The tumor–stroma ratio (TSR) was evaluated as a promising parameter for breast cancer prognostication in clinically relevant subgroups of patients. The TSR was assessed on hematoxylin and eosin‐stained tissue slides of 1,794 breast cancer patients from the Nottingham City Hospital. An independent second cohort of 737 patients from the Netherlands Cancer Institute to Antoni van Leeuwenhoek was used for evaluation. In the Nottingham Breast Cancer series, the TSR was an independent prognostic parameter for recurrence‐free survival (RFS; HR 1.35, 95% CI 1.10–1.66, p = 0.004). The interaction term was statistically significant for grade and triple‐negative status. Multivariate Cox regression analysis showed a more pronounced effect of the TSR for RFS in grade III tumors (HR 1.89, 95% CI 1.43–2.51, p < 0.001) and triple‐negative tumors (HR 1.86, 95% CI 1.10–3.14, p = 0.020). Comparable hazard ratios and confidence intervals were observed for grade and triple‐negative status in the ONCOPOOL study. The prognostic value of TSR was not modified by age, tumor size, histology, estrogen receptor status, progesterone receptor status, human epidermal growth factor receptor 2 status or lymph node status. In conclusion, patients with a stroma‐high tumor had a worse prognosis compared to patients with a stroma‐low tumor. The prognostic value of the TSR is most discriminative in grade III tumors and triple‐negative tumors. The TSR was not modified by other clinically relevant parameters making it a potential factor to be included for improved risk stratification.