The role of social capital in promoting health is now widely debated within international public health. In relation to HIV, the results of previous observational and cross-sectional studies have been mixed. In some settings it has been suggested that high levels of social capital and community cohesion might be protective and facilitate more effective collective responses to the epidemic. In others, group membership has been a risk factor for HIV infection. There have been few attempts to strengthen social capital, particularly in developing countries, and examine its effect on vulnerability to HIV. Employing data from an intervention study, we examined associations between social capital and HIV risk among 1063 14 to 35-year-old male and female residents of 750 poor households from 8 villages in rural Limpopo province, South Africa. We assessed cognitive social capital (CSC) and structural social capital (SSC) separately, and examined associations with numerous aspects of HIV-related psycho-social attributes, risk behavior, prevalence and incidence. Among males, after adjusting for potential confounders, residing in households with greater levels of CSC was linked to lower HIV prevalence and higher levels of condom use. Among females, similar patterns of relationships with CSC were observed. However, while greater SSC was associated with protective psychosocial attributes and risk behavior, it was also associated with higher rates of HIV infection. This work underscores the complex and nuanced relationship between social capital and HIV risk in a rural African context. We suggest that not all social capital is protective or health promotive, and that getting the balance right is critical to informing HIV prevention efforts.