On December 26, 2025, The Decoder reported Similarweb data showing ChatGPT’s share of generative‑AI web traffic dropping from 87.2% to 68% over the past year, while Google’s Gemini jumped from 5.4% to 18.2%. An India‑focused readout from Moneycontrol the same day highlighted Gemini’s surge from 5% to 18% as analysts framed it as a major shift in the chatbot race.
This article aggregates reporting from 3 news sources. The TL;DR is AI-generated from original reporting. Race to AGI's analysis provides editorial context on implications for AGI development.
Traffic share isn’t the same as capability, but this dataset is a clear sign that OpenAI’s once‑unchallenged distribution advantage is eroding. Similarweb’s numbers suggest ChatGPT’s share of generative‑AI web traffic has fallen nearly 20 percentage points in a year, with Gemini climbing more than three‑fold over the same period.([the-decoder.com](https://the-decoder.com/chatgpts-market-share-falls-to-68-percent-as-gemini-closes-in/)) For the race to AGI, that matters because scale—measured in active users, prompts, and real‑world edge cases—is one of the key inputs for training and refining next‑generation systems.
Google’s surge reflects both product and distribution strategy: Gemini is deeply embedded into Search, Workspace and Android, lowering friction relative to a standalone chatbot site. OpenAI, by contrast, still leans heavily on direct ChatGPT usage and API integrations. If this trend continues, Google will accumulate an increasingly rich stream of multimodal usage data, especially outside the U.S., which can feed back into better alignment, tool use and world‑modeling—core ingredients for AGI‑class systems.([similarweb.com](https://www.similarweb.com/blog/marketing/geo/gen-ai-stats/?utm_source=openai)) For OpenAI and other labs like Anthropic or xAI, the message is blunt: superior raw models are not enough if you lose the default entry points where billions of users actually interact with AI.
