The clonality of latent HIV has been inferred from viral sequences derived from various sources, including proviruses, residual viraemia of patients stable on treatment, and rebound viraemia in patients who have undergone treatment interruption. It has also been deduced from the integration sites of infected cells. It is unclear whether these data accurately reflect reactivatable latent viruses. We studied viruses reactivated from latently infected cells. Resting CD4 T cells isolated from a patient stable on treatment underwent limiting dilution and were subsequently activated with PHA, IL-2 and irradiated PBMC followed by co-culture with SupT1-CCR5 feeder cells for 21 days. The supernatant was harvested for viral RNA. Amplicons were generated from a region in Gag and one in Env and analysed by Sanger sequencing. To control for sequence variations acquired during culture, SupT1-CCR5 cells were infected with NL4-3 and underwent the same limiting dilution, culture, RNA isolation and sequencing. Pairwise comparisons were performed to obtain p-distances. The p-distances obtained from NL4-3 infected SupT1-CCR5 cells were used as references. Each pair of patient derived viral sequences was considered different if the p-distance was higher than that of the corresponding region of the reference sequences. We obtained 32 sequences of reactivated latent viruses from a single patient.19 distinct sequences could be distinguished from the Gag region. The remaining 13 sequences segregated into five groups containing up to four sequences. However, when the Env regions of these 13 sequences were analysed, only one ‘clonal’ group of two sequences remained. 30/32 reactivated latent viruses were distinct. If the threshold p-distance for two sequences to be considered distinct was set at the maximal (rather than average) p-distance observed in the reference set, 26/32 of reactivated latent viruses could still be considered distinct. These data suggests that the majority of reactivatable latent viruses are genetically distinct. Our data show that the phylogenetic structure of reactivatable latent viruses is wholly different from that of residual viraemia, where a single ‘predominant plasma clone’ can account for over 50% of all sequences observed.