What happened
SLAC was awarded a Phase I Genesis Mission grant to build a closed‑loop AI system that designs, executes, and learns from catalyst experiments, and it will collaborate on eight other AI projects spanning accelerator science, biology, quantum materials, microelectronics, fusion, and more.
On Oct. 9, 2026, the Department of Energy announced that SLAC National Accelerator Laboratory will lead an AI‑driven autonomous catalysis discovery project under the Genesis Mission, a federal program that funds AI approaches to national science and technology challenges. The project, titled “Closed‑Loop Autonomous Discovery of Mechanistic Activity‑Selectivity‑Stability Rules in Catalysis,” will integrate real‑time X‑ray measurements, electrochemistry, computational models, and literature data at the Stanford Synchrotron Radiation Lightsource (SSRL). An will evaluate competing hypotheses about catalyst behavior, identify missing evidence, and select the next experiment to run, feeding results back into its reasoning loop.
Dimosthenis Sokaras, director of SLAC’s Chemistry & Catalysis Division, will serve as principal investigator. The effort partners with Caltech’s Harry Atwater and Lawrence Berkeley National Laboratory’s Junko Yano, as well as the DOE‑funded Liquid Sunlight Alliance and the SUNCAT Center at Stanford. The Phase I award will test whether the AI‑guided workflow can achieve comparable scientific understanding with fewer experiments than traditional methods.
In addition to the catalysis effort, SLAC will partner on eight other Genesis Mission projects: one Phase I project on AI‑enabled accelerator intelligence (lead: LBNL) and seven Phase II projects covering computational enzyme design (U Washington), AI‑assisted scientific software development (Argonne), quantum magnet design (ORNL), accelerator‑facility integration (LBNL), extreme‑environment silicon design (Fermilab), AI‑driven digital twins for fusion (Commonwealth Fusion Systems), and RNA structurome decoding for the bioeconomy (UC San Diego).
Why it matters
The initiative puts AI at the core of experimental chemistry, promising faster insight into catalyst mechanisms that could lower energy use and emissions in chemical manufacturing. By demonstrating that an can autonomously choose and run experiments, the project tests a model for future autonomous laboratories, potentially reshaping how scientific discovery is conducted across many fields.
Catalysis underpins the production of fuels, chemicals, and materials; even modest efficiency gains can translate into large reductions in energy consumption and greenhouse‑gas emissions. By automating hypothesis testing, the SLAC project could accelerate the discovery of catalysts that are both more active and more durable, shortening development cycles that currently span years.
The broader Genesis Mission portfolio demonstrates a strategic push by DOE to embed AI across disparate scientific domains. Successful outcomes could validate the American Science and Security Platform as a shared infrastructure for AI‑enabled research, encouraging wider adoption by universities, national labs, and private partners.
If the closed‑loop system proves effective, it may serve as a template for autonomous laboratories in fields ranging from drug discovery to materials science, reducing reliance on trial‑and‑error experimentation and freeing researchers to focus on higher‑level design questions.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
Which component of an AI application is the machine-learning model itself?
What to watch next
Follow the Phase I pilot at SLAC’s Stanford Synchrotron Radiation Lightsource for performance metrics, watch for Phase II funding decisions on the partner projects, and monitor how DOE’s American Science and Security Platform scales AI tool sharing among national labs and industry.
Performance data from the SLAC pilot, such as the reduction in experiment count needed to reach a given confidence level, will be released later in 2026. Those metrics will indicate whether AI can deliver the promised efficiency gains.
DOE’s next round of Genesis Mission funding, expected in early 2027, will likely prioritize projects that demonstrate clear AI advantage. The outcomes of the eight partner projects will influence which research directions receive further investment.
The integration of the AI workflow with DOE’s broader data and ecosystem (the American Science and Security Platform) will be monitored for scalability, security, and openness, especially as industry partners seek to leverage the same tools for commercial R&D.