TP53 mutations are considered a surrogate biomarker of the serous‐like ‘copy number high’ molecular subtype of endometrial carcinoma (EC). In ovarian carcinoma, p53 immunohistochemistry (IHC) accurately reflects mutational status with almost 100% specificity but its performance in EC has not been established. This study tested whether p53 IHC reliably predicts TP53 mutations identified by next‐generation sequencing (NGS) in EC biopsy samples for all ECs and as part of a molecular classification algorithm after exclusion of cases harbouring mismatch repair defects (MMRd) or pathogenic DNA polymerase epsilon exonuclease domain mutations (POLEmut). A secondary aim assessed inter‐laboratory variability in p53 IHC. From a total of 207 cases from five centres (37–49 cases per centre), p53 IHC carried out at a central reference laboratory was compared with local IHC (n = 164) and curated tagged‐amplicon NGS TP53 sequencing results (n = 177). Following consensus review, local and central p53 IHC results were concordant in 156/164 (95.1%) tumours. Discordant results were attributable to both interpretive and technical differences in staining between the local and central laboratories. When results were considered as any mutant pattern versus wild‐type pattern staining, however, there was disagreement between local and central review in only one case. The concordance between p53 IHC and TP53 mutation was 155/168 (92.3%) overall, and 117/123 (95.1%) after excluding MMRd and POLEmut EC. Three (3/6) discordant results were in serous carcinomas with complete absence of p53 staining but no detectable TP53 mutation. Subclonal mutant p53 IHC expression was observed in 9/177 (5.1%) cases, of which four were either MMRd or POLEmut. Mutant pattern p53 IHC was observed in 63/63 (100%) serous carcinomas that were MMR‐proficient/POLE exonuclease domain wild‐type. Optimised p53 IHC performs well as a surrogate test for TP53 mutation in EC biopsies, demonstrates excellent inter‐laboratory reproducibility, and has high clinical utility for molecular classification algorithms in EC. © 2019 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.