Unforced Navier-Stokes Singularity And The Generalized Riemann Hypothesis
AIThis post was created with the assistance of artificial intelligence (AI).

TL;DR

Terence Tao reported on Nov. 20, 2026, three AI-assisted mathematical proof efforts: finite-time blowup for unforced 3D Euler and Navier–Stokes, and a Lean-only proof of the Generalized Riemann Hypothesis. The reports describe formalized or generated proofs, but independent checking and, for two results, clearer human-readable accounts are still needed.

Mathematician Terence Tao reported Nov. 20 that researchers have produced AI-assisted proof efforts for finite-time blowup in the three-dimensional unforced Euler and Navier–Stokes equations, alongside a Lean-only proof of the Generalized Riemann Hypothesis. Tao described the results as being at different stages of review: the Navier–Stokes argument still needs a professionally written account, while the enormous GRH formalization is only beginning to be examined and explained.

Age 18–24?Offer from Amazon

Prime made for students and young adults

  • Fast, free delivery for dorm and study essentials
  • Prime Video and Amazon Music included
  • Member-only deals
Try Prime for Young Adults Free trial for eligible 18–24 year olds
As an affiliate, we earn on qualifying purchases.

Tao said Levent Alpöge, Tristan Buckmaster and Matei Coiculescu, with substantial AI assistance, wrote a human-readable proof of finite-time blowup for the unforced incompressible Euler equations in three dimensions. The proof has also been formalized in Lean, a proof-assistant system used to encode mathematical arguments so that software can check their formal steps. Tao described this result as the most digested of the three efforts.

Alpöge and Buckmaster also produced an AI-assisted proof of finite-time blowup for the unforced incompressible Navier–Stokes equations in three dimensions, with a Lean formalization. Tao said the paper is technically human-readable and its broad strategy is clear: construct a numerical approximate discretely self-similar solution, then use rigorous perturbation theory to obtain an exact solution. He added that more work is needed to turn it into a proper, professionally written proof.

The third effort is an enormous Lean-only proof of the Generalized Riemann Hypothesis, generated by Alpöge using an internal AI model, according to Tao. Initial analysis suggests it uses modern developments in the Langlands program and arithmetic geometry, but Tao said the effort to verify the proof and extract its strategy will take time. He directed readers to a Lean Zulip discussion, whose moderators have since narrowed the thread to Lean-related aspects.

At a glance
reportWhen: Reported Nov. 20, 2026; verification an…
The developmentTerence Tao has reported three AI-assisted proof efforts addressing major open problems in fluid dynamics and number theory, while review and explanation remain in progress.

Three Major Problems Under Review

If the arguments withstand expert scrutiny, the fluid-mechanics results would address whether smooth solutions to the three-dimensional equations can develop singularities in finite time. The Navier–Stokes question is among the best-known open problems in mathematical analysis; a valid proof of blowup would settle a central question about the equations’ behavior, rather than merely report a numerical warning.

The Generalized Riemann Hypothesis is a major conjecture in number theory with implications for the distribution of prime numbers and related results. But the existence of a formal Lean artifact is not, by itself, a settled mathematical finding for the broader community: specialists must check the formalization, its assumptions and the underlying mathematical strategy. The story is also an early test of whether AI-assisted work can be made both machine-checkable and understandable to mathematicians.

Samsung 65-Inch 4K OLED Smart TV with Dolby Atmos, 120Hz Refresh Rate, and Amber Protection Bundle

Samsung 65-Inch 4K OLED Smart TV with Dolby Atmos, 120Hz Refresh Rate, and Amber Protection Bundle

As an affiliate, we earn on qualifying purchases.

From Formal Proofs to Mathematical Review

Lean allows users to express definitions and proof steps in a formal language, where a verifier can check whether the encoded argument follows from its stated premises. That can help detect gaps in a formalized proof, but it does not automatically show that the encoding captures the intended theorem or that the proof’s assumptions are appropriate. Human review and clear documentation remain important, especially for a large, technically complex formalization.

Tao’s Nov. 20 post explicitly ranked the efforts by how far they had been digested. The Euler result had both a human-readable proof and formalization; the Navier–Stokes work had a readable draft and a stated proof strategy but needed professional presentation; the GRH result was Lean-only and still under initial analysis. These are descriptions of the current state of the efforts, not independent declarations that all three results have been accepted by the mathematical community.

“I imagine the digestion process will take some time.”

— Terence Tao, on the Generalized Riemann Hypothesis formalization

Proof Status Still Needs Clarification

Tao’s report does not say that independent specialists have completed review or that the results have been accepted or published in a peer-reviewed journal. The GRH proof’s full argument and verification status remain especially unclear: Tao cited only initial analysis and an ongoing effort to extract the proof strategy. For the Navier–Stokes result, he said a more professionally written proof is still needed, but did not give a timeline or specify what further checks have been completed.

The source also does not provide full paper details, such as publication venues, dates of formal verification, or independent assessments of the Lean files. The claims should therefore be understood as reported proof efforts under examination, not as settled outcomes. The use of AI is described as substantial for the fluid results and tied to an internal model for the GRH work; further details about the tools and their role are not given.

Reviewers Must Check the Arguments

The next step is detailed mathematical and formal review. Tao pointed readers to the Lean Zulip thread for discussion of the GRH formalization, while noting that moderators had limited the discussion to Lean-related issues. For the Navier–Stokes result, a clearer, professionally prepared proof would help other researchers assess the construction and its perturbation argument. The report gives no timetable for completing either process.

Until reviewers can explain and validate the arguments, the central question is not simply whether a large proof file exists, but whether its formal statements correspond to the claimed mathematical results and whether the reasoning can be independently understood. Updates from the authors, reviewers or formalization discussions may clarify the status.

Key Questions

Has the Navier–Stokes blowup problem been definitively solved?

Not on the information in Tao’s report. He described an AI-assisted proof effort with a Lean formalization, but said a proper, professionally written proof is still needed. Independent review is not reported as complete.

What does the report say about the Euler equations?

Tao said Alpöge, Buckmaster and Coiculescu produced a human-readable, AI-assisted proof of finite-time blowup for the unforced incompressible Euler equations in three dimensions, and that it has been formalized in Lean.

Is the Generalized Riemann Hypothesis proof verified?

The report does not establish that verification is complete. Tao described an enormous Lean-only proof and said efforts to check it and extract its strategy are at an early stage.

What role did AI play?

Tao characterized the Euler and Navier–Stokes proofs as involving heavy AI assistance. He said the GRH proof was generated by Alpöge using an internal AI model. The report does not provide further technical details about the tools.

Where can readers follow the review?

Tao pointed to a Lean Zulip thread as a place to follow the GRH proof-verification effort, with moderators focusing discussion on Lean-related aspects. He also referenced Buckmaster’s Mastodon post about the fluid-equation results.

Source: terrytao.wordpress.com

You May Also Like

Japan’s Hayabusa2 Probe To Conduct Flyby Of Torifune Asteroid

Japan’s Hayabusa2 spacecraft is set to perform a flyby of the Torifune asteroid, marking a new phase in its mission to study near-Earth objects.

Common pesticide linked to more than double the risk of Parkinson’s disease

Research links a widely used pesticide to more than twice the risk of developing Parkinson’s, raising health concerns for agricultural workers and communities.

When Is The Meteor Shower Tonight

Find out the confirmed time for the meteor shower tonight, including peak viewing periods and how to observe it. Stay informed with the latest updates.

350 Years Ago, Ole Rømer Discovered That Light Has A Finite Speed

In 1676, Ole Rømer correctly predicted a delayed eclipse of Jupiter’s moon Io, showing that light takes time to travel.