Clay Math Institute weighs in on OpenAI’s Navier-Stokes claim

OpenAI announced on September 8 that an unreleased internal AI model, deploying roughly 10,000 autonomous agents over 88 hours, had produced a proof that the three-dimensional Navier-Stokes equations can develop a singularity in finite time — a result that, if validated, would resolve one of the six remaining Millennium Prize Problems posed by the Clay Mathematics Institute in 2000. The announcement has been overshadowed by an escalating dispute with NYU mathematician Tristan Buckmaster and Anthropic researcher Levent Alpöge over research credit and the possible role of private data.nature+2

A Claimed Breakthrough in Fluid Dynamics

The Navier-Stokes equations describe how fluids such as water and air move, and the question of whether their solutions always remain well-behaved or can "blow up" — producing physically impossible infinite speeds — has been open for decades. OpenAI said its proof demonstrates that an initially smooth fluid at rest can develop such a singularity under smooth external forcing, with the result verified using Lean, a formal theorem-proving language. The effort cost "millions of dollars" in compute, according to OpenAI executives on a press call.axios+3

Scientific American S&P Global Inc. ↗ 0.07% reported that the Lean verification "all but guarantees" the proof's correctness, though the broader mathematical community has yet to complete peer review. The Clay Mathematics Institute said on September 11 that it "shares in the excitement" as the result is "analysed and interrogated," stopping short of declaring the problem solved.scientificamerican+1

The Credit Dispute

The achievement arrived amid allegations from Buckmaster that OpenAI accelerated its work after learning of related research he and Alpöge had been conducting, partly using OpenAI's Codex tool. Buckmaster and Alpöge released their own results on September 7, showing finite-time blowup for related but simplified fluid equations. According to reporting by TechRepublic, Buckmaster alleges that OpenAI offered him sole authorship on its Navier-Stokes solution on the condition that Alpöge's name be removed — a claim denied by OpenAI researcher Sébastien Bubeck.techrepublic+3

OpenAI stated that its researchers and agents "did not see any of their work through any means until they released it publicly" and that "no specific user data was accessed in order to solve this problem". However, the company acknowledged it "cannot rule out that de-identified data derived from their usage of our products helped improve our models". Buckmaster publicly called Bubeck's account "wildly false", and Bubeck subsequently offered an apology for a remark he made during their exchanges, according to The Next Web.thenextweb+4

Wider Fallout

The dispute has rippled through the mathematics community. An open letter signed by leading mathematicians criticized AI companies for rushing to solve famous problems without proper attribution. Fields Medalist Deng Yu quipped on social media that if AI solves all math problems, "I'll go home and write my yuri novel," while cautioning that AI has created "an atmosphere of restlessness and impatience in mathematical circles". The Clay Mathematics Institute's formal response notably praised the potential for "new human understanding" but made no mention of prize eligibility.thenews+1