Objective To determine the optimal dose of vitamin D supplementation to achieve biochemical vitamin D sufficiency in extremely low gestational age newborns in a masked randomized controlled trial. Study design 100 infants 23 0/7-27 6/7 weeks gestation were randomized to vitamin D intakes of placebo (n = 36), 200 IU (n = 34), and 800 IU/d (n = 30) (approximating 200, 400, or 1000 IU/d, respectively, when vitamin D routinely included in parenteral or enteral nutrition is included). The primary outcomes were serum 25-hydroxy vitamin D concentrations on postnatal day 28 and the number of days alive and off respiratory support in the first 28 days. Results At birth, 67% of infants had 25-hydroxy vitamin D <20 ng/mL suggesting biochemical vitamin D deficiency. Vitamin D concentrations on day 28 were (median [25th-75th percentiles], ng/mL): placebo: 22 (13-47), 200 IU: 39 (26-57), 800 IU: 84.5 (52-99); P <.001. There were no differences in days alive and off respiratory support (median [25th-75th percentiles], days): placebo: 1 (0-11), 200 IU: 0 (0-8), and 800 IU: 0.5 (0-22); P =.63, or other respiratory outcomes among groups. Conclusions At birth, most extremely preterm infants have biochemical vitamin D deficiency. This biochemical deficiency is reduced on day 28 by supplementation with 200 IU/d and prevented by 800 IU/d. Larger trials are required to determine if resolution of biochemical vitamin D deficiency improves clinical outcomes.