Abstract Purpose To identify where suboptimal timing of antenatal corticosteroids (ACS) occurs across clinical indications, and how often ACS-exposed pregnancies continue to term. Methods Single-centre retrospective cohort of 1092 pregnancies receiving betamethasone between January 2020 and October 2025. Optimal timing was defined as delivery 1–7 days after the first dose. Bile acid concentrations in intrahepatic cholestasis and bleeding status in placenta praevia were extracted from records. A prediction model developed in 2020–2024 was evaluated in the later cohort. Results Overall, 739 of 1092 pregnancies (67.7%) delivered within the optimal window and 197 (18.0%) at term. In threatened preterm labour, 701 of 867 (80.9%) delivered within the window. In indications where delivery is planned, only 38 of 225 (16.9%) did, and 116 (51.6%) reached term. No woman with intrahepatic cholestasis (0/105) delivered within the window; 99 (94.3%) delivered at term after ACS at a median of 32 weeks, none had total bile acids ≥ 100 µmol/L, and cholestasis accounted for half of all term births after ACS. No praevia pregnancy (0/24) delivered within the window, and 19 received ACS without bleeding. ACS at ≥ 34 weeks rose from 37.0% in 2020 to 47.6% in 2025. The prediction model discriminated only moderately within threatened preterm labour (AUC 0.709). Conclusion Suboptimal ACS timing was concentrated in conditions in which clinicians decide when birth occurs, particularly cholestasis. In these conditions, ACS should be scheduled within seven days of planned delivery rather than given at diagnosis.