Objective: To evaluate, in very preterm infants, the hemoglobin (Hb) levels during the first 24 h and the neurodevelopment outcomes at 24 months of corrected age. Design, setting, and patients: We conducted a secondary analysis of the French national prospective and population-based cohort EPIPAGE-2. The eligible study participants were live-born singletons who were born before 32 weeks of gestational age, with early Hb levels who were admitted to the neonatal intensive care unit. Main outcome measures: The early Hb levels for an outcome survival at 24 months of corrected age without neurodevelopmental impairment were measured. The secondary outcomes were survival at discharge and without severe neonatal morbidity. Results: Of the 2158 singletons of <32 weeks with mean early Hb levels of 15.4 (±2.4) g/dL, 1490 of the infants (69%) had a follow-up at two years of age. An early Hb of 15.2 g/dL is the minimum receiving operating characteristic curve at the 24 months risk-free level, but the area under the curve at 0.54 (close to 50%) indicates that this rate was not informative. In logistic regression, no association was found between early Hb levels and outcomes at two years of age (aOR 0.966; 95% CI [0.775–1.204]; p = 0.758) but rather there was a correlation found with severe morbidity (aOR 1.322; 95% CI [1.003–1.743]; p = 0.048). A risk stratification tree showed that male newborns of >26 weeks with Hb of <15.5 g/dL (n = 703) were associated with a poor outcome at 24 months (OR 1.9; CI: [1.5–2.4] p < 0.01). Conclusions: Early low Hb levels are associated with major neonatal morbidities in VP singletons, but not with neurodevelopment outcomes at two years of age, except in male infants of >26 Weeks GA.