SummaryWe describe a model developed to help minimize the energy procurement costs of a New Zealand process industry that is a high user of electricity. The model accounts for stochastic prices that depend on the hydrological state of the electricity system, as well as transmission charges that are incurred during coincident electricity peaks. We describe how these are modelled and derive a stochastic dynamic programming algorithm that is used to arrange production to meet demand while minimizing the expected costs of electricity procurement.