It’s Code Red Week for OpenAI

If your Spidey senses tingled when Sam Altman got super defensive when one of his own supporters and investors gently asked where the company is getting the money to pay for its trillion-dollar commitments, then your whole body is probably vibrating right now. With Google Gemini eating into the company’s market share and interest in ChatGPT slowing, OpenAI is in full-blown panic mode and appears prepared to throw everything possible at the wall to see what sticks.

Last week, Altman reportedly declared a “code red†to employees as he was growing concerned that the company was losing the sizable head start it had built up by being first to market with an LLM chatbot. Those concerns were likely not quelled by Sensor Tower data, reported by Digital Trends, that shows ChatGPT has lost about 3% of its market share in recent months, and its steady growth has started to significantly slow down. Meanwhile, Google Gemini has seen a major uptick in usage since August. Notably, daily usage of Gemini has also been climbing while ChatGPT reportedly dropped. The bad buzz for Sam Altman has only gotten worse since Gemini 3 launched last month and has since been widely praised as superior to GPT-5.

So what’s OpenAI doing about this? Well, it started by putting out some fires. Over the weekend, the company put the kibosh on promotional messages in ChatGPT that users thought were a little too much like advertisements. The messages offered suggestions for apps or services that a user might be interested in, often unrelated to the conversation the user was having with ChatGPT. Users revolted against the supposed test feature, and the company ultimately backed down, killing the recommendations (for now).

Next step: reestablish the status quo. Some data shows that people are slowly drifting away from ChatGPT, so OpenAI dropped its own research that shows that people love using its chatbot, actually. In a survey conducted by OpenAI of about 9,000 workers across 100 companies, it found that people report saving about 40-60 minutes per day on their work by using AI tools. That data was released as part of a larger report on enterprise AI, which highlighted how weekly messages to ChatGPT Enterprise increased by about eight times over the course of the past year. So nothing to worry about here!

What’s next? Well, it’s time for OpenAI to change the narrative. The company is ready to release a new model, GPT-5.2, sometime this week, according to The Verge. That model has supposedly outperformed Gemini 3 in OpenAI’s internal tests, per The Information. Altman is also going to make an appearance on “The Tonight Show†with Jimmy Fallon on Monday, where he will probably get his hair tussled and play charades while talking up the next version of ChatGPT.

It’s rare to see any single moment as make-or-break for a company, but there is a very real threat to OpenAI that if it can’t right the ship and reestablish ChatGPT as the genericization of the AI space, the scrutiny from investors who no longer feel like the company is a sure bet will start asking questions that OpenAI doesn’t have the answers for.