Is there mold in your coffee? Ochratoxin A, the EU limit, and what the US checks
Roasting kills mold but not all of its toxin. What ochratoxin A is, the EU's 3 microgram limit, FDA's 20 microgram referral level, and what testing found.
Emily Harper
If you have seen a list of "mold-tested" coffee brands and wondered whether there is actually mold in your morning cup, here is the short version: coffee beans can carry a mold toxin called ochratoxin A, roasting removes a good share of it but not all of it, the European Union has a legal limit for roasted coffee, and the United States does not.
The short answer: Mold itself is killed by roasting. The concern is a toxin some molds leave behind, ochratoxin A (OTA), which survives roasting in part. Surveys of finished coffee find it often but usually at low levels, and the one 2026 safety assessment we found calls the risk from typical drinking low. It is worth knowing about, not worth panicking over.
Mold versus mycotoxin
A mycotoxin is a poison made by a mold. Killing the mold does not undo what it already made. A 2024 review in the Journal of Fungi says OTA in coffee comes mainly from Aspergillus species, that contamination can begin on the farm (coffee berry borers and overripe cherries are factors), and that most fungal contamination happens during post-harvest drying. It also reports that roasting cuts OTA by anywhere from about 12% to 90% depending on the roast and how much was there to begin with. That wide range is the key point: roasting helps, but you cannot count on it to finish the job.
OTA is classified by the IARC as Group 2B, "possibly carcinogenic to humans," and the same review describes it as toxic to the kidneys.
The EU has a limit. The US does not.
- EU: Commission Regulation (EU) 2023/915 sets a maximum of 3.0 micrograms per kilogram for roasted coffee beans and ground roasted coffee, and 5.0 for soluble (instant) coffee (Annex I, points 1.2.14 and 1.2.15).
- US: According to a 2026 paper by FDA scientists, FDA has not set a maximum level for OTA in foods. What it has is a referral level of 20 micrograms per kilogram: lab findings above that go to a case-by-case safety review. That is more than six times the EU's roasted-coffee limit, and it is a trigger for review, not a legal cap.
So "the US has no limit" is mostly right but not the whole story. There is a number FDA reacts to; it is just much higher than Europe's and not a rule you can point to on a bag.
What FDA actually found
The same paper summarizes FDA's testing from October 2007 to September 2022: 3,708 food samples overall, and 514 coffee bean samples. Of those, 14 (2.7%) had OTA high enough to measure, at levels from 1.2 to 116 micrograms per kilogram, and 10 of the 14 were over the EU's 3 microgram limit for roasted coffee. The authors note FDA mostly collected green (unroasted) beans rather than finished coffee, so these numbers describe raw beans, and roasting would lower them.
Two ways to read it: most coffee beans FDA checked had nothing measurable, which is reassuring, and the few that did sometimes had a lot, which is the point of having a limit.
What about the coffee you actually drink?
Other countries have tested finished coffee. A 2026 study in Toxins analyzed 65 ground, roasted Turkish coffee samples and found OTA in 53 of them (82%). It then estimated how much a drinker would take in and compared that with EFSA's risk thresholds. For a person drinking half a cup, one cup or three cups a day, the safety margins were above the thresholds for both cancer and kidney effects, and the authors concluded the risk is low under current consumption. They still recommend continued monitoring.
Put those two findings together and the honest takeaway is: OTA is common in coffee, the amounts are mostly small, and the people who reviewed it did not find a problem at normal intake. That is one country's coffee and one study, not a guarantee about any U.S. brand.
What the "mold-tested" brand lists mean
Tallow's recent video listed seven coffees that say they test their coffee for mycotoxins: Purity Coffee, Lifeboost, Bulletproof, Four Sigmatic, Kion, Lineage Provisions and Fabula. Two caveats worth keeping in mind:
- A mycotoxin test is not a mold count. It looks for the toxins, not the mold itself.
- A claim covers the coffees that were tested. It is not a promise about every product a brand sells, and the claims come from the brands themselves, which we have not independently re-tested.
How to lower your odds
- Look for published results. A brand that posts an actual OTA number for its coffee is telling you more than one that says "mold-free."
- Compare against 3. If a brand publishes a number, the EU limit for roasted coffee is the useful yardstick.
- Store coffee dry. The mold behind OTA likes moisture, and a sealed, cool, dry spot is the sensible default at home.
- Do not over-read "organic." Organic describes how the plant was grown, not whether the beans were stored dry.
Where Tallow comes in
You should not need to read a lab report to know this. Tallow shows lab results on the product page when they are on file. Purity Coffee FLOW Organic Medium Roast on Tallow lists independent-lab results for 13 contaminants, including ochratoxin A, aflatoxins and fumonisins, all reported as "not detected." Those results come from tests of its product family (Purity's Organic Whole Bean coffee), and the page says so. Not every coffee has results on file: when we checked a handful of other popular coffees, most had no mycotoxin rows yet. Scan a bag of coffee to see which kind you have, and to see ingredients and packaging in the same score.