Innovations are not distributed uniformly in space, but are trying to cluster in certain territories, which generates a differentiated capacity for resilience in these same territories. While the concept of the Regional Innovation System (RIS) has very often been mobilized to explain this phenomenon, an essential dimension of the concept is the notion of a system, which assumes that a conjunction of innovation factors interacts with each other. As a result, the impact of one factor depends on the ‘level’ of the others. However, econometric work based onModified Knowledge Production Functions (MKPF)
assumes that regional innovation is the result of specific determinants, with each explanatory factor affecting the result independently of the others. In this article, we develop a qualitative comparative analysis (QCA), which, based on fuzzy sets and the combination of necessary and sufficient conditions, we believe offers a vision that is better suited to the interactive nature of regional innovation. Moreover, QCA is particularly appropriate for small samples, which is our case since we are interested in the large French regions that have emerged since the 2016 territorial reform. The results highlight particular configurations of sufficient conditions for regional innovation, demonstrating the complexity of territorial resilience. In terms of economic policy, the most significant implication of this work is that it has shown that regional innovation is multifaceted and that the same ‘recipes’ for regional policy do not apply everywhere, given the wide variety of regional innovation patterns.