Perovskite/silicon tandem solar cells promise power conversion efficiencies beyond the Shockley-Queisser limit of single-junction devices; however, their actual outdoor performance is yet to be investigated. Here we fabricate 25% efficient two-terminal monolithic perovskite/silicon tandem solar cells and test them outdoors in a hot and sunny climate. We find that the temperature dependence of both the silicon and perovskite bandgaps-which follow opposing trends-shifts the devices away from current matching for two-terminal tandems that are optimized at standard test conditions. Consequently, we argue that the optimal perovskite bandgap energy at standard test conditions is <1.68 eV for field performance at operational temperatures greater than 55 degrees C, which is lower compared with earlier findings. This implies that bromide-lean perovskites with narrower bandgaps at standard test conditions-and therefore better phase stability-hold great promise for the commercialization of perovskite/silicon tandem solar cells. Outdoor field testing is crucial to understand how solar cells behave under operational conditions. Here, Aydin et al. show that a lower perovskite bandgap than that calculated at laboratory standard test conditions enhances the performance of perovskite/silicon tandem cells in the field.