
Langlands Program
The Langlands program is a far-reaching web of conjectures and proven theorems in mathematics that connects seemingly unrelated fields such as number theory and harmonic analysis. It is considered one of the most ambitious research programs in modern mathematics and today also influences theoretical physics.
The Langlands program is not an algorithm and not software — it is a mathematical research program. It consists of a collection of conjectures and proven statements that all claim the same thing: two areas of mathematics that developed separately for centuries are, at their core, the same. One area is number theory, the mathematics of whole numbers and their properties. The other is harmonic analysis, which studies how complicated functions can be decomposed into simpler oscillations — similar to how a piece of music can be broken down into its individual tones. The Canadian mathematician Robert Langlands proposed this connection in 1967 in a letter to a colleague. At the time, it was speculation. Today it is one of the most active research fields in mathematics worldwide.
Why a bridge between numbers and oscillations matters
In mathematics, the rule holds: whoever connects two fields doubles their toolkit. An unsolvable problem on one side can suddenly become solvable once translated to the other side. This has happened repeatedly. The most famous case is Andrew Wiles’s proof of Fermat’s Last Theorem in 1995. This theorem states that the equation xⁿ + yⁿ = zⁿ has no solution for integers and n > 2. Wiles did not prove it directly — he used a special case of the Langlands correspondence as a decisive tool.
This shows how practical the program can be, even though it sounds abstract. It is not an end in itself. Every newly proven connection between the two worlds opens new paths to solving concrete problems. Many problems in number theory that have been open for centuries are today suspected of being attackable via the Langlands program.
The correspondence: how two worlds are mapped onto each other
The central idea is a correspondence — a kind of dictionary. On one side stand so-called Galois representations. They describe what the symmetries of number equations look like. On the other side stand automorphic forms, which are special functions with a great many inner symmetries. The claim of the Langlands program is: for every Galois representation there exists exactly one matching automorphic form — and vice versa. Both objects carry the same information, just in a different language.
A comparison helps: imagine you have a recipe in German and a cooking show in Japanese. Both describe the same dish, but no one has yet noticed that they are identical. The Langlands program claims that such hidden correspondences exist systematically — and provides methods to uncover them. The difficult part is that the “translation” between the two sides is not a simple formula. It must be developed and proven separately for each new class of objects. Hundreds of mathematicians worldwide are working on this.
It is also important to note what the program is not: it is not a single statement that one proves and then checks off. It is an open research program with many sub-conjectures at various levels of difficulty. Some have been proven, many have not. The complete Langlands correspondence is considered one of the great unsolved problems in mathematics.
Langlands in current headlines and in physics
In 2024, Peter Scholze, a German mathematician, received international attention for new methods in so-called p-adic geometry — a toolkit that stems directly from the Langlands program. When reports speak of breakthroughs in number theory or new proofs concerning algebraic curves, the Langlands program is almost always lurking in the background.
Surprisingly, the program also appears in theoretical physics. Physicists working on string theory or quantum field theory encounter structures that bear a striking resemblance to the Langlands correspondences. The so-called “geometric Langlands program” is today a research field that mathematicians and physicists work on jointly. A team led by Dennis Gaitsgory and Sam Raskin published a complete proof of the geometric Langlands program in 2024 after decades of work — over 800 pages long. This was one of the biggest mathematical news stories of the year.
For AI and tech, the Langlands program has no direct application so far. It is pure basic research. But historically, basic research in mathematics has repeatedly enabled technologies that became indispensable decades later — from number theory, which today underpins internet encryption, to linear algebra, which is what makes machine learning possible in the first place.