Ashwagandha is having its longest moment yet. It's in the gummies at the checkout, the "stress" pillar of every supplement stack, the TikTok routine of every founder who wakes up at 5am. The pitch is always the same — an ancient adaptogen that lowers cortisol, so you feel calmer.
The trials mostly agree with the first half of that sentence. It's the second half that turns out to be more complicated.
What ashwagandha actually is
Withania somnifera is a small shrub used in Ayurvedic medicine for centuries. The active compounds most studied are the withanolides — a family of steroidal lactones concentrated in the root, and in some extracts the leaf as well. Modern research almost never uses raw powder; it uses standardised extracts (KSM-66, Shoden, Sensoril and a handful of others) that specify a withanolide percentage, which is what makes trial doses comparable at all.
That matters because when a headline says "ashwagandha lowered cortisol by 30%", it means one specific extract at one specific dose over one specific timeframe. Generic "ashwagandha" on a supplement label can be almost anything.
What the trials actually measured
A 2025 systematic review and meta-analysis pooled 15 randomised controlled trials covering 873 adults. After eight weeks of supplementation the pooled results showed a statistically significant drop in morning cortisol, a drop in Hamilton Anxiety Rating Scale (HAM-A) scores, and a drop in Perceived Stress Scale (PSS) scores versus placebo. Quality-of-life scores did not improve significantly.
The doses used across these trials cluster around 300–600 mg per day of a standardised extract, taken for 6–12 weeks. Effects showed up on cortisol reasonably early — often within four weeks — and continued to widen through eight.
That's a real, replicated signal. For a supplement category where most ingredients fold the moment they hit a meta-analysis, ashwagandha holds up better than most.
The cortisol-vs-perception gap
Here's where the story gets more honest. A separate 2025 meta-analysis looked specifically at whether ashwagandha's cortisol effect translates into people feeling less stressed, and found the two don't always move together. Cortisol dropped clearly. Perceived stress, measured by asking participants how stressed they felt, moved much less consistently — in some analyses, not significantly at all.
That's not a scandal. It's a real feature of stress physiology. Cortisol is one signal in a system that also includes sleep quality, blood sugar, your relationships, your inbox, and whether you had time for lunch. Lowering the number on the lab report is not the same as changing what your life feels like.
The practical read: if you're expecting ashwagandha to make a hard week feel easy, that's asking more of it than the data supports. If you're using it as one input into a stress picture that also includes sleep, movement, and doing less, the odds it contributes something real are reasonably good.
How long does ashwagandha take to work?
Most of the positive trials ran for 8 weeks, with earlier signal on cortisol around the 4-week mark and clearer subjective effects (where they appeared at all) closer to week 8. A two-week experiment is genuinely too short to tell you anything. If you're going to try it, give it a proper block — and decide in advance what you're measuring, whether that's sleep onset, resting heart rate at wake, or a weekly stress check-in with yourself. "I feel a bit calmer, I think" after three weeks is not evidence; it's optimism.
Who probably shouldn't take it
Ashwagandha is generally well-tolerated in trials, but a growing number of case reports have linked commercial ashwagandha products to clinically apparent liver injury. The US National Institutes of Health's LiverTox database now classifies ashwagandha as a "likely" cause of liver injury (likelihood score B), with cases typically appearing 2–12 weeks after starting the supplement and presenting with jaundice, itching and a cholestatic or mixed injury pattern. Most cases resolve within 1–4 months of stopping, but rare fatal cases have occurred, particularly in people with pre-existing liver disease.
That doesn't make ashwagandha broadly dangerous — millions of doses are taken uneventfully — but it does make a few groups worth flagging:
| Group | Why it's worth pausing |
|---|---|
| People with liver disease or cirrhosis | Rare but serious hepatotoxicity risk; LiverTox specifically advises against use |
| Pregnant or trying to conceive | Traditional and modern sources both flag it as best avoided in pregnancy |
| On thyroid medication | Ashwagandha can nudge thyroid hormone levels, complicating dose management |
| On sedatives or immunosuppressants | Plausible interactions worth checking with your doctor |
| Anyone who develops jaundice, dark urine, itching or unexplained fatigue on it | Stop and get liver function tested |
The other honest caveat: because supplement quality varies, a product labelled "ashwagandha" is not automatically a standardised extract at a trial-relevant dose. If you're going to try it, a product that names its extract (KSM-66, Shoden, etc.), specifies withanolide percentage, and comes from a manufacturer with proper testing is worth paying more for.
The bottom line, without the hype
The evidence for ashwagandha is better than for most stress supplements and weaker than the marketing suggests. It moves a lab number reliably. It moves how you feel more modestly. It sits in a "worth a proper 8-week trial, if the safety caveats don't apply to you, and only alongside the boring inputs that do the real work" bracket — not in the "ancient miracle" bracket the ads want.
That's still a genuinely useful place for a supplement to sit. Just go in with the right expectation.
If you want to test ashwagandha honestly, commit to eight weeks at a properly standardised 300–600 mg dose, decide in advance what one thing you're measuring (sleep onset, resting heart rate, a weekly stress score), and treat it as one input into a stress picture — not the whole picture.
