This paper aims to clarify the nature of university patents as outputs of different types of research and inputs of diverse instruments of interaction, focusing on the case of the polytechnic University of Valencia at departmental level. A patent production function tests what kind of research gives rise to patents. Several funding functions show for which instruments patents are a better input. University patents appear to be an output of costly and long-termoriented research, either publicly or privately financed, and also the input of a certain type of interaction — not through licensing but through signalling competencies. The fear that university patents negatively affect the quality of research is not justified, as they are the outcome of research at the frontiers of science. However, it is true that university patents only stimulate interaction with those firms that have enough absorptive capacity. If interaction with less capable firms is intended, other instruments are required.