Since Esping‐Andersen’s Three Worlds, it has become a truism of welfare state research that welfare states do not vary linearly along a single dimension but have to be conceptualized as multidimensional phenomena that cluster into types caused by the political economy of class coalitions. However, when moving beyond the 18 original countries of Esping‐Andersen’s analysis, the situation is less clear. Although additional worlds have been identified in the Global North and the Global South, these are usually not conceptualized along the same dimensions as the original three worlds and are rarely empirically compared with them. This paper tackles these omissions by explicitly comparing Northern and Southern countries within Esping‐Andersen’s framework. It poses the question whether the central insight of welfare state research, namely, that there are not just gradual differences between welfare states, but different types with qualitative differences, expands beyond classic welfare states. Based on newly generated data on social rights and social stratification, we employ cluster analysis with 45 Northern and Southern countries. This analysis produces mixed results. We do find different types of welfare states with qualitative differences, but these do not fully correspond to Esping‐Andersen’s Three Worlds. Moreover, our findings also point to a conceptual issue in welfare regime research: regimes are not just defined and measured in terms of different logics of welfare provision but also take into account degrees of welfare stateism. We argue that this issue is poised to become ever more pressing with the geographical expansion of welfare state research.