Herbicides are the most heavily used pesticides in farming, and governments on both sides of the Atlantic say they want fewer of them. The fastest, cheapest route is to fund the electrical, laser, robotic and bioherbicide tools that already control weeds in the field, and to help farmers choose between them – not to bet too much public money on new molecules that may never reach a farm. Safer new chemistry deserves support, but not at the expense of what works today.
A new conventional crop protection active ingredient now takes an average of 12.3 years and $301 million to discover, develop and register, according to AgbioInvestor’s industry survey. A molecule that enters the pipeline today would reach farmers in the late 2030s, if it survives at all.
Meanwhile, alternatives are already killing weeds and drying crops on commercial farms. Electrical weeding, laser weeding, autonomous mechanical robots and registered bioherbicides are past the proof-of-concept stage. What holds them back is not science. It is capital, regulation, and the lack of incentives for farmers to switch.
The cost of today’s herbicides
Herbicides and plant growth regulators make up about 62% of the conventional pesticide active ingredient applied in U.S. agriculture, according to EPA’s latest market estimates. Several of the most widely used carry well-documented risks.
- Glyphosate, the world’s most used herbicide, was classified as “probably carcinogenic to humans” by the World Health Organization’s cancer agency IARC in 2015, with human evidence pointing to non-Hodgkin lymphoma.
- Paraquat, a desiccant and burndown herbicide, has been linked in multiple studies to Parkinson’s disease. It has been banned in the EU since 2007, and in China, Brazil and Switzerland, but is still sprayed in the U.S.
- Atrazine was withdrawn from the EU market in 2004 over groundwater contamination, yet remains one of the most used herbicides on American corn.
The newest evidence on glyphosate comes from the Global Glyphosate Study, led by Italy’s Ramazzini Institute – the most comprehensive independent toxicology study of the herbicide to date. Its two latest results are:
- Cancer. In a June 2025 paper in Environmental Health, rats exposed from the womb to glyphosate, Roundup Bioflow or RangerPro developed leukaemia in every treated group, and none in the controls. Exposure started at the EU’s acceptable daily intake (0.5 mg/kg body weight per day). About 40% of leukaemia deaths in treated rats came before one year of age. The researchers also found increases in tumours of the skin, liver, thyroid, nervous system, mammary gland, kidney, bladder and other organs, many of them rare in this strain.
- Neurological effects. A three-generation study released as a preprint in August 2026 found dose-dependent, anxiety-like behaviour in exposed rats at doses from the EU acceptable daily intake upward, with stronger effects in males. The commercial formulations generally had larger effects than glyphosate alone, pointing to the role of co-formulants. “Our toxicological findings raise important concerns on the neurobehavioral effects of glyphosate-based herbicides,” said principal investigator Dr Daniele Mandrioli.
Regulators have yet to act on these findings. In September 2026 the European Chemicals Agency’s risk assessment committee said the cancer study did not change its view that glyphosate should not be classified as a carcinogen. The neurological study has not yet been peer reviewed. But both studies found effects at doses regulators currently treat as safe, which is reason enough to speed the shift to alternatives.
The damage goes beyond human health. Herbicide residues run off into rivers and drinking water, drift onto neighbouring crops and wild plants, and strip out the field-edge flora that pollinators depend on. Repeated spraying has also bred resistance: the International Herbicide-Resistant Weed Database now lists 548 unique cases across 275 weed species. Farmers end up spraying more, and more toxic, mixtures to keep up.
Green chemistry: remove the barriers, but don’t bet the budget
New-molecule research is a long shot with a long wait. Most candidates fail. The ones that succeed face a decade of toxicology studies, field trials and registration reviews before a farmer can buy them. And weeds evolve against whatever is sprayed on them repeatedly, so a new herbicide used the way glyphosate has been used will eventually meet the same resistance.
That is not a reason to sideline safer chemistry. Green chemistry – molecules designed from the start to be low-hazard to people and ecosystems, increasingly with AI-driven discovery – could give farmers valuable new options. But it faces barriers that conventional chemistry from the largest agrochemical companies does not. Registration data packages cost far more than a university spin-out or start-up can carry. Review queues are long. Early-stage capital outside the incumbents is scarce.
Governments should remove those barriers: fast-track review for genuinely low-hazard actives, publicly funded toxicology and environmental-fate data, and dedicated research grants. That money should be additional, not taken from the budget for tools that work now. Bioherbicides, electrical, laser and robotic weeders cut herbicide use this season. Each public dollar moved from their deployment to early discovery delays those gains by years.
The principle reaches beyond weeds. Biological insecticides, pheromone-based mating disruption and biofungicides are already registered and in use, yet funding and regulatory attention still flow mostly to new synthetic actives.
EPA’s $30 million desiccation prize: a test case
On 1 July 2026 the U.S. Environmental Protection Agency announced an Innovation Challenge worth up to $30 million to find alternatives to chemical crop desiccation. Desiccation is the practice of spraying herbicides such as glyphosate, diquat and paraquat on crops shortly before harvest to dry them down. It is a major route by which residues reach food: a Canadian study found glyphosate in 76% of wheat samples tested.
The agency opened the design of the prize to public comment (docket EPA-HQ-OPP-2026-3862). In its submission, the Association for Advanced Crop Protection (AACP) argued that the alternatives already exist. Electrical systems are used commercially to desiccate potatoes in Europe. Bioherbicides such as FireHawk are already used as desiccants in the U.S.
AACP’s core recommendation was that around 90% of the prize should go to technologies that are on the market or can reach it within about 18 months. It proposed three tracks:
- Scale-up (about 50%): commercially available alternatives, especially bioherbicide desiccants, with funds for grower cost-share, precision spraying equipment, financing and outreach.
- Near-market innovation (about 40%): adapting proven electrical weed-control systems to desiccation through phased, milestone-based field trials.
- University research (no more than 10%): early-stage approaches, with progress reporting to protect taxpayers.
The comment also urged EPA to judge entries on residues in finished food, worker exposure and cost per acre, not just on how closely they mimic a chemical’s dry-down speed. As AACP put it, “Public funds are best spent accelerating the field-scale expansion of what already works.”
STREAM: the right toolbox, if the balance is right
The U.S. Department of Health and Human Services takes the same problem on with a health agency’s lens. In August 2026 its research agency ARPA-H launched STREAM – Systems for Tracking and Resilience in Efficient Agricultural-input Management. It is part of a joint HHS, USDA and EPA commitment of more than $1 billion to farm modernisation and food security.
STREAM has four technical areas:
- Next-generation herbicides, discovered with AI and machine learning.
- Precision and non-chemical weed control, from laser ablation to small autonomous weeders.
- Low-cost, continuous sensors to monitor herbicide exposure.
- Ways to break down residues in soil and water and keep them out of crops.
The program is significant because it treats herbicide exposure as a public health problem, and because three of its four areas centre on things that are not new molecules. Lasers, electrical weeders, robots, bioherbicides and residue sensors exist today. What they need is the engineering push to match chemical programs on speed and cost per hectare, then independent field data that farmers trust.
STREAM’s value will depend on how it weights those areas when it makes awards. Program manager CAPT James Coburn says “farmers deserve a smarter, safer, and more effective toolbox”. Much of that toolbox is already built. The quickest health gains will come from scaling it and measuring what it removes from farm workers’ bodies and from the food supply, while new chemistry works its way through the pipeline.
Farmers need data to choose the right tool
No single alternative replaces a herbicide everywhere. Laser weeders suit high-value vegetables. Electrical systems handle tough perennial weeds and burndown. Autonomous robots work between crop rows. Bioherbicides fit burndown and pre-harvest desiccation. Which one works on a given farm depends on the weed species and their size, the crop, the soil, the season and the farm’s economics.
Today most farmers get weed-control advice from the same dealers who sell them herbicides. Independent performance data on alternatives is scattered across company trials and a handful of university plots. A farmer weighing an expensive new machine against a cheap, familiar spray has little independent evidence to go on.
Governments can close that gap cheaply. They should fund independent multi-site trials under common protocols, publish cost-per-hectare benchmarks, and back decision-support platforms that match a farm’s weed pressure and cropping system to the alternatives that have worked in similar conditions. Linking those tools to residue testing would let farmers – and their buyers – verify the result. The EU already points the way with its online Farmer’s Toolbox for integrated pest management and the IWMPRAISE integrated weed management project, which brought together 37 research, industry and advisory partners from eight countries.
Farmers will not switch without incentives
A proven alternative cuts no pesticide until a farmer buys it. Electrical and robotic systems carry a high up-front cost, and the switch carries real risk to yield and income. Conventional herbicides, by contrast, are cheap, familiar and sold through well-established dealer networks. AACP’s EPA comment names these – not technical performance – as the main barriers to adoption.
The European Union shows what happens when targets run ahead of incentives. Its 2020 Farm to Fork strategy set two goals for 2030: halve the use and risk of chemical pesticides, and halve the use of the more hazardous ones. In 2022 the Commission proposed a binding Sustainable Use Regulation to enforce them. Facing farmer protests and a split Parliament, it withdrew the proposal in February 2024.
The lesson is not that reduction targets fail. It is that farmers resist targets that arrive as obligations without the tools and money to meet them.
Where the EU has paired ambition with payment, the picture is better. Under the current Common Agricultural Policy, member states must spend at least 25% of direct payments on eco-schemes that reward farmers for going beyond legal minimums. Rural development funds can pay for non-chemical alternatives, precision sprayers and advisory services. In December 2025 the Commission also proposed faster approval routes for biopesticides.
The EU’s headline indicator suggests pesticide use and risk fell 58% by 2023 against its 2015–2017 baseline. But the Commission’s own Joint Research Centre now calls that indicator “flawed”: it is based on sales and broad hazard classes, and its refined methods produced a reduction of about 29% against an earlier baseline. Europe’s experience makes two points for the U.S. Incentives must reach farmers directly. And progress must be measured in residues and exposure, not just in sales.
What governments should do
Public funding for pesticide reduction should follow a simple rule: pay first for what can cut exposure soonest. In practice that means:
- Put most of the money into scale-up. Reserve the large majority of prize and grant funding for weed-control technologies and bioherbicides that are commercial or close to it, as AACP proposed for the EPA challenge.
- Pay farmers to make the switch. Cost-share, equipment finance and payments for verified reductions de-risk the transition. Europe’s eco-schemes are a model; its withdrawn regulation is a warning.
- Give farmers the data to choose. Fund independent trials, cost benchmarks and decision-support tools that show which alternative works for which farm.
- Clear the regulatory path. Labelling and registration rules written for spray chemistry should not block electrical, laser or biological tools.
- Measure what matters. Fund residue testing in food, soil, water and people, so that success is judged by exposure removed, not by sales.
- Back green chemistry with its own money. Fast-track low-hazard actives, fund the data they need for registration and provide research grants, in addition to, not instead of, support for tools already on farms.
- Apply the same rule to all pesticides. Biological insecticides and fungicides deserve the same deployment-first support as alternatives to herbicides.
The EPA prize and ARPA-H’s STREAM program are the most serious U.S. federal moves on herbicide exposure in years. Both can deliver real reductions within a few growing seasons, but only if they put taxpayers’ money behind the tools farmers can already use.
(1 votes, average: 5.00 out of 5)
Loading...
