In the energy-growth nexus, due to structural change that cannot be accounted for in the linear model this study considers the possibility of the nonlinear model. Using dataset for fifteen countries in Africa for the period 1980 to 2010, we found that linear model has misspecification error, thus, lending support for the nonlinear model. We provide evidence of inverted “U” shape in Benin, Cote d’Ivoire, Egypt, Togo and Tunisia. This also confirms EKC hypothesis. The low regime of energy consumption retards growth in Sudan and South Africa, while high regime of energy consumption enhances growth in Algeria, Morocco and Senegal. We also find evidence of neutrality hypothesis in Cameroon and Zambia.