Pregnancy exercise can prevent excessive gestational weight gain (EGWG), gestational diabetes mellitus (GDM) and hypertension (GH), but inter-individual variability has not been explored. We aimed to analyze the prevalence--and potential sociodemographic and medical predictors of--non-responsiveness to gestational exercise, and the association of non-responsiveness with adverse pregnancy outcomes. Among 688 women who completed a supervised light-to-moderate intensity exercise program (three ~1-h sessions/week including aerobic, resistance, and pelvic floor muscle training) until near-term, those who showed EGWG, GDM or GH were considered 'non-responders'. A low prevalence of non-responders was observed for GDM (3.6%) and GH (3.4%), but not for EGWG (24.2%). Pre-pregnancy obesity was the strongest predictor of non-responsiveness for GH (odds ratio 8.40 [95% confidence interval 3.10-22.78] and EGWG (5.37 [2.78-10.39]), whereas having a highest education level attenuated the risk of being non-responder for GDM (0.10 [0.02-0.49]). Non-responsiveness for EGWG was associated with a higher risk of prolonged labor length, instrumental/cesarean delivery, and macrosomia, and of lower Apgar scores. No association with negative delivery outcomes was found for GDM/GH. In summary, women with pre-pregnancy obesity might require from additional interventions beyond light-to-moderate intensity gestational exercise (e.g., diet and/or higher exercise loads) to ensure cardiometabolic benefits.