Using a spatial dynamic panel, the long-run industrial sector convergence rate across Mexico’s states is found to be 2%. The model is a system-General Method of Moments with correction for spatial autocorrelation and an explicit human capital input. The significant inequality between the richest and poorest states is caused by differences in factor accumulation. Physical capital accumulation dominates in richer states while the human capital accumulation is in poorer states. Regional inequality is predicted to grow unless there is government intervention to address the bipolar regional divide. More investment in human capital in non-industrialized states to draw strength from Mexico’s diversity is recommended.