To explore the effects of uncertain fuel costs on the bulk energy flows in the US, we introduce stochastic fuel costs in a generalized network flow model of the integrated electric energy system, including coal, natural gas, and electricity generation. The fuel costs are modeled as discretely distributed random variables. A rolling two-stage recourse stochastic programming approach is employed to simulate the decision process involving uncertain costs with forecast updates. All the data are derived from publicly available information for the years 2002, when natural gas prices rose much higher than forecast, and 2006, when gas prices were lower than expected. Government forecasts of the natural gas prices are adapted to generate the scenarios considered in the stochastic formulation. Compared to the expected value solution from the deterministic model, the recourse solution found from the stochastic model for 2002 has higher total cost, lower natural gas consumption and less subregional power trade but a fuel mix that is closer to what actually occurred. The comparisons are qualitatively similar but muted in the 2006 case. The stochastic model assists decision makers to simulate the future flows within the national electric energy system and better understand how they are affected by uncertain fuel costs.