We present deterministic and stochastic models for determining the optimal mix of different technologies for electricity generation, ranging from carbon, nuclear and combined cycle gas turbine to hydroelectric, wind and photovoltaic, taking into account the actual sites and the cost of investment in new sites, the cost of of mantainance, the use of emission quotas and the relative constraints as well as the green certificates one may use. The stochasticity is related to the future price of energy and to the future price of emissions, in this paper we limit our study to the variaility of fuels. The stochasticity appears in the expected costs and the probability that the total cost do not overcome a specific threshold is taken into account by considering CVaR risk measure. A comparison between the deterministic solution and the stochastic solution shows the role of using the risk the importance to use risk measure in the stochastic long run approach.