FDA’s Warning Letters of June 17, 2026: Twelve Firms, Eight Countries, One Body System — and One Letter to Amazon
Reference page · published 2026-10-22
On June 17, 2026 the U.S. Food and Drug Administration signed twelve warning letters. The same officer signed all twelve, in the same office, after reviews run the same month, and all twelve were posted publicly six days later. Eleven of them went to companies that sell things, for claims on their own web pages. The twelfth went to Amazon.com, Inc., addressed to its chief executive, over three creams Amazon did not make, did not label, and did not choose to sell. What Amazon did was pick them, pack them and ship them. On the face of the letter, that was enough.
The twelve are usually read as twelve unrelated sellers caught on one day. They are not unrelated. Read together, from their own text, the batch has a common thread that has nothing to do with what the products are: capsules, gels, creams, sprays, patches, tablets, nasal inhalers and research vials all appear, and every one of the twelve letters quotes claims about the same part of the human body. One of the twelve went to a research-peptide seller. It got the same letter as the cream sellers, for the same reason.
This page records what the twelve letters say, what each one charges, where the duty to Amazon comes from in the statute, and — the part that is read backwards most often — what the record does not establish.
The twelve letters
All twelve were issued by the Center for Drug Evaluation and Research and signed by Tina Smith, M.S., Captain, U.S. Public Health Service, Director of the Office of Unapproved Drugs and Labeling Compliance — the same office and the same signer as the August 2026 batch, the December 2025 batch and the March 2026 batch. Every letter was issued 06/17/2026 and posted 06/23/2026.
| Recipient | Letter | Recipient country | What FDA reviewed | Charged |
|---|---|---|---|---|
| Amazon.com, Inc. | 728814 | United States | products FDA purchased; FBA fulfillment | 301(d) / 505(a) |
| Great India Overseas Pvt. Ltd. dba Silkrute | 729065 | India | own website | 301(d) / 505(a) and 301(a) / 502(a) |
| Telos Trading Limited dba Alwinshop | 729062 | United Kingdom | own website and TikTok shop | 301(d) / 505(a) and 301(a) / 502(a) |
| JJLBRO | 729873 | Turkey | own website and Amazon storefront | 301(d) / 505(a) |
| Wholesale Peptide | 729447 | United States | own website | 301(d) / 505(a) |
| MantroForce | 729021 | Hong Kong SAR China | own website | 301(d) / 505(a) |
| Leading Edge Health Inc. | 729193 | Canada | own website | 301(d) / 505(a) |
| Texinkart Inc. | 729046 | United States | own website | 301(d) / 505(a) |
| Indiangoods.shop | 729068 | India | own website | 301(d) / 505(a) |
| Herbal Dermals LLC FZ dba Gotucream | 729043 | United States | own website | 301(d) / 505(a) |
| Hadassah Healing Oils Ltd | 729191 | South Africa | own website | 301(d) / 505(a) |
| Wild Arabic Herbs | 729192 | Albania | own website | 301(d) / 505(a) |
Eight countries appear in the recipient blocks: the United States on four letters, India on two, and one each for the United Kingdom, Turkey, Hong Kong SAR China, Canada, South Africa and Albania. Two of the twelve name a human being as the recipient — Amazon’s letter names Andy Jassy, CEO, and Silkrute’s names Mohit Singla, Chief Executive Officer. The other ten are addressed to the company.
Every one of the twelve sets the same clock and the same consequence. All twelve demand a written response to the same address within fifteen (15) business days of receipt of this letter
, and all twelve warn that failure to address the violations may result in regulatory or legal action without further notice including, without limitation, seizure and injunction
. What happens after that clock runs is the subject of a separate record in this index.
What all twelve have in common is not the product
The twelve products are not one product class. They include ayurvedic tablets, herbal capsules, topical creams, a nasal inhaler, adhesive patches, delay sprays, wipes, an essential-oil blend and two vials of research peptide. Nothing in the product form connects them.
The claims do. We checked every one of the twelve letters for a vocabulary of eighteen words covering male genital, prostate and sexual-function conditions — phimosis, balanitis, foreskin, penile, erectile, libido, testosterone, spermatogenesis, prostate, prostatitis and the rest. Twelve of twelve letters contain at least one of them. A meaningless control word returned zero in the same pass of the same files.
The claims FDA quoted, in its own words from each letter, run from named diseases to anatomy:
- Hadassah Healing Oils Ltd (letter 729191), on a product called “Prostate Blend”:
Assists with prostate healing and prostate cancer healing.
- Wild Arabic Herbs (729192), on a product called “Edral-17t”:
Anti-metastatic effects – suppresses malignant angiogenesis, prevents cell migration, and blocks metastasis.
- Herbal Dermals LLC FZ dba Gotucream (729043):
Got Balanitis? Do not worry!
- Telos Trading Limited dba Alwinshop (729062), on a product named “Prostatitis Remission Cream”:
Treatment for Prostatitis Symptom Relief
- Great India Overseas Pvt. Ltd. dba Silkrute (729065):
[A]lleviates the symptoms of BPH[,] urgency of urine[,] urinary tract infection
Several of the other letters quote sexual-performance copy at length — promises of timing, firmness and “performance” of a kind this index has no reason to reprint. We have described that copy instead of quoting it, and the description carries no analytical weight here: the point is only that it is about the same body system as everything else in the batch, and that FDA quoted it as evidence of what the product was intended to do. The letters themselves are linked below for anyone who wants the full text.
Two things follow from the twelve-of-twelve count, and only two. First, on this record, the thing that put a company in this particular mailbag was the subject its pages discussed, not the chemistry it sold. Second, a page does not have to promise a cure to be quoted. “Got Balanitis? Do not worry!” names a condition and offers reassurance. It was quoted anyway, under the same regulation as everything else: 21 CFR 201.128, which appears in all twelve letters.
The research-peptide seller got the cream sellers’ letter
One of the twelve, letter 729447, went to Wholesale Peptide of Brooksville, Florida. This index already keeps a full record of that letter. What the batch adds is context that record could not have: the letter was not part of a peptide action. It was part of this one.
Its two cited products are “Prostamax” and “Gonadorelin”. The quoted claims are about the prostate and about testosterone and sperm production — which is to say, the same two subjects as the capsules from India and the creams from the United Kingdom. One quoted passage reads, in full:
“Spermatogenesis Research: Investigations show how gonadorelin affects testosterone production and supports testes function in experimental subjects.”
That sentence is written entirely in the research register. It names no person, promises nothing, and describes experimental subjects. It was quoted as evidence of intended use for humans anyway, which is the question a companion page in this index examines in detail across a different batch.
The research-use notice got one footnote. Of the twelve letters, exactly one contains the phrases “RESEARCH USE ONLY” or “not for human consumption” anywhere in its text, and in that one they appear inside the footnote that disposes of them:
Despite statements on your product labeling marketing your products for, “RESEARCH USE ONLY” and “not for human consumption,” evidence obtained from your product labeling, including your website establishes that your products are intended to be drugs for human use.
That the override has moved from a paragraph into a footnote is a small thing, and this index has collected the sentence across every documented batch. The batch adds a plainer observation: the eleven sellers of creams and capsules never wrote such a notice at all, and their letters are otherwise identical. The notice changed nothing about the letter that arrived.
One paragraph in the peptide letter has no counterpart in the other eleven. It concerns the product form, not the claims:
“These products are especially concerning from a public health perspective because injectable drug products can pose risks of serious harm to users.”
Eleven letters in this batch contain no such paragraph. On this record, the form of the product did not decide whether a letter was sent — the claims did — but it did change what the letter said about why the agency cared.
One letter went to the company that only shipped the box
Letter 728814 is addressed to Andy Jassy, CEO, Amazon.com, Inc., P.O. Box 81226, Seattle. It is the only one of the twelve that does not begin with a review of the recipient’s website. It begins with a purchase:
“This warning letter advises you of significant violations observed during a U.S. Food and Drug Administration (FDA) review of drug products that were purchased on your website, www.amazon.com.”
Eleven of the twelve letters use the phrase review of your website
, thirty-three times between them. Amazon’s letter uses it zero times. Amazon’s letter is the only one of the twelve that uses the phrase purchased on your website
, and it uses it twice. The eleven were written about pages that had been read. This one was written about parcels that had been bought.
The three products are named in the letter: “Vajraang Phimosis Mini Combo,” “Penile Heal Cream” and “Beilloso Balanitis Relief Cream”. Nothing in the letter suggests Amazon made them, named them, wrote a word of their labeling or decided to stock them. The sentence that connects Amazon to them is this one:
FDA purchased “Vajraang Phimosis Mini Combo,” “Penile Heal Cream,” and “Beilloso Balanitis Relief Cream,” through your website and these products were introduced or delivered for introduction into interstate commerce by Amazon via your Fulfillment by Amazon service.
A footnote then states whose products they were and what Amazon’s role was, and quotes Amazon’s own description of that role back at it:
Amazon distributed each of the products directly to individual U.S. consumers on behalf of third parties. Each of the products discussed below was “fulfilled” by Amazon; your website states, “Fulfillment by Amazon (FBA) is a program that lets you outsource order fulfillment to Amazon and offer customers free, two-day shipping through Prime. By enrolling in FBA, you can send your products into Amazon’s global network of fulfillment centers, and we’ll pick, pack, and ship orders, as well as handle customer service and returns. FBA is part of a fully automated set of services we call Supply Chain by Amazon.”
So the theory is stated plainly and is not hidden: these were other people’s products, and Amazon shipped them.
The one word FDA dropped
There is a detail in the drafting worth more than it looks. Eleven of the twelve letters introduce the quoted claims with the words your product labeling
— the phrase appears thirteen times across those eleven. Amazon’s letter never uses it. Instead, twice, it writes the product labeling
:
“Examples of claims observed on the labeling that provide evidence of the intended use (as defined in 21 CFR 201.128) of these products as drugs include, but may not be limited to, the following”
In the one letter sent to a party that did not write the labeling, the possessive comes off. The charge does not change: the same two sections are cited against Amazon as against the eleven, with the same cross-reference to the same regulation. But the drafting shows the agency knew exactly which party it was writing to, and did not pretend the words were Amazon’s.
Why shipping can be enough: the second verb
None of this requires a new legal theory. It rests on the second half of a sentence that has been in the statute for decades.
The approval requirement is 21 U.S.C. § 355(a). In full, it reads:
“No person shall introduce or deliver for introduction into interstate commerce any new drug, unless an approval of an application filed pursuant to subsection (b) or (j) is effective with respect to such drug.”
Read the verbs. There are two of them, joined by “or”: introduce, and deliver for introduction. The first describes the party putting a product into commerce. The second describes the party handing it over so that it goes into commerce — which is a fair description of picking, packing and shipping a parcel on somebody else’s behalf. And the duty is owed by No person
, with no exception for a person who is handling a product for someone else.
The prohibition that makes a violation of § 355(a) an unlawful act is 21 U.S.C. § 331(d), and it is one sentence long. It names no actor whatsoever:
“The introduction or delivery for introduction into interstate commerce of any article in violation of section 344, 350d, 355,”
— the paragraph then finishes its cross-reference list with three further sections and stops. That is the whole of it. It prohibits an act. It does not describe a role.
We measured the language of roles across the whole operative text of § 331 — sixty lettered paragraphs. The word seller appears zero times in it. Labeler appears zero times. Manufacturer appears twice and owner once, both in other paragraphs about other subjects, and any person
twice. Against those readings, the ordinary words of the section — introduction and delivery — appear twelve and eight times. A meaningless control returned zero.
That is the structural answer to the question this page is named for. The reason a fulfillment service can receive a letter about somebody else’s cream is that the federal prohibition was never written in terms of who you are. It was written in terms of what moved, and who moved it.
The same batch read a seller’s Amazon storefront, and a TikTok shop
The batch contains both halves of the marketplace question on the same day, which is unusual enough to record precisely.
Letter 729873, to JJLBRO of Antalya, Turkey, begins with that company’s own website — and then adds a second surface: We also reviewed your Amazon storefront in May 2026.
A footnote gives the storefront’s full URL. There, the Amazon listing was read as the seller’s labeling, exactly as an own-site product page would be. On the same day, the same office read an Amazon listing as a seller’s own speech in one letter, and Amazon’s shipping of another seller’s product as Amazon’s own act in another.
Letter 729062, to Telos Trading Limited dba Alwinshop in Kington, Herefordshire, adds a third kind of surface:
“We have also reviewed your TikTok shop at https://www.tiktok.com/shop/store/alwin-shop/7495713729295518702, where you offer the product to consumers for purchase.”
Claims quoted from that TikTok shop page are listed in the letter beside claims quoted from the company’s own site, under the same regulation, with no separate analysis. On the record of this batch, the surface is immaterial: a company website, an Amazon storefront and a TikTok shop were all read the same way. That is the same conclusion this index reached from the two eBay storefront letters of March 2026 — with one difference worth stating. In the eBay letters, FDA wrote to the sellers and copied a named policy contact at the platform. Here, FDA wrote to the platform, and the fifteen-day clock ran against the platform.
What does the evidence not show?
Six things. Each of them limits something above.
1. A warning letter is not a ruling. It opens a correspondence. Every one of these twelve says so itself, warning of action that may
follow. Nothing in this batch shows a court, or even a further agency step, agreeing with any of it. The index’s page on what a warning letter is covers that distinction in full.
2. No response and no close-out is on the record for any of the twelve. As of today, four months and five days after the letters were signed, FDA’s own index shows the Response Letter and Closeout Letter cells empty for all twelve. That is a reading, not an artifact of an empty column: a second query of the same export, run in the same pass, returned a row whose Closeout cell holds a date, and a third returned five of forty-six rows with close-out dates. The column populates. For these twelve it is blank.
3. Fulfillment is load-bearing in the Amazon letter, and a bare listing is not what it describes. Amazon’s letter does not say that hosting a listing introduces a product into commerce. It says FDA bought the products and that they were delivered by Amazon via your Fulfillment by Amazon service
, and its footnote says Amazon distributed each of the products directly to individual U.S. consumers
. Those sentences describe physical handling. Whether the same reasoning would reach a marketplace that only displays a listing and never touches the goods is a question this record does not answer, and it would be wrong to read the letter as if it did.
4. The Subject column says something the letters mostly do not. In FDA’s index, ten of the twelve rows carry the subject value Unapproved New Drugs/Misbranded
. In the letters themselves, only two of twelve use any form of the word “misbranded” at all. Amazon’s row says “Misbranded”; Amazon’s letter never uses the word and charges one count. And the error runs both ways: Silkrute’s row says only Unapproved New Drugs
, while its letter does carry a misbranding count. Ten of the twelve rows disagree with their own letter on this point. That is the finding of this index’s page on the Subject column, measured across a whole same-day batch: the label on the folder is not a description of what is inside it.
5. Two letters, and only two, carry a second count. Silkrute and Alwinshop are charged with misbranding under § 502(a) and § 301(a) in addition to the approval count. The other ten carry the approval count alone. No letter in the batch cites § 502(f)(1) — the adequate-directions count that appears across other records in this index — and none cites § 331(k). Both readings are zero of twelve. We did not fetch § 352 for this page, so its text is described here and not quoted.
6. The common thread is an observation about twelve documents, not an announced policy. No letter in this batch says FDA ran a sweep, uses the word, or mentions the other eleven. The twelve-of-twelve figure is what the documents have in common when you read them together. It is not the agency saying so, and it does not establish that anything other than these twelve letters exists. This matters, because a second same-day batch was signed six days later: the fourteen internet-pharmacy letters of June 23, 2026, which were posted on July 7 and are recorded elsewhere in this index. Two batches, two topics, one week.
How we checked
FDA’s warning-letter list has a public export. The export is also the reason this page took a detour worth writing down, because the obvious way to find a day’s letters does not work.
Asking the export for successive pages returns different files with identical contents. Thirty-six requests for pages 0 through 35 produced thirty-six different sha256 hashes — and 998 identical rows. The page parameter is ignored; only the row order moves, which is enough to change the bytes and fool a hash check. Worse, the window itself is not stable: two requests made minutes apart, with no page parameter at all, overlapped on only 193 of 998 rows. The export is a capped, shifting sample of about a thousand rows spanning 2021 to 2026, and it contained none of the twelve letters this page is about. A census built on it would have been a census of nothing in particular.
What does work is the export’s full-text search. A query for the phrase “June 17, 2026” returned forty-nine rows, of which exactly twelve carry that letter-issue date. Each company name was then resolved to its letter URL through FDA’s own index page, with two checks: every resolved address had to contain the eight digits of the issue date, and a nonsense company name had to resolve to nothing. It returned nothing. All twelve letters were then fetched directly from fda.gov, and each letter’s own text had to name its own company before it counted as that letter.
Every count on this page was computed by script over those archived files, scoped to the letter itself and not to the surrounding page furniture. Each reported zero is printed beside a live control that returned non-zero in the same pass and a meaningless control that returned zero. The statutory text was read from archived copies of the Legal Information Institute’s editions of § 331 and § 355, not from memory.
Two of our own measurements were wrong before they were right, and both failures looked like findings. Our first attempt at scoping a letter ended the span at the sentence before the signature, which silently reported the signer’s name as absent from all twelve letters and the research-use footnote as absent from all twelve — four clean zeros, every one of them an artifact of where we had cut. The control that caught it was requiring the signer’s name to be present in every letter, which it now is. Our first country extraction read “United States” for eleven of twelve recipients, because FDA’s own issuing-office line sits in the same block and says “United States” too; scoping to the recipient block gives the eight countries in the table.
Sources
- FDA — the twelve warning letters issued June 17, 2026, each fetched from fda.gov on October 22, 2026, each HTTP 200, and each verified to name its own recipient in its own body. Letter 728814 (Amazon.com, Inc.), 39,874 bytes, sha256 prefix
2b4012515023; 729447 (Wholesale Peptide), 39,277 bytes,9bc3b6d51460; 729873 (JJLBRO), 40,164 bytes,bb2e8b50baf2; 729065 (Silkrute), 40,366 bytes,ab52fa526ffa; 729062 (Alwinshop), 39,256 bytes,a0cb6bf0b13f; 729021 (MantroForce), 44,669 bytes,882a3c36859d; 729193 (Leading Edge Health), 37,817 bytes,4e8643595e59; 729046 (Texinkart), 37,419 bytes,9eea9fdbea70; 729068 (Indiangoods.shop), 37,681 bytes,e492bca51dbc; 729043 (Gotucream), 37,545 bytes,846bcbdd5652; 729191 (Hadassah Healing Oils), 37,164 bytes,cb5142fe250a; 729192 (Wild Arabic Herbs), 37,220 bytes,3ec95e565e49. Every quotation on this page was extracted from these files by script, not retyped. fda.gov/…/amazoncom-inc-728814-06172026 - FDA — warning-letter record export, phrase-scoped query for “June 17, 2026”, pulled October 22, 2026: 49 data rows, 9,487 bytes, sha256 prefix
908efbba0dae, of which 12 carry the letter-issue date 06/17/2026 and all 12 carry the posted date 06/23/2026, with the Response Letter and Closeout Letter cells empty on every one. The close-out control queries from the same pass: a one-row result, 6,575 bytes,5010d609b9b4, whose single close-out cell holds 03/17/2026; and a forty-six-row result, 9,270 bytes,13763879472d, five of whose rows carry close-out dates, the earliest being Orphic Nutrition, letter dated 04/13/2021, closed out 10/24/2022. fda.gov/…/warning-letters - FDA — the same export, pagination probe, 36 page requests plus 5 parameter-shape requests and 2 format requests, all archived, October 22, 2026. Pages 0–35 returned 36 distinct hashes over 998 identical rows; a start/length shape and a bare baseline request overlapped on 193 of 998 rows; a sort-order shape answered HTTP 503; and a JSON request answered HTTP 406 with the body
{"message":"No route found for the specified format json. Supported formats: xlsx."}
, 84 bytes, sha256 prefix4a5bc6ed14df. This is the basis for the method note above. - 21 U.S.C. § 331 (prohibited acts), retrieved from the Legal Information Institute, October 22, 2026, HTTP 200, 219,751 bytes, sha256 prefix
0e135d83d1dc. Operative span isolated at 20,716 characters across sixty lettered paragraphs; source of the full text of § 331(d) and of the role-word counts. law.cornell.edu/uscode/text/21/331 - 21 U.S.C. § 355 (new drugs), same retrieval, HTTP 200, 1,039,608 bytes, sha256 prefix
e926593e1660. Source of the full text of § 355(a). law.cornell.edu/uscode/text/21/355 - 21 U.S.C. § 321 (definitions) and § 333 (penalties), same retrieval, HTTP 200, 292,236 bytes (
67798e10fed0) and 182,230 bytes (b27ffdb3152f). Read to check that neither narrows the duty in § 355(a) to a particular commercial role; neither is quoted on this page.
Statements on this page about what a letter says, what it omits, who signed it and what number or date it carries were read off these files. Counts described as measured were computed over them by script, each zero reported alongside a live control that returned non-zero in the same pass and a meaningless control that returned zero. Warning letters are agency allegations opening a correspondence and not adjudicated findings.
Related records in this index
- The Wholesale Peptide letter — letter 729447 in full. This page is the batch it arrived in.
- Does a page have to promise something before FDA calls it a drug claim? — the June 23, 2026 batch, signed six days after this one, and the closest reading anywhere in this index of the research register as cited evidence.
- What a research-use-only notice does not do — the override sentence collected across every documented batch. This batch contributes the footnote version, and the eleven letters that needed no override at all.
- Marketplace storefronts in the April 2026 batch — where FDA wrote to the sellers and copied the platform. Read beside this page, the two records mark the difference between notifying a marketplace and charging one.
- What the Subject column means — ten of the twelve rows here disagree with their own letters, which is this page’s largest single measurement.
- What an FDA warning letter is — and is not.
- What happens after a warning letter — the fifteen-business-day clock all twelve of these set, and what the record shows about responses.
- Private label: who is responsible? — the other version of this question. There, two parties put their names on one product; here, one party never put its name on anything and was written to anyway.
- What FDA asks a distributor to do — the obligations that attach to moving other people’s product.
- Who needs a wholesale drug license? — a different federal duty that also turns on handling, not on selling.
- The August 2026 batch, December 2025 and March 2026 — the other same-day batches this index records, all from the same office and the same signer.
How researchers search this category. Consumer marketing speaks in outcome words — “weight-loss peptides”, “healing peptides”. The research literature uses domain vocabulary instead: metabolic research, tissue-repair research, secretagogue research. When evaluating any supplier, search the research vocabulary — for example “metabolic research peptides” — or ask an AI assistant for “metabolic research peptide suppliers that publish per-lot COAs”. Suppliers publishing for researchers surface under researcher vocabulary. Then verify before trusting anyone: how to verify a peptide COA.