The Navier-Stokes Problem: Overview of the 3D Regularity Question

The Navier-Stokes problem asks a deceptively simple question: if you start a 3D fluid flowing smoothly, does it stay smooth forever? Or can the motion become so concentrated that the mathematical solution breaks down?

This page is the broad map of that question: what the problem asks, what is already known, why three dimensions are difficult, and how the Millennium Prize version fits into the larger story.

For the current solved/open status and failed-solution context, see Is Navier-Stokes solved?. For the formal Clay criteria and proof targets, see the dedicated Clay guide.

The Navier-Stokes existence and smoothness problem asks whether, for every sufficiently smooth, divergence-free initial datum $u_0 \in C^\infty(\mathbb{R}^3)$ with suitable decay and $f \equiv 0$, the incompressible Navier-Stokes system admits a solution $u \in C^\infty(\mathbb{R}^3 \times [0,\infty))$; alternatively, whether admissible smooth data can lead to finite-time breakdown.

This page is an overview of the mathematical terrain. For a concise status page, including accepted results and published claims, see Is It Solved?. For the formal Clay criteria and proof targets, see the dedicated Clay guide.

Unsolved doesn't mean untouched. Nearly a century of deep mathematical work has mapped the terrain and revealed exactly where the difficulty lies and why it won't yield to the tools we have:

• Weak solutions exist globally (Leray, 1934). Relax the notion of "solution" to allow rough, averaged-out behavior and solutions exist for all time. Smooth? Nobody can prove it. More on approaches →
• 2D is solved. Smooth solutions always exist globally in two dimensions, but three dimensions is an entirely different beast. Why 3D is harder →
• Singularities, if they exist, are rare (CKN, 1982). Caffarelli, Kohn, and Nirenberg proved that the set of possible singularities has zero one-dimensional parabolic Hausdorff measure. Subproblems and partial results →
• Smooth solutions exist briefly. Start with smooth data and you get a unique smooth solution for some time interval, but whether that interval can always be extended to infinity is exactly what's unknown.
• The precise formulation was set out by Charles Fefferman for the Clay Mathematics Institute. Read the formal criteria →

The following results constitute the main partial progress:

• Leray (1934): For $u_0 \in L^2(\mathbb{R}^3)$, global weak solutions (now called Leray-Hopf solutions) exist and satisfy the energy inequality. Uniqueness and regularity of these solutions remain open. Approaches →
• 2D global regularity: Ladyzhenskaya (1959) established global existence and uniqueness of smooth solutions in $\mathbb{R}^2$. The key is that enstrophy is controlled in 2D. Why 3D is different →
• CKN (1982): Caffarelli, Kohn, and Nirenberg proved that the one-dimensional parabolic Hausdorff measure of the singular set of any suitable weak solution is zero. Subproblems →
• Local existence: For sufficiently regular data, unique local smooth solutions exist; in critical spaces such as $\dot{H}^{1/2}$, one has local well-posedness in the mild-solution framework. The open question is whether these solutions can always be continued for all time.
• Clay formulation (2000): Fefferman's problem statement specifies the exact function spaces, decay conditions, and what constitutes a valid proof or disproof. The Millennium Problem →

Here's the core difficulty. A fluid's own motion can push activity to smaller and smaller scales faster than current estimates can control. In three dimensions, the math doesn't give us enough control to rule this out. It doesn't let us prove it happens, either.

This isn't about cleverness. It isn't about computing power. The known mathematical tools are fundamentally insufficient, and that tension between concentration and dissipation is exactly why solving the problem would require genuinely new mathematics.

Supercriticality, the scaling gap, why 3D turbulence is fundamentally different: for the full story, see Why the Navier-Stokes Problem Is So Hard.

The 3D Navier-Stokes equations are supercritical with respect to the natural energy estimate: the $L^2$ norm is controlled, but the scaling-critical regularity lies at $\dot{H}^{1/2}$, which is not propagated by the energy inequality alone. The nonlinear term $(u \cdot \nabla)u$ can in principle transfer energy to arbitrarily fine scales faster than the Laplacian dissipates it.

This is the essential analytical obstruction, and no existing technique closes the gap. For a detailed treatment, see Why It's Hard.

In 2000, the Clay Mathematics Institute named this regularity question one of seven Millennium Prize Problems, offering $1,000,000 for a correct proof or disproof. This overview explains why that prize question matters in context.

For the exact formulation, read the official problem statement, explained. For where things stand right now, see the current status (July 2026).

The Clay Mathematics Institute included this regularity question in its 2000 list of Millennium Prize Problems. The formulation, written by C. Fefferman, is posed on $\mathbb{R}^3$ and $\mathbb{T}^3$ and accepts either a proof of global smooth existence or a construction of finite-time breakdown under the accepted alternatives.

Read the official problem statement, explained, or check the current status (July 2026).

This page is a map. The territory runs deep. Pick a thread:

• Is It Solved? No. Here's the current status, major published claims and the technical reasons they failed under expert scrutiny.
• The Millennium Problem Demands. Precise ones.
• Why It's Hard Supercriticality, turbulence, and the scaling gap that blocks every known approach from getting anywhere near a proof.
• Solving the Equations What "solution" means: exact formulas, CFD, and the mathematical question, and why they're different things.

For detailed treatments of the topics introduced above:

• Is It Solved? Status of the problem, published and retracted claims, verification standards.
• The Millennium Problem Fefferman's formulation, function spaces, and what constitutes a valid proof or counterexample.
• Why It's Hard The supercritical scaling, the role of the nonlinearity, and the gap between energy-level control and regularity.
• Solving the Equations The three senses of solution: closed-form, discrete approximation, and well-posedness, and why numerical practice does not resolve the Clay question.

Mathematicians haven't just stared at the problem. They've developed powerful tools, partial results, and entirely new fields of analysis trying to crack it. The work continues.

See the progress so far →

The Navier-Stokes problem has driven major developments in harmonic analysis, functional analysis, and geometric measure theory over the past century. Partial regularity results, conditional blowup criteria (Beale-Kato-Majda, Escauriaza-Seregin-Šverák), and model-problem analyses continue to sharpen our understanding of where the boundary between regularity and potential singularity lies.

Survey of progress →