We examine how recombining university-developed technology and recombining firm-developed technology differently impact the development of breakthrough inventions. Given that university-developed technology reflects the focus on basic science from university research, and firm-developed technology reflects its applied and context-dependent nature, we propose that recombining them has opposite effects on inventions. Recombining university-developed technology increases the likelihood of generating a breakthrough invention, while recombining firm-developed technology decreases the likelihood. Boundary crossing recombination interacts with the two types of sources, positively moderating the effects of recombining university- and firm-developed technology on the likelihood of generating breakthrough inventions. Drawing on a sample of U.S. nanotechnology patents, we find broad support for our hypotheses. The findings imply that aggregating the effects of recombined technologies from different sources may cause us to overlook the negative effect of recombining firm-developed technology in generating breakthrough inventions.