Background Outcomes of HIV-infected children before widespread use of antiretroviral therapy (ART) for treatment and prevention of mother-to-child transmission (PMTCT) have been well characterized but less is known about children who acquire HIV infection in the context of good ART access. Methods We enrolled newly-diagnosed HIV-infected children ≤24 months (mos) of age at 3 hospitals and 2 clinics in Johannesburg, South Africa. We report ART initiation and mortality rates during 6 months from enrollment and factors associated with mortality. Results Of 272 children enrolled, median age 6.1mos, 69.5% were diagnosed during hospitalization. By 6mos post-enrollment, 53(19.5%) died and 73(26.8%) were lost-to-follow-up. Using Kaplan-Meier analysis, the probability of death by 6mos after enrollment was 23.5%. The median age of death was 9.1mos (95% confidence interval[CI]: 8.6; 12.0). Overall, 226(83%) children initiated ART which was associated with a 71% reduction in risk of death (Hazard Ratio[HR] =0.29 [95%CI: 0.15; 0.58]). In multivariable analysis of infant factors, weight-for-age z-score <-2SD (HR=2.43[95%CI: 1.03-5.73]), CD4<20% (HR=3.29[95%CI: 1.60-6.76]) and identification during hospitalization (HR=2.89[95%CI: 1.16-7.25]) were independently associated with mortality. In multivariable analysis of maternal factors, CD4≤350/no maternal ART was associated with increased mortality risk (HR=2.57[95%CI: 1.19-5.59]) vs. CD4>350/no maternal ART; exposure to maternal/infant antiretrovirals for PMTCT was associated with reduced mortality risk (HR=0.53[95%CI: 0.28-0.99]) vs. no PMTCT. Conclusions ART initiation is highly protective against death in young children. However, despite improved access to ART, young children remain at risk for early death; innovative approaches to rapidly diagnose and initiate treatment as early in life as possible are needed.