A successful HIV vaccine would have a substantial impact on acquisition of infection, progression of disease among the infected, or infectiousness of the infected. Current vaccine candidates are anticipated to have their major effect on viremia, however, with the expectation that this would induce or be concordant with a reduced rate of AIDS, death, or infectiousness. Although direct assessment of disease progression or infectiousness may be impractical, available potential surrogates for these endpoints may be misleading. This article summarizes the proceedings of a National Institute of Allergy and Infectious Disease-sponsored workshop to explore the use of surrogate endpoints for licensure of an HIV vaccine. Early, medium, and late endpoints were discussed, along with challenges such as surrogate validity, the confounding effect of antiretroviral therapy initiation, and potential selection bias in the vaccine and placebo recipients who become infected. Results from 5 hypothetic HIV vaccine clinical trials with ambiguously successful results were presented to an expert panel for interpretation and discussion of next steps. Key recommendations included assessing magnitude and durability of surrogate effects, generalization across populations, and directed improvement of vaccines. Use of acquisition and a postinfection surrogate as coprimary endpoints was supported, along with use of composite endpoints and exploration of heterogeneity in vaccine efficacy by characteristics of the host and virus.