Latest World Health Organization guidelines recommend weight-based nevirapine prophylaxis for all HIV-exposed infants in resource-limited settings, yet low birth weight (LBW) infants (< 2500 g) have been understudied. Using data from the NIH-funded India six-week extended-dose nevirapine (SWEN) study, a randomized clinical trial of SWEN versus single-dose nevirapine (SD) for prevention of breast-milk HIV-1 transmission, we examined the relative impact of SWEN among 737 mother-infant pairs stratified by infant birth weight. Birth weight groups were defined as very LBW (VLBW) ≤ 2000 g, moderate LBW (MLBW) >2000 g and ≤ 2500 g, and normal birth weight (NBW) > 2500 g. Outcomes were HIV-1 infection, HIV-1 infection or death by 12 months, and severe adverse events (SAEs). The Kaplan-Meier method was used to estimate probability of efficacy outcomes in birth weight groups, and differential effects of SWEN by birth weight group were examined using Cox proportional hazards models adjusting for independent risk factors for HIV maternal-to-child transmission and significant covariates. Among 50 VLBW, 249 MLBW, and 433 NBW infants, 50% were randomized to SWEN; median gestational age was 36, 38 and 38 weeks, respectively; and there was no difference in breastfeeding duration (p = 0.99). Compared to SD: SWEN-treated VLBW had lower estimates of HIV-1 infection (13% vs. 38%, p = 0.004) and HIV-1 infection or death (13% vs. 41%, p = 0.002); SWEN-treated MLBW had lower estimated HIV-1 infection (13% vs. 17%, p = 0.042); and efficacy endpoints were similar by treatment arm in NBW. In multivariate analysis, SWEN was associated with reduced risk of HIV-1 infection or death by 83% (p = 0.03) in VLBW versus 45% (p = 0.05) in MLBW. SAE frequency was similar by treatment arm in VLBW (68% vs. 76%, p = 0.53) and MLBW (37% vs. 36%, p = 0.93). SWEN may safely increase HIV-free survival among HIV-exposed LBW infants with greatest protective advantage among infants ≤ 2000 g.