This paper examines the growth performances of Indian States from 1970 to 1994. First, using principal components analysis, we proceed at a grouping of the States following their endowments in physical, social and economic infrastructure. We then combine principal components analysis with panel data estimation techniques in order to assess the contribution of various types of infrastructure to growth. Our estimations account for reverse causality arising from the endogeneity of infrastructure investment, as well as for spatial autocorrelation of residuals and for spatial autoregression. We do find evidence of conditional convergence across States. This does not rule out persistent income inequalities due to the dispersion of steady-state income levels. Such disparities are accounted for by differences in the structure of production and in infrastructure endowments, as well as in State-specific fixed effects int he growth regression. We highlight finally spatial spillover effects of growth across States.