•Ground and truck-launched air delivery robot operations are analyzed.•For low densities, truck-launched air drones present the lowest operation costs.•For high densities, ground delivery robots present the lowest operation costs.•In any case, ground delivery robots generate less externalities. The e-commerce boom has increased the complexity of last-mile logistics operations in urban environments. In this context, unmanned aerial vehicles (UAVs), also known as delivery drones, and ground autonomous delivery devices (GADDs) show great potentialities. The objective of this paper is to provide strategic insights to adequately match these autonomous technologies with some given characteristics of cities and help define relevant decision variables. Using continuous approximation equations, the operations costs as well as the externalities induced by a) GADDs in association with an urban consolidation center (UCC) and b) truck-launched UAVs are estimated. Then, the developed mathematical formulations are applied in two different use cases: a part of the Paris suburbs (France) and the historical center of Barcelona (Spain). In less dense and larger service regions such as the Paris suburbs, truck-launched delivery drones seem more suitable to reduce the carriers’ operations costs. In denser neighborhoods such as the Barcelona historical center, GADDs are expected to be more economically profitable. In both use cases, GADDs would generate less externalities. Finally, considering the high uncertainty of some input parameters, a sensitivity analysis of the models is done.