Survival rates in excess of 25% at 23 weeks' gestation and in excess of 50% at 24 weeks' gestation have been reported among live births in the 1990s within tertiary perinatal care centres in the USA and Australia. Decisions about medical management at these gestational ages can no longer be based merely on whether survival is possible. Relevant moral considerations include the primacy of the newborn's best interests, parental autonomy, physicians' duties of beneficence and non-maleficence, and distributive justice. There is significant variability between developed nations in the survival of extremely premature infants among cohorts born within perinatal tertiary care centres. This is, at least to some degree, the result of differences in the aggressiveness of obstetrical and neonatal management at these gestational ages. There is also great variability in the prevalence of major neurodevelopmental disability among survivors. Moreover, the prevalence of major disabilities does not inform quality-of-life considerations adequately. Despite similar gestational age ranges over which the benefit:burden ratio of aggressive obstetric and neonatal care is questioned in developed countries, there is marked variation in the frequency with which it is provided within these ranges. This is understandable given the relevant moral values and the different ways in which competing values will be balanced by different individuals, cultures and societies; the increasing survival of extremely premature infants, but the persistence of high (but widely variable) prevalences of major disabilities reported among survivors and even higher prevalences of mild-to-moderate neurodevelopmental sequelae; our imperfect ability to estimate an individual extremely premature infant's prognosis; and the complexities of estimating the quality of life from the individual's own perspective.