OpenAI Claims AI Agents Solves Millennium Prize Problem on Fluid Equations

San Francisco, California – September 8, 2026

OpenAI announced Tuesday that it has produced a proof showing the Navier-Stokes equations—fundamental to fluid dynamics—can develop a singularity in finite time, resolving one of mathematics' most enduring open problems.

The AI company said its internal model, working through 10,000 coordinated agents over 88 hours, generated a solution demonstrating that fluid velocity can grow unbounded even when total energy remains finite. The result was independently verified through Lean, a formal proof-checking system, over 17 hours.

The Navier-Stokes existence and smoothness problem is one of seven Millennium Prize Problems posed by the Clay Mathematics Institute. It asks whether three-dimensional fluid flow can "blow up"—where velocity becomes infinite while energy stays finite. For nearly 90 years, mathematicians have worked to determine if such singularities can occur.

"This problem has remained unsolved for 200 years because the Navier-Stokes equations are just so enormously complex, and the pen-and-paper calculations you need to do in order to solve this problem are just mind-bogglingly intricate," said Venkat Chandrasekaran, OpenAI computer scientist, at a press briefing.

OpenAI said its approach followed a multi-stage process. First, approximately 100 agents worked for 50 hours on a related Euler equations problem—a frictionless version of Navier-Stokes—producing a disproof of regularity. Then 10,000 agents tackled the full Navier-Stokes system over 88 hours. The agents sent 2.7 million messages and used 130 billion output tokens specifically for Navier-Stokes.

The solution takes the form of a vortex—a swirling fluid structure that spirals inward while stretching axially. As the vortex core contracts, fluid accelerates along the axis while energy remains bounded. This configuration shows the equations can break down through internal fluid motion rather than external forcing.

The announcement came amid controversy. Tristan Buckmaster, a mathematics professor at New York University, and Levent Alpöge at Anthropic had been working on related problems using AI assistance. On September 7, they released partial results for the Euler equations. OpenAI said it began its Navier-Stokes effort on September 1 after hearing rumors of their work, but denied accessing their prompts or proofs.

"We did not use their prompts or proofs to prompt our models or direct our agents," said Sébastien Bubeck, OpenAI researcher who led the project, at a press conference. "We, whether it's the researchers or the agents, did not see any of their work until they were released publicly."

Charles Fefferman of Princeton University, who wrote the Clay Institute's official problem description, called the Euler work by Buckmaster and Alpöge the "heroic" achievement. Terence Tao, Fields Medalist and mathematician at UCLA, said he would prefer to see the work published through peer-reviewed channels.

The Navier-Stokes equations govern fluid motion in aerospace design, weather forecasting, and blood flow modeling. A proof that solutions can break down has implications for computational fluid dynamics and the reliability of numerical simulations.

The proof is available as a preprint on OpenAI's website. Independent mathematicians are now working to verify the result. Formal verification in Lean provides initial confidence, but peer review remains the standard for mathematical acceptance.

The other six Millennium Prize Problems—including the Riemann Hypothesis and P vs NP—remain unsolved. Solutions to each carry a $1 million prize.

Source: OpenAI, September 8, 2026. The Guardian, September 8, 2026. Nature, September 8, 2026. Quanta Magazine, September 8, 2026. New Scientist, September 8, 2026.