Abstract: Fighting in internal armed conflicts does not occur arbitrarily across a countries territory. Therefore, this paper addresses the question where fighting takes place in sub-state wars with at least two non-state actors and how we can explain spatial patterns of armed conflict in areas of limited statehood. In order to predict conflict behavior in space, this analysis, first, builds upon the theoretical argument that how actors finance their fighting explains patterns of violence in sub-state wars. Second, the paper uses a new database on violent events in Somalia from 1990-2006 to test the implications of our spatial theory. The findings show that strategic and financially important grids experience the most violent events.