As global temperatures go up and incomes rise, air conditioner sales are poised to increase dramatically. Recent studies explore the potential economic and environmental impacts of this growth, but relatively little attention has been paid to the implications for inequality. In this paper we use household-level microdata from 16 countries to characterize empirically the relationship between climate, income, and residential air conditioning. We show that both current and future air conditioner usage is concentrated among high-income households. Not only do richer countries have much more air conditioning than poorer countries, but within countries adoption is highly concentrated among high-income households. The pattern of adoption is particularly stark in relatively low-income countries such as Pakistan, where we show that the vast majority of adoption between now and 2050 will be concentrated among the upper income tercile. We use our model to forecast future adoption, show how patterns vary across countries and income levels, and discuss what these patterns mean for health, productivity, and educational inequality.