Igusa Stacks and the Langlands program

Objective

The Langlands program, often called a ``grand unified theory of mathematics'', predicts reciprocity laws that relate very different kinds of mathematical objects: automorphic forms and Galois representations. Shimura varieties play a fundamental role in constructing and propagating instances of Langlands reciprocity. In the past decade, advances in p-adic geometry have revolutionised the study of Shimura varieties. This recently led to the introduction of certain p-adic analytic varieties called Igusa stacks, which connect Shimura varieties to ideas from geometric Langlands via the work of Fargues--Scholze. This gives access to powerful new tools and makes Igusa stacks seem as fundamental as the Shimura varieties themselves. This proposal will realise the potential of p-adic Igusa stacks and of the closely related Igusa varieties to give a systematic understanding of the cohomology of Shimura varieties with l-torsion coefficients and of congruences modulo l between automorphic forms on Shimura varieties. There are two components to this proposal, depending on whether or not l = p. When l is not p, I will construct the relative intersection cohomology of the Igusa stack and describe it in terms of purely local (generalised) eigensheaves, expanding on and proving a conjecture of Fargues. This will lead to a description of torsion in the intersection cohomology of Shimura varieties, reminiscent of Arthur's conjectures in characteristic 0, and to an axiomatic approach to congruences, such as level-raising, when the tame level varies. When l is p, I will prove an integral Eichler--Shimura comparison theorem for ordinary p-adic automorphic forms. Initially, my approach will exploit the ordinary part of the Igusa stack, then move into deeper Newton strata for generalisations. This theorem will unify geometric and representation-theoretic perspectives on ordinary p-adic automorphic forms and will have major applications to Euler systems and Iwasawa theory.