Penny wise, pound foolish

How much does it really cost to answer a tough scientific question?

We recently attended a glyphosate symposium. It was fascinating to hear the state of modern science around glyphosate from toxicologists, epidemiologists, regulatory scientists, and advocates. One theme that caught our attention was how "expensive" these experiments are. But how much does good science actually cost, compared to all of the expenses around it, especially when one considers legal fees and settlements.

Is science really that expensive?

First, let's estimate how much science costs.

Most of what gets published is relatively cheap. Cell culture experiments that expose human or animal cells to a substance in the lab, meta-analyses that pool results from existing studies, review papers, and computational studies — these run in the tens or hundreds of thousands of dollars, sometimes less. A meta-analysis might cost $50,000–$200,000. A cell line study can be done for under $100,000.

Genotoxicity assays, which test whether a chemical damages DNA or causes mutations in cells or bacteria, cost as little as $5,800 for a bacterial reverse mutation test. Reproductive and developmental toxicity studies — which expose pregnant animals to a substance and track effects on fertility, fetal development, and offspring across one or more generations — run $129,000–$432,000 each.

The most expensive individual studies are the two-year rodent carcinogenicity bioassays — where groups of rats or mice are exposed to a substance at various doses and monitored for tumor development over their lifetimes. For better or worse, these are often considered the gold standard.

The EPA's own published cost schedule puts a single such bioassay at $1.77 million (2016 dollars), and a combined chronic toxicity/carcinogenicity study at up to $4.07 million. For glyphosate, the exact number depends on who is counting: IARC Monograph 112 reviewed eight studies (six rat, two mouse); the EPA used fifteen acceptable carcinogenicity studies in its evaluation; and a subsequent analysis by Portier et al. (2020) identified twenty-one chronic exposure studies, of which only thirteen were suitable for reanalysis. At $1.77–4.07 million each, the core cancer evidence — the studies that regulators and IARC actually fight about — cost somewhere in the range of $25–85 million total.

Then there is population-level epidemiology. The Agricultural Health Study — 89,000 participants tracked since 1993 — has cost an estimated $100–200 million over its lifetime, but glyphosate is one of roughly 50 pesticides in the study. The National Toxicology Program ran only 13-week toxicity studies in 1992 and never conducted a full two-year bioassay on glyphosate.

To prove causation, regulators typically insist on all of these — genotoxicity, reproductive toxicity, chronic bioassays, epidemiology — layered on top of one another. No single study type is sufficient on its own.

So let's be generous and overestimate. Bayer claims there are 2,400 studies on glyphosate over 50 years. If we assign each an average cost of roughly $200,000 — which is high, given that most are literature reviews, in vitro work, or short-term assays — we arrive at about $480 million. That's real money. But now let's be even more conservative, and round up to a billion.

A billion dollars for the entire scientific corpus on which global regulatory decisions rest. Over fifty years. For a chemical applied to every continent and has been found in most of our bodies.

Fine, you say. A billion dollars. Is that a lot?

Well, consider this: Bayer has already paid more than $11 billion in settlements of lawsuits alleging harms from Roundup, whose major active ingredient has historically been glyphosate. As of early 2026, its balance sheet carries an additional €11.8 billion in litigation provisions, of which approximately €9.6 billion is specifically for glyphosate. In 2026 alone, the company expects approximately €5 billion in litigation payouts. Recently, it offered a settlement in a major class action suit of up to $7.25 billion.

Just the settlements are 20× all science. And this is not counting legal fees.

Plaintiff attorneys typically work on contingency fees of roughly 33–40% for individual mass tort cases, and often 22–33% in class actions. That means just the plaintiff lawyers have already received more total compensation for arguing about glyphosate studies than all scientists combined received for producing them.

Meanwhile, global glyphosate revenue runs $8–10 billion per year. Monsanto's Roundup and other glyphosate-based herbicides peaked at $3.53 billion in revenue in fiscal year 2009 with a 52% gross margin.

A single year of Roundup gross profit exceeded the total investment in glyphosate safety science.

And then there is executive compensation. Bayer's current CEO earns about $7 million per year. But Monsanto CEO Hugh Grant earned $19.5 million in his final year and walked away from the Bayer acquisition with an estimated $77–123 million in total gains, including a $2.5 million "merger recognition bonus". The top five Monsanto executives collectively gained an estimated $170–250 million at acquisition, based on SEC proxy filings and contemporaneous reporting.

Last but not least, there are shareholders. After Bayer purchased Monsanto in 2018 for $63 billion, its market capitalization fell from approximately €120 billion to roughly €37.6 billion. Approximately $90 billion in shareholder value was lost.

That is 90× all the science.

Clearly something is wrong with this picture.

An obvious solution would be to put a tax on all pesticide sales, and use it to fund independent science that would answer the key questions once and for all. But we have yet to really figure out how to do that.

In 2014, a Russian initiative called Factor GMO was announced at a London press conference as what would be the "world's largest" independent study of genetically modified organisms and glyphosate. It had a high-profile advisory board. But almost immediately, there were concerns with the design of the study — it was to be conducted "at undisclosed locations in Russia and Western Europe." And then it simply disappeared. As BuzzFeed News documented in a 2018 investigation, the entire project evaporated, leaving behind only questions about where the funding went.

Of course, the failure of one attempt doesn't mean the concept is wrong. The real work is twofold: redirecting even a fraction of available resources toward scientific funding, and then doing objective, convincing science that can actually answer the difficult questions. It is yet another challenge the scientific community should take on board.