Sam Altman Says OpenAI Won’t Stop Releasing AI Math Results, Hints at Larger Stockpile
Altman says OpenAI has more math discoveries beyond the hundreds of proofs released publicly, and that it spent hundreds of millions of dollars on the work. About $10 million went to its Navier–Stokes solution, he said.
OpenAI CEO Sam Altman said in October 2026 that he sympathizes with mathematicians overwhelmed by the company’s stream of AI-generated results — but has no plans to slow down, and suggested the publicly released findings represent only a fraction of what the company holds.
Speaking on the Cheeky Pint podcast released Thursday, Altman said OpenAI has more mathematical discoveries beyond the hundreds of proofs already made public. Asked directly whether the company is sitting on additional results, he replied: “We do. Of course.” He also estimated OpenAI has spent hundreds of millions of dollars on mathematical problems, including roughly $10 million — “maybe a little more” — on its solution to the Navier-Stokes problem, one of mathematics’s Millennium Prize problems.
Altman argued that the public’s right to new scientific knowledge “outweighs” the disruption caused by OpenAI’s releases. “Humanity deserves fundamental knowledge, and we should not be sitting on a major fundamental science discovery,” he said. He acknowledged, however, that the company needs to improve how it communicates its findings.
Mathematicians speaking to The Verge have repeatedly criticized OpenAI for inadequate attribution, little context, and insufficient verification in its announcements. This month alone, OpenAI released hundreds of mathematical results, often accompanied by writeups researchers described as poorly written and hard to follow.
The company is also facing criticism for withholding details on one of three recently announced breakthroughs — a result involving Yang-Mills theory, the mathematical foundations of modern particle physics. OpenAI says it is following the recommendations of a hand-picked group of experts, an approach that contradicts advice from a previous group of mathematicians the company consulted.
Altman acknowledged that OpenAI’s pursuit of mathematical problems “got out of control” for a period, describing it as deliberate. He referenced an internal Slack channel called “Model Name-Hold My Beer,” where employees tested AI models against the hardest problems they could find. The spending involved suggests researchers may face growing difficulty competing with the resources AI labs are directing at scientific problems.