For more than a century, the Navier-Stokes equations have posed a seemingly simple yet very difficult question to answer: if a fluid starts moving smoothly, can it develop a singularity in finite time or will it remain mathematically stable forever? OpenAI claims to have found an affirmative answer for one formulation of the problem, although the mathematical community still has to review the proof before considering one of the millennium problems solved.
Read more Artificial intelligence accelerates the transformation of the professional office
To arrive at this hypothesis, the company has presented a 166-page paper describing a mathematical construction capable of generating a velocity that becomes unbounded at a certain instant. The work focuses on the three-dimensional Navier-Stokes equations and proposes a scenario in which an initially regular solution ends up developing a singularity.
The OpenAI team claims to have identified a situation where the Navier-Stokes equations cease to provide a regular solution
It is worth noting that this is not a purely abstract problem. These equations describe how fluids like water or air behave and are present in very different phenomena. Among them, the movement around an airplane, ocean currents, or turbulence, for example. Precisely for this reason, understanding whether their solutions can remain smooth indefinitely is one of the great challenges of mathematics.
In this regard, OpenAI’s proposal revolves around a rather striking phenomenon, a vortex that becomes increasingly narrow as its velocity increases. As the process approaches the moment of singularity, the core of the mentioned vortex shrinks and concentrates the motion in an increasingly smaller region. That is, the velocity can grow without limit even though the total energy associated with the system remains controlled.
The image provided by OpenAI helps to understand the peculiarity of the result. Instead of imagining an amount of fluid that simply explodes, the proof describes a spinning structure that stretches and compresses at the same time. Thus, the motion concentrates in a very small area, while the vortex evolves into a larger structure.
OpenAI claims that about 10,000 AI agents worked for 88 hours to build and verify the proof
For this construction to be valid, the researchers introduce a smooth and localized external force both in space and time. That is, the result presented by OpenAI holds that a solution can be constructed that starts from rest and ends up reaching an unbounded velocity in finite time, while keeping the total kinetic energy limited.
Read more Más Madrid and PSOE attack Ayuso over the penthouse and mock the president’s new residence
At the same time, the technological aspect of the announcement is also extraordinary. According to OpenAI, the proof was obtained after 88 hours of coordinated work by about 10,000 artificial intelligence agents. The company states that the system used is an experimental model still in training and significantly more capable than GPT-6 Astra. Instead of merely providing isolated answers, the agents would have worked in a coordinated way on different parts of the problem until reaching the final solution.
Furthermore, the company assures that the proof has not remained only on paper, but part of the work has been formalized using Lean, a mathematical proof assistant that allows translating arguments into a formal language and verifying them with computer tools. This formalization offers an additional layer of verification, although it does not replace independent examination by mathematicians.
And this is where the main nuance of the announcement appears. OpenAI presents the result as a proposed solution, but Navier-Stokes continues to be an unsolved problem for the Clay Mathematics Institute. The reason is that a proof of this magnitude needs to be analyzed very thoroughly by independent specialists. Errors in complex mathematical arguments can be extremely difficult to detect, and scientific validation requires more than just publishing a proof.
Not only that, but this matter gains even greater interest due to the parallel race of other researchers working on problems related to the formation of singularities in fluid equations. For this reason, OpenAI has pointed out that its work presents important differences compared to previous results and has defended the independence of its development.
Therefore, the discovery is still pending and cannot be considered finalized. If the proof passes the validation of the mathematical community, the result would not only represent a historic advance in the study of fluids but also an extraordinary demonstration of the capacity of artificial intelligence systems to participate in frontier mathematics. Until then, Navier-Stokes remains an open problem and OpenAI’s proposal will have to earn its place in history.
Read more Epidemic outbreaks increased in Madrid in 2025 with the flu driving cases and hospitalizations
