The local p-adic section conjecture and global consequences

Points on p-adic curves of genus two or more are claimed to match exactly the sections of their fundamental group.

2 papers September 24, 2026 Explained by AI

Grothendieck predicted in 1983 that, for curves of genus two or more, points can be read off from symmetry groups alone. Each point gives a splitting of a certain group, called a section. The conjecture says every section comes from exactly one point. The second paper claims to prove the local version, over the p-adic numbers. It also claims the global version for some curves over number fields, including the modular curves X₀(N) and X₁(N).

First step

1983

A. Grothendieck, in his letter to G. Faltings

Significance

Major

The local p-adic section conjecture was a central open problem in anabelian geometry, if this proof holds.

Trust

Claimed, not checked

Field

Number theory

2 papers

The explanation

Take the curve y² = x⁵ + 1. A few points with whole-number coordinates sit on it. For example, (0, 1) works because 1² = 0 + 1. So do (0, −1) and (−1, 0). The right-hand side has degree 5, which gives this curve genus 2. Genus counts the holes in the curve’s complex picture, like the two holes of a double doughnut.

Now change the numbers you allow. The p-adic numbers, for a prime p, measure size by divisibility by p instead of by distance. A curve still has points with p-adic coordinates. The curve also has a fundamental group, written π₁(X) (a group recording how loops and coverings of the curve fit together). It sits over the Galois group of the base field (the symmetries of its numbers). A section is a way to lift each base symmetry into π₁(X), compatibly.

Every point gives a sectionSource 1. Grothendieck’s section conjecture, from his 1983 letter to Faltings, predicts the converse for curves of genus at least twoSource 2. Every section should come from a point, and different points should give different sections. Then the points could be recovered from groups alone. The local p-adic form asks this over a finite extension of the p-adic numbersSource 1.

Earlier work came close. Koenigsmann proved a birational version, using a larger Galois group of the curve’s function fieldSource 3. Pop and Stix showed that each section sits at a valuation extending the p-adic oneSource 4. A valuation here is a way to measure size on the curve’s functions. The open step was passing from such a valuation to an actual pointSource 5.

The second paper claims the local conjecture in fullSource 6. For every prime p, every finite extension of the p-adic numbers and every such curve, points and sections match one to oneSource 7. Back on y² = x⁵ + 1 over the 5-adic numbers, each section would mark exactly one 5-adic point.

The key step is a statement about completions of fields of functions. It claims a section at a valuation lifts to the larger Galois groupSource 8. Then Koenigsmann’s theorem and a classical argument about two valuations finish the case. The geometry is a gluing problem. Near a branch point, a cover can be replaced by a trivial cover and glued back. But that needs roots on every sheet above a ring around the point, not just oneSource 9. The paper builds those roots in two stages.

The paper also draws a global consequence for curves over number fieldsSource 10. There, the section conjecture follows when a descent condition holds: the descent locus must equal the rational pointsSource 10. Stoll’s criteria give that condition for the modular curves X₀(N) and X₁(N) of genus at least two over the rationalsSource 11 Source 12. For other curves the condition stays a real extra hypothesisSource 13. The first paper builds a related cover with prescribed behavior on small p-adic disksSource 14.

How sure is this? No Lean formalization is listed for these papers in the openai/math release. So no computer proof checker has checked them. The argument draws on many deep published results in p-adic geometry. No specialist review has been published.

What it doesn't show. This does not prove Grothendieck's global section conjecture for every curve over a number field; that still needs an extra descent condition.

Where it came from

This result comes from an unreleased OpenAI model, published in the openai/math release on 6 October 2026.

Hype check

Major only if the proof holds. No Lean formalization is listed and no specialist review is published. The global result needs an extra descent condition, met for the modular curves named.

Words used

Genus: The number of holes in a curve’s complex picture; a double doughnut has genus 2.
p-adic numbers: A number system for a prime p where a number is small when it is divisible by a high power of p.
Fundamental group: A group that records how loops and finite coverings of a curve fit together.
Galois group: The group of symmetries of a field of numbers that respect addition and multiplication.
Section: A way to lift every symmetry of the base field into the curve’s fundamental group, compatibly.
Valuation: A rule that measures how large or small the functions on a curve are.
Modular curve: A curve, such as X₀(N), whose points describe elliptic curves with extra level structure.

The original papers (2)

Étale covers with a prescribed exterior sheet (PDF, opens in a new tab)
September 24, 2026
The p-adic section conjecture (PDF, opens in a new tab)
October 6, 2026

All quoted lines (14)

The local $p$ -adic section conjecture asserts that this construction accounts for all sections, and distinguishes all rational points.
The p-adic section conjecture Back to text
It is a local analogue of the section conjecture in Grothendieck's letter to Faltings of 27 June 1983, formulated over fields finitely generated over the prime field.
The p-adic section conjecture Back to text
Koenigsmann proves that every birational section over a finite extension of $\mathbb Q_p$ is contained in the decomposition group of a rational point.
The p-adic section conjecture Back to text
Pop–Stix localize every \'etale section at a valuation extending the $p$ -adic valuation.
The p-adic section conjecture Back to text
The passage from a fixed valuation to a rational point remains the geometric issue addressed here.
The p-adic section conjecture Back to text
We prove the local $p$ -adic section conjecture for smooth proper geometrically connected curves of genus at least two, over every finite extension of $\mathbb Q_p$ .
The p-adic section conjecture Back to text
For every prime $p$ , every finite extension $k/\mathbb Q_p$ , and every smooth proper geometrically connected curve $X/k$ of genus at least two, the map $X(k)\longrightarrow \left\{\begin{gathered} s:G_k\to\pi_1(X)\text{ a continuous group homomorphism}\\ \text{such that }(\pi_1(X)\to G_k)\circ s=\mathrm{id}_{G_k} \end{gathered}\right\}\big/\pi_1(X_{\bar k})$ is bijective.
The p-adic section conjecture Back to text
A section localized at a rank-one valuation therefore lifts to a birational section.
The p-adic section conjecture Back to text
The roots must exist on the entire inverse image of the annulus: a root on only one sheet does not trivialize the full cover needed for gluing.
The p-adic section conjecture Back to text
Combined with established finite-descent theorems, this also yields the global section conjecture for smooth proper geometrically connected curves of genus at least two over number fields when their finite-cover descent locus equals their rational points; this includes $X_0(N)$ and $X_1(N)$ of genus at least two over $\mathbb Q$ .
The p-adic section conjecture Back to text
Stoll's finite-descent criteria supply the required equality when the curve has a nonconstant morphism over its number field to an abelian variety with finitely many rational points and trivial divisible Tate–Shafarevich subgroup, or when its Jacobian has these properties.
The p-adic section conjecture Back to text
Combined with established finite-descent theorems, this also yields the global section conjecture for smooth proper geometrically connected curves of genus at least two over number fields when their finite-cover descent locus equals their rational points; this includes $X_0(N)$ and $X_1(N)$ of genus at least two over $\mathbb Q$ .
The p-adic section conjecture Back to text
The general implication retains the finite-cover equality as a genuine arithmetic hypothesis; local point-theoreticity alone does not supply it.
The p-adic section conjecture Back to text
Given finitely many disjoint small open disks, we construct one connected finite \'etale cover with a sheet isomorphic to their entire exterior and with degree divisible by $p$ on every connected component above each disk.
Étale covers with a prescribed exterior sheet Back to text

Other sources

031Uchida’s conjecture for open homomorphisms of Galois groupsNumber theory
001Milne’s rationality conjecture and algebraic specializationNumber theory
018The Margulis–Platonov conjecture over global fieldsNumber theory
022The weak inhomogeneous Duffin–Schaeffer conjectureNumber theory
013Ostmann’s inverse Goldbach conjectureNumber theory

Browse Number theory Named conjectures Open since the 1980s Claimed, not yet checked