Computing Cox rings via the cone conjecture

Abstract

We initiate a program to study the Cox ring of Calabi-Yau varieties, employing the notion of Morrison-Kawamata dream spaces. In this setting, we establish an analogue of the Hu-Keel GIT constructions for Mori dream spaces. More precisely, for a Morrison-Kawamata dream space $X$, we establish a correspondence between the small $\mathbb{Q}$-factorial modifications of $X$ and the GIT quotients of $\operatorname{Spec}\operatorname{Cox}(X)$. We further show that the Cox ring of a Morrison-Kawamata dream space is a filtered direct limit of subalgebras, each of which is an inverse limit of finitely generated $\mathrm{Cl}(X)$-graded $\mathbb{K}$-algebras. As an application, we give an explicit presentation of the Cox ring of a very general hypersurface of multidegree $(2,\dots,2,n+1)$ in $(\mathbb{P}^1)^m\times \mathbb{P}^n$. Furthermore, we prove that the Cox ring of such a hypersurface is of dense $F$-pure type.