This study addresses the following research questions: what happens to the propensity to collaborate with other firms, once the researcher creates her own venture? Do her collaborations decrease or do they grow? These questions have been overlooked by the current literature, even though they carry important policy implications. Our key suggestion is that the effectiveness of a technology transfer tool can be better assessed by taking into account the possible crowding-out effects with other channels of knowledge transfer. We do so for Italy, by comparing the behavior of single researchers, before and after the establishment of their own firm, with the behavior of a control group. We assess whether those academics that founded their own firm significantly change their attitude to perform research collaborations with other firms by means of co-publication and co-patenting. We also verify whether creating a firm has an impact on the overall patenting and publication performance. Our results suggest that there is a negative effect on the overall publishing performance, while the patenting activity does not change significantly. Regarding copublications, our results confirm the existence of a substitution effect between spin-offs and co-publication with firms, while we observe an increase in the case of co-patenting. A closer look at the data, however, suggests that the latter is mainly triggered by the patenting activity that the researcher performs with his or her own company.