Summary Background Low‐dose aspirin is a risk factor for peptic ulcer disease but previous, population‐based cohort studies may have underestimated the low‐dose aspirin risk because they did not use a new‐user design. Gastrointestinal bleeding occurs more frequently early after initiation of low‐dose aspirin therapy than in later years. Aim To assess the associations of low‐dose aspirin with gastric and duodenal ulcer incidence in prevalent‐ and new‐user design. Methods Multivariate Cox regression models in the German ESTHER study (N = 7737) and the UK Biobank (N = 213,598) with more than 10 years of follow‐up. Results In the prevalent‐user design, there was no significant association between low‐dose aspirin and gastric ulcer observed in both cohorts. Furthermore, low‐dose aspirin was weakly, statistically significantly associated with prevalent duodenal ulcer in the UK Biobank (hazard ratio [95% confidence interval]: 1.27 [1.07–1.51]) but not in the ESTHER study (1.33 [0.54–3.29]). When restricting the exposure to only new users, the hazard ratios for incident gastric and duodenal ulcer disease were 1.82 [1.58–2.11] and 1.66 [1.36–2.04] in the UK Biobank, respectively, and 2.83 [1.40–5.71] and 3.89 [1.46–10.42] in the ESTHER study, respectively. Conclusions This study shows that low‐dose aspirin is an independent risk factor for both gastric and duodenal ulcers. The associations were not significant or weak in the prevalent‐user design and strong and statistically significant in the new‐user design in both cohorts. Thus, it is important to weigh risks against benefits when low‐dose aspirin treatment shall be initiated and to monitor adverse gastrointestinal symptoms after the start of low‐dose aspirin therapy.