School underachievement is a persistent problem in the United States. Direct-to-student, computer-delivered growth-mindset interventions have shown promise as a way to improve achievement for students at risk of failing in school; however, these interventions benefit only students who happen to be in classrooms that support growth-mindset beliefs. Here, we tested a teacher-delivered growth-mindset intervention for U.S. adolescents in Grades 6 and 7 that was designed to both impart growth-mindset beliefs and create a supportive classroom environment where those beliefs could flourish (N = 1,996 students, N = 50 teachers). The intervention improved the grades of struggling students in the target class by 0.27 standard deviations, or 2.81 grade percentage points. The effects were largest for students whose teachers endorsed fixed mindsets before the intervention. This large-scale, randomized controlled trial demonstrates that growth-mindset interventions can produce gains when delivered by teachers.