Indoor air pollution (IAP) - predominantly from solid fuel use for cooking - is a global health threat, particularly for women and young children, and one of the leading causes of infant deaths worldwide in developing and emerging countries. This paper estimates the causal effect of cooking fuel choice on infant mortality, focusing on children under 5 years of age, through the channel of IAP in India using National Family Health Survey (NFHS) data during the period 1992-2016. The main empirical framework quantifies how solid cooking fuel, an indirect measure of IAP, affects under-five mortality across all 36 states in the country during this period. To address the potential endogeneity issue in the relationship between type of cooking fuel and mortality, I instrument for cooking fuel choice using forest cover and agricultural land ownership, which induce significant variations in fuel type. The non-IV results show that the use of solid fuel for cooking increases the risk of mortality in children aged under five by 0.8 percentage points, indicating that previous studies over-estimated the marginal impact by about 0.3-0.6 percentage points or 76,000-152,000 deaths per year nationally. Analysis based on IV strategy shows that cooking fuel choice has a significant impact on under-five mortality mainly through its effect on neonatal mortality, and the result is robust to a set of alternative specifications with inclusion of various controls and different estimation methods. I also develop a theoretical model of fuel use to demonstrate how this causal effect of cooking fuel choice on infant mortality can come about.