The work compares current approaches for evaluating and planning Brazilian energy mixes for future power generation, based not only on energy cost components contribution to a portfolio, but on their contribution to portfolio risk, as well. Energy planners have traditionally used least-cost as a basis for generating capacity additions, understating the true value of non traditional renewable technologies for decreasing risk. This project applies widely accepted finance theory, Mean-Variance Portfolio Theory, developed by Harry Markowitz, to provide an economic basis for selecting alternative generating scenarios. First, Brazil`s expected future generating mix for 2030 as predicted by Empresa de Pesquisa Energética in its 2030 Brazilian Power Planning, published in 2006, is evaluated. This mix is referred to as the reference EPE scenario. The risk-return properties of Brazil`s expected EPE mix for the year 2030 is compared to other possible mixes on the projected efficient frontier. The model finds solutions that are superior to the EPE mix in that they reduce risk or cost or both, while including a greater share of wind and biomass from sugar cane in the mix. The basic findings of this analysis seem quite robust, and do not materially change the shape of the efficient frontier, where it is technically feasible, even when the risk parameter estimates and cost covariations are changed significantly in the sensitivity analysis.