Abstract Aims Transcatheter aortic valve replacement (TAVR) as an alternative to surgical aortic valve replacement (SAVR) has transformed severe aortic stenosis (AS) management. Our aim was understand AS cost drivers from referral to 1-year post-procedure. Methods and results We identified patients referred for either TAVR/SAVR between 1 April 2015 and 31 March 2018, with follow-up until 31 March 2019 in Ontario, Canada. We stratified costs into (i) a referral phase, (ii) a procedural phase from the procedure date to 60 days post-procedure, and (iii) post-procedure phase from 61 days to 1 year. Multivariable regression modelling using generalized linear models with a log link gamma distribution was used to identify cost drivers in each phase. The study cohort included 12 086 AS patients; 4832 were referred for TAVR and 7254 were referred for SAVR. The median cost for TAVR was higher than SAVR in the referral ($3593 vs. $2944) and post-procedural ($5938 vs. $3257) phases. In contrast, for the procedural phase, SAVR had a median cost of $29 756 vs. $27 907 for TAVR. Predictors of high cost in the referral phase were longer wait-time, and an urgent in-hospital procedure. In the procedural phase, procedural complications were the major drivers of higher cost. In the post-procedural phase, patient co-morbidities were the major drivers, specifically dialysis, liver disease, cancer, peripheral vascular disease, and diabetes mellitus. Conclusion We identified distinct patterns of cost accumulation and modifiable drivers for SAVR compared with TAVR; these drivers may guide clinical and health policy decisions to make AS care more efficient.