The Department of Energy´s Lawrence Berkeley National Laboratory is driving a shift in how research is conducted by combining Artificial Intelligence, robotics and powerful data systems. By integrating these tools across disciplines from energy and materials to particle physics, Berkeley Lab is shortening development cycles and increasing experimental throughput. The article frames this work as strengthening the U.S. scientific enterprise by pioneering AI-enabled discovery platforms and sharing them across the research community to accelerate innovation.
One major focus is automating materials discovery. At the automated materials facility called A-Lab, algorithms propose candidate compounds while robots prepare and test them, creating a tight loop that reduces the time needed to validate materials for batteries and electronics. Exploratory robotic systems such as Autobot at the Molecular Foundry extend that automation to flexible lab workflows, enabling faster investigation of materials relevant to energy and quantum applications. These automated pipelines are presented as a core way Berkeley Lab scales experimental work.
Artificial Intelligence is also embedded in instrumentation and data infrastructure. Machine learning models optimize and stabilize beams at the Berkeley Lab Laser Accelerator, and deep-learning controls are being applied to the Advanced Light Source and its ALS-U upgrade to improve synchrotron performance. On the data side, a web-based platform called Distiller streams microscopy data to the Perlmutter supercomputer at NERSC for near-instant analysis, allowing scientists to refine experiments in real time. Berkeley Lab researchers are also using machine learning at NERSC to predict particle behavior in fusion plasmas and deploying AI at ESnet to predict and troubleshoot network traffic for high-speed collaboration.
Beyond optimization and throughput, Artificial Intelligence is treated as a collaborator. The lab uses its light sources to validate AI-generated designs, such as a recently reported AI-designed enzyme, bridging prediction and experiment. The piece concludes that by embedding AI across robotics, instrumentation and computing, Berkeley Lab and its DOE Office of Science user facilities are building a smarter scientific infrastructure designed to solve complex, large-scale problems more quickly and efficiently.
