Self-driving laboratories (SDLs) speed up scientific discovery by automating experiment execution and optimization. Recent AI agents can propose new experiments, yet existing SDLs are limited to the procedures they were originally built for; adapting them to new experiments still requires extensive manual engineering, making laboratory flexibility a scaling bottleneck.
This paper introduces the Self-Building Laboratory (SBL) concept: given a high‑level scientific goal, the system automatically designs, configures, and operates its own lab setup. Starting from any existing assets—an empty room or a functional SDL—SBL builds missing capabilities such as lab layout, a digital twin, robot skills, and experimental workflows.
We demonstrate SBL’s scalability across different robots and configurations, scaling from a single workstation to laboratories where several robots collaborate. An end‑to‑end validation is performed with a real‑world quantum‑dot synthesis experiment.
The key contribution of SBL is shifting the researcher’s role from manually configuring and programming device automation to merely specifying scientific objectives, thereby making scientific automation more accessible, flexible, and scalable.
Review: SBL offers a practical pathway for rapid lab reconfiguration, potentially lowering the barrier to automated experimentation and accelerating interdisciplinary research.