How Long Does Pot Last? The Order to Work It Out In

Cannabis flower that was dried properly and kept airtight, dark, and at room temperature stays chemically usable for one to two years, and it does not acquire a safety problem on a schedule. Ross and ElSohly, writing in the United Nations Bulletin on Narcotics, measured Δ9-THC in flower held at 20–22 °C and reported average losses of 16.6% after one year, 26.8% after two, 34.5% after three, and 41.4% after four. Those numbers answer a potency question. The safety question runs on a different variable entirely: flower becomes vulnerable to mold once its water activity climbs above 0.65, roughly 65% relative humidity inside a sealed container, which is the ceiling written into ASTM D8197. Under that ceiling, a two-year-old jar can be sound. Over it, a three-week-old jar can be a problem. So work it out in this order: what form the product is, how it was stored, whether the package held, and what you can see and smell right now.

Decide which question you are actually asking

Three questions hide inside "how long does pot last," and they run on separate clocks.

Potency is the slow one. The Ross and ElSohly curve is smooth, and after the second year the loss rate settles to roughly 7% annually. A jar down a quarter of its Δ9-THC is a jar you use slightly more of.

Aroma is the fast one, and most intuitions here are wrong by an order of magnitude. Bueno and colleagues, publishing in the Journal of Cannabis Research in 2020, recorded a 51.6% drop in total terpene content after a single month of storage. Ross and ElSohly measured 55.2% gone at three months against freshly harvested material, a benchmark that folds drying losses in with storage losses. Monoterpenes such as myrcene leave first; the heavier sesquiterpenes hold on. That is why nine-month-old flower smells like hay.

Contamination has no curve. It is a threshold, and it either happened or it did not.

Your nose grades quality and cannot clear a sample

For about two years I told people a careful nose was a sufficient instrument for old flower. Then I bought a water-activity meter, the cheap kind that sits in the container overnight, and measured four jars I had cleared by smell. Two were above the 0.65 ceiling. One read 0.71. The cost was not the flower. It was that I had put the nose advice in writing, and had handed a friend a jar I cleared that way.

Dryness, harshness and a flattened smell are quality signals tracking the terpene loss above. Visible fuzz or webbing is a late signal, meaning growth advanced enough to see without magnification. A musty or ammonia note means active microbial metabolism. None of those gives you a colony count, and the absence of all three tells you little about a jar that spent six months at high humidity. Visual inspection cannot quantify contamination, which is why laboratories plate samples instead of looking at them.

What storage actually changes in the chemistry

The cleanest experimental work on this is old. Fairbairn, Liebmann and Rowan stored herbal cannabis, resin and pure cannabinoid solutions under varied conditions for up to two years, publishing in the Journal of Pharmacy and Pharmacology in 1976. Light was the greatest single factor in cannabinoid loss. Temperature, up to 20 °C, made an insignificant difference. Air oxidation in the dark caused meaningful losses, reduced when the resin glands were left undamaged, which the authors called "well filled, well closed containers." Grinding a week's supply in advance opens every one.

Their third finding dismantles a story you have heard. Δ9-THC lost to light does not turn into cannabinol, while air oxidation in the dark does produce it. The popular claim that aged flower reliably becomes sedating because THC converts to CBN assumes a single degradation route. A jar that sat in sunlight took the other route, losing THC without accumulating much CBN, and you rarely know the starting composition well enough to compute the ratio anyway. CBN is real chemistry, not a promise about how the evening will go.

The pattern is regular enough to be forensically useful: Zamengo and colleagues, in Forensic Science International in 2019, estimated a sample's storage time from its THC and CBN figures to within roughly 5% to 10%. Chemists use CBN to date a sample, never to clear one.

Mold is a threshold, not a curve

ASTM D8197, currently in its 2022 revision under Subcommittee D37.04, specifies a water activity range of 0.55 to 0.65 for dry cannabis flower. The two limits exist for opposite reasons: under 0.65 to guard against microbial growth, above 0.55 so the material does not shatter in routine handling. Past 0.68 the standard calls for further drying.

What a passing laboratory result means depends on which state issued it. Colorado's marijuana rules at 1 CCR 212-3 set total yeast and mold below 10,000 CFU/g and separately require the absence of four Aspergillus species. Massachusetts and Nevada use the same 10,000 CFU/g for unprocessed flower. Michigan applies 10,000 CFU/g to medical product and a limit ten times looser to adult-use flower, a gap its regulator addresses in its bulletin on equivalent-license transfers. California sets no yeast and mold count for inhalable goods at all. Under Title 4, section 15719 of the California Code of Regulations, Aspergillus flavus, A. fumigatus, A. niger and A. terreus must simply not be detected in one gram, and section 15721 caps total aflatoxins B1, B2, G1 and G2 at 20 µg/kg with ochratoxin A at the same figure. Canada's limit is 50,000 CFU/g. The American Herbal Pharmacopoeia monograph several of these frameworks borrowed from recommends under 10,000 CFU/g for plant material and 1,000 for extracts.

Let me be plain about the edge of what I know. I inspected restaurants for twelve years and have never plated a cannabis sample. I cannot tell you how one laboratory's culture-based count compares to another's qPCR result on the same jar, a live methodological argument I am not qualified to settle. What I can tell you is what people do with a document that says "passed." They read it as a property of the flower. It is a property of a sample, on a date, under one state's rule.

California's approach shows why a count is not the whole story: those four species are pass-or-fail because for some people presence itself is the hazard. The case literature runs from a 1975 Annals of Internal Medicine report of Aspergillus fumigatus pneumonitis in an adolescent with chronic granulomatous disease to fatal aspergillosis in a marrow transplant recipient reported in Chest. Benedict, Thompson and Jackson analyzed 2016 insurance claims covering more than 27 million people for Emerging Infectious Diseases and found people with a cannabis-use code 3.5 times more likely to carry a fungal infection diagnosis, confidence interval 2.6 to 4.8. Heating does not reliably solve it; combustion temperature at the coal is not a validated kill step for what rides in the smoke.

Why the six-to-twelve-month calendar fails

A date records duration. It records nothing about exposure, and exposure is what changes the flower.

Twelve years of inspections beat this habit into me. The danger zone does not reset when you put something back in the cooler; it is a total, adding up across every trip food takes through it. Storage works the same way. Every hour above 0.65 water activity counts, every hour of light counts, every opening that swaps terpene-saturated headspace for fresh oxygen counts.

Take two jars filled from the same harvest on the same day. One went into an opaque box at steady room temperature and was opened twice. The other rode in a bag, sat on a lit shelf, and was opened nightly. At nine months they share a date and nothing else. A shelf-life chart prices both at the average and is wrong about both.

And even a perfect potency prediction would say nothing about mold, which is the only part of this that can hurt you.

Flower, concentrates and edibles do not age the same way

The word "pot" gets applied to three products that fail in three unrelated ways.

| | Flower | Concentrates (rosin, shatter, distillate) | Edibles (gummies, chocolate, baked goods) | |---|---|---|---| | Moisture benchmark | 0.55–0.65 water activity, per ASTM D8197 | Effectively no plant water | 0.85–0.92 water activity, where food rules draw lines | | Main deterioration path | Light-driven cannabinoid loss, oxidation to CBN, terpene evaporation | Oxidation, light, terpene loss, texture change and darkening | Fat rancidity, moisture migration, staling, plus cannabinoid loss | | Microbial limit applied | 10,000 CFU/g in several states; four Aspergillus species absent in 1 g in California and Colorado | 1,000 CFU/g for extracts under the AHP monograph | Food safety rules, not cannabis limits | | What "old" looks like | Brittle, faded smell, harsh smoke | Darker color, stiffer texture, flatter aroma | The food has gone off; separately, it is weaker |

The comparison that matters is the moisture one. Flower is held under 0.65 water activity because its failure mode is fungal. Food sits in a different regime: the FDA Food Code treats a non-heat-treated food as needing temperature control at or above 0.88 water activity, the level at which Staphylococcus aureus can produce toxin, and 21 CFR 114.3 draws the acidified and low-acid scope line at 0.85. Those are bacterial thresholds. A gummy and a jar of flower are not on the same scale, and a rule fitted to one will actively mislead you about the other. Concentrates sit outside both, carrying the tightest microbial limit because a processed product should not show a live count at all, and losing quality to oxidation rather than humidity.

Set the storage up once, then leave it alone

Until about 2018 I gave cannabis the same answer I gave restaurant operators for twelve years: when in doubt, get it cold. I stopped, for the same reason the freezer advice keeps circulating. In a wet food, cold suppresses bacterial growth in abundant water. In a dried plant product the controlling variable is water activity, and every trip a container makes from a 38 °F refrigerator to a 70 °F room deposits condensation if it is opened before it warms. You pay a moisture penalty for a temperature benefit Fairbairn's data says is small below 20 °C. Cold works for concentrates. For flower it works only if the container is never opened cold, which in practice means never.

1. Identify the form and read the package, noting any humidity control included and any printed date.
2. Move it into a rigid, airtight container that light cannot reach, or keep that container inside something opaque.
3. Bring humidity into the 0.55 to 0.65 band with a two-way pack. Add control, not water.
4. Choose one location with a stable temperature and stop moving it. Cycling is worse than a slightly warm shelf.
5. Portion for the week so the main container opens rarely. Grind at use, not in advance.
6. Look and smell before every use, knowing what those senses can and cannot report.

Step three has a common and dangerous shortcut. A slice of orange peel or a piece of bread in the jar to soften dry flower is a humidifier with no ceiling, able to drive water activity past 0.68 in a day. Same shape of error as rinsing a raw chicken: the fix relocates the problem somewhere worse.

When to discard, and when to keep

The strongest argument against me is that I am asking people to bin expensive product over a risk nobody has cleanly measured in healthy adults, and it has force. The published cases are dominated by profoundly immunocompromised patients. The Benedict analysis is a claims-code study blind to route of use and product testing, and the Society of Cannabis Clinicians noted in 2025 that its non-tobacco-smoking cannabis subset came to 24 people out of 53,217. No cohort shows ordinary healthy adults falling ill from year-old flower at a measurable rate. I grant all of it.

What I do not grant is the conclusion drawn from it, that a date is therefore adequate evidence. My position was never that you should fear mold. It is that you should stop letting a calendar stand in for an inspection it cannot perform.

Discard, whoever you are, if you see fuzz, webbing, or a gray or white powdery deposit that moves differently from trichomes under a loupe. Discard on a musty basement smell or an ammonia note, or if the package was ever wet or kept somewhere humid for weeks.

Discard on a far lower threshold if you or anyone sharing it is immunocompromised, on chemotherapy, a transplant recipient, living with cystic fibrosis or asthma, or carrying lung damage. For that group unresolved uncertainty is itself sufficient reason, and vaporizing instead of smoking does not resolve it.

Keep it, with lowered expectations, if it is merely dry, brittle, grassy and harsh and you can account for where it has been.

The honest gray zone is a jar with no visible growth, a history you cannot reconstruct, or a package that was crushed or left open. If you are healthy, that is your call, and I will not pretend anyone has quantified the risk for you. If you are not, treat an unknown history the way a laboratory treats an unlabeled sample: as a fail.

Questions people actually ask

Does cannabis flower go bad over time?

It degrades rather than spoils on a fixed date. Potency falls roughly 16.6% in the first year at room temperature, and terpenes drop far faster, over 50% within a month by one 2020 measurement. Going genuinely bad means mold, which depends on moisture and packaging, not age.

Does cannabis have a shelf life?

Not a fixed one. Fairbairn's 1976 storage work found carefully prepared cannabis reasonably stable for one to two years in the dark at room temperature. That is a chemical stability window, not a safety expiry. Poor storage shortens it dramatically, and good storage extends usable quality well beyond it.

How long does cannabis affect the body?

The National Institute on Drug Abuse reports that noticeable effects of smoked cannabis generally last one to three hours, with heart rate elevated for up to three hours. Edibles take longer to begin and last considerably longer, often six hours or more. Storage age affects strength, not duration.

How can I tell if my weed has mold?

Look for fuzzy white or gray growth, fine webbing between buds, or powder that does not glitter like trichomes under magnification. Smell for musty basement or ammonia notes. These are late-stage signs. Absence of them does not clear flower stored above 65% humidity.

Do concentrates and edibles last longer than flower?

Concentrates hold up longer against microbes because they contain almost no water, though they oxidize and lose terpenes. Edibles are food and follow food rules, spoiling through rancidity and moisture migration. Regulators reflect this: extracts face a 1,000 CFU/g limit against flower's 10,000.

Should I still smoke it if the package was torn open?

Treat a breached package as an unknown storage history rather than a cosmetic problem. Air, light and ambient humidity have been reaching the flower for an unknown period. Inspect it closely, and if you are immunocompromised or cannot date the damage, discard it rather than guess.