Highly active antiretroviral therapy (HAART) has dramatically improved survival after AIDS. The benefits of HAART have not been equally realized for all communities, however. We characterize the association of socioeconomic status (SES) with survival after AIDS diagnosis in San Francisco in the period before (1980-1995) and after (1996 - 2001) the wider use of HAART. Using citywide surveillance data, we examined differences in survival after AIDS diagnosis by neighborhood household income using Kaplan-Meier survival analysis and Cox proportional hazards analysis to adjust for significant covariates. Residing in higher SES neighborhoods significantly predicted better survival after AIDS from 1996 to 2001 (hazard ratio = 0.92 per $10,000 increase in neighborhood household income, 95% CI: 0.85-0.99) after adjusting for CD4 count at diagnosis, age, and injection drug user status. Persons living in poorer neighborhoods were less likely to use HAART at any time in the past compared with persons in wealthier neighborhoods. Moreover, no association between survival and neighborhood SES was evident in the era prior to the wide use of HAART. Finally, the difference in survival by neighborhood income level disappeared after controlling for the use of HAART, suggesting that use of or access to treatment explained the association. From 1996 to 2001, survival with AIDS was worse for people living in poorer neighborhoods compared with those living in wealthier neighborhoods of San Francisco as a result of unequal access to or use of HAART.