Experimental evidence suggests that perinatal light imprinting of circadian clocks and systems may affect downstream physiology and cancer risk in later life. For humans, the predominant circadian stimulus is the daily light-dark cycle. Herein, we explore associations between perinatal photoperiod characteristics (photoperiod: duration of daylight as determined by time-of-year and location) and childhood cancer risk. We use pooled data on 182,856 mothers and babies from prospective birth cohorts in six countries (Australia, Denmark, Israel, Norway, UK, USA) within the International Childhood Cancer Cohort Consortium (I4C). Cox proportional hazards regression was used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs). In line with predicted differential dose–responses, restricted cubic splines indicate a potential non-linear, non-monotonic relationship between perinatal mean daily photoperiod (0–24 h) and childhood cancer risk. In a restricted analysis of 154,121 individuals who experienced third trimester photoperiods exclusively within the 8–16-h range, the relative risk of developing childhood cancer decreased by 9% with every hour increase in third trimester mean daily photoperiod [HR: 0.91 (95%CIs: 0.84–0.99)]. In conclusion, in this first study of perinatal photoperiod and childhood cancer, we detected an inverse [“protective”] linear association between third trimester mean daily photoperiod and childhood cancer risk in the 8–16-h set of the total study population. Limited statistical power impeded the investigation of risks with individuals exposed to more extreme photoperiods. Future studies are needed to confirm differential photoperiod-associated risks and further investigations into the hypothesized circadian imprinting mechanism are warranted.