We study the effects of the geographic expansion of a Neonatal Intensive Care Unit (NICU) system and a Newborn Emergency Transportation System (NETS) on neonatal and infant mortality and long-term impairments. We utilize gradual expansion in Hungary, we use administrative and census data, and we identify the effects from longitudinal variation in access, using changing distance as an instrument. Improving access to delivering in a city with a NICU decreases 0-6-day mortality by 153/1000 (<1500g) and 24/1000 (<2500g). NETS effects are positive but smaller. Improved access saves lives in the long run, with zero overall effects on long-term impairments.