De novo mutations of the gene sodium channel 1α (SCN1A) are the major cause of Dravet syndrome, an infantile epileptic encephalopathy. US incidence of DS has been estimated at 1 in 40 000, but no US epidemiologic studies have been performed since the advent of genetic testing. In a retrospective, population-based cohort of all infants born at Kaiser Permanente Northern California during 2007-2010, we electronically identified patients who received ≥2 seizure diagnoses before age 12 months and who were also prescribed anticonvulsants at 24 months. A child neurologist reviewed records to identify infants who met 4 of 5 criteria for clinical Dravet syndrome: normal development before seizure onset; ≥2 seizures before age 12 months; myoclonic, hemiclonic, or generalized tonic-clonic seizures; ≥2 seizures lasting >10 minutes; and refractory seizures after age 2 years. SCN1A gene sequencing was performed as part of routine clinical care. Eight infants met the study criteria for clinical Dravet syndrome, yielding an incidence of 1 per 15 700. Six of these infants (incidence of 1 per 20 900) had a de novo SCN1A missense mutation that is likely to be pathogenic. One infant had an inherited SCN1A variant that is unlikely to be pathogenic. All 8 experienced febrile seizures, and 6 had prolonged seizures lasting >10 minutes by age 1 year. Dravet syndrome due to an SCN1A mutation is twice as common in the United States as previously thought. Genetic testing should be considered in children with ≥2 prolonged febrile seizures by 1 year of age.