Theoretical literature acknowledges the interplay between knowledge creation and network collaboration, but related formal models are scarce. Empirical studies show conflicting evidence about the relation between knowledge creation and network collaboration ( ositive, none, or negative), perhaps because they take little account of their endogeneity. The simulation model in this paper deepens formal theoretical understanding and reconciles the evidence by analysing feedbacks between both processes. The model also highlights some aspects of these feedbacks that may not be desirable. Collaboration can increase disparities between agents because of the cumulative nature of knowledge creation. Collaboration may not boost knowledge creation, if collaborators are an innocuous by product of the attractiveness of most productive agents. Most productive agents may attract many collaborators and become less productive if the cost of networking is high, and even improductive if the cost of networking is higher than its contribution to knowledge creation. The model offers some keys to design policies that avoid these scenarios.