The Clay Mathematics Institute offers $1 million each for seven of the hardest problems in mathematics. Only one, the Poincaré conjecture, is listed as solved. This week brought two new claims, both from work with AI help: a Riemann Hypothesis proof that Lean’s maintainers certified and a Navier-Stokes blow-up claim that has not been checked yet.
This explainer lists the seven problems in plain words, explains Clay’s rules for paying a prize and sets out where each claim stands. For background, see our explainers on the Riemann Hypothesis and on Lean, the proof checker.
We track what AI systems can do in mathematics and other tests on our AI model benchmarks page.
Business Pill: The Cost of Checking
A short Business Engineer explainer on verification cost, using the Lean-certified Riemann proof as the week’s example.
The key insight: Clay offers $1 million for each of seven problems and lists only Poincare as solved; this week brought a Lean-certified Riemann proof and an unchecked Navier-Stokes claim, and Clay’s rules require publication, two years and general acceptance before it will consider either.
What Are the Millennium Prize Problems?
The Clay Mathematics Institute announced seven Prize Problems at a meeting in Paris on 24 May 2000. Its board designated “a $7 million prize fund for the solutions to these problems, with $1 million allocated to the solution of each problem.”
Clay says the problems were chosen to record “some of the most difficult problems with which mathematicians were grappling at the turn of the second millennium.” One of them, the Riemann Hypothesis, was formulated in 1859 and also appeared on David Hilbert’s 1900 list of problems, Clay notes.

What Are the Seven Problems, in Plain Words?
P vs NP. “If it is easy to check that a solution to a problem is correct, is it also easy to solve the problem?”, as Clay puts it. Clay lists it as unsolved.
The Riemann Hypothesis. It is about how prime numbers are spread out. Formulated in Riemann’s 1859 paper, it asserts that all the “non-obvious” zeros of the zeta function have real part 1/2, Clay says.
Navier-Stokes. These equations govern the flow of fluids such as water and air. Clay says that for many decades no proof answered the most basic questions: do solutions exist, and are they unique?
Yang-Mills and the mass gap. Experiments and computer simulations suggest a “mass gap” in the quantum versions of the Yang-Mills equations of particle physics, but no proof of this property is known, according to Clay.
The Hodge conjecture. It asks how much of the shape of the solutions to a system of algebraic equations can be described by further algebraic equations. Clay says it is known in some special cases but not in dimension four.
The Birch and Swinnerton-Dyer conjecture. It links the number of points on an elliptic curve, a curve defined by a cubic equation, to the structure of its rational points. Such curves arise in Wiles’ proof of Fermat’s Last Theorem and in cryptography, Clay notes.
The Poincaré conjecture. Posed by Henri Poincaré in 1904, it asks whether the three-dimensional sphere is the only simply connected three-dimensional space. Grigoriy Perelman solved it in preprints posted in 2002 and 2003, Clay says.
Has Anyone Ever Won?
Poincaré is the only problem Clay lists as solved. Clay announced on 18 March 2010 that Perelman met the criteria for the first prize, and on 1 July 2010 he turned down the $1 million, the Associated Press reported. He said Clay’s decision was unfair because his contribution was no greater than that of Richard Hamilton, whose Ricci flow theory his proof built on.
How Does Clay Decide?
Clay’s rules, revised on 26 September 2018, say it “does not accept direct submission of proposed solutions.” Before it will consider one, three conditions must be met: the solution “must be published in a Qualifying Outlet”, “at least two years must have passed since publication”, and it “must have received general acceptance in the global mathematics community.”
By our arithmetic, that means a proof published in late 2026 could not be considered before late 2028.
What Happened This Week?
Riemann. The Lean FRO and the Mathlib Initiative, which run Lean and its shared maths library, said on 21 November: “We hereby certify that both the statement and the proof of the Riemann Hypothesis are correct.” The 17-million-line proof is by Levent Alpöge, a mathematician at Anthropic, who said he found it working with Claude, as we reported. There is no human-readable paper yet.
Navier-Stokes. Alpöge and NYU mathematician Tristan Buckmaster claim, in a paper dated 20 November, that the equations can break down in finite time with no outside force, OfficeChai reported, as we covered. The paper comes with a Lean formalization of about 600,000 lines, but the numerical certificate the proof depends on has not been released, and the result has not been peer reviewed.
What Does a Lean Check Settle?
Lean is software that checks every logical step of a proof down to basic rules, as our Lean explainer sets out. For Riemann, the Lean statement says the problem as stated was Mathlib’s own version, “carefully vetted by experts in Lean and number theory”, and that several independent checkers accepted the proof.
What a computer check does not do is meet Clay’s rules. Those still require publication in a qualifying outlet, two years and general acceptance by mathematicians.
The Structural Read
As we read it, machine checking changes how fast a proof can be trusted, but not Clay’s timetable, which still runs on publication and two years.
As we read it, the two claims sit at different stages: Riemann has a machine certificate and no paper yet, while Navier-Stokes awaits the certificate its proof relies on.
Clay Mathematics Institute, rules
“The proposed solution must have received general acceptance in the global mathematics community.”
Three Implications
FOR MATHEMATICIANS Clay’s rules still require publication, two years and general acceptance.
FOR AI LABS Both new claims came from work with AI help, per their authors.
FOR READERS Clay lists one problem as solved, Poincare.
What Is Not Established
Neither claim can meet Clay’s two-year condition before late 2028, by our arithmetic. The Riemann proof has a machine certificate but no human-readable paper yet. The Navier-Stokes claim depends on a numerical certificate that has not been released. On Clay’s site, Riemann is still labelled unsolved and Navier-Stokes is labelled active.
The Bottom Line
Seven problems carry $1 million each, and Clay lists one as solved. This week added a Lean-certified Riemann proof and an unchecked Navier-Stokes claim, both from work with AI help, and Clay’s rules mean neither can be considered for a prize before at least two years after publication.
97,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
Frequently Asked Questions
What are the Millennium Prize Problems?
They are seven mathematics problems chosen by the Clay Mathematics Institute and announced on 24 May 2000, with $1 million offered for the solution of each: P vs NP, the Riemann Hypothesis, Navier-Stokes, Yang-Mills and the mass gap, the Hodge conjecture, the Birch and Swinnerton-Dyer conjecture and the Poincare conjecture.
Which Millennium Prize Problems have been solved?
Only the Poincare conjecture, according to Clay. Grigoriy Perelman solved it in preprints posted in 2002 and 2003, and in 2010 he turned down the $1 million prize, the Associated Press reported.
How does Clay award the prize?
Clay does not accept direct submissions. Under its rules, a solution must be published in a qualifying outlet, at least two years must pass after publication, and it must have received general acceptance in the global mathematics community.
Has AI solved a Millennium Prize Problem?
Not by Clay's rules. In November 2026 the Lean FRO and the Mathlib Initiative certified a Lean proof of the Riemann Hypothesis that Anthropic mathematician Levent Alpoge said he found working with Claude, and Alpoge and Tristan Buckmaster claimed a Navier-Stokes result. Under Clay's two-year rule, neither can be considered before late 2028, by our arithmetic.
A note on sourcing. The list of problems, the descriptions in quotation marks, the 24 May 2000 announcement, the $7 million fund and each problem’s status label are from the Clay Mathematics Institute’s own pages; the prize conditions are quoted from Clay’s rules, revised in 2018. Perelman’s refusal of the prize is from the Associated Press, 1 July 2010. The Riemann certification is from the Lean FRO and Mathlib Initiative statement of 21 November, as quoted in our coverage; the Navier-Stokes claim is OfficeChai’s reporting and the project’s GitHub page, as in our 21 November piece. The earliest-consideration date is our arithmetic on Clay’s two-year rule.
Sources: Clay Mathematics Institute, The Millennium Prize Problems, read 22 Nov 2026 · Clay Mathematics Institute, Rules for the Millennium Prize Problems (revised 26 Sep 2018), read 22 Nov 2026 · Clay Mathematics Institute, Poincare Conjecture page, read 22 Nov 2026 · Clay Mathematics Institute, Riemann Hypothesis page · Clay Mathematics Institute, Navier-Stokes Equation page









