AI-Assisted Proofs Crack Navier-Stokes and Riemann
Original: Unforced Navier-Stokes Singularity and the Generalized Riemann Hypothesis
Why This Matters
Three of math's hardest open problems may have fallen in a single week, with AI playing a central role.
Mathematicians Alpöge, Buckmaster, and Coiculescu, using heavy AI assistance, have produced Lean-formalized proofs of finite-time blowup for 3D Euler and Navier-Stokes equations, plus a Lean-only proof of the Generalized Riemann Hypothesis, as of November 20, 2026.
In what Fields medalist Terence Tao described as results that 'will be reported widely in other outlets,' three separate breakthroughs were announced on his blog. First, Levent Alpöge, Tristan Buckmaster, and Matei Coiculescu produced a human-readable proof of finite-time blowup for the unforced incompressible Euler equations in three dimensions, formally verified in Lean. Second, Alpöge and Buckmaster generated a proof of the same blowup phenomenon for the unforced Navier-Stokes equations—one of the original Millennium Prize Problems—following the expected path of numerically constructing an approximate discretely self-similar solution and applying rigorous perturbation theory. Tao notes the paper is 'technically human-readable' but needs further professional polish. Third, and most opaque, Alpöge alone used an internal AI model to produce an enormous Lean-only proof of the Generalized Riemann Hypothesis. Initial analysis suggests it draws on the full machinery of the Langlands program and arithmetic geometry. All three results relied heavily on AI. The Lean community is now working to verify the GRH proof and extract its underlying strategy.