OpenAI Solved Navier-Stokes Millennium Prize Problem

The AI identified where fluid equations fail by predicting physically impossible infinite speeds in air.

Updated on Sept. 18, 2026 in Mathematics

Isometric editorial illustration featuring a complex layered geometric structure representing a microscopic vortex, rendered in deep teal and oxblood tones. — OpenAI’s AI model has identified a fundamental discrepancy in the Navier-Stokes equations, proving they predict impossible physical states at the 70-nanometre scale. AI Illustration. Upload story photo >

Live Poll

Do you trust artificial intelligence models to accurately solve complex scientific and mathematical problems?

OpenAI has announced that its artificial intelligence model successfully addressed the Navier-Stokes Millennium Prize Problem. The AI demonstrated how the equations predict impossible infinite speeds when a vortex of fluid shrinks to a width of 70 nanometres.

Why it matters

This finding clarifies the breakdown of fluid dynamics models at the molecular level, highlighting a fundamental discrepancy between continuous flow assumptions and discrete molecular reality. It resolves a long-standing mathematical impasse by showing where the standard model ceases to be accurate.

The AI identified that Navier-Stokes equations fail when air vortex widths approach 70 nanometres, a scale where molecular spacing renders continuous models inaccurate. This follows 2024 supercomputer simulations of 155 billion water molecules, which confirmed that discrete particle behavior diverges from the continuous flow predicted by classic equations.

The players

OpenAI

A San Francisco-based artificial intelligence research organization focused on large-scale machine learning and generative models.

Yu Deng

A researcher who evaluated the implications of the AI findings for the Boltzmann equation.

The details

The Navier-Stokes equations represent fluid as a continuous substance, an approximation that fails when physical dimensions reach the scale of individual air molecules. OpenAI’s AI model analyzed how these fluid vortices stretch into thin shapes, discovering that the equations produce infinite speeds at a threshold of 70 nanometres. By comparison, the Boltzmann equation — a mathematical framework that treats gas as a collection of individual particles — avoids these inaccuracies by accounting for discrete molecular interactions.

2024: Researchers performed supercomputer simulations of 155 billion water molecules.
July 2026: Yu Deng discussed implications for the Boltzmann equation.
September 11, 2026: OpenAI announced the Navier-Stokes solution.

The Tech Race

This development represents a departure from traditional mathematical approaches to the seven Millennium Prize Problems, which have long been dominated by human theoretical proofs. By successfully using AI to identify the precise limits of the Navier-Stokes equations, OpenAI has established a new standard for computational research in classical mathematics.

This finding primarily influences the development of high-fidelity simulations used in aerospace and weather forecasting, where fluid dynamics are critical. Scientists will likely shift toward particle-based simulations for small-scale flows to avoid the errors identified in the continuous models.

The takeaway

The solution demonstrates that current fluid dynamic modeling relies on outdated assumptions at the nanometre scale. Observers should watch for the official review of this solution by the Millennium Prize committee to confirm the eligibility for the US$1-million award.