2024年伊始，OpenAI气势低迷

OpenAI needs more compute. It’s not clear what it expects you or anyone else to do about that, but it would very much like you to know that it needs more compute. In a strange video posted by the company on X, OpenAI President Greg Brockman explained with a sense of desperation that demand for its products, like image generation, is making it hard for the company to launch new features and invest in research.

In the video—which reads like one part justification for the company’s massive investments into data center projects despite having nowhere near the revenue needed to fund the buildout, and one part an attempt to justify falling behind in the AI race on the grounds of being too popular—Brockman basically flags compute as the key to the company’s success.

“OpenAI did not set out with the thesis that compute was the path to progress. It’s that we tried everything else, and the thing that worked was compute, was scale,” he says, in what could be read as the head of the company trying to insist to investors that if they just let OpenAI spend endlessly, it’ll eventually crack the code that has made generating meaningful revenue so elusive.

Compute enabled our first image generation launch (and a +32% jump in WAU over the following weeks) as well as our latest image generation launch yesterday. We have a lot more coming… and need a lot more compute. pic.twitter.com/rHfQv1aLKS— OpenAI (@OpenAI) December 17, 2025For now, OpenAI’s compute is capped, and it’s put the company in the unfortunate position of having to make a Sophie’s Choice-style decision on which parts of its model it cares about the most.

“When we had our image generation launch in March that went viral, we did not have enough compute to keep that going. And so we made some very painful decisions to take a bunch of compute from research and move it to our deployment to try to be able to meet the demand. And that was really sacrificing the future for the present,” Brockman explained.

It’s certainly not lost on OpenAI that the general perception is that it’s fallen behind in the AI race. Following the launch of Google’s Gemini 3 model, which became the talk of the AI space, OpenAI CEO Sam Altman declared a “code red” and pushed out a new model to try to make up for lost ground. With that knowledge, it’s hard not to hear Brockman’s explanation as something of an excuse, as if to say, “We would have loved to have made some incredible breakthroughs, but we had to prop up our very popular slop machine, remember how much everyone loved that?

”To drive home the point of how important computing power is to the company’s prospects, OpenAI followed up the video with an infographic titled “OpenAI’s Compute Flywheel,” showing how more compute leads to better products, which leads to more revenue. That does seem like a better flywheel than its current one, in which it shifts money around between a handful of companies while boosting their bottom line with no obvious indication that money was actually used for anything.

“We want to be ahead of the curve,” Brockman said. “And the truth is, I don’t think we will be, no matter how ambitious we can dream of being right now. I think demand will far exceed what we can think of.” Of course, the prospect of just not pouring resources into making slop like Sam Altman with many collars or Sam Altman as a sexy fireman on an inaccurate calendar is apparently not feasible.