Traditional energy performance calculated using building simulation with the typical meteorological year (TMY) weather data represents the energy performance in a typical year but not necessarily the average or typical energy performance of a building in long term. Furthermore, the simulated results do not provide the range of variations due to the change of weather, which is important in building energy management and risk assessment of energy efficiency investment. This study analyzes the weather impact on peak electric demand and energy use by building simulation using 30-year actual meteorological year (AMY) weather data for three types of office buildings at two design efficiency levels across all 17 climate zones. The simulated results from the AMY are compared to those from TMY3 to determine and analyze the differences. It was found that yearly weather variation has significant impact on building performance especially peak electric demand. Energy savings of building technologies should be evaluated using simulations with multi-decade actual weather data to fully consider investment risk and the long term performance.