Recent reports have demonstrated racial disparities in the prevalence of HIV infection among men who have sex with men (MSM). The objectives of this study are to investigate whether racial disparities exist in HIV incidence among young MSM in Baltimore, MD and to examine potential explanations for differences. Data were collected by the Baltimore Young Men's Survey, a cross-sectional venue-based survey (1996 to 2000) enrolling MSM aged 15 to 29 years. HIV incidence was ascertained using the serologic testing algorithm for recent HIV seroconversion. HIV incidence was 4.2% per year (95% confidence interval [CI]: 1.2 to 10.5) among 843 participants. There were substantial racial differences in HIV incidence, ranging from 0 among Hispanics to 11.0% per year (95% CI: 5.5 to 19.7) among non-Hispanic blacks. In multivariate analysis, among MSM at risk for HIV acquisition, race was not associated with unprotected anal intercourse. Independent risks included having more than 4 recent male sexual partners (adjusted odds ratio [AOR] = 1.6, 95% CI: 1.0 to 2.4) and being under the influence of drugs while having sex (AOR = 1.6, 95% CI: 1.1 to 2.3). Non-Hispanic blacks were no more likely than non-Hispanic whites to report these risk behaviors. Possible alternative explanations for the observed racial disparities in HIV incidence and implications for prevention are explored.