This paper examines the growth trajectories of Indian states. We find that not only is the disparity among states increasing, there is evidence that the distribution of the per-capita income of states is becoming bimodal. We use the methodology of Phillips and Sul (2007) to identify convergence clubs among Indian states and find three clubs. Furthermore, in a novel application of this methodology we use the identified clubs to run separate growth regressions for different clubs of states and find significant differences in the regression coefficients. This vindicates our approach and shows that combining data from all states at the same time to study growth dynamics can be misleading in a heterogenous country like India.