Today’s scientific endeavors in research and education are increasingly team-driven. As a small example of the rise of team science, the percentage of multiple-authored scientific publications has risen up from 55% in the late 1960s to 90% in the early 2010s. It is thus crucial to understand the team dynamics of knowledge creation in order to design more effective funding and other support policies for today’s science. In this regard, we note two challenges facing team science, as echoed by a recent report by the National Research Council on the effectiveness of team science (NRC 2015). The first challenge is how to manage team diversity. On one hand, diversity is the very source of team creativity as it leads to greater interaction of different ideas and people. On the other hand, too much diversity may be a centrifugal force risking fragmentation of team work and constant discord of team members. The other challenge is how to manage task interdependence inherent to increasingly large-scale, complex scientific projects. Key to this challenge is a question of knowledge transfer within and beyond the team, as the acquisition of tacit knowledge becomes ever more important in increasingly open knowledge networks. We examine these challenges in our analysis of forty-six laboratories of our own institution was created on the mandate to conduct scientific and technological research projects of national strategic importance. This focus on university labs is largely due to the relatively scant attention given to university labs compared to company labs despite the former’s distinctive role in the triple helix of university-industry-government relations. Through both in-depth interviews with team leaders and members as well as statistical analyses of the data gathered from our surveys, we explore how team diversity and the process of knowledge transfer affect lab performance. Preliminary findings indicate that while the effect of team diversity on lab performance is curvilinear as predicted, the interactive effect of team diversity with tacit knowledge transfer is rather negative in teams facing greater task interdependence. The latter finding implies that the funding or other support policy for collaborative science programs must take into account a kind of coordination cost that is incurred to manage team diversity.