Theorem 18.6.5. Given \(C\) and \(D\) two categories with finite colimits, the assignment \[ \varphi\mapsto ((x,y)\mapsto\varphi(x\otimes y)) \] defines, for any category \(E\) with finite colimits, an equivalence \[ \RexCat(C\otimes D,E)\simeq\BRex(C,D;E) \]

Proof

The localization functor \(\gamma\colon\rex{C\times D}\to C\otimes D\) is right exact by Corollary 17.1.9. By Corollary 17.3.7, there is therefore an equivalence: \[ \Rex(C\otimes D,E)\simeq \RexCat^W(\rex{C\times D},E)\, . \] On the other hand, we have \[ \RexCat(\rex{C\times D},E)=\Fun(C\times D,E) \] by Theorem 18.5.2. This produces a canonical full embedding \[ \Rex(C\otimes D,E)\hookrightarrow\Fun(C\times D,E) \] whose image at the level of objects obviously is \(\BRex(C,D;E)\), by definition of \(W\).

From Section 18.6, Tensor product of finitely cocomplete categories.