Climate changes induce warmer climate with stronger and more frequent extreme events. Due to the uncertain nature of climate, accurate simulation of future conditions is impossible and a major challenge is the selection of climate data in the impact assessment. This work compares application of three climate data sets in an energy simulation of the EPFL campus: i) Regional Climate Models (RCM data), ii) statically representative RCM data, and iii) morphed data. The energy behavior of the campus is analyzed, including its future thermal behavior, as well as its dynamic hourly variation due to the climatic data. The objective of this paper is to understand and quantify the energy transition, from 2010 to 2100, by focusing on the thermal behavior of buildings, as well as their energy demand for heating and cooling. Results explain the difference between three cases, underling the important impact related to a sound selection of the weather data.