We examine the optimal licensing strategy of a research lab selling to firms who are product market competitors. We consider an independent lab as well as a research joint venture. We show that (1) demands are interdependent and hence the standard price mechanism is not the profit-maximizing licensing strategy; (2) the seller's incentives to develop the innovation may be excessive; (3) the seller's incentives to disseminate the innovation typically are too low; (4) larger ventures are less likely to develop the innovation, and more likely to restrict its dissemination in those cases where development occurs; and (5) a downstream firm that is not a member of the research venture is worse off as a result of the innovation.