Abstract Energy storage devices can facilitate more efficient energy management by regulating the peak of generated power. Managing the stored energy usually presents a complicated optimization problem. In this paper, we show an optimal “peak shaving” strategy, that enables minimization of the power peak and derive an analytic design method for attaining optimal peak shaving. The analysis reveals the lowest possible peak, given only the load's demand profile and the storage capacity. The effects of losses in the storage device are analyzed numerically, showing the increase of power peak associated with the increase of loss.