We examined correlates of late and delayed initiation of antiretroviral therapy (ART) in British Columbia, Canada. From December 2013 to December 2015 we recruited treatment-naïve people living with HIV who initiated ART within the previous year. 'Late initiation' was defined as CD4 cell count ≤500 cells/µL at ART initiation and 'delayed initiation' as ≥1 year between HIV diagnosis and initiation. Multivariable logistic regression assessed independent correlates of late and delayed initiation. Of 87 participants, 44 (51%) initiated late and 22 (26%) delayed initiation. Delayed initiation was positively associated with older age (adjusted odds ratio [AOR]: 1.06 per year, 95% confidence interval [95% CI]: 1.01-1.12) and inversely associated with wanting to start ART at diagnosis (AOR: 0.06, 95% CI: 0.02-0.21). Variables associated with late initiation were older age (AOR: 1.09 per year, 95% CI: 1.03-1.15) and medical reason(s) for initiation (AOR: 5.00, 95% CI: 1.41-17.86). Late initiation was less likely among those with greater perceived ART efficacy (AOR 0.94, 95% CI: 0.90-0.98) and history of incarceration (AOR: 0.12, 95% CI: 0.03-0.56). Disparities in timing of initiation were observed for age, perceived ART efficacy, and history of incarceration. Enhanced health services that address these factors may facilitate earlier treatment initiation.