The success of combination antiretroviral therapy for HIV-1 infection has generated interest in mechanisms by which the virus can persist in the body despite the presence of drugs that effectively inhibit key steps in the virus life cycle. It is becoming clear that viral reservoirs established early in the infection not only prevent sterilizing immunity but also represent a major obstacle to curing the infection with the potent antiretroviral drugs currently in use. Mechanisms of viral persistence are best considered in the context of the dynamics of viral replication in vivo. Virus production in infected individuals is largely the result of a dynamic process involving continuous rounds of de novo infection of and replication in activated CD4(+) T cells with rapid turnover of both free virus and virus-producing cells. This process is largely, but not completely, interrupted by effective antiretroviral therapy. After a few months of therapy, plasma virus levels become undetectable in many patients. Analysis of viral decay rates initially suggested that eradication of the infection might be possible. However, there are several potential cellular and anatomical reservoirs for HIV-1 that may contribute to long-term persistence of HIV-1. These include infected cell in the central nervous system and the male urogenital tract. However, the most worrisome reservoir consists of latently infected resting memory CD4(+) T cells carrying integrated HIV-1 DNA. Definitive demonstration of the presence of this form of latency required development of methods for isolating extremely pure populations of resting CD4(+) T cells and for demonstrating that a small fraction of these cells contain integrated HIV-1 DNA that is competent for replication if the cells undergo antigen-driven activation. Most of the latent virus in resting CD4(+) T cells is found in cells of the memory phenotype. The half-life of this latent reservoir is extremely long (44 months). At this rate, eradication of this reservoir would require over 60 years of treatment. Thus, latently infected resting CD4(+) T cells provide a mechanism for life-long persistence of replication-competent forms of HIV-1, rendering unrealistic hopes of virus eradication with current antiretroviral regimens. The extraordinary stability of the reservoir may reflect gradual reseeding by a very low level of ongoing viral replication and/or mechanisms that contribute to the intrinsic stability of the memory T cell compartment. Given the substantial long-term toxicities of current combination therapy regimens, novel approaches to eradicating this latent reservoir are urgently needed.