John Deere See & Spray: independent university trials cut post-emergence herbicide 43% to 62%

The problem

Broadcast spraying applies post-emergence herbicide across an entire field regardless of where weeds actually are. That wastes product, raises grower cost, and, with herbicide- resistant weeds like Palmer amaranth spreading, keeps chemical pressure high across ground that has no weeds on it at all.

What was built

John Deere’s See & Spray is a computer-vision targeted-spraying system: cameras and on-board processors on the sprayer boom detect green weeds against soil and residue in real time and fire only the nozzles over a weed, instead of broadcasting herbicide across the whole field (source). The version measured in the Iowa State study is the See & Spray Ultimate (source).

What makes this story checkable is who measured it: not John Deere’s marketing, but independent university weed-science programs, whose peer-reviewed and extension results are quoted below (source).

The outcome

Every figure below is quoted verbatim from an independent source, re-fetched on 2026-08-13; each number is attributed to its own source and none are merged.

• Peer-reviewed savings: 28.4% to 62.4%. The Weed Technology 2024 paper by Avent, Norsworthy et al., from experiments run two years at Keiser, AR and Greenville, MS, reports that “targeted applications provided comparable control to traditional broadcast applications across all weeds, resulting in a 28.4% to 62.4% POST herbicide savings” (source).
• University of Arkansas three-year trial: 43% to 59% reduction. The March 2025 Division of Agriculture release reports a “43 to 59 percent reduction in herbicide use compared to broadcast applications” (source). It adds a modeled Arkansas grower saving of 13% to 80%, or $1.6 million to $48 million per year on post-emergence herbicide expenditures, presented here as a projection rather than a measured outcome (source).
• Iowa State field-scale study: 76% average product savings. Across five soybean fields totaling 415 acres, “average product savings across all fields was 76%”, with 4,700 gallons of tank mix and about $6,500 ($15.7/acre) of herbicide product saved, per-field savings ranging 90.6%, 87.6%, 87.2%, 71.2% and 43.9% (source).

The honest tradeoff (shown, not merged away)

The same University of Arkansas trial found a real, setting-dependent risk. At a low sensitivity setting, where only larger weeds get sprayed to maximize short-term savings, missed small Palmer amaranth let the pigweed population increase 280% each year, which the researchers warn “could result in accelerated herbicide resistance or additional in-season applications, costing Arkansas producers upwards of $60 million per year” (source). Lead researcher Jason Norsworthy, Distinguished Professor of weed science, is on record: “All of my data would indicate that using the high-sensitivity setting is best in most scenarios”, at which control is comparable to broadcast and herbicide savings stay substantial (source).

How this was verified

Independently validated by TIN against the public record on 2026-08-13. The headline rests on independent academic measurement rather than a vendor attestation: a peer-reviewed Weed Technology 2024 paper plus two independent university field studies. All three primary figures were re-fetched verbatim from the live pages this session and are shown separately, never merged. The honest limit: the Arkansas and Iowa State releases are university extension research communications reporting their own trials, which is Tier 1 for who states the number; the herbicide-cost-savings dollar ranges are modeled projections and are labeled as such in the prose above, not carried as measured results.

Related case files

The value of this story is entirely about who measured the number, which is clearest against its opposite. In the SEC’s AI-washing case against Delphia and Global Predictions, regulators fined two advisers for AI capability claims they could not substantiate; here the claim survives precisely because independent universities, not the vendor, ran the trial. The same contrast runs through the SEC’s action against Presto Automation, where a voice-AI drive-thru vendor overstated how much its system did without human agents. For a quantified automation outcome measured in production rather than in marketing, see Omega Healthcare’s UiPath document automation.

Sources

Tier legend: Tier 1 primary (the party stating its own metric on record, an audited artifact, regulatory data, or a court document); Tier 2 strong secondary (independent reputable press naming the parties); Tier 3 weak (vendor marketing or a data aggregator). Figures re-checked live on 2026-08-13.

1. Weed Technology (Cambridge University Press), “Comparing herbicide application methods with See & Spray technology in soybean,” 2024 (Tier 1, peer-reviewed study by Avent, Norsworthy et al.; the 28.4% to 62.4% figure and the Keiser, AR / Greenville, MS trial locations). https://www.cambridge.org/core/journals/weed-technology/article/comparing-herbicide-application-methods-with-see-spray-technology-in-soybean/05967C11D1BF0B174A13BE687B865AF2
2. Tristen Avent, “Utility of John Deere’s See & Spray Ultimate in Southern U.S. Row Crop” (M.S. thesis, University of Arkansas), ScholarWorks (Tier 1, primary thesis carrying the verbatim 28.4% to 62.4% savings sentence). https://scholarworks.uark.edu/etd/5771/
3. University of Arkansas System Division of Agriculture (AAES), “Precision Agriculture Research Measures Effectiveness of See & Spray Technology,” March 2025 (Tier 1, university research release; the 43% to 59% reduction, the 13% to 80% / $1.6M to $48M modeled savings, the 280% pigweed increase, and the $60M resistance-risk figure). https://aaes.uada.edu/news/see-and-spray-research/
4. University of Arkansas Cooperative Extension (UAEX), same release, March 2025 (Tier 1, second official host; carries the Norsworthy high-sensitivity quote). https://www.uaex.uada.edu/media-resources/news/2025/march/03-10-2025-see-and-spray-research.aspx
5. Iowa State University Integrated Crop Management, “Precision Spraying Technology,” August 2024, by Doug Houser, Matt Darr and Ryan Huffman (Tier 1, university field study; 415 acres, 76% average savings, 4,700 gallons and $6,500 / $15.7 per acre, and the per-field breakdown). https://crops.extension.iastate.edu/cropnews/2024/08/precision-spraying-technology
6. Weed Science Society of America, “Machine-Vision Technology Shows Promise to Reduce Herbicide Use,” December 2024 (Tier 2, society press release relaying the peer-reviewed study). https://wssa.net/2024/12/machine-vision-technology-shows-promise-to-reduce-herbicide-use/