Type 2 diabetes (T2D) is a global epidemic that currently affects 347 million people worldwide. Many of the abnormalities that characterize this disorder, particularly perturbations in glucose levels, appear to begin as early as 12 years of age. One proposition to explain its pathogenesis, the ‘fetal insulin hypothesis’, suggests that genetically determined impaired insulin secretion or action leads to poor insulin-mediated fetal growth and impaired insulin function in post-fetal life. Elevated glucose levels and an increased risk for T2D is a consequence of these impairments. While the hypothesis is compelling, to date little empirical evidence has been presented in its support. We used a family-based design to explore genes responsible for elevated blood glucose at birth and in early infancy. 610 newborns, 643 mothers and 366 fathers enrolled in the Family Atherosclerosis Monitoring In EarLY Life (FAMILY) study were included. Plasma glucose and anthropometric characteristics were collected at birth, 3, and 5 years. Quantitative trait disequilibrium tests and mixed-effects regressions were conducted to uncover genes associated with glucose. A genotype score was also tested. Risk alleles for six loci increased glucose levels from birth to 5 years of age (ADCY5, ADRA2A, CDKAL1, CDKN2A/B, GRB10, and TCF7L2, 4.85×10−3 ≤ P ≤ 4.60×10−2). The six genes are also associated with impaired pancreatic function, low birth weight (a marker of poor fetal growth), and T2D, supporting a pattern consistent with the fetal insulin hypothesis. Effect sizes for these genes were similar to those in adults, suggesting age-independent effects. Ours is the first study to systematically explore genetics of glucose in early childhood and among a few to provide support for the fetal insulin hypothesis. Our results suggest that a unique subset of genes, involved in beta cell dysfunction, low birth weight, and increased T2D risk, modulate plasma glucose at birth and in early infancy.