With the increasing sophistication and use of computer codes in the nuclear industry, there is a growing awareness of the need to identify and quantify the uncertainties of these codes. Work is now being performed at Battelle-Pacific Northwest Laboratories to study the uncertainties in steady-state stored energy calculations by using linear propagation of uncertainties. This method predicts the uncertainty of variables by propagating input variances through models. Comparison of Monte Carlo analysis to linear propagation shows good agreement and verifies the adequacy of linear propagation. Linear power, radial gap width, fuel thermal conductivity, flux depression, and fuel heat capacity are all shown to be parameters of major importance when calculating both stored energy and its uncertainty. The uncertainty for stored energy at beginning-of-life is ∼17% (99% confidence level) and rises to a maximum of 37% during a simulated two-cycle power history.