A number of potential advantages of industry agglomeration—or spatial clustering—have since long been identified in the research literature, notably related to shared costs for infrastructure, the build-up of a skilled labor force, transaction efficiency, and knowledge spill-overs leading to firm learning and innovation. We identify two shortcomings of existing research on the clustering phenomenon. First, the abundance of theoretical concepts stands in sharp contrast with the general lack of work aiming to validate these mechanisms empirically and the contradictory evidence found in recent empirical work in the field. Secondly, there is a lack of a unified theoretical framework for analyzing spatial clustering. In an attempt to remedy the second shortcoming, this paper investigates the nature of the cluster from a knowledge creation or learning perspective. We argue for the need to establish a specific theory of the cluster where learning occupies center stage. The basic requirements for such a theory of the cluster are discussed. Two main components of such a  theory are identified: it must explain the existence of the cluster on the one hand and its internal organization on the other.