The literature reveals that linear models do not accurately represent the asymmetric relationship between economic growth and energy consumption (EC). To fill this gap, we examined the asymmetric relationship between EC and economic growth in a non-linear panel autoregressive distributed lag (ARDL) framework of 85 countries as a whole sample and of Brazil, Russia, India, China, and South Africa (BRICS), the Next Eleven, Big Four in Western Europe, Asia-Pacific region, Group of Seven, South Asian Association for Regional Cooperation (SAARC), Economic Cooperation Organization (ECO) and the Arab League as a sub-sample analysis from 1977 to 2014. A second generational unit root test has been applied to check the problem of cross-sectional dependency. Asymmetric co-integration was employed to analyse the co-integration between the variables of interest. Long-run and short-run estimates have been calculated using the non-linear panel ARDL method. Results indicate that positive shocks to energy use tend to have a growth-enhancing effect in ECO and the Next Eleven while in the rest of the regions, the effect is growth contraction. Moreover, economic recovery from a positive shock to energy use is the case in the Arab League, Asia-Pacific region, Group of Seven and in the whole sample. However, a negative shock to EC is observed in the Group of Seven, Asia Pacific region, Big Four in Western Europe and ECO, and the whole sample worsens the economic contraction. We can deduct from this study’s results that information on the asymmetric relationship between the subject variables is needed to design sound economic policy decisions across different economic regions.