Ashwagandha: what it is, how it works, and what research shows

What ashwagandha actually is

Ashwagandha is the root of Withania somnifera, a small shrub in the nightshade family, related to tomatoes and potatoes. It grows in India, the Middle East and parts of Africa, and it has a long history in Ayurvedic practice. That history tells you how it has been used, not what a capsule does today.

On the shelf, it comes in a few forms: dried root ground into powder, concentrated extracts of the root, and extracts that blend root and leaf. These are different products. An 8:1 extract, for example, is concentrated from about eight parts of plant, so 600 mg of extract and 600 mg of plain powder are not the same dose. Forms compared walks through the differences in more detail.

Labels deserve some skepticism. When researchers tested 24 commercial ashwagandha products, the measured withanolide content, the group of compounds usually used as a marker, differed from the label by a factor of 2 to 35. A printed percentage is only as good as the testing behind it.

Two clarifications help. "Adaptogen" is a marketing and herbalism term with no formal medical or regulatory definition. And ashwagandha is a dietary supplement, not an approved drug, and not a stimulant in the caffeine sense. If anything, drowsiness shows up among its reported side effects.

How it may work in the body

The mechanism most often discussed is the stress-hormone system. Cortisol is the body's main stress hormone, and several trials track it. A 2025 meta-analysis of 7 trials (488 participants) found ashwagandha lowered cortisol compared with placebo.

The same analysis is a useful reality check, though. A matched analysis of 6 trials found no significant effect on self-reported perceived stress. A 30-day trial in 60 healthy adults found lower cortisol at the lower dose only, while self-reported stress and mood improved just as much with placebo. A 60-day trial in 64 adults with chronic stress, using 600 mg/day of root extract, did find lower stress scores and cortisol than placebo. So the cortisol signal is fairly consistent, and the feeling-of-stress signal is less so.

Other hormones move too. In 57 overweight men aged 40 to 70, a low-dose extract (21 mg of withanolide glycosides a day) raised DHEA-S and testosterone but did not improve fatigue or vigor. In 50 adults with subclinical hypothyroidism, 600 mg/day for 8 weeks shifted TSH, T3 and T4 versus placebo. That last result is not a treatment claim. It is a reason people with thyroid conditions or thyroid medication should be careful.

Beyond these measurements, how ashwagandha produces any effect in people is not settled. Treat the explanations you read online as hypotheses.

What the evidence supports, and how firmly

Most human trials are small, short and often funded by extract manufacturers. Reviews repeatedly flag bias risk and inconsistency between studies. With that in mind, here is a rough map.

On anxiety, the pattern is encouraging but loose. Meta-analyses agree on direction, yet they describe the underlying trials as heterogeneous, small or preliminary, and none can name a dose that works.

On sleep, one trial used a root extract in 80 adults, half with diagnosed insomnia. Another gave 27 elderly men with insomnia 6 g/day of root powder for 4 weeks and found better self-reported sleep, but it compared them only with their own baseline. A before-and-after design cannot separate a real effect from expectation.

Blood pressure shows no clear effect either way. In 80 healthy adults, 600 mg/day for 8 weeks did not significantly change blood pressure. In 60 adults with hypertension and anxiety, anxiety and cortisol fell, but blood pressure was reported only as a favorable trend.

Where the evidence is thin, animal-only, or early

Ashwagandha shows up in a lot of research areas where the findings do not support any practical claim. A few examples, to keep expectations calibrated:

• Asthma and lung health: a 2026 review called the effect on lung-function tests "significant but inconsistent" without specifying the underlying trials. The one concrete study found was in mice, using an isolated compound called withaferin A.
• Bowel disease: a 2025 review lists it among plants with preclinical activity, with no ashwagandha-specific dose, and its authors say more research is needed.
• Parkinson's and Alzheimer's: the work described is in mouse models of a combined ashwagandha and CoQ10 formulation. There is no human trial of ashwagandha for either condition.
• Seizures: a review mentions a theoretical pathway but cites no trial testing ashwagandha on seizures.
• Cancer: a small phase 1 trial tested the safety of an investigational leaf-extract drug in advanced lung cancer, not whether it works. A non-randomized study in 100 breast cancer patients linked a 6 g/day extract to less fatigue, but with no significant survival difference.
• Hair and skin: a manufacturer-funded trial of 600 mg/day for 75 days found improved hair measures. Psoriasis findings come from mice and from seed oil, which is not a standard capsule ingredient.

There are also human findings that are real but narrow. A trial in 58 children with mild ADHD reported better symptom scores at 300 mg/day for 8 weeks, and a trial in 66 adults with schizophrenia found reduced symptoms when 1,000 mg/day was added to usual care. Each is a single trial, and neither is a basis for self-treating a diagnosed condition.

Doses and forms used in studies

There is no established effective dose. A 2026 review of 23 trials (2,317 people) looked at standardized root extract at 125 to 600 mg/day, and the most commonly cited human dose in single trials is 600 mg/day of root extract for 8 weeks or so. Studies of plain root powder use far more: 5 g/day for three months in infertile men, and 6 g/day in the insomnia trial above, both without placebo groups.

Because extract and powder are so different in concentration, a dose from one trial does not transfer cleanly to another product. Before comparing, check which form the study used and what your product contains. Most trials run between 4 and 12 weeks, so there is little to say about benefits or risks beyond a few months.

More on the topic is collected on the ashwagandha topic page.

Safety and who should be careful

For most healthy adults in trials, ashwagandha has been tolerated well. A review of 30 human trials of root preparations found no serious adverse events, with drowsiness, stomach discomfort and loose stools the most common mild effects. The 23-trial review of standardized extract reported mild stomach upset, headache and drowsiness infrequently, and about as often in placebo groups. In 18 healthy adults taking an 8:1 extract up to 1,250 mg/day for 30 days, one person withdrew for an adverse reaction.

The exceptions matter. Case reports describe liver injury judged probably linked to ashwagandha: a man taking 1,350 mg/day for 6 months, another case at 450 mg for 45 days, and a healthy 22-year-old whose injury appeared 30 hours after starting. The first two resolved after stopping. These cases are rare against the number of people who take it, but they are why anyone with liver disease, or who develops yellowing skin, dark urine or unusual fatigue, should stop and seek care.

Who it might suit, and how to approach it

The people most likely to get something from ashwagandha are healthy adults curious about modest changes in perceived stress, sleep or exercise recovery, who accept that the effect, if any, will be small. It is not a replacement for care of anxiety, insomnia, thyroid disease or diabetes.

If you try it, a few habits keep the experiment honest:

• Pick a product that names its form (root powder, or a named extract) and offers third-party testing.
• Start with one product, one dose and one goal, so you can tell what changed.
• Give it at least 4 to 8 weeks, which is roughly how long trials ran, and track something specific such as sleep or stress ratings.
• Stop if you notice stomach trouble, unusual drowsiness, palpitations or signs of liver trouble, and tell your clinician.
• Do not stack it with other mood- or sleep-affecting supplements without checking for overlap.

The honest summary is a supplement with a plausible signal on cortisol and some small, inconsistent benefits, a decent short-term safety record, and a long list of claims that outrun the evidence.