According to CNBC, OpenAI CEO Sam Altman sent a memo to staff on Monday, September 23, 2025, declaring a “code red” effort to improve the ChatGPT chatbot. The company will pull back on investments in areas like health, shopping, and advertising to prioritize ChatGPT development. This comes as OpenAI faces stiff competition, despite ChatGPT having over 800 million weekly users. Google recently announced its Gemini 3 AI model, which topped benchmarks and has been widely praised. Google also revealed its Gemini app has 650 million monthly active users, while its AI Overviews in search serve 2 billion users monthly. Altman even publicly congratulated Google on Gemini 3 in a post on X last month.
OpenAI’s Strategic Retreat
So, a “code red.” That’s serious language for a company that’s been the undisputed frontrunner. Here’s the thing: it signals a strategic retreat. They’re basically saying, “We spread ourselves too thin.” Projects in health and shopping? Shelved, or at least deprioritized. The entire focus is now on defending the core product. Nick Turley, head of ChatGPT, echoed this in a post on X, talking about making ChatGPT “more intuitive and personal.” But that’s table stakes now. The memo, first reported by The Information, feels like a reaction to market pressure, not a visionary leap.
The Google and Anthropic Problem
And that pressure is very real. Look at Google’s numbers. 650 million monthly users for the Gemini app is a massive installed base to leverage. 2 billion monthly users seeing AI Overviews? That’s an insane distribution channel OpenAI can only dream of. Gemini 3 isn’t just a good model; it’s a good model hooked directly to the world‘s largest search engine and built into Android phones. Then there’s Anthropic, which has been steadily building a reputation for robust, reliable models. OpenAI’s lead in mindshare is eroding, and its technical moat might be, too. Altman congratulating Google is a classy move, but you have to wonder what’s being said behind closed doors.
What “Code Red” Actually Means
What does a “code red” execution look like? It probably means all hands on deck for the next major ChatGPT model update. It means prioritizing raw capability and speed over everything else. But it also hints at potential internal turmoil or at least a major reassessment. Are they worried about user growth stalling? Are power users drifting to other platforms? The pullback from advertising is especially telling. That was a potential revenue goldmine they’re now putting on ice. The entire strategy has shifted from expansion to consolidation in a matter of months. That’s a jarring pivot for a company of this scale.
A New Phase for the AI Race
We’re entering a new, brutal phase of the AI wars. The era of OpenAI having the field mostly to itself is over. It’s now a straight-up slugfest with well-resourced, deeply integrated giants like Google and focused, high-quality competitors like Anthropic. For businesses integrating this tech, that’s good news—more competition drives innovation. But for OpenAI, the challenge is existential. Can they innovate fast enough on a single front to stay ahead? Or is this the moment the pack finally catches the leader? The “code red” memo is their answer: they know they can’t afford to lose this sprint.
