One of the main reasons of the development of cluster policies relies on the growing awareness that network failures have to be merged with traditional market ones in the design of public incentives to foster innovation. Cluster policies aim at strengthening knowledge networks through R&D collaborative incentives in order to stimulate the expected benefits of knowledge spillovers. They rely on two related network failures. (i) The potentialities of knowledge spillovers from science to industry can be inefficiently exploited due to the cultural divide between the two communities. (ii) Entrepreneurship matters in clusters. Their performance can be assessed by SMEs and spinoff birth and entry. The aim of this research is to go more in depth into these policy guidelines, by offering an original view of dealing with the links between the implementation of these public collaborative incentives and the resulting structuring of clusters. We will focus on the Aerospace Valley cluster in Toulouse – France, over a 10 years-period (2005-2015). Our basic starting assumption is related to the fact that in spite of their control on the selection of R&D collaborations at the micro-level, cluster policy practitioners cannot have a perfect control of the structure as a whole. In network theories, these micro-macro scales problems is typical: the “macro-behavior” of the network and its structural properties, both resulting from the aggregation of ties built by agents, largely escape their own intention. Since clusters are networks, such a research can be useful to highlight the strengths and weaknesses of cluster policies.