AI system helps design plants that extract critical minerals from soil
Researchers at Oak Ridge National Laboratory have developed an agentic AI platform that uses the Frontier supercomputer to accelerate phytomining research. The system trains and fine-tunes models to guide experiments on plants and microbes that hyperaccumulate critical minerals from the soil. This approach could help reduce U.S. dependence on foreign sources for essential minerals.
The OPAL co-scientist platform pairs the Frontier supercomputer with automated laboratory systems across four national laboratories. In one demonstration, researchers studying nickel absorption in pennycress plants compressed analysis of over 1,000 physical traits from hundreds of manual hours to minutes of AI interaction. Twelve pennycress lines from different global regions were examined.
Phytomining targets marginal lands unsuitable for conventional agriculture or mining, including rehabilitation zones. The approach addresses U.S. supply chain vulnerabilities for minerals essential to electronics and energy technology, offering a low-cost alternative to traditional extraction.
This technology could reshape mineral supply chains by enabling domestic production from previously unusable land. Rural communities may gain new economic opportunities, while industries dependent on imported critical minerals could see reduced vulnerability. However, widespread deployment depends on scaling from laboratory experiments to field conditions, and the timeline for commercial viability remains uncertain. The approach's environmental footprint and economic feasibility will ultimately determine its real-world adoption.