TL;DR: On October 8, 2026, the U.S. Department of Energy selected a SLAC-led collaboration with Caltech and Lawrence Berkeley National Laboratory to test a closed-loop AI system that chooses the next catalyst experiment at SLAC's Stanford Synchrotron Radiation Lightsource.
DOE Adds Catalyst Autonomy Test To Genesis Mission
On October 8, the U.S. Department of Energy (DOE) announced a second round of funding for its Genesis Mission, a nationwide initiative to use AI to tackle pressing challenges in science and technology, according to Caltech's October 8 announcement.
Among the projects that will receive funding is a Caltech-involved effort that will use AI to help identify new and better catalysts, materials that speed up chemical reactions.
The catalysis project is led by the DOE's SLAC National Accelerator Laboratory in collaboration with Caltech and Lawrence Berkeley National Laboratory.
It is one of six newly announced Phase I projects that will receive short-term seed funding to explore feasibility.
The DOE also announced 12 new Phase II funding awards for larger projects that are ready to scale up, Caltech reported.
The broader DOE announcement said the 12 Phase II Genesis Mission project awards total $159 million, according to the Department of Energy's October 8 announcement.
That DOE announcement said the 12 newly selected Phase II and six Phase I projects advance the frontiers of Super Intelligence (SI) for science, capping a first-year cohort of 297 projects across the Genesis Mission portfolio.
Building on the Genesis Mission's growing portfolio of projects, first announced in July, these Phase II project awards will unleash scientific discoveries with Super Intelligence (SI), the DOE said.
What The Closed-Loop Project Will Build
The catalysis project is titled "Closed-Loop Autonomous Discovery of Mechanistic Activity-Selectivity-Stability Rules in Catalysis."
It will build a closed-loop system in which an AI agent will compare competing ideas about how catalysts work, identify what evidence is missing to validate or disprove those ideas, and then choose the next experiments that should be conducted.
At the Stanford Synchrotron Radiation Lightsource (SSRL), the system will combine real-time X-ray measurements with electrochemistry, computer models, and published knowledge.
Each result will feed back into the system and guide the next measurement.
Phase I of the project will test whether this approach can enable scientific understanding with fewer experiments than standard approaches, Caltech reported.
The Genesis-funded project aims to use AI and automated experimentation to speed up the process, enabling the rapid testing and identification of the best catalysts for particular chemical reactions, including the efficient and economically viable creation of liquid fuels.
Faster insight into how catalysts work can help researchers develop more efficient and durable processes for energy and chemical production while advancing the broader vision of autonomous laboratories that can help scientists test ideas on faster timescales.
The SLAC-led project will connect SI reasoning directly with experiments to speed the scientific understanding of catalytic processes important for energy and chemical production, according to SLAC's October 8 announcement.
How AI Will Select The Next X-Ray Measurement
At SSRL, a DOE Office of Science user facility, the system will combine real-time X-ray measurements with electrochemistry, computer models and published knowledge.
The AI weighs the scientific evidence, decides which measurement would tell researchers the most, and that decision becomes a real experiment at SSRL.
Dimosthenis Sokaras, the project's principal investigator and director of the Chemistry and Catalysis Division at SSRL, described that loop in a SLAC press release.
"The AI weighs the scientific evidence, decides which measurement would tell us the most, and that decision becomes a real experiment at SSRL," said Sokaras.
He added: "The result comes straight back into the reasoning process. That closed loop is a key step toward autonomous laboratories."
SLAC said the SLAC-led project will build a closed-loop system in which an SI agent compares competing ideas about how catalysts work, identifies what evidence is missing and chooses the next experiment that can best test them.
SLAC also leads the DOE Basic Energy Sciences ISAAC AI Pathfinder, a Genesis Mission project that brings together measurements from DOE user facilities across the country, theoretical calculations run on DOE supercomputers, and the scientific literature so SI can weigh evidence and rank competing scientific hypotheses.
SLAC said it has a long-standing catalysis program that combines theory, catalyst testing and X-ray measurements of working catalysts.
SLAC also said it has deep expertise in machine learning and artificial intelligence for autonomous diagnostics, control and high-speed data processing for critical systems.
Why Catalyst Libraries Are Central To The Test
To produce useful products from chemical reactions, scientists are always searching for the most effective catalysts, Caltech reported.
The Liquid Sunlight Alliance (LiSA), a DOE-funded Energy Innovation Hub led by Caltech's Harry Atwater, the Otis Booth Leadership Chair of the Division of Engineering and Applied Science, is specifically looking for catalysts that can drive reactions to convert water, carbon dioxide, and nitrogen into liquid fuels and valuable chemicals.
Atwater, who is also the Howard Hughes Professor of Applied Physics and Materials Science, said the Genesis mission project brings together laboratory-based accelerated materials discovery with X-ray beamline measurements to speed the pace of knowledge generation about how catalysts enable pathways for chemical fuel synthesis.
To fully understand how certain catalysts work and whether they can be improved, scientists with LiSA have used X-ray beamlines at SLAC's Stanford Synchrotron Radiation Lightsource to measure changes in the physical structures of catalysts as they become active and spur on chemical reactions.
That work is extremely time-consuming, Caltech reported.
Although millions of possible catalyst compositions exist — even miniscule changes in the chemical composition can dramatically affect how a catalyst works — only a handful have been tested in this way.
As part of the new project, Caltech researchers will prepare libraries of hundreds to thousands of catalyst compositions to test with SLAC's X-ray beam, said Caltech's Joel Haber, a member of the professional staff and team lead with LiSA.
"We can measure how well the material is performing as an electrocatalyst at the same time we're measuring the structure of the catalyst," Haber said.
He added that researchers will be able to see what changes when it turns on and becomes effective as an electrocatalyst and then correlate the structure with that performance.
That data will be combined with electrochemical measurements from labs at Caltech along with information from computational chemists in the collaboration to constantly update the AI model.
Who Leads The Collaboration
The project's principal investigator is Dimosthenis Sokaras, director of the Chemistry and Catalysis Division at SLAC's Stanford Synchrotron Radiation Lightsource.
The project is a collaboration with the California Institute of Technology led by Harry Atwater, professor of engineering and applied science, and Lawrence Berkeley National Laboratory led by Junko Yano, division director of Molecular Biophysics and Integrated Bioimaging, SLAC reported.
The collaboration includes researchers at the DOE-funded Liquid Sunlight Alliance, a multi-institution consortium that includes SLAC and is led by Atwater.
It also includes the SUNCAT Center for Interface Science and Catalysis, a joint venture between SLAC and the School of Engineering at Stanford University.
Caltech's Haber is a member of the professional staff and team lead with LiSA.
Atwater is the Otis Booth Leadership Chair of the Division of Engineering and Applied Science at Caltech and the Howard Hughes Professor of Applied Physics and Materials Science.
The article about the award was written by Kimm Fesenmaier, with contact listed as Kimm Fesenmaier at (626) 395-1217.
Caltech's address was listed as 1200 East California Boulevard in Pasadena, California 91125.
How Phase I Seed Awards Differ From Phase II Scale-Ups
The SLAC-led catalysis project is a Phase I award from the Genesis Mission: Transforming Science and Energy with AI Request for Application (RFA), SLAC reported.
Of eight new partner awards involving SLAC, one is a Phase I RFA award and seven are Phase II RFA awards.
The goal of the Phase I RFA awards is to identify promising pathways toward transformative scientific capabilities and establish a foundation for future investment and scale, SLAC said.
Project teams will design and demonstrate research workflows that integrate SI with scientific investigation, while rigorously evaluating whether those approaches can accelerate discovery, improve predictive capabilities, enhance experimentation or generate new scientific insights.
In addition to leading the catalysis project, SLAC will partner in eight newly announced projects led by national labs as well as universities and companies in areas such as fusion energy, biosciences, quantum materials and more.
The SLAC-led project will explore the use of Super Intelligence to accelerate understanding of a process central to energy and chemical production.
SLAC Laboratory Director John Sarrao said SLAC is proud to lend leadership and expertise across new projects that illustrate DOE's commitment to bringing together laboratories, universities and the private sector to address critical energy challenges.
"Together, these nine new projects will leverage SI combined with SLAC expertise to accelerate our scientific and technological impact," Sarrao said.
What The $159 Million Phase II Cohort Will Pursue
Through DOE's Office of Science, the 12 newly awarded projects address fusion, biology, particle physics, geothermal energy, separations, enzymes, microelectronics, accelerators, quantum computing, supercomputing software, quantum magnets and isotope operations.
Led by Commonwealth Fusion Systems, one project will build a digital twin to simulate and optimize operations of a fusion demonstration device, focused on accelerating the path to commercial fusion energy via an AI-enabled digital twin platform for SPARC.
Led by the University of California San Diego, the project on decoding the RNA Structurome will expand the RNA structure database five-fold for training SI models to unlock the ability to engineer resilient crops and sustainable microbes.
Led by MIT, the Lattice Quantum Chromodynamics project will integrate Super Intelligence with supercomputing calculations of fundamental particle physics to dramatically accelerate discoveries about the basic building blocks of the universe.
Led by the University of California Irvine, the Multi-agent AI Expert for Subsurface Reasoning and Optimization, known as MAESTRO, will establish an SI platform to accurately map, stimulate, and manage deep underground reservoirs to unleash vast resources of reliable geothermal energy.
Led by the University of Illinois Urbana-Champaign, the AI-extended Interfacial Separations project, known as AXIS, will harness SI to design more efficient electrochemical processes for extracting critical rare-earth elements from domestic sources.
Led by the University of Washington Seattle, one project will build foundational SI tools to predict and design functional enzymes by overcoming barriers in computational enzyme design through improved sequence-structure ensemble modeling.
Microchips, Accelerators, Quantum Systems And Software
Led by Fermilab, the project on Accelerating Extreme Environment Specs-to-Silicon, known as AXESS, will use SI to rapidly design rugged, high-performance microchips capable of operating in extreme conditions such as space, high-radiation, and ultra-cold environments.
Led by Lawrence Berkeley National Laboratory, the Multi-office Accelerator Team Core, known as MOAT-Core, will develop a shared SI platform across National Laboratories to autonomously run and design particle accelerators faster and more efficiently.
Led by Harvard University, the project on Application-Aware Error Correcting Codesign for Scientific Quantum Computing, known as ASQC, combines SI with advanced quantum hardware to tackle quantum computing error correction.
Led by Argonne National Laboratory, the Iterative Framework of AI-Assisted Scientific Development and Optimization, known as AI4HPC, will create an SI framework to modernize, optimize, and verify supercomputing scientific software.
Led by Oak Ridge National Laboratory, the project on AI-empowered Design of Functional Quantum Magnets will develop a physics-informed SI framework to design new quantum magnets by working backward from desired properties, with applications in quantum sensing, low-power electronics, and quantum technologies.
Led by Northeastern University, the project on Agentic Assistance for Advanced Rare Isotope Separator Operations at the Facility for Rare Isotope Beams will develop an intelligent, automated assistant to safely streamline the operation and tuning of the FRIB particle separator, unlocking hundreds of hours of research time annually.
These new selections bring the total number of Phase II projects to 14, alongside previously announced awards for GridFM 2.0 under the Office of Electricity and Prometheus under the Office of Nuclear Energy.
DOE is also issuing six new Phase I awards, expanding the Genesis Mission's initial round of selections.
How Genesis Mission Organizes Data, Compute And Partners
The Genesis Mission is a national initiative to mobilize government, industry, academia, nonprofits and international partners to advance SI for science and technology, SLAC reported.
Using the DOE-built American Science and Security Platform, partners work together on shared infrastructure that connects researchers with data, compute and SI tools to accelerate scientific discovery.
Partners are focusing on the most critical challenges facing the nation in energy, national security, scientific discovery, health, space and more.
The Phase II awards are developing new, powerful capabilities to tackle the Nation's most complex National Science and Technology Challenges, the DOE said.
By connecting America's leading scientists with advanced SI, supercomputing, scientific data, and DOE's world-class research infrastructure, the Genesis Mission is strengthening U.S. leadership at the frontiers of science and technology, the DOE said.
President Trump launched the Genesis Mission in November 2025 to accelerate discovery and strengthen U.S. leadership in science, medicine, and technology.
DOE Under Secretary for Science Dr. Darío Gil said the projects represent a critical step in turning the promise of Super Intelligence for science into transformative scientific capability.
"The Phase II teams have already demonstrated what is possible when SI is integrated with scientific expertise, data, and computing," Gil said. "These awards give them the opportunity to take that work to the next level."
Gil added: "By bringing together our National Labs, universities, and industry partners, we are building on extraordinary work already underway and accelerating the pace of discovery."
Caltech's Earlier Genesis And Cloud-Lab Roles
Two Caltech-led Genesis Mission projects were announced in the first round of DOE funding for the initiative in July, Caltech reported.
At the same time, a programmable cloud laboratory led by Caltech and funded by the National Science Foundation (NSF) was announced as part of a network of AI-enabled automated laboratories that represent a core NSF contribution to the nationwide Genesis Mission.
SLAC's premier facilities produce datasets that illuminate the world through research in physics, biology, energy and other fields, making discoveries from the grand scale of the cosmos to the smallest and fastest scales of the motions of electrons, atoms and molecules, SLAC said.
More information about the newly awarded projects and previously awarded projects is available through the DOE announcement, which listed media inquiries at (202) 586-4940 or [email protected].