If localized knowledge spillovers are present in the university setting, higher rates of both start-up and/or survival would be observed in areas that are geographically proximate to the university. Using a detailed industry data set for Texas for 1999:3–2005:2, we analyze start-up and exit rates for high-tech firms. Based on a Poisson quasi-maximum likelihood estimation, we find evidence that the level of R&D and proximity of a research institution positively affects the likelihood of technology start-ups. However, using both the Cox proportional hazards model and Kaplan–Meier approach, our results suggest that geographic proximity to knowledge centers does not reduce hazard rates. (JEL R12, R53, O18)