The Nutrition Geeks investigation

I sent two rounds of supplements from the UK's fastest-growing supplement brand to accredited third-party labs. The ashwagandha measured less than 1% of its claimed active compound. The milk thistle measured roughly half. Then came the public statement, the legal threat, and a question that goes far beyond one brand: who decides how a supplement gets tested? Here's the whole thing, with every number and every source.

The brand and the question

Nutrition Geeks have been one of the fastest-growing supplement companies in the UK for the last few years - selling millions of pouches, going viral on TikTok, and according to their own website serving over 4 million customers. Their magnesium is the best-selling supplement on Amazon UK.

Their prices are very low. Which made me ask the obvious question: is this too good to be true?

To answer it, I bought their products at retail like any customer, and sent multiple batches to independent labs to determine whether what's on the label is actually in the capsules.

The email trail

Before testing anything, I emailed Nutrition Geeks as a customer and asked for their third-party test results.

They replied that they don't have third-party tests for their products, saying that Eurofins - one of the world's leading labs, and one of the labs I used - takes up to 12 weeks to test. In my experience that is not accurate: my results from Eurofins came back in under 30 days. They said they were working towards third-party testing batches in the future, and that for now all they could send was basic COAs and microbiological tests.

A microbiological test is technically a third-party test, but all it tells you is that the product won't make you ill. It's the bare minimum. It doesn't tell you whether the product is what the label says it is, or whether the dose is right.

• December 2025First email asking for third-party tests. Response: none available, microbiological COAs only. (screenshot)
• 1 December 2025First batch of four products shipped to the labs.
• April 2026Asked again for third-party tests. Same response. (screenshot)
• Mid 2026First video published; passed 200,000 views. Nutrition Geeks released a public statement disputing the testing methods, and threatened to sue.
• 16 July 2026I sent Nutrition Geeks specific questions about the magnesium maths, their choice of lab and method, and the in-house ashwagandha method.
• 22 August 2026As of filming the response video: no reply received.

What I tested, and how

Four products went to the labs: Magnesium 3-in-1, Milk Thistle, Lion's Mane and Ashwagandha (KSM-66). Testing was done through Eurofins and Alkemist Labs - Alkemist being an industry leader for botanical testing, based in California.

The chain of custody matters, so here's how it works: products are purchased as a normal retail customer would buy them, and shipped to the lab with the manufacturer's seal still intact. The lab documents this - including photographic evidence of the unopened packets - before testing. The only way to trick that system would be for the lab itself to fake results.

Where a validated pharmacopoeial method exists, that's what was used: the USP method for ashwagandha withanolides, and the pharmacopoeial HPLC method for milk thistle silymarin. A pharmacopoeia is the official rulebook for medicines and supplement testing - written by government-appointed scientists, legally recognised, and it specifies the exact test you have to use to measure a compound. Tying third-party testing to a pharmacopoeial method removes the argument about method choice. That's exactly why I used them.

Why "active compounds" and not just the big number on the front

Many supplements use extract ratios. A 10:1 extract ratio theoretically means 10kg of raw material was used to make 1kg of extract - which lets a brand print 1,000mg on the label even though there's only 100mg of extract in the capsule. This is 100% legal in the UK.

Nutrition Geeks' lion's mane says 4,000mg. But you cannot fit 4,000mg in a tablet that size - you can fit 267mg of a 15:1 extract. For context, a capsule that size usually holds 500-700mg of actual material. In the USA, advertising the big number that way isn't permitted, because it's so easy to misread.

What actually matters is the specification - the active compound inside the plant that carries the claimed benefit. Silymarin in milk thistle. Withanolides in ashwagandha. That's what I tested wherever possible, because that's what you're paying for.

The results

Both batches, all four products, active compound measured against what the label implies:

Ashwagandha: the one that made me call the lab

The ashwagandha uses KSM-66 - the most popular ashwagandha ingredient in the world, sold by a supplier that has built its brand on quality. Seeing it on the label was a green flag.

Which is why the result shocked me. The label describes 250mg of a 12:1 extract at 5% withanolides - so a serving should carry about 12.5mg of withanolides. The lab measured 0.118mg. The second batch measured 0.1mg. That's a fraction of one percent of the implied claim - so far off that I got the lab on a video call to make sure it wasn't a mistake:

"Yes, we are sure. We stand by our results. We have a very, very amazing team of chemists back there, double and triple-checking work over and over again - and yes, we are sure of those results." Elan Sudberg · CEO, Alkemist Labs (degree in chemistry), on the ashwagandha and milk thistle results

Worth adding: 250mg of an ashwagandha extract is a fairly weak dose to begin with, in my opinion.

Milk thistle: roughly half

Milk thistle is taken for liver support, and the compound with the research behind it is silymarin. The label describes a 100mg extract at 80% specification - so about 80mg of silymarin per serving. Batch 1 measured 43.5mg; batch 2 measured 45mg. Both roughly half the implied claim.

Lion's mane: over the claim - with a caveat

Standardised to 50% polysaccharides, 267mg of extract should carry 133.5mg; the lab measured 159mg, which is over the claim. Credit where due. (For mushroom supplements, beta-glucans are a better quality marker than polysaccharides - but that's a topic for another day.)

The magnesium label maths

The magnesium result is the interesting one, because it passed the lab test - 385mg measured against 384mg claimed - and yet the label still doesn't add up.

Magnesium supplements are never pure magnesium: the mineral is bound to a carrier - glycinate, malate, citrate - that helps your body absorb it. Each form has a chemical maximum for how much elemental magnesium it can carry: roughly 13% for glycinate, 15% for malate, 16% for citrate. That's not an opinion, it's chemistry.

Run the numbers on the three forms listed on this label and the maximum possible elemental magnesium comes to about 254mg. The label claims 384mg. Something has to be making up the missing ~130mg - and it isn't any of the forms declared on the label.

In the industry this is called buffering: quietly topping up with a form like magnesium oxide, which carries a high elemental percentage but is poorly absorbed compared to the chelated forms - so the label number looks great, while a chunk of the dose does much less for you. To be clear about what the evidence shows: the total elemental magnesium matched the label. What the maths shows is that it cannot all be coming from the forms printed on the label. I put this question to Nutrition Geeks on 16 July 2026; as of 22 August, no response.

But why should you listen to me on this? I'm not a PhD, and I haven't studied magnesium for 40+ years. So I took the question to someone who has: Prof. Dr. Jürgen Vormann, who has published around 280 studies on magnesium since 1983 and is the former president of the German Society for Magnesium Research. I asked him directly whether these limits are real - whether it's a chemistry fact that you can only get so much magnesium into citrate, malate or glycinate - and what it means when a label hits a number those forms can't reach:

Me: "Why is that? Is it a chemistry fact that you can only get so much magnesium into magnesium citrate, malate or glycinate?"

Prof. Vormann: "Well, that's simple chemistry - the citrate anion is larger than the magnesium; roughly only about 15% of the whole is magnesium. You can't change it. You cannot fill more magnesium into the citrate. You can't do this."

Me: "In the industry we know a lot of brands will put a higher number on the label… You send it to a lab and it does test at that much elemental magnesium - but we know from industry insiders that they're adding oxide to the tablet and then not putting it on the label."

Prof. Vormann: "This is, I would say, of course not correct. If you claim it's magnesium citrate, the magnesium should only come from the citrate and not from the oxide. This can't work. You can't fool the chemistry - this is not possible."

Prof. Dr. Jürgen Vormann · ~280 published studies on magnesium since 1983; former president of the German Society for Magnesium Research · full interview on the channel

In other words: the forms printed on this label cannot carry the dose printed on this label. The magnesium is real - the lab confirmed the total - but some of it has to be coming from a form that isn't declared.

Double-checking everything

I could have published after the first round. Instead, I gave the benefit of the doubt: maybe these were just bad batches. So I bought a second round of products - two units of each this time, keeping one of each sealed and unopened in case I ever need to prove anything later - and sent them off.

The second batch results are in the chart above: ashwagandha still at a fraction of a percent of its claim, milk thistle still around half, magnesium within range of its label total.

And if you're wondering how you know I didn't swap the capsules before shipping: the labs document whether the manufacturer's seal is intact on arrival, with photographs. The packets were unopened.

Nutrition Geeks respond

After the first video passed 200,000 views, Nutrition Geeks published a public statement saying they had consulted industry experts "who believe that inappropriate testing methods were used", provided their own lab reports, and threatened to sue me if I didn't take the video down.

They don't say who those experts are, and none appear to have come forward publicly - they would be welcome to. Here's their response, point by point, with what I found when I dug into it:

On magnesium: nothing

Their statement doesn't mention magnesium at all. The question about the label maths - how a 384mg claim can come from forms that max out around 254mg - remains open and unanswered.

On milk thistle: a different batch, an unaccredited lab, and a method argument

• The lab report they provided is for a completely different batch from the two I tested - so it doesn't actually answer my results.
• The lab they used is not ISO 17025 accredited - their own website says the accreditation application is in process. ISO 17025 is the international standard under which a lab is independently audited, method by method, for technical competence.
• They argue the HPLC method I used under-reports compared to the UV method they use. But HPLC is the method specified for silymarin by the UK, US and European pharmacopoeias. UV is not. More on what the actual experts say below.

On ashwagandha: a same-batch report, and a secret method

Here they did provide a third-party test for the same batch I tested - with very different results. How can the same batch produce very different numbers? The answer is the method.

KSM-66's supplier issued a declaration stating that they don't claim their product meets specification under the USP method, that their product contains different types of withanolides than those the USP method measures, and that they validate withanolides using their own in-house method.

So on 16 July 2026 we asked Nutrition Geeks to share that in-house method, so we could rerun the tests with it and - most importantly - understand what it actually records. From what we can tell, the method appears to be shared only with labs under non-disclosure agreements. As of 22 August: no reply. The offer stands - if they share the method, I will rerun the tests and publish the results openly, even if they invalidate mine.

It's normal for ingredient suppliers to develop in-house methods. But those methods should be transparent and open to scrutiny, the way the USP method is. Both lab reports Nutrition Geeks provided are what's known as black-box reports: they state a number without showing the peaks measured to reach it. Mine show the peaks - you can see exactly which withanolides and which silymarin compounds were identified. A black-box report is a little like saying "trust us, we tested it" without showing your workings.

What the experts say

I'm not a PhD. So rather than argue method choice myself, I put the questions to people who've spent their careers on exactly this chemistry. Full interviews are on the channel.

On UV vs HPLC for silymarin

A plain-English way to understand the difference. UV testing shines light through a sample and measures the "glow" - but silymarin isn't the only thing that glows, so everything similar gets added into the number. HPLC first separates the mixture into individual compounds, so you measure just the silymarin itself.

"UV spectrometry is a quite simple and cheap method but it has some limitations… some other compounds that have similar absorption to silymarin compounds will also be measured together with silymarin. This is some kind of very severe bias, which makes the amount that you measure with UV usually higher or much higher than what you measure by HPLC… [HPLC] is what the European Pharmacopoeia and other modern pharmacopoeias are using for the quality control of silymarin." Prof. Dezső Csupor · one of Hungary's foremost authorities on medicinal plants; pioneered supplement quality testing in Hungary

"The one that is more in line with technology in 2026 is the HPLC method… So yeah, I'm with you. And I disagree with that point." Prof. Nicholas H. Oberlies, PhD · medicinal chemistry & pharmacognosy; has studied milk thistle chemistry for decades - responding to the claim that UV, not HPLC, is the right method

On the industry's testing games

"The first way is lab shopping - send the same sample to different labs and pick the result you want. The second is method… Some companies have their own proprietary method for strength, which is fine under the law - but you have to disclose the method and it has to be replicable. What you often get instead is: 'You can't have our method, and you can't replicate it.' That's a problem… Anyone talking about 5% withanolides - I'd ask to see the COAs, and I'd ask to see the methodology. The 5% claim is always the one I want to ask more questions about." Blake Ebersole · co-founder, Ashwagandha Standards Alliance - speaking about the industry generally, not about any specific brand

A wider pattern

The withanolides disagreement isn't new, and it isn't unique to this brand:

• ConsumerLab, 2024: tested 13 ashwagandha products using the same USP method I used. Only 5 passed. One failing KSM-66 brand made the same argument Nutrition Geeks made - wrong method. ConsumerLab's answer was essentially mine: the USP method is the current published standard, and the supplier's own method wasn't peer-reviewed, published or validated.
• Nootropics Depot - a US brand with its own in-house lab - removed withanolide claims from its KSM-66 product after its own testing. The founder wrote publicly that they compared their data with Alkemist Labs - the same lab I used - "and it essentially matched."
• Several major brands - MyProtein, Wild Nutrition, NOW Foods, Holland & Barrett, Nutravita - sell KSM-66 products without advertising a withanolide content on the label, despite the ingredient historically being marketed at 5% withanolides.
• KSM-66's own website today makes no mention of 5% withanolides that I could find - but the 2017 archived version shows "up to 5% withanolide" on their specification sheet.

Prompted by all this, I tested a handful more ashwagandha products with the USP method: two more KSM-66 brands (both measured very low withanolides, consistent with the Nutrition Geeks results) and three non-KSM-66 brands - two of which measured significantly more withanolides, with only a gummy failing.

Draw your own conclusions - and to be clear about mine: KSM-66 has a large body of research behind it (much of it supplier-funded), and I am not claiming their product is bad, or that Nutrition Geeks or KSM-66 are gaming their tests. I'm saying that without transparent, replicable testing methods, nobody can know. That's the whole problem.

Corrections and open offers

Standing offers, all open as of publication:

• To Nutrition Geeks and KSM-66: share the in-house withanolide method, and I'll rerun the tests with it and publish the results - whatever they show.
• To the unnamed experts who advised that my methods were inappropriate: come forward publicly and make the case.
• To anyone at all: buy my supplements, or any brand's, send them to an independent lab, and check the results yourself. Every batch that leaves my facility is tested before it ships - the reports are in a public database, searchable by batch number.

One more thing worth saying plainly: I don't think Nutrition Geeks are deliberately underdosing their supplements. The most likely explanation is that the ingredients they buy are underdosed on arrival - and a brand of their size should be testing ingredients and finished products to catch exactly that. There is currently no law in the UK requiring supplement brands to third-party test anything. Most don't. That's the actual scandal.

What this costs

In the last three months I've spent £35,191.76 on third-party testing - both my own supplements and other brands' for these videos. Testing a product costs a couple of hundred pounds. A brand doing six-figure batch runs has no cost excuse not to test every batch.

The bigger cost is legal. Every video now comes with an "URGENT LEGAL LETTER" and a take-it-down-in-24-hours threat, and each one means thousands in lawyers' fees to answer properly. Publishing honest third-party testing should be encouraged by this industry, not litigated against. That's the path I've chosen anyway.

More tests are in the works: vitamin D, protein, omega-3, creatine and more over the coming months. Subscribe on YouTube to catch them, or join the email list below.

Watch both videos

Documents and sources

1. StatementNutrition Geeks: "A message to our customers" - their public response, in full
2. Their CoAMilk thistle lab report provided by Nutrition Geeks (different batch; unaccredited lab; UV method)
3. Their CoAAshwagandha lab report provided by Nutrition Geeks (same batch; peaks not identified)
4. SupplierKSM-66 declaration statement on USP method applicability
5. SupplierKSM-66 quality control page ("we welcome third-party verification")
6. ArchiveKSM-66 specification sheet, 2017 (Wayback Machine - "up to 5% withanolide")
7. EmailNutrition Geeks reply, December 2025 - no third-party tests available
8. EmailNutrition Geeks reply, April 2026 - same response
9. RecordFSA food hygiene rating check - no rating found at time of filming
10. 3rd partyConsumerLab ashwagandha review, 2024 - 5 of 13 products passed the USP method
11. IndustryNootropics Depot founder on KSM-66 withanolide testing (Reddit)
12. StandardsAshwagandha Standards Alliance - Blake Ebersole's organisation
13. DatabaseNF Supplements public batch-testing database - my own brand's reports, searchable by batch number
14. LabAlkemist Labs and Eurofins Scientific - the laboratories used

My own lab reports for every result quoted above are shown in the videos; the full certificate PDFs are available on request via the contact page.