The development of a safe and effective HIV vaccine remains the best hope to control the global HIV epidemic. So far, the different strategies tried for vaccine development have led to disappointing results. The first attempted strategy involved trying to raise neutralizing antibodies to inactivate the virus and prevent infection. Both of Vaxgen's VAX004 and VAX003 phase 3 trials made use of this approach but ultimately failed. Given the difficulties encountered, the focus then shifted to the cell mediated arm of the immune system, the T lymphocytes. However, the phase 2 STEP study, which was aimed to stimulate cell-mediated immunity, was halted in 2007 because it failed to prevent infection and there was an increased incidence of HIV infection in vaccinated individuals. Many researchers now believe that vaccine candidates need to induce both sustained broadly neutralizing antibodies and a strong cell-mediated response. Therefore, attention is now focused on the prime-boost approach: a DNA or vector vaccine to elicit cytotoxic T cells that destroy infected cells followed by a subunit vaccine to induce neutralizing antibodies. RV144, the largest ever HIV vaccine trial, used a prime-boost combination vaccine, which was shown to be safe and modestly effective. The ongoing RV152 study will provide more information on the modest degree of efficacy of the RV144 vaccine with results expected in 2013. Finally, the ongoing HVTN 505 trial also makes use of the prime-boost strategy and is expected to provide a better understanding of T-cell-based vaccines. In this review, we discuss the results of all the above-mentioned trials and consider whether an HIV vaccine needs to induce both humoral and cellular immunity to be effective.