This article summarizes the historical background for the use of oxygen during newborn resuscitation and describes some of the research and the process of changing the previous practice from a high-to a low-oxygen approach. Findings of a recent Cochrane review suggest that more than 100 000 newborn lives might be saved globally each year by changing from 100 to 21% oxygen for newborn resuscitation. This estimate represents one of the largest yields for a simple therapeutic approach to decrease neonatal mortality in the history of pediatric research. Available data also suggest that, for the very low birth weight infant, use of the low-oxygen approach should be considered with the understanding that some of the smallest and sickest preterm neonates will need some level of oxygen supplementation during the first minutes of postnatal life. As more data are needed for the very preterm population, creation of strict guidelines for these infants would be premature at present. However, it can be stated that term and late preterm infants in need of resuscitation should, in general, be started on 21% oxygen, and if resuscitation is not started with 21% oxygen, a blender should be available, enabling the administration of the lowest FiO 2 possible to keep heart rate and SaO 2 within the target range. For extremely low birth weight infants, initial FiO 2 could be between 0.21 and 0.30 and adjusted according to the response in SaO 2 and heart rate.