Glyphosate might be the most argued-about bottle in the shed. To some people it's a harmless tool that's been used safely for fifty years. To others it's a corporate cover-up waiting to be exposed. We wanted to do something different this time and actually walk through the whole story: the history, the cancer classification and why it's controversial, who's tried to replicate the cancer findings and what they found, how long it actually lingers in soil and water, and which countries have banned it and why.

This is the companion piece to our episode on the topic. No belief systems here, just what's actually been measured and published.

What Glyphosate Actually Is

Glyphosate is a broad spectrum herbicide, meaning it kills almost any green plant it touches rather than targeting one specific weed. It works by blocking an enzyme plants need to build certain amino acids, part of what's called the shikimate pathway. Humans and animals don't have that pathway, which is the core reason it's considered low toxicity to us directly.

It was developed by a Monsanto chemist named John Franz and hit the market as Roundup in 1974. It became the most widely used herbicide on the planet once Monsanto released "Roundup Ready" crops in the 1990s: soybeans, corn, and cotton genetically engineered to survive being sprayed with it directly. That's the point where this stopped being just a chemistry question and got tangled up with GMOs and seed industry consolidation, and a lot of the emotional charge around glyphosate today is really inherited from that separate fight.

The 2015 Classification That Started It All

In March 2015, the International Agency for Research on Cancer (IARC), the cancer research arm of the World Health Organization, classified glyphosate as "probably carcinogenic to humans," Group 2A. That single classification is the reason we're all still talking about this a decade later.

For clarity. IARC isn't asking "how much glyphosate would actually give a person cancer at real-world exposure levels." It's asking a narrower question: is there evidence this substance could cause cancer under any studied circumstance. Group 2A sits below Group 1, which includes known carcinogens like tobacco and asbestos, and IARC based its classification on what it called limited evidence in humans, sufficient evidence in animal studies, and strong evidence of genotoxicity.

For context, Group 2A also includes red meat and working the night shift. Group 1, the "known carcinogen" tier, includes alcohol and processed meat like bacon. IARC's whole system is built around hazard, meaning could this theoretically cause cancer, not risk, meaning how likely is that at the doses people actually encounter. That distinction is where almost every downstream argument about glyphosate splits.

Nearly Every Other Major Agency Disagreed

IARC's 2015 finding stands almost alone among major regulatory and scientific bodies.

The European Food Safety Authority (EFSA) reviewed the same and additional evidence and didn't confirm IARC's conclusion. Germany's BfR, the institute that led the EU's official evaluation, concluded glyphosate is unlikely to pose a carcinogenic risk to humans. The European Chemicals Agency reached the same conclusion. The U.S. EPA has repeatedly found it "not likely to be carcinogenic to humans." Brazil's health regulator reached a similar conclusion in 2019. A joint WHO and FAO expert panel, using additional evidence IARC didn't have, also didn't confirm IARC's finding.

That doesn't automatically make IARC wrong. Regulatory agencies have gotten things wrong before, lots of times, and industry pressure on regulators is a real, documented problem in pesticide history. But it does mean the "the cancer agency said it causes cancer" headline skips over just how isolated that conclusion actually is among expert bodies.

Who Tried to Replicate the Cancer Link, and What They Found

The independent re-review. After IARC's classification, four independent expert panels were convened specifically to re-examine the same evidence IARC used: animal studies, genetic toxicology, epidemiology, and exposure data, and compare their conclusions to IARC's. Those panels concluded the data did not support IARC's "probable carcinogen" call. They found the animal tumor data lacked consistent dose-response relationships and statistical strength, and the genetic toxicology data didn't support classifying glyphosate as a genotoxic carcinogen.

The Agricultural Health Study. This is the largest and longest-running human study on the subject, tracking more than fifty thousand farmers and pesticide applicators in the U.S. It's the closest thing we have to a real-world replication attempt in an actually exposed population, and it did not find a clear statistical link between glyphosate and non-Hodgkin lymphoma. IARC acknowledged this study but argued its follow-up period might have been too short to fully rule out an effect, which is a legitimate limitation to raise. But it means the largest human dataset we have came back negative or inconclusive, not confirmatory.

The meta-analyses. On the other side, several meta-analyses pooling smaller case-control studies from the U.S. and Europe have found a modest increased risk of non-Hodgkin lymphoma among people with heavy occupational exposure: farmers and applicators using it many days a year, not casual homeowner use. This is genuinely the strongest evidence on the "there might be something here" side of the debate.

The honest tally: the largest prospective study did not replicate a clear cancer signal. The four independent panels reviewing IARC's own source material did not replicate IARC's conclusion. Several smaller retrospective meta-analyses did find an elevated risk specifically tied to heavy, prolonged occupational exposure. Nobody has produced a clean replication showing casual or low-level exposure causes cancer.

The Part That Makes This Messier Than a Clean "Science Says X" Story

In 2017, Reuters obtained an early draft of IARC's glyphosate review and compared it to the final published version. They found roughly a dozen changes to the animal studies chapter between draft and publication, and in every case, a negative or neutral conclusion in the draft became more supportive of a carcinogenic link in the final version. Reuters couldn't determine who made the changes or why. Christopher Portier, a special advisor to the IARC working group, later acknowledged in a deposition that an internal subgroup had initially concluded there was only "limited evidence of animal carcinogenicity," a phrase that later became "sufficient evidence" in the final report, and he said he didn't know exactly how or when that shift happened. That's a real, documented irregularity, and it's the strongest ammunition critics of IARC have.

But the industry response wasn't a clean, disinterested defense of science either. Internal documents later revealed Monsanto ran a coordinated campaign to discredit the IARC finding, including ghostwriting scientific literature later attributed to independent academics, a tactic critics have compared directly to the tobacco industry's old playbook. And the same Portier who flagged IARC's process irregularities was later paid as a consultant for plaintiffs' attorneys in Roundup litigation, a conflict critics point to when questioning his neutrality.

Both sides of this fight have documented credibility problems. That's not us dodging the question, that's the actual state of the record.

The Lawsuits and Where the Money Is

Since Bayer acquired Monsanto in 2018, it has faced well over one hundred thousand Roundup lawsuits in the U.S. alone, mostly from people who developed non-Hodgkin lymphoma after long-term occupational or heavy personal use. Bayer has paid out billions in settlements while continuing to maintain the product doesn't cause cancer, and it stopped selling glyphosate-based Roundup for the U.S. residential lawn and garden market back in 2023, a decision it's been explicit was about litigation exposure, not new safety findings. Commercial agricultural glyphosate is still sold and used.

As of 2026, the U.S. Supreme Court has agreed to hear Monsanto v. Durnell, a case that could decide whether federal pesticide law shields Bayer from these state-level lawsuits altogether. That ruling could reshape the entire legal landscape either way it goes.

How Long Does It Actually Stick Around, and What Does It Touch on the Way Out

This is the environmental side of the story, and it's less argued about than the cancer question but doesn't get talked about nearly as much.

"Breaks down fast" depends entirely on where it lands. A commonly cited average half-life, the time it takes for half the chemical to break down, is about 47 days in soil. But the real published range is huge: anywhere from 2 days to almost 200 days, depending on soil type, temperature, and how much microbial life is doing the work of breaking it down. One test site in Texas measured just 3 days. A site in Iowa, same chemical, measured 142 days. Cold soil, heavy clay soil, or soil that isn't very biologically active holds onto it much longer than warm, sandy, active soil, and in ground that freezes for part of the year, like ours here in New England, that breakdown process basically stalls. Studies have found it still present the following spring.

When it does break down, it turns into something that lasts even longer. Glyphosate doesn't just vanish, soil microbes break it apart into a compound called AMPA (aminomethylphosphonic acid). Think of it as simply what glyphosate turns into. It's a simpler molecule, and it consistently lasts longer in the environment than the glyphosate it came from.

One wrinkle worth knowing: AMPA isn't only a glyphosate byproduct, it's also produced when certain detergents and water-treatment chemicals break down. So finding AMPA in a soil or water sample doesn't automatically prove herbicide use, it could be from spraying, from something unrelated, or both.

Published studies put AMPA's half-life in soil anywhere from about 76 days to well over a year, and in cold, frozen-winter soils like Scandinavia's or ours, some studies have found it still detectable almost two years later. A survey of farm soil across eleven European countries found glyphosate or AMPA in nearly half the samples tested, and AMPA actually showed up more often than glyphosate itself. A soil test that comes back "no glyphosate found" can still mean the field is carrying its longer-lasting leftover.

Is AMPA itself dangerous? Honestly, nobody's studied it nearly as much as glyphosate, so be skeptical of confident claims either direction. What research exists generally suggests similar low toxicity to glyphosate in most testing, but a few lab studies looking at effects on individual cells have found AMPA can be more reactive than glyphosate in some tests and less in others. That's a genuinely unsettled question.

Here's where the actual safety numbers help. In the EU, drinking water has a strict cutoff of 0.1 micrograms per liter for glyphosate and for AMPA, each on their own, an extremely conservative cleanliness standard, not a health-danger line. Separately, French regulators set an actual health-based safety threshold for glyphosate and AMPA combined at 900 micrograms per liter, roughly nine thousand times higher than that strict cutoff. When French water was actually tested, both compounds together turned up in only 3 to 4 percent of groundwater samples, with even the ultra-strict 0.1 standard exceeded less than 1 percent of the time. At the levels actually showing up in real water testing, nobody's finding numbers anywhere near a genuine health risk by current standards.

Where there's a real open question is at the cell level. One recent lab study, testing cells in a dish rather than a living animal or person, found glyphosate and AMPA both caused measurable damage to cell DNA on their own, and that together they seemed to cause even more damage than either alone. That's worth taking seriously, but a cell-dish result doesn't automatically mean the same thing happens inside a living body at real-world exposure levels. It's a reason for more research, not proof of harm.

Bottom line on AMPA: at the levels actually turning up in water testing, it isn't showing up as a safety problem by any current regulatory standard. The genuine unresolved question is at the lab-bench level, cell studies suggesting it's not simply harmless leftover, combined with the fact that it hasn't gotten the research attention glyphosate has. Not "obviously dangerous," not "definitely nothing to worry about" either.

Water tells a similar story. Glyphosate breaks down faster in water than in soil, commonly a few days up to about 90 days, but it clings to pond and waterway sediment, and studies have found it lasting over a year down there even after the water above tested clean. A clean water test doesn't necessarily mean the sediment underneath is clean too.

Earthworms and bees show the clearest real-world effects. This is where the data is more consistent than anything in the cancer debate. A 2022 study of French farmland found glyphosate and AMPA in most of the soil samples and most of the earthworms living in that soil, with the chemical building up in the worms at two to three times higher levels than expected. On the pollinator side, several independent studies, including one published in PNAS, found that glyphosate disrupts the gut bacteria honeybees depend on to fight off disease, and that bees with disrupted gut bacteria died at higher rates when later exposed to common pathogens. It's not universal across every species though: one study on bumblebees found no effect on their gut bacteria under the same kind of exposure that consistently hurt honeybees. It's a real, repeatable effect, just not one that applies evenly across every pollinator.

Where It's Banned, and Where It Isn't

People assume "banned in Europe" means banned everywhere overseas. It's more fractured than that.

Glyphosate is not banned across the EU as a whole. In 2023 the European Commission renewed its approval for another ten years, through the end of 2033, based on EFSA and ECHA reviews. Individual member states, including France, Germany, Belgium, Italy, Austria, Luxembourg, and the Netherlands, maintain their own partial bans, mostly for home and public-space use rather than commercial agriculture.

Outside Europe, a handful of countries have gone further. Sri Lanka, Vietnam, and a few Gulf states have had full or near-full bans at various points, though some, like Sri Lanka's, have been reversed and reinstated more than once as farmers pushed back over crop losses. Mexico announced plans to fully ban it by 2024 but postponed that once it was clear there wasn't a viable large-scale alternative ready. Colombia and El Salvador both banned it and later reversed course. Brazil, one of the largest agricultural economies in the world, has kept it legal after its own health agency found no clear risk.

In the U.S., there's no federal ban. The EPA maintains its position that glyphosate isn't likely carcinogenic and hasn't approved a cancer warning label. What exists instead is a patchwork of city and county restrictions, mostly in California, plus places like Portland and Boulder, that restrict or ban it on public land like parks and schools without touching agricultural use. California's Prop 65 requires a cancer warning label on the product itself, which is a state disclosure law, not a safety finding, a distinction that gets lost in a lot of the online conversation.

The EPA is currently in the middle of a mandated fifteen-year registration review, expected to wrap up in late 2026, the first full review since the IARC controversy began.

Where We Land on This

We try to run everything on this site through the same filter: separate what was actually observed from what somebody is interpreting for you.

One international agency, using a hazard-based standard that also puts red meat and hot beverages in a similar tier, called glyphosate a probable carcinogen based on limited human evidence and animal data that four independent panels later said didn't hold up under closer review. Nearly every national regulatory body that did its own full review disagreed. The largest, longest human study we have didn't find a clear cancer signal. Smaller studies of people with heavy, sustained occupational exposure have found an elevated risk of one specific cancer type. The company that makes it ran a documented influence campaign, and one of IARC's own advisors later took money from plaintiffs' lawyers, so neither side gets to claim the moral high ground on transparency.

Separately, and on firmer ground, the environmental persistence data is pretty consistent. Glyphosate typically clears out in weeks to a couple months under good conditions, but its leftover compound, AMPA, often sticks around for months to years, especially in cold ground like ours. And its effects on earthworms and honeybees are more consistently reproduced across independent labs than anything in the human cancer debate.

If you're a homeowner spraying a driveway crack twice a summer, the evidence for meaningful personal cancer risk is thin. If you're a farmer or applicator using it dozens of days a year for decades, the evidence is genuinely mixed enough that caution, proper PPE, minimizing drift and skin contact, isn't fear-mongering, it's what the actual data supports. And if you're gardening organically or trying to protect pollinators near your beds, the environmental case for caution is honestly on firmer scientific footing than the cancer case is.

Not a clean, satisfying answer. But that's what the evidence actually gives us here, and we'd rather tell you that than pretend otherwise.

Sources

  • IARC Monograph on Glyphosate (2015), iarc.who.int
  • Reuters investigation into IARC's draft review, October 2017
  • "A review of the carcinogenic potential of glyphosate by four independent expert panels and comparison to the IARC assessment," Critical Reviews in Toxicology, 2016
  • "Glyphosate toxicity and carcinogenicity: a review of the scientific basis of the European Union assessment and its differences with IARC," Archives of Toxicology
  • Agricultural Health Study, NIH-funded cohort of 50,000+ U.S. farmers and applicators
  • U.S. EPA glyphosate registration review materials
  • USTR summary of EU member-state glyphosate restrictions; EU Commission renewal decision (2023)
  • "Expert Review Under Attack: Glyphosate, Talc, and Cancer," American Journal of Public Health
  • ANSES (French food and environmental safety agency) glyphosate and AMPA drinking water monitoring data
  • Silva et al., EU topsoil glyphosate and AMPA survey across eleven member states
  • Pelosi et al. (2022), glyphosate and AMPA in French farmland soil and earthworms
  • Motta et al., PNAS, "Glyphosate perturbs the gut microbiota of honey bees"

Have a source you think we got wrong or missed? Let us know in the comments, we'd rather correct this than leave it wrong.