Geometric Langlands Conjecture: Proof Explained
Quick answer: The geometric Langlands conjecture relates categories of sheaves built from bundles on a curve to categories built from local systems for a dual symmetry group. In 2024, Dennis Gaitsgory, Sam Raskin and seven collaborators announced a proof of its global, unramified form in characteristic zero, which includes complex curves. The proof spans five papers totaling roughly 1,000 pages. The wider arithmetic Langlands program remains open.
The achievement is a translation between entire mathematical structures: objects on one side correspond to objects on the other, together with the maps between them. A useful starting analogy is Fourier analysis, which describes a signal through its frequency components. Below we explain the two sides of the correspondence, exactly which version was proved, and how the five papers fit together.
Which Version Was Proved?
- Global: it concerns the whole curve at once.
- Unramified: no extra singular behavior is prescribed at marked points.
- Categorical: it matches objects together with the maps and structure connecting them, not just lists of objects.
- Characteristic zero: it covers curves over the complex numbers, among others.
What Is the Langlands Program, in One Paragraph?
In 1967, Robert Langlands wrote to André Weil proposing deep links between number theory and highly symmetric functions called automorphic forms. The proof of Fermat’s Last Theorem used one case of those links. For a gentle start, see the Langlands program explained with a worked example.
Three Settings for Langlands Ideas
| Version | Setting | Status |
|---|---|---|
| Arithmetic (original) | Whole numbers and number fields | Mostly open; many important cases proved |
| Function field | Curves over finite fields, a halfway house | The correspondence for \(GL_n\) is proved (Drinfeld for rank 2, Lafforgue in general) |
| Geometric | Curves over the complex numbers (Riemann surfaces) | Global, unramified categorical correspondence in characteristic zero proved in 2024; further variants are active research |
The geometric version replaces numbers with shapes. It grew in the 1980s and 1990s from work of Vladimir Drinfeld, Gérard Laumon, Alexander Beilinson and others. Building on that work, Dima Arinkin and Dennis Gaitsgory formulated the modern categorical statement in a paper first posted in 2012.
What Does the Geometric Langlands Conjecture Say?
Start with a smooth complex algebraic curve. It has one complex dimension, which is two real dimensions, so it can be pictured as a surface such as a sphere or a doughnut. Choose a symmetry group \(G\), such as the invertible \(n \times n\) complex matrices. Langlands attaches a “dual” group \(G^\vee\) to \(G\).
- Two spaces: one organizes \(G\)-bundles over the curve. The other organizes local systems for \(G^\vee\), which describe how data can be transported consistently around loops in the curve.
- Two categories: the theorem compares carefully defined categories of sheaves on these spaces. In the de Rham version, the bundle side uses D-modules (objects related to systems of differential equations) and the other side uses certain sheaves on the space of local systems.
- The claim: the two categories are equivalent, preserving both the objects and the maps between them.
A helpful analogy is Fourier analysis, which rewrites a sound wave through its frequency components, including their amplitudes and phases, without losing information. In geometric Langlands, the role of pure frequencies is played by Hecke eigensheaves: objects that respond in a controlled, predictable way to certain geometric operations, as a pure tone responds predictably to a filter. You can explore ordinary Fourier decomposition with our Fourier series calculator.
Who Proved It, and How?
The proof was announced in 2024 and presented across five papers totaling roughly 1,000 pages, which have since been revised. The project involved nine mathematicians, with different subsets authoring individual papers: Dennis Gaitsgory, Sam Raskin, Dima Arinkin, Dario Beraldo, Justin Campbell, Lin Chen, Joakim Faergeman, Kevin Lin and Nick Rozenblyum. The first paper constructs the Langlands functor, the translation from the automorphic side to the spectral side. Other papers establish the structures and compatibility results needed. The final paper completes a crucial multiplicity-one step about the Hecke eigensheaves attached to irreducible local systems.
| Paper | Topic |
|---|---|
| I | Construction of the Langlands functor |
| II | Kac–Moody localization and the fundamental local equivalence |
| III | Compatibility with parabolic induction |
| IV | Ambidexterity |
| V | The multiplicity one theorem |
The papers are research-level mathematics; links to them are collected on the authors’ proof page listed in the sources.
Why It Matters
- It settles a central conjecture that guided a whole field for 30 years.
- It strengthens the bridge to number theory. Ideas often travel from the geometric version to the arithmetic one; earlier, Fargues and Scholze had recast parts of the local arithmetic program in geometric terms.
- It connects to physics. Kapustin and Witten showed in 2006 that geometric Langlands appears as a duality in quantum field theory.
What it does not do: it does not give a general method for solving equations in whole numbers, and it does not complete the arithmetic or ramified Langlands problems.
What Happened After 2024?
In 2025, Gaitsgory and Raskin extended part of the theory to positive characteristic. Gaitsgory received the 2025 Breakthrough Prize in Mathematics for a body of work that includes the proof. On March 28, 2026, Peter Scholze surveyed geometric Langlands and its consequences at the Bourbaki seminar in Paris.
Common Misconceptions
“The Langlands program is solved.” No. The geometric version over the complex numbers is proved. The original arithmetic version, about whole numbers, is still largely open.
“Nobody can check a 1,000-page proof.” Experts check long proofs in parts, through seminars and refereeing. Scholze’s 2026 Bourbaki lecture is one public example of that process.
Watch: Geometric Langlands in Video
If you learn better by watching, these videos are worth your time.
Check Your Understanding
Five quick questions to see how much of this has stuck. Pick an answer and press Submit; you will see a short explanation either way.
Answer each question and press Submit to see your result right away.
Frequently Asked Questions
What is the geometric Langlands conjecture?
It predicts that two categories attached to a curve are equivalent: one built from G-bundles on the curve (the automorphic side) and one built from local systems for the dual group (the spectral side). It is the geometric counterpart of the Langlands program in number theory.
Who proved the geometric Langlands conjecture?
A team of nine mathematicians led by Dennis Gaitsgory and Sam Raskin, with Dima Arinkin, Dario Beraldo, Justin Campbell, Lin Chen, Joakim Faergeman, Kevin Lin and Nick Rozenblyum, in five papers completed in 2024.
Which version of geometric Langlands was proved?
The global, unramified, categorical geometric Langlands conjecture in characteristic zero, which includes curves over the complex numbers.
Where can I read the geometric Langlands proof?
The five papers are on arXiv and linked from the authors’ proof page. They total roughly 1,000 pages and are written for research mathematicians.
Does this prove the Langlands program?
No. It proves the geometric version over the complex numbers. The original arithmetic Langlands conjectures about number fields remain largely open, although many important cases are proved.
Keep Exploring
- Langlands Program Explained: A Simple Worked Example
- Fields Medal 2026 Winners: Their Mathematics Explained
- 7 Millennium Prize Problems Explained: 2026 Status
Sources and Further Reading
- Quanta Magazine: Monumental Proof Settles Geometric Langlands Conjecture (July 19, 2024)
- Gaitsgory, Raskin et al., Proof of the geometric Langlands conjecture I (arXiv 2405.03599)
- Dennis Gaitsgory: Geometric Langlands proof page (Max Planck Institute)
- Arinkin and Gaitsgory, Singular support of coherent sheaves and the geometric Langlands conjecture (arXiv 1201.6343)
- Peter Scholze, Geometric Langlands, Séminaire Bourbaki exposé 1252 (March 2026)
- Wikipedia: Geometric Langlands correspondence
Related to This Article
More math articles
- Patterns of Intergroup Relations
- 7th Grade OAA Math Worksheets: FREE & Printable
- DNA Strand Direction: Complement and Reverse Complement
- Shop Math: Fractions, Scale, and Dimensions
- Grade 5 Math Flashcards (Free Online: Formulas, Terms & Concepts)
- STAAR Algebra 1 Passing Score: Raw Points You Need in 2027
- Discovering the Magic of ASA and AAS Congruence in Triangles
- How to Write Two-variable Inequalities Word Problems?
- Work Calculator
- How to Estimate Limits from Tables


What people say about "Geometric Langlands Conjecture: Proof Explained - Effortless Math"?
No one replied yet.