Cannabis Got Roughly Ten Times Stronger Since the 1990s. Nobody Rewrote the Warning Label to Match.

Ask most professional football players privately why they use cannabis, and the answer isn’t rebellion. It’s pain, sleep, and getting through a Tuesday practice on a body that’s absorbed hits it was never built for — and until recently, saying so out loud could cost a career, not just a reputation. A new study funded by the NFL and NFLPA, published in Sports Health, found elite contact-sport athletes use cannabis deliberately for recreation, therapy, and recovery, but still navigate a fog of misinformation and inconsistent team policy. The authors’ recommendation was blunt: leagues should replace abstinence-only messaging with honest, harm-reduction-oriented education. It’s good advice, and almost nobody in a position to fund cannabis science at scale has caught up to the product actually being sold.

Cannabis got roughly ten times stronger since the 1990s, and nobody rewrote the warning label to match — being honest about that is not the same thing as telling people to stop.

Start with the number everyone in the industry knows and almost no consumer does: average flower potency has climbed from roughly 3-4% THC in the 1990s to 15-25% today, and that’s before you get to concentrates — wax, shatter, rosin — running 60-90% THC, a category that didn’t meaningfully exist for consumers a generation ago. A systematic review of high-potency cannabis and health outcomes models roughly one additional first-time treatment-seeker per 100,000 people for every three-percentage-point rise in average THC potency. That’s not a moral judgment about cannabis. It’s a dose-response relationship, the same kind researchers track for alcohol and opioids, and it’s been mostly absent from the public conversation about legalization because the conversation got stuck on “legal or not” a decade before potency data caught up.

The biology behind the psychosis link just got more specific

Cannabis’s association with psychosis risk isn’t new, but it’s mostly been epidemiology — bigger studies of more people, showing a correlation without a clean mechanism. That changed some this month. A new imaging study published in Molecular Psychiatry used a brain-scanning technique called magnetic resonance spectroscopy — think of it as a chemistry readout of a specific brain region rather than just a picture — to measure glutamate, a signaling chemical, in the anterior cingulate cortex of 79 people across a spectrum from healthy controls to those already experiencing psychosis. The finding: lower glutamate in that region and higher cannabis use were each independently tied to worse psychosis symptoms, but the two effects didn’t just add up separately — glutamate levels only predicted symptom severity in the people who used cannabis. In non-users, the same glutamate measurement told researchers nothing. Cannabis use appears to be switching on a specific biological pathway that, without it, stays quiet.

That finding lands alongside a major cross-methodology review in The Lancet Psychiatry, which pooled epidemiological, genetic, experimental, and animal evidence and concluded that daily use and cannabis use disorder in the mid-teens-to-mid-20s age range carry elevated rates of psychosis, bipolar disorder, anxiety, depression, and suicidal behavior — and that the evidence across those very different research methods now coheres toward cannabis playing a causal role, not just riding alongside some other explanation. If you’re the person these studies are actually about — someone who started young, uses daily, and has wondered whether the mood swings or the paranoid stretch last month had anything to do with what you were smoking — this is the research saying: probably, yes, and it’s not a personal failing, it’s a known biological pathway with your specific pattern of use sitting right in the middle of it.

We still can’t measure what “too high to drive” means

Compare that biological specificity to a much older, much simpler safety question the field still hasn’t solved: how do you tell if someone is too impaired by cannabis to drive? Alcohol has a number — blood alcohol concentration, understood by every driver, enforced at every traffic stop. Cannabis doesn’t. A recent review of the science of cannabis-impaired driving confirms there is still no validated, court-ready impairment standard, because blood THC levels simply don’t reliably predict real-world impairment — THC can linger in blood long after intoxication has worn off, or barely register in someone acutely impaired by a fresh, high-potency dose. This isn’t a minor technical gap. It’s the same kind of honest-information failure the NFL study flagged for athletes, just aimed at everyone with a driver’s license: a decades-old promise that legalization would come with real safety infrastructure, still unmet for the single most common real-world risk question people actually ask.

Alcohol has a number — blood alcohol concentration, understood by every driver, enforced at every traffic stop.

The market isn’t waiting for the science to catch up

None of this is slowing normalization down. Georgia’s newly expanded medical cannabis law took effect July 1 and patient enrollment jumped 22% in about seven weeks — from roughly 34,000 to more than 41,500 — partly because patients can now legally vaporize their medicine for the first time, and the state replaced its old 5%-potency cap with a 12,000-milligram possession limit that says nothing about potency at all. Canada’s mature legal market posted CA$2.353 billion in cannabis sales through the first five months of 2026, still climbing year over year. And a striking generational shift is underway alongside it: roughly 40% of Gen Z drinkers also use cannabis or THC products, and more than 60% of that overlapping group say cannabis directly affects how much they drink — a real substitution effect, not just parallel use, that’s reshaping what “harm reduction” even means for a 24-year-old choosing between a joint and a fourth beer.

Here’s the part that should complicate any easy “kids these days” framing: teen cannabis use itself keeps declining, even in states that legalized years ago — Colorado’s past-year teen use fell from 21% in 2015 to 10% in 2025; Minnesota’s combined 8th/9th/11th-grade rate dropped from 14.9% to 6.3% over roughly the same window. The risk this piece is describing isn’t more teenagers using. It’s a smaller, specific population using a dramatically more concentrated product, with a research base that’s still catching up to what that product actually is — and a testing infrastructure that isn’t always trustworthy even when it exists: New York regulators recalled 54 product lots this year after a contract lab was found to have misreported mold-contamination results, meaning some consumers were told their product was safe when nobody had actually verified that.

This isn’t the first time the science lagged the shelf on purpose. For decades, the only cannabis American researchers were legally permitted to study came from one federal contract farm at the University of Mississippi — a single, low-potency supply that held a research monopoly from the 1960s until protocols finally began loosening in the 2020s. An entire generation of the addiction and psychiatric literature — the studies still cited in today’s policy debates, in courtrooms, in insurance denials — was built on a product that had already stopped resembling what anyone could actually buy. The glutamate study, the impairment test that doesn’t exist, the mold recall: each one is a fresh version of the same old failure, honest information arriving years behind the product it’s supposed to describe.

If you use concentrates, or you’re the parent of someone who does, the useful takeaway isn’t fear — it’s specificity. A gummy, a joint, and a dab are not the same exposure, in the same way a beer and a shot of grain alcohol aren’t the same exposure, and treating them as interchangeable is exactly the information gap this piece is about. Ask what the potency actually is before you ask whether it’s legal. Legality was never the safety question.