PurposeThe present study aims to provide new evidence regarding the factors that determine the survival of firms in the Spanish wine industry and to improve the understanding of sector dynamics.Design/methodology/approachThe empirical analysis, conducted over a representative sample of wineries in the DOC Rioja wine industry, is based on non-parametric (Kaplan–Meier graph) and semi-parametric survival models (Cox proportional hazard model).FindingsThe empirical model finds that wineries with a higher number of networks with institutions enjoy better survival prospects. This study also shows that a winery’s previous performance affects the winery’s survival probability; therefore, successful wineries in the past encounter a smaller hazard of exit. Although spending on R&D and exporting are factors likely to improve wineries' efficiency and competitiveness, these factors did not contribute significantly to the survival of DOC Rioja wineries.Originality/valueThis paper makes a significant contribution to the understanding of the determinants of wineries' survival and has important policy implications. In order to raise the probability of survival, policy makers should promote the networks that link wineries and institutions. Moreover, this study is based on survival analysis which, although frequently used in medical and behavioural sciences, has rarely been applied to wine economics. Finally, it uses a unique data set obtained from primary data collection, which previous studies have not analysed in relation to the probability of winery survival.