the consumer AI market's infancy.
For you, the user, this is excellent news. More competition means more choice, pressure on prices, and companies forced to earn your loyalty instead of taking it for granted. The dominance of a single player has never benefited anyone except that player. A market with three or four heavyweights fighting it out means you win.