This is the first study that uses a direct measure of skyglow, an important aspect of light pollution, to examine its impact on infant health at birth. We find evidence of reduced birth weight, shortened gestational length and even preterm births. Specifically, increased nighttime brightness, characterized by being able to see only one-third to one-fourth of the stars that are visible in the absence of artificial light, is associated with an increase in the likelihood of a preterm birth by as much as 12.8 percent, or an increase of approximately 45,000 preterm births nationwide annually. Our findings add to the literature on the impact of in utero and early-life exposure to pollution, which thus far has focused primarily on air pollution. The unique feature of our identification strategy to determine a causal effect is the application of Walker's Law in physics, which provides a scientific basis to estimate skyglow. We use estimated skyglow as an instrumental variable to address the endogeneity problem associated with the skyglow variable. In addition, our study shows that increased skyglow is associated with less sleep, indicating a likely biological mechanism that links sleep deprivation to light-pollution induced circadian disruption. This result, combined with the literature on the adverse effects of sleep disorders, completes the causal chain underlying our finding on the adverse health impact of skyglow. Our study has important policy implications for current installation of LED streetlights in many U.S. municipalities, highlighting the necessity of minimizing skyglow contributed by streetlights.