From High-Efficiency Devices to Reproducible Manufacturing: Self-driving Laboratories Are Reshaping Perovskite Photovoltaics

Abstract As perovskite solar cells (PSCs) enter the 28% efficiency era, the field must move beyond champion-device metrics toward reproducible and manufacturable performance. Self-driving laboratories offer a route for this transition by integrating automated fabrication, high-throughput characterization, machine learning-guided decision-making, and closed-loop optimization. This Viewpoint discusses how Autonomous Materials and Devices Acceleration Platforms (AMADAPs) can shift perovskite research from human-centered empirical discovery to data-driven global optimization. The need for digital sample passports, cross-platform validation, manufacturing-aware learning systems, digital twins, and multi-objective optimization across efficiency, stability, reproducibility, cost, process tolerance, and environmental impact is highlighted. Such autonomous infrastructures may define the next phase of perovskite photovoltaics by linking molecular design, process control, device fabrication, characterization, and model interpretation into a holistic learning framework.