We merge experimental data on competitiveness of a large sample of students with their complete educational history for up to ten years after the initial assessment. Exploiting quasi-random class assignments and controlling for other non-cognitive peer characteristics, we find that having competitive peers as classmates makes students choose and secure positions in higher-paying occupations. These occupations are also more challenging and—among male students—more popular. On the cost side, competitive peers do not lead to a lower probability of graduating from the subsequent job-specific education, but they significantly increase the probability of requiring extra time to do so.