Background: HIV is transmitted through sexual or drug injection networks that connect the members of a population. The extent, structure and timing of such networks are thus important determinants of epidemic dynamics and the success of preventive interventions. In this presentation, I describe approaches to collecting network data; I provide a brief overview of the key concepts of network analysis (i.e., mapping, measuring and analyzing patterns of relationships between individuals at risk of HIV infection), and I review recent and ongoing work on HIV risks and transmission networks. In doing so, I focus on studies of heterosexual networks in HIV endemic settings of eastern and southern Africa. In such populations, networks are often characterized by a low average number of partnerships. Nonetheless, dense and decentralized networks still emerge from the formation and dissolution of multiple and possibly concurrent partnerships. Coupled with high infectivity during acute infection, these networks often expose large numbers of individuals to the risk of acquiring HIV. Conclusions: I identify two gaps in our understanding of HIV transmission networks: 1. A narrow understanding of the complex interactions between network dynamics and the scale-up of biomedical HIV prevention; 2. Limited data on key network parameters that are not easily captured in standard sexual behavior surveys. Finally, I describe emerging opportunities to learn about HIV transmission networks through sociocentric network studies, data routinely collected by health services during partner notification, or phylogenetic studies.