NSF awards $19.5M for AI-driven autonomous chemistry lab

NSF funds cloud-accessible chemistry laboratory

The National Science Foundation has awarded $19.5 million over four years to establish a remotely accessible, fully automated chemistry laboratory that researchers across the United States can use to design and run experiments without being physically present. The facility, called the Chemistry Node, will be housed at Scripps Research's Automated Synthesis Facility in La Jolla, California.

Scripps Research is leading the project in collaboration with UCLA and Sunthetics, an AI-for-chemistry company. The Chemistry Node is one of 20 facilities in the NSF Test Bed: Toward a Network of Programmable Cloud Laboratory initiative, which aims to demonstrate how automation and artificial intelligence can accelerate scientific discovery across multiple disciplines.

Brandon Orzolek, scientific director of the Automated Synthesis Facility at Scripps Research and lead principal investigator, said the project treats synthetic chemistry as a data science problem. The system will translate researcher ideas into executable experiments, run them on robotic infrastructure, analyze results with analytical instruments in real time, and use AI to propose follow-up experiments.

Why it matters

Chemistry research has historically required hands-on laboratory work, limiting access to institutions with expensive equipment and specialized staff. A cloud-based system that accepts experiment proposals from anywhere and returns results could democratize advanced research capabilities, particularly for smaller universities and colleges. The facility also addresses a persistent challenge in chemistry: reproducibility. By standardizing experimental conditions and capturing comprehensive data, the node could help establish which reactions transfer reliably across different laboratory settings.

Tackling complex molecular synthesis

The Chemistry Node will focus initially on optimizing processes for creating organic small molecules—compounds used in pharmaceuticals, agricultural chemicals, and materials science. The facility will specifically target reactions involving multiple catalytic cycles, where changing one variable affects several interconnected processes. Manual optimization of these reactions is time-intensive; automation can test combinations of variables far more efficiently.

Keary Engle, professor and John and Susan Diekman Dean of Graduate & Postdoctoral Studies at Scripps Research, is a co-principal investigator leading catalytic experimental design. His laboratory will be among the first to propose experiments for the system. Other co-principal investigators include Abigail Doyle, a UCLA professor directing data-rich experimental design and quality control, and Daniela Blanco, CEO of Sunthetics, who will lead AI integration and web interface development.

Educational access and workforce development

Beyond serving as a discovery engine, the Chemistry Node will provide training opportunities for students at institutions that lack advanced research infrastructure, including R2 universities, primarily undergraduate institutions, and two-year colleges. Engle noted the facility will help educate chemists who combine expertise in synthetic chemistry, catalysis, AI, and automation—a skill set he believes will be essential for the field's future.

Orzolek said the project progresses through multiple phases over four years, with increasing accessibility to the broader research community. By the final year, the team expects to operate a closed-loop system capable of taking an idea from a remote user and producing an end product, compressing work that would otherwise require years.

The details were first reported by Scripps Research.