There seems to be a widespread impression that OpenAI has somehow “solved the Navier–Stokes equations,” and that this represents some enormous breakthrough for fluid dynamics and engineering.
It doesn’t.
The Navier–Stokes equations have been solved numerically for decades. We use them every day to simulate airflow around aircraft, weather systems, combustion, ocean currents, pipelines, turbines and pretty much anything else involving fluids.
Computational fluid dynamics is an enormous field, and the equations already do an extremely good job of describing real-world phenomena within the regimes where the continuum approximation is valid.
The Clay Millennium Prize problem is something much narrower and more abstract.
It asks whether, under a very specific set of mathematical assumptions, smooth solutions to the three-dimensional incompressible Navier–Stokes equations always remain smooth, or whether they can develop a singularity in finite time.
That is a deep mathematical question. But it is not the same thing as asking whether we can calculate how air flows over a wing.
An engineer does not need to know whether an ideal continuous velocity field remains infinitely well behaved for all time and at arbitrarily small spatial scales. They discretize the equations onto a finite mesh, run the simulation for a finite amount of time and compare the results against experiments and observations.
There is an additional irony here: actual fluids aren't continuous all the way down anyway. Air and water are made of molecules. Navier–Stokes is a continuum approximation that emerges at sufficiently large scales, and eventually stops being the appropriate description once you get down toward molecular scales.
So the Millennium problem is essentially a question about the mathematical properties of an idealized continuous model, not an engineering roadblock that humanity has been waiting 90 years for somebody to remove.
This doesn't make the result unimpressive. If the OpenAI proof survives scrutiny, it is an extraordinary achievement in mathematics and an extraordinary demonstration of what AI can now do.
But your aircraft isn't going to become more aerodynamic. Your CFD solver isn't suddenly obsolete. Weather models aren't about to become dramatically better.
“Navier–Stokes has been solved” makes for a great headline.
It just means something very different from what most people think it means.
