The Navier-Stokes equations describe how fluids like water and air flow, and are used successfully every single day by engineers designing aircraft, predicting weather, and modelling ocean currents. Yet a startlingly basic mathematical question about these same equations remains completely unanswered: does a smooth, well-behaved solution always exist for all future time, given smooth starting conditions, in ordinary three-dimensional space? This is one of the seven Clay Millennium Prize Problems, carrying a $1 million reward.
Every time you watch a weather forecast, board an airplane, or see a computer-animated ocean in a film, you're benefiting from the Navier-Stokes equations — a set of equations describing how fluids (liquids and gases) move and interact, named after Claude-Louis Navier and George Gabriel Stokes, who developed them in the 1820s–1840s. Engineers solve these equations numerically every day with tremendous practical success. And yet, remarkably, mathematicians cannot prove a seemingly basic fact about them: that a smooth, sensible solution to these equations always continues to exist, forever, without suddenly "blowing up" into some kind of mathematical nonsense.
Engineers don't need a complete mathematical proof of long-term existence to use these equations productively — they typically only need numerical approximations valid over comparatively short, practically relevant timeframes (a weather forecast for the next week, an aircraft's behaviour during a specific flight). The open mathematical question is a much more demanding, purely theoretical one: whether a fully rigorous, exact, smooth solution is mathematically guaranteed to exist for all time, without exception, under completely general starting conditions — a question of pure mathematical completeness that doesn't need to be answered for the equations to remain enormously practically useful in the meantime.
If a solution could genuinely "blow up" (some physical quantity like fluid velocity becoming truly infinite in finite time, rather than merely very large), it would suggest the equations themselves fail to fully capture real fluid behaviour at extreme scales — perhaps signalling that some additional physics (not captured by the idealised mathematical model) must take over in such extreme situations. Resolving this question, either way, would represent a landmark achievement in understanding the fundamental mathematical limits of one of physics's most practically important equation systems.
The Navier-Stokes equations are a system of nonlinear partial differential equations (see the Differential Equation entry) expressing, essentially, Newton's second law (force equals mass times acceleration) applied to a continuously flowing fluid, incorporating pressure, viscosity (internal friction), and external forces. Their nonlinearity — arising from the fluid's velocity affecting its own future rate of change — is precisely the source of the deep mathematical difficulty in question.
The official Clay Mathematics Institute formulation (authored by Charles Fefferman, 2000) asks for a proof — or disproof via explicit counterexample — that smooth solutions exist for all time in three-dimensional space with smooth, sufficiently well-behaved (technically, "Schwartz-class") initial data, incorporating a physically reasonable energy bound. Either a full proof of existence, or a specific counterexample showing solutions can genuinely blow up in finite time, would resolve the problem and claim the prize.
In two spatial dimensions, this existence-and-smoothness question was fully resolved decades ago (essentially by the 1960s, through work of Olga Ladyzhenskaya and others) — smooth solutions are proven to exist for all time in 2D. It's specifically the physically realistic three-dimensional case that remains stubbornly open, illustrating how a seemingly minor change in dimension can transform a settled question into one of the hardest open problems in mathematics.
Jean Leray proved in 1934 that so-called weak solutions (a more relaxed, less strict mathematical notion of "solution," not required to be fully smooth) do exist for all time in three dimensions. This is a genuinely important partial result, but weak solutions are explicitly permitted to have certain mathematical irregularities and potentially non-unique behaviour — meaning Leray's 1934 result, while a foundational and celebrated contribution, is considerably weaker than what the Clay Millennium Prize actually demands: the existence specifically of fully smooth, physically well-behaved solutions.
A major partial result, proven by Luis Caffarelli, Robert Kohn, and Louis Nirenberg in 1982, shows that if a weak solution to the 3D Navier-Stokes equations does develop singularities (points where the solution fails to be smooth), the set of such "bad" points must be extremely small in a precise measure-theoretic sense — effectively confining any potential blow-up behaviour to a very restricted, low-dimensional set of points in space-time. This remains among the strongest rigorous partial-regularity results available, and any full resolution of the Millennium Problem will very likely need to substantially extend or circumvent techniques from this line of research.
In two dimensions, a certain physical quantity called vorticity (roughly, the local spinning or rotational tendency of the fluid) obeys a comparatively simple conservation-like behaviour that provides the key mathematical control needed to rule out blow-up. In three dimensions, vorticity can be dramatically "stretched" by the fluid's own three-dimensional motion (a phenomenon called vortex stretching), potentially amplifying without bound — and it is precisely this vortex-stretching mechanism, entirely absent in the simpler 2D case, that is widely suspected (though not proven) to be the essential source of the full 3D problem's mathematical difficulty. This effect is loosely analogous to how a spinning ice skater speeds up by pulling in their arms.
Physically, unresolved fine-scale behaviour in fluid flow is closely associated with turbulence — the famously complex, chaotic behaviour of real-world fluids at high flow speeds, still not fully understood even at the purely physical (let alone rigorously mathematical) level despite over a century of intensive study by physicists and engineers. Werner Heisenberg reportedly remarked, apocryphally but tellingly, that he would ask God about turbulence first, expecting a real answer, since he doubted even quantum mechanics would prove as difficult. Whether the Navier-Stokes existence problem is deeply connected to a genuine mathematical understanding of physical turbulence, or is instead a somewhat separate purely mathematical question, remains itself a topic of ongoing discussion among specialists.
Terence Tao (2016) constructed a modified, "averaged" version of the Navier-Stokes equations — differing from the true equations in a specific technical way — that provably DOES exhibit finite-time blow-up, demonstrating that if the true 3D Navier-Stokes equations do turn out to have smooth solutions for all time, this fact must depend on some genuinely special mathematical structure specific to the real equations, and not merely on generic features shared by any superficially similar nonlinear fluid-type system. This result is widely regarded as significant evidence that resolving the genuine Millennium Problem, in either direction, will likely require substantial new mathematical ideas beyond currently available techniques.