Agentic AI Boosts Hunt for Mineral‑Harvesting Crops, Aiming to Cut U.S. Dependence on Imports

The United States is turning to advanced artificial intelligence to speed up the discovery of plant species capable of extracting critical minerals from soil, a move that could lessen the nation's reliance on foreign sources for the metals that power smartphones, electric‑vehicle batteries and semiconductor chips.

Researchers are deploying a new class of “agentic” AI systems that can autonomously design, run, and analyze experiments on candidate hyper‑accumulator plants. Unlike traditional machine‑learning tools that simply predict outcomes, these agents can formulate hypotheses, select test conditions, and iteratively refine their approach without direct human instruction, dramatically shortening the research cycle.

The effort builds on a growing body of work showing that certain plants naturally concentrate elements such as lithium, cobalt, nickel and rare‑earth metals in their tissues. By identifying and optimizing such species, scientists hope to create a low‑impact, renewable source of these strategic materials, potentially harvesting them from agricultural lands or marginal soils where conventional mining would be uneconomic.

Early trials have already yielded promising leads. The AI agents screened thousands of genetic variants and environmental parameters, pinpointing a handful of grasses and legumes that consistently amassed higher concentrations of target minerals than previously recorded. Laboratory analyses confirmed that the plants’ root systems were pulling the metals from the surrounding substrate, a process that could be scaled up with appropriate agronomic practices.

Policy makers view the development as a strategic hedge against supply chain disruptions that have plagued the tech sector for years. By diversifying the sources of critical inputs, the United States aims to bolster national security and reduce exposure to geopolitical volatility in mineral‑rich regions abroad.

Funding for the project comes from a blend of federal research grants and private‑sector partnerships, reflecting broad interest in sustainable mineral sourcing. The next phase will involve field tests on larger plots, assessment of extraction efficiency, and evaluation of economic viability. If the approach proves cost‑effective, it could usher in a new era where farms double as mineral mines, blending food production with resource recovery in a circular economy model.