For three and a half years, “ChatGPT” was basically a synonym for AI. That’s officially over. New data from Sensor Tower shows OpenAI’s chatbot has dropped below 50% of the global AI assistant market for the first time, with Google’s Gemini and Anthropic’s Claude eating into its lead. Here’s what actually moved, why it happened faster than anyone modeled, and what you should do with this information depending on how you actually use AI.
Let’s start with the headline number, because it’s a big one.
According to Sensor Tower’s State of AI 2026 report, ChatGPT held more than half the global AI assistant market through January. By March, it had slipped below 50%. By the end of May, its share had fallen to 46.4%. In that same window, Google’s Gemini climbed to 27.7% and Anthropic’s Claude rose to 10.3%. Everyone else, Grok, Perplexity, DeepSeek, Meta AI, splits what’s left, each holding under 5%. Here’s the thing though. ChatGPT is not shrinking. It just crossed 1.1 billion monthly users, becoming the fastest app in history to reach a billion monthly users, faster than TikTok, YouTube, or Instagram ever managed. OpenAI’s last reported weekly active user figure was 900 million back in February. The pie is just getting cut more ways now. Gemini sits at 662 million monthly users and Claude at 245 million.
So this isn’t really a story about ChatGPT losing customers. It’s a story about ChatGPT losing its monopoly on the category it created.
Why ChatGPT’s lead shrank faster than anyone expected
Two forces are doing most of the work here, and they call for two different reactions from you.
Gemini is winning on default access, not necessarily on quality. Google has baked Gemini into Android, Search, Chrome, and Workspace. Most people didn’t “choose” Gemini the way they once chose to download ChatGPT. It just showed up, already signed in, already sitting inside apps they were using anyway. That’s a structural advantage that has nothing to do with which model writes better code or argues a point more clearly.
Claude is winning on trust, and one specific event proves it. In late February, OpenAI announced a partnership with the U.S. Department of Defense. Anthropic had been offered a similar arrangement and walked away from it over concerns about surveillance and autonomous weapons use. The market reacted within 48 hours. ChatGPT’s U.S. uninstalls spiked roughly 295% day over day, against a typical daily fluctuation closer to 9%, while Claude’s U.S. downloads jumped enough to take the number one spot on the App Store’s Top Free Apps chart. Sam Altman later admitted on social media that the rollout was “rushed” and “looked opportunistic and sloppy,” and called the U.S. government’s decision to exclude Anthropic from official agency work “a very bad decision.” Claude’s U.S. user share has roughly tripled over the past year as a result, according to Sensor Tower data reported across multiple outlets.
That single event is a useful reminder that people no longer evaluate AI tools on capability alone. They’re weighing who built it, what it’s being used for, and whether a company’s choices line up with their own values. That’s a genuinely new dynamic in software, and it isn’t going away.
The real story underneath the user numbers is money
Most headlines about this report are skipping past the part that actually matters. The user numbers are interesting, but the revenue numbers are where this market is heading.
Consumer spending on AI apps is on pace to hit $4.2 billion in the first half of 2026, nearly double the $1.83 billion spent in the same period of 2025. Total hours spent inside AI apps are projected to climb from 17.2 billion to roughly 36 billion over that same stretch. The top three assistants alone account for close to 89% of all that time.
Inside that growth, one number stands out: 13% of Claude’s users are paying subscribers, the highest conversion rate of any assistant tracked in the report. Claude also starts at a higher price point than ChatGPT’s or Gemini’s entry-level paid plans, which tends to pull its average revenue per user up. For context, ChatGPT still leads on raw retention, with Sensor Tower putting its return rate noticeably ahead of Claude’s. Different companies are winning different games here, and that matters when you’re deciding which one is actually worth paying for.
Translation: download counts make headlines, but conversion rate tells you which company is actually building something people are willing to pay for.
ChatGPT is quietly becoming an ads and shopping platform
One detail in this report deserves more attention than it’s getting.
OpenAI started testing ads inside ChatGPT in February. By May, an average of 17% of ChatGPT’s daily users were seeing them, and the leading categories are exactly what you’d guess: software and shopping first, followed by media and entertainment, then food and dining.
At the same time, ChatGPT is sending growing volumes of shopping referral traffic to retailers like Target, Walmart, and Costco. Amazon has blocked ChatGPT’s web crawlers, so it gets almost none of that referral traffic back. That’s opened a lane for retailers to build their own AI shopping assistants instead of waiting on ChatGPT to send customers their way. Walmart’s Spark assistant is gaining ground while Amazon’s own Rufus has stayed roughly flat, even though shoppers who do use Rufus convert at noticeably higher rates than those who don’t.
I covered this exact shift in more detail a few weeks ago when ChatGPT first started acting as a brand discovery channel. This report confirms that trend is accelerating, not slowing down. If you sell anything online, it’s worth paying close attention to.
What this actually means for you
Enough context. Here’s what to actually do with this, broken out by how you use AI day to day.
If you use an AI assistant casually, for writing, research, or quick questions
- Stop treating “ChatGPT” and “AI” as the same word. Pick one task you do every week, drafting an email, summarizing an article, brainstorming a list, and run it through ChatGPT, Gemini, and Claude’s free tiers once. Ten minutes will tell you more about which tool fits your style than any review ever will.
- Check what you’re already paying for before you sign up for anything new. If you already have a Google One or Workspace plan, there’s a good chance Gemini access is already bundled in. Look before you add a separate subscription elsewhere.
- If trust and company values matter to you, factor that in on purpose. The DoD episode showed that a meaningful number of users now pick AI tools partly the way they pick brands, on principle as much as features. Nothing wrong with that. Just make the choice deliberately instead of by default.
If you run a business, a store, or handle marketing
- Audit how your product pages show up to AI shopping agents, not just search engines. If ChatGPT, Gemini, or Claude can’t read your pricing and product details cleanly, you’re invisible at the exact moment an AI is comparing you to a competitor on a customer’s behalf.
- Decide on purpose whether you want AI crawlers indexing your site. Amazon’s choice to block ChatGPT’s crawlers cost it referral traffic. That might be the right call for your business too, or it might not be. Either way, make it a decision, not an accident.
- Watch which assistant your specific customers actually use. A younger consumer audience skews toward different assistants than an enterprise B2B audience does. Don’t assume ChatGPT is still the default for your customers just because it was in 2023.
If you’re deciding which subscription to actually pay for
- Match the tool to the job, not the headline. Claude’s high conversion rate suggests strong satisfaction among people doing focused, complex work. Gemini’s strength is integration if you already live inside Google’s tools. ChatGPT still has the largest plugin ecosystem and the highest retention. None of these is universally “best.” They’re each best at different things.
- Try before you commit to a year. All three offer usable free tiers. Spend a week running your actual, real tasks through each one before you pick a paid plan.
- Revisit your choice every few months. A market that moved this much in five months isn’t done moving. The tool that’s right for you in June might not be the best fit by December.
The honest caveats
A few things worth keeping in your back pocket before you treat this as gospel.
This data comes from Sensor Tower, a market intelligence firm, not from OpenAI, Google, or Anthropic directly. Different outlets covering the same report cited slightly different figures for the same metrics in the days after release, anywhere from 46% to 46.4% for ChatGPT’s share, for example. The overall trend held steady across every source. The third decimal point isn’t worth arguing over.
It’s also worth remembering that OpenAI still leads by a wide margin on absolute users and on retention. Losing market share percentage while your total user base keeps growing is not the same thing as losing. ChatGPT created this entire category and is still the biggest player in it by far. What’s changed is that it’s no longer the only serious option, and people are finally acting like it.
If you’ve only ever used one AI assistant, this is your sign to actually compare. Share this with whoever in your group chat still says “let me ChatGPT that” for everything.
Sources and further reading
- ChatGPT’s market share slips below 50% for first time, TechCrunch (June 16, 2026)
- Sensor Tower State of AI 2026 Report, PR Newswire / Sensor Tower (June 16, 2026)
- ChatGPT uninstalls surged by 295% after DoD deal, TechCrunch (March 2, 2026)
- ChatGPT reaches 900M weekly active users, TechCrunch (February 27, 2026)
- ChatGPT app hits 1 billion monthly active users in record time, Reuters (June 2, 2026)
- ChatGPT just became a brand discovery channel, here’s what that means for your business, Just Being Resourceful (May 31, 2026)
Just Being Resourceful! 
Leave a comment