Three years after ChatGPT pushed Google into panic mode, the situation has flipped. OpenAI CEO Sam Altman has now called a “code red” emergency because Google’s Gemini 3 model is pulling ahead in performance, engagement, and market share. OpenAI has stopped work on ads, shopping tools, health agents, and the Pulse assistant to focus only on making ChatGPT faster, more reliable, and more personal.
The rush comes as Google’s full-stack infrastructure and product integration give it a strong advantage, creating pressure on OpenAI’s business and finances.
Google Gemini 3 Vs ChatGPT: Why Google Is Suddenly Winning
OpenAI’s internal memo shows that all teams must pause revenue projects and join daily calls until ChatGPT becomes “more intuitive and personal.” This is happening because Gemini 3 is beating GPT-5 in benchmark tests and attracting more users.
Google’s Gemini platform jumped from 450 million monthly users in July to 650 million in October, and data from Similarweb shows users now spend more time on Gemini than on ChatGPT. Even though OpenAI says ChatGPT has 800 million weekly users, Google’s growth and higher user engagement are shifting momentum.
Google’s advantage comes from its “full-stack” control. DeepMind researchers train models on Google’s custom TPU chips, then deploy them across Google Cloud and products like Search, YouTube, Gmail, and Android phones.
This makes training cheaper and distribution easier. OpenAI, in comparison, depends on Microsoft cloud services and Nvidia chips, which are expensive. This creates a cost disadvantage at a time when OpenAI is already facing huge financial pressure.
OpenAI Code Red & What Future Holds For ChatGPT
The new “code red” alert is OpenAI’s highest emergency level, similar to the “code red” Google used in 2022 after ChatGPT launched. Now the situation has reversed. OpenAI must stop projects that can make money, like ads and personal shopping assistants, because improving ChatGPT has become more important for survival. Reports say OpenAI needs to reach around $200 billion in yearly revenue by 2030 to become profitable, while Google already makes $120 billion per year and has nearly $100 billion in cash to spend freely on AI.
Google’s growth is also affecting Nvidia. After news that Meta might buy Google TPU chips instead of Nvidia GPUs, Nvidia’s stock dropped more than 6%, losing almost $250 billion in value. If companies switch to Google’s TPU chips, Nvidia could lose around 10% of its yearly revenue. Meanwhile, OpenAI must continue paying extremely high compute costs to Nvidia for model training.
Sam Altman has told employees that a new reasoning model coming next week will beat Gemini 3. He says OpenAI would not trade positions with any other company. But the emergency decisions, the halt of monetisation plans, and the pressure from Google’s aggressive AI growth show how serious the situation has become.
