Importance: Racial/ethnic disparities in quality of care among extremely preterm infants are associated with adverse outcomes. Objective: To assess whether racial/ethnic disparities in major outcomes and key care practices were changing over time among extremely preterm infants. Design, Setting, and Participants: This observational cohort study used prospectively collected data from 25 US academic medical centers. Participants included 20 092 infants of 22 to 27 weeks' gestation with a birth weight of 401 to 1500 g born at centers participating in the National Institute of Child Health and Human Development Neonatal Research Network from 2002 to 2016. Of these infants, 9316 born from 2006 to 2014 were eligible for follow-up at 18 to 26 months' postmenstrual age (excluding 5871 infants born before 2006, 2594 infants born after 2014, and 2311 ineligible infants including 64 with birth weight >1000 g and 2247 infants with gestational age >26 6/7 weeks), of whom 745 (8.0%) did not have known follow-up outcomes at 18 to 26 months. Main Outcomes and Measures: Rates of mortality, major morbidities, and care practice use over time were evaluated using models adjusted for baseline characteristics, center, and birth year. Data analyses were conducted from 2018 to 2019. Results: In total, 20 092 infants with a mean (SD) gestational age of 25.1 (1.5) weeks met the inclusion criteria and were available for the primary outcome: 8331 (41.5%) black infants, 3701 (18.4%) Hispanic infants, and 8060 (40.1%) white infants. Hospital mortality decreased over time in all groups. The rate of improvement in hospital mortality over time did not differ among black and Hispanic infants compared with white infants (black infants went from 35% to 24%, Hispanic infants went from 32% to 27%, and white infants went from 30% to 22%; P = .59 for race × year interaction). The rates of late-onset sepsis among black infants (went from 37% to 24%) and Hispanic infants (went from 45% to 23%) were initially higher than for white infants (went from 36% to 25%) but decreased more rapidly and converged during the most recent years (P = .02 for race × year interaction). Changes in rates of other major morbidities did not differ by race/ethnicity. Death before follow-up decreased over time (from 2006 to 2014: black infants, 14%; Hispanic infants, 39%, white infants, 15%), but moderate-severe neurodevelopmental impairment increased over time in all racial/ethnic groups (increase from 2006 to 2014: black infants, 70%; Hispanic infants, 123%; white infants, 130%). Rates of antenatal corticosteroid exposure (black infants went from 72% to 90%, Hispanic infants went from 73% to 83%, and white infants went from 86% to 90%; P = .01 for race × year interaction) and of cesarean delivery (black infants went from 45% to 59%, Hispanic infants went from 49% to 59%, and white infants went from 62% to 63%; P = .03 for race × year interaction) were initially lower among black and Hispanic infants compared with white infants, but these differences decreased over time. Conclusions and Relevance: Among extremely preterm infants, improvements in adjusted rates of mortality and most major morbidities did not differ by race/ethnicity, but rates of neurodevelopmental impairment increased in all groups. There were narrowing racial/ethnic disparities in important care practices, including the use of antenatal corticosteroids and cesarean delivery.