Groupes de surface dans les réseaux des groupes de Lie semi-simples [d'après J. Kahn, V. Marković, U. Hamenstädt, F. Labourie et S. Mozes]

A cocompact lattice in a semisimple Lie group $G$ is a discrete subgroup $\mathrm{\Gamma }$ such that the quotient $G\mathrm{/}\mathrm{\Gamma }$ is compact. Does such a lattice always contain a surface group, i.e. a subgroup isomorphic to the fundamental group of a compact hyperbolic surface? If so, does it contain surface subgroups close (in a precise quantitative sense) to Fuchsian subgroups of $G$, i.e to discrete subgroups of $G$ contained in a copy of $\mathrm{(P)SL}(2,\mathbf{R})$ in $G$? The case $G=\mathrm{PSL}(2,\mathbf{C})$ corresponds to a famous conjecture of Thurston on 3-dimensional hyperbolic manifolds, and the quantitative version of the case $G=\mathrm{PSL}(2,\mathbf{R})\times \mathrm{PSL}(2,\mathbf{R})$ implies a conjecture of Ehrenpreis on pairs of compact hyperbolic surfaces; these two conjectures were proved by Kahn and Marković around ten years ago. Motivated by a question of Gromov, Hamenstädt solved the case that $G$ has real rank one, except for $G=\mathrm{SO}(2n,1)$. In a recent preprint (arXiv:1805.10189), Kahn, Labourie, and Mozes treat the case of a large class of semisimple Lie groups, including in particular all complex simple Lie groups; the surface groups they obtain are images of representations that are Anosov in the sense of Labourie. We present some of the ideas of their proof.