We study how extreme weather exposure impacts infant survival in the developing world. Our analysis overcomes the absence of vital registration systems in many poor countries by extracting birth histories from household surveys. Studying 53 developing countries that span five continents, we find impacts of hot days on infant morality that are an order of magnitude larger than corresponding estimates from rich country studies, with humidity playing an important role. The size and implied geographic distribution of harms documented here have the potential to significantly alter assessments of optimal climate policy.