This research investigates the time-varying causality between energy consumption (renewable and non-renewable) by the U.S. energy sectors and real GDP covering the period 1973:Q1-2018:Q4. Unlike previous literature, this study does not only use a conventional causality technique, but also incorporates a time-varying causality approach to see whether the relationship between renewable and non-renewable energy consumption by sector and real GDP changes over the study period. The results obtained from the Toda-Yamamoto approach reveal a causality relationship in the case of industrial and transportation sectors. The results from the time-varying causality technique, however, indicate that the causality relationship is detected for all sectors over different time periods. This study therefore addresses the shortcomings of the conventional causality test, which emphasizes the importance of time-varying in determining the underlining relationship. Given the energy-GDP causality, it is worth noting that the validity of a causality hypothesis proposed by a conventional method does not necessarily dominate the whole sample period. Policy implications are discussed.