CHATGPT JUST LOST ITS MAJORITY

For the first time since it launched the entire category in late 2022, ChatGPT holds less than half the AI assistant market. Sensor Tower's State of AI 2026 report puts its global share at 46.4% by May, down from a commanding majority a year earlier, with Google's Gemini at roughly 27.7% and Anthropic's Claude at about 10.3%. The headline writes itself, and it is also slightly misleading: this is the end of ChatGPT's monopoly on attention, not the start of its decline. Understanding the difference is the whole story, because what is actually happening underneath the number is the AI market growing up — moving from a single default everyone reached for to a field where users compare, switch, and choose. That shift changes what it takes to win.

The numbers, read correctly

The most important thing to hold onto is that ChatGPT did not shrink. Its absolute user base grew to more than 1.1 billion monthly users — still the fastest product ever to reach a billion — even as its share fell. What changed is the denominator: Gemini climbed from roughly 533 million to 662 million monthly users in five months, and Claude roughly quadrupled from about 60 million to 245 million over the same stretch. ChatGPT's slice got smaller because the pie grew faster than it did. As TechCrunch reported, the more telling finding in the data is behavioral: users are now willing to switch between assistants rather than defaulting to one. That willingness is the thing that ends an era.

A caveat worth stating: different trackers disagree on the exact level — methodologies that weight desktop, mobile app, and web traffic differently put ChatGPT somewhat higher or lower. But across all of them the direction is the same, which is what matters.

Two different playbooks

The two challengers are not winning the same way, and the contrast is the most useful thing in the report. Gemini's rise is a platform economics story: Google embedded it at the operating-system level across Android and wired it into Search, Gmail, and Workspace, so hundreds of millions of people meet Gemini by default instead of deciding to install it. When a product becomes the default, users stop evaluating it against alternatives — they simply use what is in front of them. That is distribution as a moat, and it compounds quietly without requiring Gemini to win a single head-to-head comparison.

Claude's growth runs on the opposite engine. It has no default surface on a billion phones; it wins on reputation — productivity, coding, and research — and on trust. The signal investors should watch is monetization: roughly 13% of Claude's users pay for a subscription, the highest conversion rate in the field, and its US revenue per user has climbed sharply. That same productivity reputation is why developers increasingly reach for it directly through tools like Claude Code. High conversion on a smaller base is a fundamentally healthier business signal than enormous free-tier reach, and it is the metric that will matter when these companies are valued.

Trust became a product feature

The single most revealing data point is what happened around values. When OpenAI announced a US defense deal in early 2026, Sensor Tower recorded ChatGPT uninstalls spiking to roughly 200% above their normal rate in one week, with a corresponding surge in Claude downloads. That put a number on something that had been anecdotal: for a meaningful slice of users, a company's policy and ethical positioning now influences product choice as directly as features do. Combine that with the friction of ads arriving on ChatGPT's free tier — and a churn rate that ticked up over the same period — and you get a market where switching costs are low and trust is a feature you can lose. In a network-effects business, the assumption was always that the leader's lead compounds. What this episode shows is that when switching is one download away, that compounding is far weaker than it looked.

Why it matters for OpenAI

The timing sharpens all of this. OpenAI filed a confidential S-1 in June, the first formal step toward a public offering, which means investors are about to weigh two true facts that point in opposite directions: a product with 1.1 billion users and the fastest growth in software history, against a share trajectory that has moved in one direction — down — for eighteen straight months. Both are real. The question a public market will ask is which one is the trend and which is the legacy, and the answer depends on whether ChatGPT can convert its enormous reach into the kind of paying loyalty Claude is demonstrating at a fraction of the scale.

The Bottom Line

ChatGPT losing its majority is not a story about a product failing; it is a story about a market maturing past its default phase. The era when "AI assistant" meant "ChatGPT" is over, replaced by users who comparison-shop and pick a model for the job. In that world the durable advantages are the unglamorous ones — distribution you own, like Google's, and trust you can keep, like Anthropic's — not the novelty of being first. The companies that internalize that the large language model is becoming a commodity, and that the contest has moved to distribution and trust, are the ones that will hold share. For more on how platform competition actually plays out, start with the artificial intelligence hub.