DOE-Backed AI Agents Accelerate Critical Mineral Recovery from Batteries
Researchers at SLAC National Accelerator Laboratory are using AI to bypass thousands of manual experiments to achieve high-purity battery recycling.
Researchers at the Stanford University-led SLAC National Accelerator Laboratory are deploying a team of AI agents to optimize the recovery of critical metals from used lithium-ion batteries. The project, part of the Department of Energy's (DOE) Genesis Mission, aims to streamline the extraction of nickel and cobalt to secure domestic mineral supplies.
The AI-driven workflow addresses the technical challenge of separating metals at high purity levels, specifically utilizing solvent extraction methods. Rather than relying on traditional trial-and-error, the AI agents perform specialized roles: they scan vast amounts of scientific literature, access datasets such as the American Science Cloud, and develop hypotheses to identify the most promising experimental paths. This creates a tight feedback loop where human researchers conduct a limited number of physical experiments and feed the results back to the AI to further refine the process.
The Genesis Mission Framework
This initiative is a cornerstone of the broader Genesis Mission, a DOE effort designed to build an integrated scientific platform. By connecting supercomputers, experimental facilities, and specialized datasets with AI systems, the mission seeks to address designated national science and technology challenges. The goal is to transition away from legacy research methods that often require thousands of manual iterations, moving instead toward an autonomous research model applicable to chemistry, physics, fusion energy, and energy materials.
Implications for Scientific Discovery
Beyond the immediate goal of battery recycling, the project serves as a proof-of-concept for AI as a collaborative research partner. By drastically reducing the time and cost associated with physical experimentation, this model can be scaled to other complex scientific problems involving massive datasets. Steve Eglash, director of SLAC’s Applied Energy Division, noted that the ability of AI agents to look broadly across different disciplines, including biology and other physical sciences, is a particularly powerful capability.
Strengthening the Supply Chain
Improving the efficiency of battery recycling has direct economic and environmental consequences. By enhancing the recovery of nickel and cobalt, the U.S. can strengthen its domestic supply of critical minerals and reduce reliance on external sources. Furthermore, the process prevents hazardous battery materials from entering landfills, aligning environmental protection with industrial security.
The Path Forward
As the project progresses, the focus remains on validating the AI's hypotheses through targeted physical testing. Ahamed Irshad Maniyanganam, SLAC associate scientist and lead researcher, emphasized the shift in methodology, stating that researchers will now perform far fewer experiments because the AI agent handles the initial heavy lifting of exploration. The success of this loop will determine how quickly the DOE can deploy similar AI-enabled workflows across other critical energy sectors.