In this study, we reexamine the impact of foreign direct investment inflows on carbon emissions in middle-income and OECD countries over the period 1992–2017. For that purpose, we employ a two-step system GMM dynamic panel data estimator controlling for endogeneity, omitted variable, and simultaneity biases in our panels. Employing a dynamic panel estimation methodology, we introduced some new findings and believe that these have important policy implications. The empirical results from the analyses show that FDI increases carbon emissions in middle-income countries and provide evidence of the pollution haven hypothesis in developing countries. Our findings suggest that FDI has a small halo effect on advanced economies. Our study also provides evidence of the Environmental Kuznets Curve hypothesis as we consistently find an inverted-U-shaped relationship between carbon emissions and per capita income across different panel samples. Policymakers planning to attract FDI in middle-income countries should do a cost-benefit analysis by considering its damage to the environment and its positive impact on economic growth.