The increasing number of people living with human immunodeficiency virus, HIV, (PLWH) have an elevated incidence of risk for noncommunicable comorbidities, the aetiology of which remains incompletely understood. While sleep disturbances are often reported in PLWH, it is unknown to what extent they relate to changes in the circadian and/or sleep homeostatic processes. We studied the relationship between sleep characteristics, circadian phase, and HIV status in older adults from the HAALSI (Health and Ageing in Africa: a Longitudinal Study of an INDEPTH Community in South Africa) subsample of the Agincourt Health and Demographic Surveillance System in South Africa (n = 187, 36 human immunodeficiency virus positive [HIV+], age: 66.7 ± 11.5 years, range 45—93 years), where HIV prevalence is high and (in contrast to the global north) does not associate significantly with potentially confounding behavioural differences. In participants with valid actigraphy data (n = 172), regression analyses adjusted for age and sex indicated that HIV+ participants had slightly later sleep onset (β = .16, p = .039), earlier sleep offset times (β = −.16, p = .049) and shorter total sleep times (β = −.20, p = .009) compared to the HIV negative (HIV−) participants. In a subset of participants (n = 51, 11 HIV+), we observed a later dim light melatonin onset (DLMO) in HIV+ (21:16 ± 01:47) than in HIV− (20:06 ± 00:58) participants (p = .006). This substantial difference remained when adjusted for age and sex (β = 1.21; p = .006). In 36 participants (6 HIV+) with DLMO and actigraphy data, median phase angle of entrainment was −6 min in the HIV+ group and +1 h 25 min in the HIV− group. DLMO time correlated with sleep offset (ρ = 0.47, p = .005) but not sleep onset (ρ = −0.086, p = .623). Collectively, our data suggest that the sleep phase occurred earlier than what would be biologically optimal among the HIV+ participants. This is the first report of a mistimed circadian phase in PLWH, which has important potential implications for their health and well-being, especially given the well-established relationships between circadian asynchrony and sleep deprivation with poorer health outcomes.