Google’s Gemini reported a 28% month-over-month growth in December 2025, commanding close to 40% of rival ChatGPT’s web audience size. Data by Similarweb showed that ChatGPT’s seven‑day average visitors fell by about 22% over six weeks, dropping from nearly 203 million to 158 million. Responding to the numbers, Google AI CEO Demis Hassabis wrote: “a lot more hard work still to do of course, but making relentless progress…” highlighting Google’s continued push to expand Gemini’s reach and usage as competition in the AI chatbot space intensifies.To recall, Google launched its Gemini 3.0 model in November 2025. Days after the launch, OpenAI CEO Sam Altman declared an internal “code red” emergency, ordering employees to halt work on advertising and revenue-generating features to fix ChatGPT’s deteriorating competitive position.
Google’s Gemini forces OpenAI CEO Sam Altman send ‘Code Red’ warning
In December last year, OpenAI CEO Sam Altman sent an internal ‘Code red’ warning to its employees. According to internal memos obtained by The Wall Street Journal and The Information, Altman told employees to abandon plans to introduce advertising, AI shopping assistants, health agents, and its Pulse personal assistant. Instead, he asked them to focus on ChatGPT. “We are at a critical time for ChatGPT,” Altman wrote in the memo.
How Google is crushing ChatGPT
While ChatGPT still leads in over AI traffic, it is facing a downward trend. As mentioned above, the AI chatbot dropped below the 65% mark earlier this month. It has a market share of 68% in December last year which fell to 64.5% on January 2, 2026, Similweb Data shows. Gemini on the other hand, has registered a substantial growth from 5.7% a year ago to now commanding 21.5% Generative AI market share.
Google’s infrastructure advantage proves decisive
Enterprise trust, integration advantage – Gemini is seamlessly embedded across Google’s ecosystem, and hardware synergy – Google’s TPU v5 chips are optimised for Gemini workloads, reducing costs and boosting performance, are some of the factors that may have worked in favour of Google. More fundamentally, Google controls what insiders call the “full stack.” DeepMind researchers build models, proprietary TPU chips train them cheaply, Google Cloud hosts them efficiently, and seamless integration delivers them across Search, YouTube, Gmail, and 3 billion Android devices. OpenAI depends on Microsoft’s cloud infrastructure and Nvidia’s expensive chips, creating structural cost disadvantages.“One of the most important things for us at Google is this is possible because we have a very differentiated full-stack approach,” DeepMind CTO Koray Kavukcuoglu told reporters last year.
About the Author
