BACKGROUND: Men who have sex with men (MSM) have an increased risk of anogenital human papilomavirus (HPV) infection, which can lead to HPV-related anogenital lesions such as warts, anal intraepithelial neoplasia, and anal cancer. Some of these HPV types are preventable with vaccines. We aimed to describe the incidence of anal, penile, and oral HPV infection, and to estimate the site-specific transmission probability per partner, for teenage MSM. METHODS: In our observational cohort study, we enrolled teenage MSM (aged 16-20 years) with low sexual exposure and a low prevalence of HPV in Melbourne (VIC, Australia). At baseline, 3, 6, and 12 months, we took a swab from the anal canal, and participants self-collected a swab from the penis and an oral rinse. Our primary outcome was definite and probable incident HPV infection of the anus, penis, or mouth at any time in the 12 months from baseline, assessed through the presence of HPV DNA. We defined definite incident HPV infection as the same HPV type detected more than once from the same site in men who had a negative HPV test at baseline. We defined probable incident HPV infection as only one positive test. We estimated the probability of HPV transmission per partner using HPV prevalence in MSM with a similar age to partners of men in our cohort. This study is registered at the Australian New Zealand Clinical Trials Registry and ClinicalTrials.gov, numbers ACTRN12611000857909 and NCT01422356. FINDINGS: We enrolled 200 MSM aged 16-20 years (median 19 years [IRQ 18-20; range 16-20]) between Sept 20, 2010, and Aug 24, 2012. Over the 12 month follow-up period, we detected 48 definite (107 possible) HPV infections in the anus, ten definite (34 possible) HPV infections on the penis, and no definite (six possible) infections in the mouth. Definite incidence rate per 100 person-years for any anal HPV infection was 57 (95% CI 46-68), and for any anal HPV type in the quadrivalent vaccine was 33 (23-44). Definite incidence rate per 100 person-years for any penile HPV was 12 (6-21) and for any HPV type in the quadrivalent vaccine was 5 (1-12). Estimated probabilities of HPV transmission from the penis to the anus were significantly higher than were those from the anus to the penis (p<0·05 for all HPV types in the quadrivalent vaccine). INTERPRETATION: High incidence rates suggest that the vaccination coverage in MSM will need to be high. The transmission estimates will inform HPV modelling. FUNDING: Merck.