Utilities need to know the grid topology and loads to perform any meaningful grid optimization task. Rather than passively collecting data from sensors and meters, this talk puts forth an active data acquisition scheme termed grid probing. The idea is to leverage smart inverters to probe an electric grid and record its voltage response at probed buses, with the purpose of grid learning. Grid probing can be performed by commanding inverters to perturb their power injections. Multiple probing actions can be induced within a few tens of seconds. We will explain how probing can be used to infer: a) non-metered loads; and b) feeder topologies. For non-metered loads, we assume the topology to be known and that non-metered loads do not change significantly during probing. By coupling the power flow equations across time, we are able to identify non-metered loads under specific placement conditions, which can be easily checked beforehand. Our analysis and solvers can handle phasor and/or magnitude-only voltage data. We have also designed inverter injections to improve load recovery and ensure safe feeder operation. Grid probing can be also used for recovering grid topologies even without knowing the loads at non-probing buses: Probing all terminal buses on a radial feeder and measuring voltages at all buses, the topology and line impedances can be uniquely recovered. If voltages are recorded only at probing buses, then a reduced grid bearing several similarities to the original grid can be recovered. Speaker(s): Vasesilis, Location: Seattle, Washington