How to build an effective self-driving laboratory

Self-driving laboratories can accelerate materials discovery by achieving scientific objectives hundreds of times faster than automation alone. However, a self-driving laboratory can take years to adapt to a new study. We contend that effective self-driving laboratories must (1) work faster than automation alone, and (2) be readily adapted to new studies. Here, we examine the characteristics of effective self-driving laboratories and highlight the challenges of reusing these laboratories for new studies. We then provide strategies for building effective self-driving laboratories that accelerate the discovery of new materials. Graphical abstract