OpenAI Solved Navier-Stokes Millennium Prize Problem

An artificial intelligence model identified the specific conditions under which the equations fail.

Updated on Sept. 18, 2026 in Mathematics

Isometric editorial illustration showing a swirling vortex structure transitioning into a cluster of distinct geometric spheres, representing fluid dynamics at the molecular scale. — OpenAI announced that its artificial intelligence model successfully identified specific physical limits where Navier-Stokes fluid-flow equations lose accuracy at the molecular level. 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 an artificial intelligence model successfully addressed the Navier-Stokes Millennium Prize Problem. The system identified instances where the fluid-flow equations predict physically impossible infinite speeds.

Why it matters

The findings explain why Navier-Stokes equations fail at extreme scales, demonstrating that the assumption of continuous substances becomes inaccurate when analyzed at the molecular level.

The study utilized simulations of 155 billion water molecules conducted in 2024 to map fluid behavior. The AI model identified a vortex singularity that emerges in air at a width of 70 nanometres.

The players

OpenAI

This San Francisco-based artificial intelligence company develops large-scale machine learning models and research applications.

Yu Deng

Yu Deng is a mathematician who explored the implications of molecular-level equations for fluid dynamics research.

The details

The AI solution demonstrates that fluid vortices stretch into thin, elongated shapes until they hit the molecular scale, where the continuity assumption breaks down. The model suggests that the Boltzmann equation, which treats gas as individual molecules, is more accurate than traditional equations in these specific scenarios.

Researchers simulated 155 billion water molecules throughout 2024.
Yu Deng discussed implications for the Boltzmann equation in July 2026.
OpenAI announced the Navier-Stokes solution in September 2026.

The Big Picture

This discovery marks a potential resolution to one of the seven foundational challenges defined by the Clay Mathematics Institute. It shifts the paradigm of fluid dynamics from assuming continuous matter to accounting for discrete molecular interactions.

By validating the breakdown of continuous fluid models at the 70-nanometre scale, this research could refine future micro-engineering and nanotechnology designs. Understanding these limits improves the precision of atmospheric and industrial flow simulations.

The takeaway

The solution highlights the critical importance of considering molecular granularity when applying macro-scale physical laws. Readers interested in physics should note that even established fundamental equations reach boundaries where their underlying assumptions no longer hold true.