ChatGPT’s market share just dropped below 50% for the first time.
Sensor Tower’s State of AI Report for 2026, out today via TechCrunch, puts ChatGPT at 46.4% of the AI assistant market in May 2026, down from a clear majority in January.
The rest went to Google’s Gemini (27.7%) and Anthropic’s Claude (10.3%). For the first time since ChatGPT launched in November 2022, the AI assistant market is genuinely competitive.
The new numbers
Three numbers tell the story.
ChatGPT at 46.4%. Down from above 50% in January 2026. Still 1.1 billion monthly active users, still the fastest app ever to reach that mark. But no longer a majority.
Gemini at 27.7%. 662 million MAU. Google’s ecosystem integration (Search, Workspace, Android default) is the main driver.
Claude at 10.3%. 245 million MAU. Smaller than ChatGPT, but the conversion rate is the interesting number (more on that below).
Grok, Perplexity, DeepSeek, and Meta AI each hold under 5%. The long tail is still fragmented, but the top three now control 89% of total time spent on AI assistants.
Why ChatGPT is losing share
Three forces, in order of how cleanly they show up in the data.
Gemini’s growth is the biggest. Google’s ecosystem integration means Gemini shows up everywhere. Android default, Workspace sidebar, search results. You don’t have to install it. You have to opt out. 662 million MAU is the result.
Claude’s productivity reputation is real. Anthropic is closing on ChatGPT’s user retention rate, per Sensor Tower. People who try Claude for coding or long-form work tend to stick.
The DoD trust hit is the most interesting data point. OpenAI’s Department of Defense deal in February 2026 triggered a 295% spike in ChatGPT uninstalls. That’s not a benchmark number, that’s a behavior number. Brand trust and values alignment matter to users.
That last one is the kind of stat that tells you more about where the AI market is going than any leaderboard ranking. The market is mature enough that users are voting with the delete button.
Claude’s quiet win
Here’s the number that matters: 13% of Anthropic’s users pay for a subscription. That’s the highest conversion rate in the industry.
Compare that to ChatGPT, where the paid user base is much larger in absolute terms but a smaller percentage of the 1.1B MAU. Claude has fewer users. But the ones it has are more likely to pay.
Last week, I ranked every AI subscription at $20 a month. The data now backs that up. Users who try Claude tend to pay for it. The market is finally letting you pick, and Claude’s 13% is the proof that the “premium AI” tier has real demand.
Worth noting: OpenAI’s $150M partner network announcement a day earlier was about extending distribution. This data says the user base is still dominant, but the competitive ground is shifting underneath.
The market is maturing
The other big story in the Sensor Tower report isn’t share, it’s growth rates.
$4.2 billion in AI app spending in H1 2026. That’s 2.3x the $1.83B from H1 2025. Hours spent: 17.2 billion to 36 billion, also 2.1x. Sounds like a gold rush.
But the growth rates are decelerating. The Sensor Tower report flags this explicitly. The 2025 numbers were off a smaller base. The 2026 numbers are off a much larger one. The percentage growth will keep shrinking.
Asia recorded its first download decline in Q1 2026, down 3.3% year over year. China and India drove the dip. Asia still leads in total downloads but trails North America and Europe in in-app spending, which is where the premium features will go.
The “every app is a rocket ship” phase is over. Now it’s about who holds the high-value users.
The OpenAI ad experiment
OpenAI started running ads in ChatGPT in February 2026. By May, 17% of daily users were seeing them.
Top ad categories: Software & Shopping first, then Media & Entertainment, then Food & Dining. ChatGPT is sending real referral traffic to Target, Walmart, and Costco.
One wrinkle: Amazon blocked ChatGPT’s web crawlers. Result: Amazon referral traffic from ChatGPT has been flat since the block. Amazon’s own Rufus assistant has flat user growth, but the users who engage with Rufus convert at higher rates. Walmart’s Spark is gaining ground.
The shopping layer is its own competitive war. ChatGPT is winning on reach. Amazon is defending. Walmart is gaining.
What this means for your subscription pick
The data is the data. Here’s what it says.
If you tried ChatGPT and stuck around, you’re in the 1.1B MAU. The default is still the default.
If you tried Claude and paid for it, you’re in the 13%. Smaller club, higher intent.
If you tried Gemini, you got there through Google. Whether you stay is a different question.
The market is finally letting you pick. The numbers above are what the “pick” actually looks like at scale. None of this is “ChatGPT is dying” or “Gemini is winning.” It’s “for the first time, it’s not a one-horse race.”
Bottom line
ChatGPT is still the biggest player. But for the first time, it’s not the only one that matters.
The market is maturing. Brand trust and ecosystem integration are the new battlegrounds. The data backs up what TRT readers saw in the Jun 13 guide: Claude’s value is real, Gemini’s reach is undeniable, and ChatGPT’s lead is no longer unassailable.
The 295% uninstall spike after the DoD deal is the stat to remember. Users vote with the delete button. The market is finally listening.
