This paper develops microeconomic models of electricity storage. The full economic potential of price arbitrage with nodal storage hinges not only on battery characteristics (energy-to-power ratio) but crucially also on the quality of price forecasting. Economic theory can also quantify the intrinsic profitability limitations as increased storage deployment smoothes price variations. The profitability of supply-side storage for intermittent power producers depends on the exposure to curtailment risk and thus the empirical distribution of negative price episodes. Demand-side storage can be economical when improved utilization of existing power lines avoids expanding them, and this trade-off potential is captured in a theoretical model. Economic magnitudes are explored with zonal prices from Ontario, which has some of the most volatile electricity prices in North America. Wider use of locational marginal prices in wholesale electricity markets will incentivize efficient deployment of electricity storage and also induce intermittent power producers to internalize curtailment risk appropriately.