MIT Lands Up to $52.9M ARPA-H Award to Design Safer Herbicides With AI

MIT will lead a project backed by up to $52.9 million from ARPA-H to design safer herbicides with generative AI, the institute announced on November 10, 2026. The award, listed by the U.S. Advanced Research Projects Agency for Health with a September 18, 2026 start date, makes MIT the prime awardee.

Christopher Voigt, head of MIT's Department of Biological Engineering, serves as principal investigator. The project, called Precision Learning for On-target Weed control (PLOW), sits inside ARPA-H's STREAM program — Systems for Tracking and Resilience in Efficient Agricultural-input Management.

Why does weed control need new chemistry?

U.S. agriculture applies more than 1 billion pounds of herbicide each year to prevent significant crop losses and enable regenerative practices that protect soil, MIT's announcement states. Repeated use has allowed some weeds to acquire resistance mutations.

Glyphosate shows the strain. It kills weeds by interfering with EPSP synthase, an enzyme plants need to produce several critical amino acids, and resistance makes that mechanism less effective.

Health adds a second design constraint. Herbicide exposure may contribute to increased risks of cancer, metabolic and endocrine disorders and other adverse outcomes, according to MIT's announcement. Farm workers face exposure through frequent use, while many U.S. residents encounter small amounts through food and drinking water. PLOW will weigh human and ecological harm alongside weed-killing performance when selecting candidates.

Supply security matters too. The alternatives are intended to reduce U.S. dependence on imported phosphorus, a critical ingredient in glyphosate. Voigt said resistance plus insecure precursor supply chains could leave the country without agrochemicals that are "absolutely essential for getting the high crop yields and economical prices that the United States now sees."

How does the AI pipeline work?

PLOW's first research group will build models that propose chemical compounds or small interfering RNA (siRNA) molecules to disrupt the plant enzymes targeted by glyphosate and other herbicides. Connor Coley and Regina Barzilay of MIT and Alexander Tropsha of the University of North Carolina lead that work.

The models will optimize candidates for:

• low synthesis cost
• low risk of inducing resistance
• low risk of harming humans or causing ecological damage
• effectiveness against weeds

The team expects billions of possible candidates. That scale puts the next stage at the center: Angela Koehler of MIT and Jaime Cheah of the Broad Institute will lead high-throughput screening to determine which molecules hit their targets. Screening data flows back to the computational group to improve the models, an iterative design-and-test loop rather than AI output treated as a finished herbicide.

After the pool narrows, a third group evaluates each molecule under different agricultural conditions in greenhouse studies. Neal Stewart and Scott Lenaghan at the University of Tennessee at Knoxville, Patrick Tranel at the University of Illinois at Urbana-Champaign, Tim Willson at UNC and Robigo founder and CEO Andee Wallace lead that work. The first greenhouse tests of initial RNA interference-based candidates should be complete by the end of the program's first year.

What happens to chemicals already in use?

PLOW also builds living microbial systems in which engineered bacteria and fungi convert agrochemicals and formulation ingredients into biomass before those substances reach food, soil or waterways. The approach aims to scale down environmental and human exposure.

The discovery and degradation work covers two of STREAM's four technical areas: AI-assisted discovery of next-generation herbicides, and removal or degradation of agricultural chemicals. The other areas cover precision and nonchemical weed control, including laser ablation and miniature autonomous weeders, plus improved monitoring of herbicides and associated chemicals. The U.S. Department of Health and Human Services launched STREAM on August 17, 2026, framing it as protection for farm workers, consumers and the environment while maintaining crop yields.

The award ties generative AI's value to a demanding experimental test. A billion-candidate library counts only if screening and greenhouse studies find molecules that work in the field while meeting cost, resistance and safety constraints — and PLOW's degradation track targets the exposure problem from existing agrochemicals in the meantime.