Characterization of HIV-1 subtype diversity in regions where vaccine trials are conducted is critical for vaccine development and testing. This study describes the molecular epidemiology of HIV-1 within a tea-plantation community cohort in Kericho, Kenya. Sixty-three incident infections were ascertained in the HIV and Malaria Cohort Study conducted in Kericho from 2003 to 2006. HIV-1 strains from 58 of those individuals were full genome characterized and compared to two previous Kenyan studies describing 41 prevalent infections from a blood bank survey (1999–2000) and 21 infections from a higher-risk cohort containing a mix of incident and prevalent infections (2006). Among the 58 strains from the community cohort, 43.1% were pure subtypes (36.2% A1, 5.2% C, and 1.7% G) and 56.9% were inter-subtype recombinants (29.3% A1D, 8.6% A1CD, 6.9% A1A2D, 5.2% A1C, 3.4% A1A2CD, and 3.4% A2D). This diversity and the resulting genetic distance between the observed strains will need to be addressed when vaccine immunogens are chosen. In consideration of current vaccine development efforts, the strains from these three studies were compared to five candidate vaccines (each of which are viral vectored, carrying inserts corresponding to parts of gag, pol, and envelope), which have been developed for possible use in sub-Saharan Africa. The sequence comparison between the observed strains and the candidate vaccines indicates that in the presence of diverse recombinants, a bivalent vaccine is more likely to provide T-cell epitope coverage than monovalent vaccines even when the inserts of the bivalent vaccine are not subtype-matched to the local epidemic.