We examine the trade-off between the benefits of allowing firms to cooperate in R&D and the corresponding increased potential for product market collusion. For that we utilize a dynamic model of R&D whereby we consider all possible initial marginal cost levels (technologies), including those that exceed the choke price. This global analysis yields four possibilities: initial marginal costs are above the choke price and this technology is, or is not, developed further, and initial marginal costs are below the choke price and the technology is, or is not, (eventually) taken off the market. We show that an extension of the cooperative agreement towards collusion in the product market is not necessarily welfare reducing: if firms collude, they (i) develop further a wider range of initial technologies, (ii) invest more in R&D such that process innovations are pursued more quickly, and (iii) abandon the technology for a smaller set of initial marginal costs. We also dis cuss the implications of our analysis for antitrust policy.