For most of the past three years, ChatGPT and AI assistant were essentially the same word. That is changing, and the data from Sensor Tower's State of AI 2026 report puts a precise number on it.
ChatGPT's share of global AI assistant users fell below 50% for the first time in March 2026, landing at 46.4% by the end of May. Two years ago, that figure was above 80%. The tool still has the most users by a wide margin, over 1.1 billion monthly, but Google's Gemini now has 662 million and Claude has 245 million. The gap is closing.
Gemini's growth is mostly structural. Google built it into Search, Gmail, and Android, products that hundreds of millions of people already use every day. That is a distribution advantage almost no competitor can match. Claude's growth is different. It earned its users, primarily through a reputation for being genuinely useful for work tasks. Claude is now closing in on ChatGPT's user retention rate, meaning the people who try it are increasingly likely to stick with it.
The more telling number is not who has the most users. It is who is turning users into revenue. Claude earns more per user than ChatGPT does on mobile in the United States, climbing from under $0.50 per user in September 2025 to $2.76 by May 2026. ChatGPT earns around $1.74 per user by comparison. And 13% of Claude's users are paying subscribers, the highest conversion rate in the market. That figure matters because it suggests Claude has found a specific audience that finds it worth paying for, even when free alternatives exist.
OpenAI is taking a different path to revenue. The company started running ads inside ChatGPT in February 2026, and by May, roughly 17% of daily users were seeing them. The ad pilot reached $100 million in annualized revenue in under six weeks, with over 600 advertisers participating. That is a meaningful signal: OpenAI has 900 million weekly active users but only around 50 million paying subscribers. Advertising is the clearest way to make the other 850 million financially productive.
Shopping is where this gets directly relevant for business operators outside the tech sector. ChatGPT is already sending referral traffic to retailers like Walmart, Target, and Costco. Amazon, which has blocked ChatGPT's web crawlers, is missing out on that traffic. Walmart's own AI shopping assistant, Sparky, built in partnership with OpenAI, is gaining ground: shoppers who use it have average order values 35% higher than those who do not. Amazon's Rufus tells a similar story: customers using it are roughly 60% more likely to complete a purchase.
The pattern across all of this is consistent. AI assistants that are embedded in places where people already spend time, whether that is Google's apps, a retailer's checkout flow, or a workplace productivity tool, are winning. Standalone tools with no integration advantages are having a harder time.
For business operators, two things are worth watching. First, the AI assistant your team uses today may not be the right one in twelve months. The market is moving fast enough that switching is becoming normal behavior, not exceptional. Second, if you sell products or services online and have not yet thought about how AI assistants describe or recommend you, that window is narrowing. The tools that decide what to recommend to shoppers are maturing quickly, and product data quality, catalog completeness, and review coverage are increasingly what those tools reward.