Networks of interaction have assumed particular significance in recent years because of their presumed importance for learning and innovation. Alliances between related firms are thought to encourage interactive learning between participating organizations through the sharing of knowledge and information, which is itself facilitated through trust, shared values and ways of working. The vast body of literature that has emerged is, however, incredibly fragmented, encompassing an array of theoretical positions and perspectives. This paper focuses upon two issues which are believed to be of particular significance and which need clarification in order to move to a clearer understanding of the ways in which networks of interaction evolve, and of their capabilities and limitations in relation to economic performance and competitiveness: (1) the importance of network structure, arguing that innovative activity requires flexibility with regard to network formation. (2) The role of geography in relation to the construction and functioning of alliances. It is the contention here that networks are likely to be increasingly international in scope.