The projected number of Internet of Things (IoT) sensors makes battery maintenance a challenging task. Although battery-less IoT is technologically viable, the sensors should be somehow energized, either locally or remotely. Unmanned aerial vehicles (UAVs) can respond to this quest via wireless power transfer (WPT). However, to achieve energy neutrality across the IoT networks and thus mitigate the maintenance issues, the UAVs providing energy and connectivity to IoT sensors must be supplied by recharging stations having multi-source energy harvesting (EH) capability. Yet, as these sensors rely solely on UAV-transferred power, the absence of UAVs causes sensor outages and hence loss of coverage when they visit recharging stations for battery replenishment. Hence, besides the UAV parameters (e.g., battery size and velocity), recharging duration and station density must be carefully determined to avoid these outages. To address that, this paper uses stochastic geometry to derive the coverage probability of UAV-powered sensors. Our analysis sheds light on the fundamental trade-offs and design guidelines for energy-neutral IoT networks with recharging stations in regard to the regulatory organization limitations, practical rectenna and UAV models, and the minimum power requirements of sensors.