Using panel data from 8,650 infants in the first two waves of the Early Childhood Longitudinal Study–Birth Cohort, the current study found that the negative effects of alternative families on children’s cognitive development started to emerge when children were as young as 9 months of age. Meanwhile, coresidence with grandparents was not disadvantageous and, in some cases, even beneficial for the cognitive development of 9-month-old infants. Our analyses also found that some forms of family instability were actually more beneficial to children’s cognitive growth between 9 months and 2 years of age than all types of stable family structures. However, continuous coresidence with grandparents did not further benefit children’s cognitive growth rate. Overall, the analyses found evidence for the resource deprivation and attachment perspectives but cast some doubt on instability theory.