A key challenge to HIV vaccine development is the integration of HIV proviral DNA into the host genome upon infection. Therefore, an optimal vaccine should block infection within hours of viral exposure, providing 'sterilizing immunity' at mucosal sites and in blood via potent, broadly reactive antibody to the HIV envelope glycoprotein. This is difficult due to the envelope's conformational complexity and sequence diversity. Antibodies that do not completely prevent infection nevertheless could reduce the viral infectious burden, allowing strong cellular immunity to control viremia, delay disease progression and prevent viral transmission, while also providing help for T- and B-cell responses. Rapidly responsive, potent, persistent immunity might best be achieved using prime-boost strategies incorporating a replicating vector and an optimally designed envelope subunit.