What are some mathematical consequences of the study of 6D $\mathcal N = (2,0)$ SCFT?

If you take the (2,0) theory and put it on a manifold which is $T^2 \times M_4$, it is known to reduce to $\mathcal{N} = 4$ super-Yang Mills theory on $M_4$. That theory exhibits S-duality, which has been shown to be related to geometric Langlands. In particular, S-duality is part of an $SL(2,\mathbb{Z})$ symmetry, and the $G \leftrightarrow \widehat{G}$ duality in both geometric Langlands and $\mathcal{N}=4$ SYM is related to the $\left({0\ -1 \atop 1\ \ 0}\right)$ element in the $SL(2,\mathbb{Z})$ symmetry of the torus above. The name S-duality arises because this operates as the $\tau \leftrightarrow -\frac{1}{\tau}$ Mobius transformation on the coupling and exchanges strong and weak coupling. The somewhat mysterious S-duality in four dimensions then becomes something geometric in six dimensions. Presumably, then, a mathematical understanding of the six dimensional theory would give new insight into the mathematics of geometric Langlands (and maybe, if one is lucky, the Langlands program more broadly).

The rub, however, is that the interesting case is when the gauge group in four dimensions is non-Abelian, and even on the physics side, there is no good understanding of the six dimensional theory that gives rise to this upon compactification. String theoretically, it should arise as the theory on a stack of M5-branes. It must be something novel, because it has to reduce to a theory with two different (Langlands dual) gauge groups simply by changing one's perspective on the torus.

In general, mathematical outputs of SUSY field theories often become more accessible after performing some twist, and the same is true of the 6d (2,0) SCFT. Considering the theory on $\Sigma\times M_4$, it admits a twist (first studied by Beem-Rastelli I believe) that's holomorphic along $\Sigma$ and topological along $M_4$. Some discussion of the difficulties in mathematically describing this twist in perturbation theory can be found in these notes from a talk of Kevin Costello: https://math.berkeley.edu/~qchu/Notes/6d/Day%205,%20Talk%202,%20Costello.pdf

One interesting application of this twist to representation theory is the AGT correspondence. Specializing $M_4=\mathbb{R}^4$, the rough idea is that after putting suitable $\Omega$-backgrounds in the $\mathbb{R}^4$ direction, the degrees of freedom of the 6d theory localize to $\Sigma$ where we find a Toda field theory. We then get several relationships between this Toda field theory living on $\Sigma$ and the 4d $\mathcal{N}=2$ gauge theory gotten by compactifying on $\Sigma$. In particular, the conformal blocks of the Toda field theory should match the partition function of the 4d $\mathcal{N}=2$ theory, and vertex operator insertions on $\Sigma$ should correspond to including certain defects in the 4d $\mathcal{N}=2$ theory. In the math literature, this connection (at least for pure gauge theory) was established through work of Maulik-Okounkov and Schiffman-Vasserot who showed that the affine W-algebra (i.e. the local observables of the above Toda field theory) acts on the equivariant cohomology of the moduli of instantons on $\mathbb{A}^2$.

Let me also very briefly comment that there's a 6d string theory (dubbed the little string theory) that becomes the (2,0) SCFT in the limit of infinite string mass. This string theory appears to be just as mathematically rich as its SCFT limit. These slides: https://member.ipmu.jp/yuji.tachikawa/stringsmirrors/2016/main/Aganagic%20v2.pdf from Mina Aganagic's 2016 String-Math talk contain some statements in this direction.