We sought to examine the relationship between housing status and risk of HIV-infection among injection drug users in Vancouver, Canada. Using Kaplan-Meier survival analysis, we found an elevated HIV incidence rate among those who reported residing in unstable housing (log-rank p=0.006). In Cox's regression survival analysis, unstable housing remained marginally associated with elevated risks of HIV infection (relative hazard=1.40 (95% confidence interval: 0.09-2.00); p=0.084) after adjustment for potential confounders including syringe sharing. Adjusted generalized estimating equations analysis that examined factors associated with unstable housing demonstrated that residing in unstable housing was independently associated with several HIV risk behaviours including borrowing used needles (adjusted odds ratio (OR)=1.14) and sex-trade involvement (adjusted OR=1.19). Our findings suggest that unstable housing environments are associated with elevated risk of HIV- infection due to risk behaviours that take place in these environments. Implications for policy including more comprehensive housing interventions (e.g. 'floating support') are discussed.