The results of the largest-ever field trial of AI-guided precision agriculture were published this week by the International Food Policy Research Institute, in partnership with John Deere, BASF, and the UN Food and Agriculture Organisation.
The trial covered 14,000 farms across eight countries - the US, Brazil, India, Australia, France, Kenya, Germany, and Argentina - over the 2025 growing season. The technology is conceptually simple: cameras mounted on spraying equipment use computer vision to identify individual weeds in real-time, and the system activates spray nozzles only when a weed is detected, rather than blanketing the entire field.
The headline result: a 60% average reduction in herbicide use, with no statistically significant reduction in crop yield compared to conventionally managed control plots on the same farms.
The reduction varied by crop and region. Soybean and corn fields in the US Midwest saw 55-65% reductions. Wheat fields in France and Germany saw 50-58%. Cotton fields in India saw the highest reductions at 62-70%, partly because weed pressure in tropical climates is higher and the AI system's advantage over broadcast spraying is greater.
"This is not a laboratory result," said Dr. Robert Chen, director of the IFPRI's agricultural technology programme. "This is 14,000 farms, across eight countries, over a full growing season, with control plots on every farm. The methodology is about as rigorous as field agriculture allows."
The economic impact is direct. Herbicide costs represent 15-25% of variable costs for most row-crop farmers. A 60% reduction translates to $40-80 per hectare in savings for US corn and soybean farmers. For smallholder farmers in India and Kenya, where pesticide costs consume a larger share of thin margins, the savings are proportionally more significant.
The environmental implications are substantial. Herbicide runoff is a major contributor to water pollution, algal blooms, and soil microbiome degradation. A 60% reduction in application means 60% less chemical entering waterways and soil. The system also reduces farmer exposure to herbicides - a significant occupational health issue, particularly in developing countries where protective equipment is often unavailable.
The technology is not new - John Deere's See & Spray system has been available since 2022 - but adoption has been limited by cost and scepticism. This trial is the first rigorous, large-scale, independent validation of the approach across diverse geographies and farming systems.
John Deere has announced a financing programme that allows farmers to acquire the system with payments tied to pesticide savings, eliminating the upfront cost barrier. BASF, which sells the herbicides that the system reduces the need for, has pivoted to framing the technology as enabling "smarter application" rather than less application - a positioning that acknowledges the trajectory while protecting revenue.
The UN FAO is recommending the technology for inclusion in national agricultural extension programmes in Sub-Saharan Africa and South Asia.
What we know for certain
A field trial across 14,000 farms in 8 countries found AI-guided precision spraying reduced herbicide use by 60% without reducing crop yields. John Deere is offering savings-linked financing. The UN FAO is recommending the technology for developing countries.
What we are inferring
Precision spraying will become standard practice within 5-10 years as costs decrease and independent validation builds confidence. The environmental benefits - reduced water pollution, soil health preservation - may prove more significant than the economic savings.
What we couldn't verify
Long-term effects on weed resistance patterns under precision spraying regimes. Whether the technology performs equally well across all crop types and farming scales. The financing programme's terms were not independently audited.