Google's Gemini hits 900m users

Google used its I/O 2026 conference on May 19 to put a number on the scale of Gemini: more than 900 million monthly users. That figure, disclosed by Chief Executive Sundar Pichai and Gemini app chief Josh Woodward in separate Google posts, was up from 400 million at last year’s I/O. The update gave Google a fresh way to argue that its AI position is being built not only on model performance, but on reach across products people already use. The company paired that user figure with new agent features in Search, Android and the Gemini app, even as a Search glitch this week showed how brittle product integration can still be. ### Where did the 900 million figure come from? Sundar Pichai said in his I/O keynote transcript that the Gemini app had “surpassed 900 million” monthly active users, more than doubling from 400 million a year earlier. Josh Woodward, vice president of Google Labs and the Gemini app, wrote separately that more than 900 million people in 230 countries and more than 70 languages now use Gemini every month. (blog.google) The May 19 disclosures came in official Google blog posts, making them the clearest primary-source confirmation of the user milestone. Indian Express, which highlighted the same figure in a broader analysis of the AI race, framed the number as evidence that Google’s installed base across Search, Android and Workspace is becoming a distribution advantage. That reading is consistent with Google’s own emphasis on products “that touch billions of people,” though Google did not present the user number as a direct comparison with OpenAI or Anthropic. (blog.google) ### Why does distribution matter more than a benchmark chart? Google said on May 19 that it now has 13 products with more than 1 billion users each, and Pichai tied Gemini adoption to that broader product base. He also said Google is processing more than 3.2 quadrillion tokens per month across its surfaces, up from roughly 480 trillion at last year’s I/O, and that daily Gemini requests have grown more than sevenfold in the same period. (blog.google) Those numbers matter because Google is embedding AI into places where users already are, rather than asking them to adopt a standalone chatbot first. Woodward’s post described Gemini as a service people use “for help every month” across web, Android and iOS, while Pichai’s keynote repeatedly tied AI usage to Search and other Google products. The result is that Google can grow Gemini through placement and defaults as much as through model upgrades. (blog.google) That is an inference from Google’s rollout pattern, not a statement the company made in those words. ### What exactly did Google announce beyond the user number? Google announced Gemini Spark, which Woodward described as a “24/7 personal AI agent” designed to proactively manage tasks under user direction. He also announced Daily Brief, a personalized morning update, and said the MacOS app will integrate Gemini Spark so it can operate on a local machine. (blog.google) Bhaskar English, citing reporting from The Economist, said Google is positioning these tools as systems that can keep handling tasks even when users are not actively on their devices. That report said Search-based information agents are being built to monitor sports, shopping deals and stock market activity. The article’s framing that AI will work on “switched-off phones” appears to be a characterization of persistent agent behavior, not language quoted from Google’s official I/O materials. (blog.google) ### What did the “disregard” glitch show? Google Search users on May 22 and May 23 reported that AI Overview responded to the word “disregard” as if it were an instruction rather than a request for a definition. Techlusive said the search result returned a chatbot-style reply — “Understood. Let me know whenever you have a new prompt or question” — instead of dictionary information. Merriam-Webster then posted the proper definition on X alongside a screenshot of the result. (bhaskarenglish.in) The episode was minor compared with Google’s I/O announcements, but it showed the tradeoff in pushing conversational AI deeper into Search. When AI features are attached to a product used for basic lookups, even a small parsing error becomes highly visible. That is an inference based on the public reaction and the placement of AI Overview in Search. (techlusive.in) ### How much of this fight is now about politics? The White House in March released a national AI legislative framework, and Trump had already signed an executive order in December 2025 aimed at preventing what the administration called a patchwork of state AI rules. Those documents show that AI policy is already a federal political issue, separate from company product launches. (techlusive.in) Axios, Forbes, USA Today and TechCrunch reported this week that President Donald Trump pulled back from signing a separate AI executive order after objections from industry figures and concerns inside the administration about slowing the U.S. position against China. Those reports indicate that the next phase of the AI contest is also being shaped by lobbying and regulation, not only by model releases and user growth. (whitehouse.gov) Google’s next visible milestones are the rollout of Gemini Spark, Daily Brief and other agent features announced at I/O 2026 on May 19. Search, Android and the Gemini app are the main surfaces Google named for those releases, while the company’s official posts say some features are rolling out now and others are coming soon. (blog.google) (forbes.com)