Knowledge creation is increasingly a collaborative process, but empirical studies provide conflicting evidence on whether the relation between knowledge creation and number of collaborators is positive, negative, or nonexistent. The simulation model developed in this paper offers a deeper formal theoretical understanding and analyzes the feedback between the processes of knowledge creation and network collaboration. The model is formed by two functions, one for the formation of the network and another for the creation of knowledge, that suffice to reproduce the three coevolution scenarios described in the empirical literature. Due to the feedback mechanisms between the two functions, changes in one of the parameters deeply affect the outcome of the model, both in the amount of knowledge produced and the structure of the resulting network, as well as in the relation between them. Analyses of collaborative knowledge creation would benefit from taking into account this feedback.