While the benefit of coronary revascularization in patients with stable coronary artery disease (CAD) is debated, data assessing the potential interest of stress CMR to guide coronary revascularization are limited. To assess the long-term prognostic value of stress CMR in consecutive patients from a large registry. Between 2008 and 2018, a retrospective cohort study with a median follow-up of 6.0 years (IQR: 5.0–8.0) included all consecutive patients referred for stress CMR. The primary outcome was all-cause death based on the National Death Registry. Among the 31,752 consecutive patients (mean age 63.7 ± 12.1 years and 65.7% males), 2679 (8.4%) died at 206,453 patient-years of follow-up. Inducible ischemia and late gadolinium enhancement (LGE) by CMR were associated with death (both P < 0.001). In multivariable Cox regression, inducible ischemia and LGE were independent predictors of death (HR = 1.61; 99.5% CI: 1.41–1.84; HR = 1.62; 99.5% CI: 1.41–1.86, respectively; P < 0.001). The addition of LVEF, LGE and inducible ischemia significantly improved the C statistic to 0.71 (99.5% CI: 0.70–0.72; C statistic improvement for model 1: 0.03; NRI = 0.035). The addition of the presence of LGE and the number of segments of LGE and inducible ischemia significantly improved the C statistic to 0.74 (99.5% CI: 0.73–0.75; C statistic improvement for model 1: 0.06; NRI = 0.061) ( Fig. 1 ). In this large observational series of consecutive patients, stress perfusion CMR had important incremental long-term prognostic value to predict death over traditional risk factors.