In this study, Markowitz mean-variance portfolio theory is applied to electricity-generating technologies of the United States and Switzerland. Both an investor (focused on changes in return) and a current user (focused on return in levels) view are adopted to determine efficient frontiers of electricity generation technologies in terms of expected return and risk as of 2003. Since shocks in generation costs per kWh (the inverse of returns) are correlated, Seemingly Unrelated Regression Estimation (SURE) is used to filter out the systematic components of the covariance matrix. Results suggest that risk-averse investors and risk-neutral current users in the United States are considerably closer to their efficiency frontier than their Swiss counterparts. This may be due to earlier and more thorough deregulation of electricity markets in the United States.