Using the nonlinear autoregressive bounds approach, the proposed study highlights that not only upside but also downside (i.e., positive and negative) variations in GDP, globalization index, and capital formation significantly affect energy utilization in India. The present study is based on time series data (i.e. from 1978 to 2014). Therefore, in the energy function, a dummy variable has also been included, which represents the possibility of the series discontinuity or structural break. The NARDL results reveal that the upside variations in national output (GDP) have amplified the scope of energy consumption in the long run, whereas the impact of downside variations is negative and substantial. Similarly, the study has separately captured the impact of upside and downside variations in the globalization index (i.e., economic and socio-political factors) on energy consumption in the country. It is evident from the results that improvement in these factors has intensified energy consumption in the long run, whereas the impact of decreased globalization index is found negative and significant. Further, the study confirms that the upside variation in capital formation has not significantly reduced energy utilization in the region. However, the downside movements (i.e., negative shocks in capital formation) have significantly increased the demand for energy in India. It is evident from the results that besides the increase in national output (GDP), the increase in the socio-political arena has also contributed to raising energy consumption in India. Contrarily, the downside movements in economic and socio-political factors have led to decreased energy consumption. The weak substitutability between upside movements in capital formation and energy consumption reveals that the country needs to generate energy-efficient production techniques. Based on the outcomes, it can be proposed that the government should promote research and development in all spheres of life (i.e., economic and socio-political) where energy is used as an input.