Google's artificial intelligence implementation has reached unprecedented scale, according to data revealed at the company's I/O 2025 developer conference. The tech giant is now processing over 480 trillion tokens monthly across its products and APIs, representing a staggering 50-fold increase from the 9.7 trillion tokens processed this time last year.
This exponential growth extends beyond raw processing numbers. The Gemini app has accumulated more than 400 million monthly active users, while developer engagement has increased dramatically, with over 7 million developers now building with Gemini—five times more than a year ago. Usage of Gemini on Vertex AI, Google's enterprise AI platform, has jumped 40 times year-over-year.
Particularly strong growth is evident with Google's 2.5 series of models, with usage increasing by 45% among those using 2.5 Pro in the Gemini app. These advanced models feature enhanced reasoning capabilities, allowing them to think through complex problems before responding.
Google's AI Overviews, which provide AI-generated answers at the top of search results, have also seen remarkable adoption, now reaching over 1.5 billion users across 200 countries and territories. The company reports that these overviews are driving over 10% growth in search queries that display them in major markets like the U.S. and India, making it "one of the most successful launches in Search in the past decade."
According to Google, this progress signals a new phase in the AI platform shift, where "decades of research are now becoming reality for people, businesses and communities all over the world." The company's infrastructure strength, including its seventh-generation TPU called Ironwood, has been crucial in delivering faster models while simultaneously reducing costs.
With this rapid growth in AI adoption across its ecosystem, Google appears well-positioned to maintain its leadership in the increasingly competitive AI landscape, where it faces challenges from rivals like OpenAI's ChatGPT and Anthropic's Claude.