Two “code red” alerts — the first from a veteran tech giant worried about a buzzy AI upstart, the second from the AI upstart after the tech giant gained ground.
What a difference three years can make.
News of a recent Sam Altman memo to OpenAI employees, first reported by The Information, is reverberating around the tech world and highlighting the competitive heat it’s facing as Google narrows the gap in the AI race.
On Monday, Altman reportedly told OpenAI employees in an internal Slack memo that he was issuing a “code red” and that the company would be putting more resources into ChatGPT and delaying other products as a result.
Altman’s memo illustrates just how much the AI race has changed. In 2022, Google’s management issued its own “code red” in the wake of ChatGPT’s launch, a moment that illustrated in sharp relief just how far behind the search giant was in the AI race despite financing the breakthrough research that paved the way for AI’s development.
Three years later, it’s clear that OpenAI’s throne is under threat. Here are some of the pressure points it’s facing as Google nips at its heels.
Google is catching up
The elephant in the OpenAI room is Google’s Gemini 3 AI model, which debuted to widespread praise.
The model’s capabilities demonstrated that Google is no longer far behind in the AI race. It’s not just OpenAI that’s unnerved, either. Nvidia, the world’s most valuable company by market cap, recently found itself defending its AI chips after a report about Google’s own chip progress.
The search giant said in November that Gemini had more than 650 million monthly active users, a large increase from the 450 million such users it reported in July. In comparison, OpenAI has said nearly 800 million weekly active users.
Salesforce CEO Marc Benioff recently said that he was ditching ChatGPT in favor of Gemini 3 because of Gemini’s “insane” improvement.
“Holy shit,” Benioff wrote on X last month. “I’ve used ChatGPT every day for 3 years. Just spent 2 hours on Gemini 3. I’m not going back. The leap is insane — reasoning, speed, images, video… everything is sharper and faster. It feels like the world just changed, again.”
Last month, Google launched “Nano Banana Pro,” its AI image generator, showcasing hyper-realistic images that users quickly used to imagine tech CEOs hanging out together or pretend famous Thanksgiving dinner table guests.
Altman’s “code red,” according to The Information’s report, specifically mentions Gemini 3 and teases a coming OpenAI model that it says tested “ahead” of Google’s flagship model, as well as mentions prioritizing OpenAI’s Imagegen image generation model for ChatGPT users.
Google’s advertising cash cow can fund its AI — while OpenAI faces a $1.4 trillion bill
The AI game is an expensive one, and Google has the advantage of being a cash-generating advertising juggernaut.
Sure, Google plans to spend between $91 billion and $93 billion this year on cap ex, much of which is going toward AI costs. But it also brought in $100 billion in revenue in just the last quarter alone — $74.18 billion of which came from its advertising business.
And unlike OpenAI, Google can leverage its massive size for a full-stack advantage, allowing it to control AI development from research to chip manufacturing to its in-house cloud, which hosts everything.
Meanwhile, some on Wall Street have raised concerns about OpenAI’s mounting AI spending commitments, which tally at least $1.4 trillion over the next eight years. In response, Altman has said OpenAI is on track to bring in $20 billion in revenue this year, and expects its annualized revenue to grow to hundreds of billions in the coming years.
But OpenAI is still figuring out its own ads business — the launch of which could be delayed by Altman’s “code red,” according to The Information.
OpenAI has a head start — but Google has a platform advantage
OpenAI hasn’t squandered its head start, and it’s landed some major wins this year.
In recent months, OpenAI has made significant plays into other industries, including social media with Sora, its TikTok-esque AI video generation app. In a direct shot at Google Chrome, OpenAI also launched Atlas, its own web browser.
And it sounds like OpenAI has more up its sleeve as it battles the bottleneck of lining up enough compute and energy to power its developments.
OpenAI executives have said compute constraints are holding back other initiatives, like making ChatGPT Pulse, a personalized update feature within the chatbot for Pro users, available to everyone. Last week, Bill Peebles, OpenAI’s head of Sora, announced that free users would face significant cuts in the number of videos they could generate per day.
ChatGPT also remains synonymous with AI — not unlike Google and online search. That will likely help continue to drive app downloads and usage and could also stave off Google’s attempts to convince users to switch to Gemini or Google’s other AI-infused products.
But humans are creatures of habit, and many already use a Google product or service everyday — a platform advantage that the tech giant is already utilizing to siphon away ChatGPT users.
Silicon Valley’s history is built on startup disrupting the status quo.
Now, with OpenAI (smartly) looking over its shoulder, we get to watch the AI race heat up as Google, a former startup, gets its AI legs and hits its stride.
For OpenAI, it’s a reminder that tech giants can put up quite a fight when facing the prospect of being disrupted — and sometimes, can turn the tables.
“I try not to think about competitors too much,” Altman said last May before critiquing Google’s aesthetic.
It sounds like those days are gone.
