Using a randomized control trial, this paper studies an at-scale preschool construction program that serves poor communities in rural Mozambique. In addition to the construction of preschools, the program hired local instructors and provided parenting education sessions. The findings show that the program had high take-up rates, significantly increasing access to preschool education. Compared to a small base of 2 percent of children in control communities enrolled in preschool, the intervention increased preschool enrollment rates in treated communities by 73 percentage points. The program also had significant positive effects on enrollment in and progression through primary school, with an increase of 6 percentage points in enrollment in first grade at age 6, and a 0.16 standard deviation impact on an index of cognitive and social-emotional skills. Using machine learning tools, the paper estimates substantial heterogeneity by child development skills at baseline. Moreover, the program caused parents in treated communities to invest more time in supporting their primary school-aged children. [Additional funding provided by the Lee Kuan Yew School of Public Policy, National University of Singapore.]