Equipment manufacturers are increasingly selling complementary services such as remanufacturing services to their equipment customers. This servitization trend mandates that the remanufacturing supply chain network is optimized accordingly. In order to set up such network, investment decisions have to be made, not only regarding the number, locations and types of remanufacturing facilities, but also with respect to the appropriate capacity and inventory levels in order to guarantee a specific service level. These network decisions are influenced by the way remanufacturing services are offered. We consider here two service delivery strategies: either a quick exchange of the used part by an available refurbished one or re-installing the original part when all corresponding remanufacturing processes are finished. Given the high level of uncertainty in this context, we build a stochastic, profit maximizing model to simultaneously determine the optimal layout and the optimal service delivery strategy for a multi-level logistics network with single indenture repairable service parts. Model results for this case study are obtained by the differential evolution algorithm.