In many industries, investment is part of the most important planning decisions. In past decades, mathematical programming models have been widely used in capacity planning and facility location to support investment decisions. Such initial techniques evolved to the use of enterprise portfolio management, very common in the energy industry. Increasing concern on risk made of risk management one of the top priorities on the planning processes of companies nowadays. As a consequence, state of the art planning models are stochastic, and usually consider some kind of risk measure as well as financial instruments to hedge the investment portfolio. Several developments on Portfolio Theory were achieved on the last years, most of them related to the study of new risk measures. The most representative measure is Conditional Value at Risk (CVaR), which not only belongs to the class of Coherent Risk Measures, but is also a Spectral Risk Measure. We propose a two stage stochastic investment model which incorporates CVaR as a constraint on risk. The model is applied to the case of an integrated company that invests on Oil, Gas and Energy. In order to reduce the need of computational resources, we make use of some decomposition techniques. The results for the case study are presented.