Although a positive association is found between HIV prevalence and intimate partner violence, a causal interpretation is hard to establish due to eneogeneity of the HIV prevalence. Using the distance from the origin of the virus as an instrument, I find that an exogenous increase in HIV prevalence in a cluster increases the risk of physical and sexual violence against women within marriage. A positive effect on emotional violence is also found in some sub-sample. I explore three channels that potentially mediate the effect of HIV on spousal violence. While I do not find convincing evidence on the effects through changing marriage market conditions or shorter expected lifespan, I find evidence that suggests the increased conflicts over condom use as a channel through which HIV increases the risk of spousal violence.