CFU Thresholds for Cannabis Cultivation: State Limits, What Drives Them, and How to Stay Under

A CFU threshold is the maximum colony-forming units per gram of yeast and mold your harvested flower can carry and still pass state compliance testing. Across North America those limits run from 1,000 to 100,000 cfu/g, and the exact number that applies to you depends on your state and whether you are selling dried flower, an extract, or an infused product. Most states cap dried flower total yeast and mold at 10,000 cfu/g; extracts and infused products usually sit at 1,000 cfu/g. Aspergillus is frequently handled as a separate presence or absence test rather than a count. Miss the number and a mature lot gets quarantined or destroyed.

Here is the part nobody puts next to the threshold: passing is normal. In the largest grow-side study to date, Punja and colleagues assayed more than 2,000 samples over three years and found that under good handling, the majority of dried commercial cannabis samples contained under 1,000 to 5,000 cfu/g. So a failure is not bad luck. It is a signal that something in the grow environment or the post-harvest process moved the count up. This page gives you the number you have to clear, explains what drives the count, and lays out the sanitation and air-testing protocol that keeps you under it.

What CFU means on a cannabis test result

A colony-forming unit is one viable microbe, or one clump of them, that grows into a single visible colony on an agar plate. Labs take a gram of your flower, suspend it, plate the suspension, incubate it, then count the colonies. The count is reported as cfu per gram. Total yeast and mold count, or TYMC, is the number that governs most pass or fail decisions. It is a bulk measurement of all the yeast and mold present. It does not tell you which fungi are there.

That last point matters more than most growers realize. The incumbent science is blunt about it: the TYM current testing does not identify specific species of fungi nor confirm presence/absence of other potentially toxigenic fungi, such as Fusarium and Penicillium, that can be present on dried cannabis. A clean TYMC number is a pass. It is not proof that a toxigenic species is absent. That is why several states bolt a species-specific Aspergillus test onto the general count, and why your sanitation program has to target the environment, not just chase a single number.

TYMC is also distinct from TAMC, the total aerobic microbial count that captures bacteria. TAMC limits tend to run an order of magnitude higher, in the 10,000 to 100,000 cfu/g range, because bacteria are ubiquitous and the human concern is different. If you want the lab-method depth, how plates are poured, dilution series, incubation windows, that belongs on our TYMC testing method explainer. This page stays at the threshold-and-action level.

One more distinction that trips up new QA leads: the air CFU reading you pull off a settle plate in the grow room is not the same measurement as the product CFU a third-party lab reports on harvested flower. Same unit, different sample, very different purpose. We get to that in the air-testing section, because it is the single biggest blind spot we see in grow-room monitoring.

CFU thresholds by state and product form

Start with the range, then narrow to your state. Peer-reviewed work puts the North American spread at 1,000 to 10,000 cfu/g of dried product on the low end, climbing to 50,000 to 100,000 cfu/g in looser jurisdictions. Canada sits at 50,000 cfu/g. Inside the United States it varies state by state, and the number also shifts by product form.

Here is the operator-facing matrix for the states that move the most volume. Treat these as the structure of the rule, and confirm the live figure with your state regulator before you build an SOP around it, because testing rules are amended often.

  • Dried flower, most states: 10,000 cfu/g total yeast and mold is the common ceiling. New York loosened this for one category (see below), but the 10,000 cfu/g pattern holds across the majority of states.
  • Extracts and fully processed products: typically 1,000 cfu/g. The logic is that a concentrate or oil has been through a process step and should carry a far lower microbial load than raw flower.
  • Infused products, oils, edibles, tinctures: also typically 1,000 cfu/g, mirroring the extract rule.
  • Aspergillus: in states like California, Colorado, Michigan, Massachusetts, and Nevada, Aspergillus is a separate presence-or-absence test in 1 gram of flower, not a CFU count. A detectable colony is a fail regardless of your TYMC number. Our Aspergillus testing coverage goes deep on the species-level pass or fail.

Two real codified examples make the pattern concrete.

New York rewrote its adult-use rule in a way most state guides still get wrong. On the medical side, unextracted flower must have 10,000 cfu/g or less and extracted or infused products must have 1,000 cfu/g or less. On the adult-use side the state changed course: unextracted cannabis flower now has no official limit. Instead, as the Office of Cannabis Management put it, labs must report the results, and regulators may require extra testing if there are safety concerns. That shift landed in January 2023, when New York changed the rule for adult-use cannabis flower, so that instead of setting a limit, they now only require labs to report the results. Report-only is not the same as no standard. Your flower result still prints on the certificate of analysis, and a high number still invites scrutiny and brand risk even where it is not an automatic fail.

Washington State runs the opposite model: a codified numeric standard in rule. WAC 314-55-102 lays out colony-forming-unit-per-gram limits for unprocessed plant material, including a 10,000 cfu/g line for bile-tolerant gram-negative bacteria. Washington is the clean example of a state where the number is written into administrative code and the lab tests to that code. Pull the current text of WAC 314-55-102 directly before you cite a specific line, because the citation is live regulation.

The takeaway for a multi-room or multi-state operator: there is no single national number. You build your SOP to the strictest threshold in your footprint, usually the 1,000 cfu/g extract line and any Aspergillus presence-absence requirement, and you hold every room to it. That is how a Cultivation Director standardizes one protocol across a state footprint instead of maintaining a patchwork.

Why samples fail (the Punja variables, translated for operators)

When a lot fails, the count did not appear at harvest. It built up over weeks, driven by variables the Punja study isolated across more than 2,000 samples. Here is what moved the number, translated out of the academic paper into grow-room terms.

Genotype and inflorescence structure.The strain you run is a variable, not a constant. Genotypes with a higher abundance of stigmatic tissues and inflorescence leaves carried more yeast and mold. Dense, leafy buds trap moisture and give fungi more surface to colonize. If a specific cultivar fails repeatedly while others pass in the same room, the genetics are part of your problem, not just your process.

Environmental drivers.Four environmental factors drove counts up: leaf litter left in the grow space, harvesting activity by workers that disturbs and spreads spores, and higher temperature and relative humidity inside the inflorescence microclimate. The fix is housekeeping and climate discipline. Clean plant debris, control traffic during sensitive windows, and hold canopy-level temperature and humidity down. The study also confirmed the counter-move: air circulation achieved by fans during inflorescence maturation significantly reduced counts.

Post-harvest drivers.How you dry decides a large share of the outcome. Wet-trimming and rack-drying held more moisture against the bud and raised counts. Hang-drying of entire inflorescence stems lowered them. The hard target the researchers landed on: drying to a moisture content of 12 to 14 percent, a water activity of 0.65 to 0.7 or lower, which was inversely correlated with CFU levels. Below that water activity, yeast and mold cannot readily grow. Buy a water-activity meter and make 0.65 a release gate.

Seasonal effect.Time of year was significant. The May-to-October window raised counts; November-to-April lowered them. If you grow through summer, your contamination pressure is higher and your sanitation cadence has to tighten to match. This is where a prevention program pays for itself. One failed lot of premium indoor flower is a five- to six-figure destruction event depending on state and volume. Twelve months of a disciplined ClO2 air-quality program costs a fraction of a single failed harvest. Frame CFU threshold as a revenue-protection number, not a compliance-only number, and the budget conversation changes.

Air CFU vs product CFU (the leading indicator most grows miss)

The product CFU on your certificate of analysis is a lagging indicator. By the time the lab reports it, the crop is harvested and the money is already committed. Air CFU, the count you pull off settle plates in the grow room, is a leading indicator. It tells you where the environmental load is trending weeks before harvest lab day.

Punja demonstrated the method in the study, sampling air for yeasts and molds alongside the flower, but the paper stops at methodology. Nobody translates it into a grower protocol. That translation is the GroClarity air-testing service, and it is the gap in every top-ranking page on this topic.

Here is the practical version. Expose settle plates at canopy height in each flower room on a fixed weekly schedule, at the same locations every time, incubate, and count. You are not chasing an absolute pass or fail on air plates. You are watching the trend. A stable baseline week over week means your environment and sanitation are holding. The escalation rule is simple: if air plate counts rise for two consecutive weeks, something changed, find it before it reaches the flower. That is the whole value of a leading indicator. It buys you time to act while the crop is still correctable.

Build the cadence into your harvest calendar, log every count, and you now have both a control system and a data trail. Our air-sampling field guide covers plate placement, exposure time, and reading counts. If you want the program run for you, on a schedule tied to your harvest calendar, that is what GroClarity's air-testing service delivers, with detailed online reports you can hand to ownership and to regulators.

Holding the line with ProKure ClO2 (EPA-compliant sanitation)

Knowing your number is half the job. Holding under it means sanitizing the environment the fungi live in, between crops and during the grow, without creating a compliance problem of its own. This is the Q40 question every CFU guide dodges: what do you actually deploy, and is it registered. The honest answer separates three roles, and the EPA status is different for each, so keep the claims matched to the right product.

Surface disinfection between cycles: ProKure V.ProKure V is a liquid, water-activated chlorine dioxide formula, EPA Reg. No. 87508-3-89334. It is a hospital-type disinfectant that kills bacteria, viruses, mold, fungi, and odor-causing bacteria, and it kills SARS-CoV-2 on hard non-porous surfaces in 30 seconds. The reason cultivators pick it over bleach or quats comes down to one combination: it is suitable for use with plants in the room and it leaves no residue. You add water to activate, apply it, and there is no wipe, no rinse, no residue touching the next crop. For between-cycle turnover of grow beds, benches, trays, tools, and trim-room surfaces, that is the disinfectant half of the system. Deep product detail, the full kill list, and the SDS live on prokuresolutions.com.

Fast air reset in an empty room: ProKure G.ProKure G is a fast-release gas, EPA Reg. No. 87508-2-89334, the deodorizing delivery system for mold and mildew, odor-causing bacteria, and chemical odors. It uses deep-penetrating ClO2 vapor and works in 4 to 6 hours. One hard rule governs its use, straight off the label: it is not for use with/on plants/crops/feed. G is an empty-room tool for a deep clean between harvests to reduce downtime. People and plants come out, G goes to work, the room is ventilated, then you re-enter.

Ongoing air quality during the grow: ProKure D.ProKure D is a slow-release gas that provides ClO2 protection for up to 30 days. It improves air quality and controls odors. It is not an EPA-registered product and not labeled for sale as a pesticide under 40 CFR, so it carries no kill claims, it is an air-quality and odor layer, and that disclaimer rides with it verbatim. The trifolds hang 18 to 24 inches above the canopy and run with plants in the room across the cycle. Replace each pouch before its up-to-30-day window closes to keep ClO2 levels steady. This is the set-and-forget air layer that drives most of the volume on a real grow.

What the EPA registration actually guarantees is narrow and worth stating plainly: a registered product may make the claims on its label, under the directions on its label, and nothing beyond. ProKure V carries disinfectant kill claims because its label supports them. ProKure D does not, because it is not registered, and anyone selling it with kill claims is overreaching the label. Dose on the air CFU trend, not a calendar. When your weekly plates climb, tighten coverage. When they hold flat, hold steady. To buy or reorder the line, route to the store at shop.groclarity.com. It is a violation of Federal Law to use any of these products in a manner inconsistent with the labeling, so read the entire label and use strictly in accordance with its directions.

What to do after a failed CFU test

A failing result is a fire with a defined response. Work it in order.

First, quarantine the lot and pull a re-test. Lab results vary, and the cfu counts may be influenced by the methods used, the agar medium, the extraction method, incubation time and temperature, and dilution. A single marginal fail can be lab variance. Segregate the lot so it cannot move, then re-sample and re-run before you write it off.

Second, run root cause across three buckets: environment, post-harvest, or lab variance. Pull your air plate logs for the room and window the lot came from. Rising air counts point at the environment. A clean air trail with a drying-stage problem points at post-harvest, check your water-activity records against the 0.65 to 0.7 target. A one-off with clean records on both sides points at the lab. The data trail you kept is what lets you answer this instead of guessing.

Third, choose remediation with eyes open on the compliance implications. Options range from reconditioning to destruction depending on your state and the product form, and some remediation steps carry their own regulatory reporting. Our post-fail remediation playbook walks the options and their compliance consequences. Whatever you choose, the environment that produced the fail still needs a between-cycle reset before the next crop goes in, which puts you back in the sanitation protocol above.

Building a defensible QA data trail

Steven the head grower and the QA lead face the same demand from ownership and auditors: prove the air quality held. A number on a certificate is not proof of a program. A logged trail is. Three records make a defensible file.

Weekly air plate log.Date, room, plate location, exposure time, incubated count, and the initials of whoever read it. Week over week, this is the document that shows a stable environment and catches the two-week rising trend early. We publish a weekly air plate log template you can adopt as-is, paired with the ProKure SDS and EPA label pack for your procurement and QA files.

Batch-level moisture and water activity records.Record the water activity at dry-down release for every batch, against the 0.65 to 0.7 ceiling. This closes the post-harvest loop and gives you the single strongest predictor of a passing TYMC in writing.

Sanitation SOP with the details that survive an audit.Product name, dose, contact time, operator initials, and the EPA registration number for each registered product, ProKure V at 87508-3-89334. For ProKure D, record the up-to-30-day placement and replacement dates and carry the not-an-EPA-registered-product disclaimer in the SOP so the file is accurate on its face. A Cultivation Director standardizing across facilities runs one version of this SOP in every state and defends it with one data trail. That is what GroClarity's cultivation-IPM program is built to deliver.

Frequently asked questions

The questions below are the ones growers and QA leads actually ask about thresholds. The numbers throughout this page trace to peer-reviewed research and primary state regulation: Punja et al. 2023 is the grow-side anchor on what drives yeast and mold counts, and Washington WAC 314-55-102 plus the New York Office of Cannabis Management testing limits are the two codified-standard examples in the reference list below. Confirm any specific state figure with that state's regulator before building an SOP around it. See product labels for all specific claims and instructions.

Related resources