Abstract Research Summary While multi‐project work (MPW) is becoming an increasingly popular work arrangement, its relationship with project performance is understudied. On the one hand, MPW is deployed to increase employee worktime utilization and productivity, which should be reflected in more timely project completion. On the other hand, MPW also brings switching costs due to attention residue and cognitive setup. Based on this trade‐off, we derive an inverted U‐shaped relationship between MPW and project performance. We find support for this relationship in a longitudinal dataset containing 9,649 project‐month‐employee observations. More specialized experience, project similarity, and employee familiarity positively moderate the inverted U‐shape. Furthermore, the results are robust to a host of model specifications, data structures, assumptions, and alternative explanations. Managerial Summary How many projects can you work on simultaneously? We study this question in the context of new product development (NPD) projects in a multinational organization. We suggest that multi‐project work (MPW) might be a double‐edged sword. On the one hand, MPW academics or engineers can be more productive by filling the gaps in their schedules and developing time management practices. On the other hand, MPW also carries switching costs. This trade‐off creates an inverted U‐shaped relationship between MPW and project performance. So, how can MPW be more beneficial or less costly? We find that more specialized employees can benefit more from productivity gains while working with familiar members or similar projects can alleviate switching costs.