The Grey Market That Grew in the Gap

On 12 August 2026 Eli Lilly announced six lawsuits against businesses it said were selling retatrutide — a molecule that was, at that moment, still in late-stage clinical trials and approved by no regulator anywhere. The defendants ranged from medical spas to online sellers. The company said it had referred more than two hundred individuals and entities to the Food and Drug Administration, and had reported more than 14,000 websites, advertisements, social-media posts and product listings that it said were unlawfully marketing the drug 1. An investigational medicine, in other words, had a retail market before it had a label.

This piece is about how that happened. It is not about counterfeiting, which is older than any of these molecules, nor about how to read a regulator's public notice; both are told elsewhere on this site. It is about economics: how a quiet trade in laboratory reagents and a loud market in wellness came to share the same sellers, what the scattered data say about its growth, and what regulators saw, and when. It describes a market. It endorses nothing sold in it, and it names no sellers.

Two markets, one shelf

The research-reagent trade is old, respectable and dull. Laboratories buy small quantities of synthetic peptides from catalogue suppliers for receptor assays, antibody work, method development and a hundred other purposes, and the label on such material — research use only — is an honest description of the customers it serves. Nobody writes magazine features about it, which is roughly how its participants like it.

The wellness market is newer, louder and far larger in its ambitions. It sells to individuals, through clinics, spas, influencers and websites, on the promise of changes to the body. For most of the past decade the two markets touched only at the edges. From the early 2020s they began to merge, and the reason was structural rather than conspiratorial: the same molecules, from the same contract manufacturers, could be sold into either market, and one of the two markets had far fewer rules.

The research-use label is what made the merger possible, and it is also what regulators came to reject. The label is a statement of intended use, and intended use is judged from everything around a product — its marketing, its instructions, the claims made for it. The US regulator's public guidance on unapproved GLP-1 products describes products 'falsely labeled' for research purposes or as not for human consumption being sold directly to consumers with dosing instructions, and warns that they are of unknown quality 2. Lilly's complaints made the same argument about retatrutide: sellers presenting it as a research product while plainly marketing it to be taken by their customers 1.

Three gaps

Grey markets do not grow at random. They grow in gaps — places where demand exists and the legitimate supply is absent, expensive or legally closed — and by 2026 there were three of them, overlapping.

Each gap feeds a different part of the trade. The price gap draws people who would buy the approved product if they could afford it; the argument over who pays for it is the subject of another article in this cluster. The access gap draws the former customers of the compounding era, whose supply was cut off when a shortage listing ended; that story, too, has its own article.

The approval gap is the strangest of the three. It is a market in molecules whose trial results were published, discussed and admired before any regulator had finished reviewing them, so that the publicity of success arrived years ahead of the product. A phase 3 readout that would once have interested only clinicians and investors now circulates within hours as a social-media clip. For the first time, a late-stage investigational drug had a consumer following large enough to support a trade in something sold under its name.

What the numbers can and cannot say

Nobody measures a grey market directly; that is close to the definition of one. What exist are shadows cast on instruments built for other purposes, and two of them are unusually informative.

The first is testing. Buyers in this market, distrustful of what they receive, send samples to independent analytical laboratories, and the volume of that testing is a rough proxy for the volume of trade. According to data one such laboratory in Central Europe shared with Bloomberg Businessweek, the number of peptide tests it processed rose by more than 1,200 per cent between 2023 and 2025 3. That is a single laboratory, and demand for testing can rise for reasons other than market growth. It is still an extraordinary number.

The second is payments. As banks and card networks declined to process these sales, a share of the trade moved onto cryptocurrency, where transactions are public even when identities are not. A blockchain-analytics firm estimated in June 2026 that this slice alone had reached about $100 million a year; that it had jumped 159 per cent quarter on quarter, to $32 million, in the first three months of 2026; and that it was on pace for $39 million in the second quarter 4. The same analysis found that independent testing per buyer had fallen by 88 per cent as the market went mainstream, and that a noticeable part of the new demand came from young people, some of them under age, drawn in through an appearance-focused online subculture 4.

Both figures are floors, not censuses. Neither counts trade conducted by card, bank transfer or cash, nor the clinics and spas that buy in bulk and sell by the appointment. Taken together they establish something narrower and more reliable: the market grew several-fold in three years, and as it grew, the share of buyers checking what they had bought went down rather than up. That second finding is the one a public-health official would underline.

What regulators saw, and when

Regulators saw it in pieces, and mostly from the supply side. In October 2025 the UK medicines regulator dismantled what it described as the first illicit production facility for weight-loss medicines found in the country: a warehouse outside Northampton holding tens of thousands of empty injection pens ready to be filled, raw chemical ingredients, and more than 2,000 finished pens labelled as retatrutide and tirzepatide awaiting dispatch. The agency believed it to be the largest seizure of trafficked weight-loss medicines ever recorded by a law-enforcement body anywhere 5.

In the United States, the regulator's attention through 2025 and early 2026 was concentrated on GLP-1 copies and on the research-use pretext described above 2. But in the same months it moved in the opposite direction on a different group of peptides. In April 2026 it announced that 12 peptides — among them several long popular in the wellness trade — would come off the category of bulk substances it had flagged in 2023 as raising significant safety concerns for compounding, after the nominations for them were withdrawn, and it scheduled advisory-committee meetings to consider whether they should join the list pharmacies may compound from 6. On 23 July 2026 that committee voted to back compounding of four of them. Its votes are advisory, and the final decision rests with the agency 7.

It is tempting to read that as inconsistency. It reads better as two different theories of the same gap. For GLP-1 copies, where an approved product exists and was getting cheaper, the regulator's approach was to close the gap by pushing buyers back toward the approved medicine. For peptides with no approved version at all, the direction of travel in 2026 was to ask whether some of the demand might be better met inside pharmacy regulation than outside it. Whether that shrinks the grey market or lends its vocabulary new respectability is one of the open questions of the year.

Where the story stops

The grey market will outlast any single enforcement action, because the gaps that feed it are features of how medicines are priced, approved and distributed rather than accidents. What remains is a measurement problem. How do you estimate the size and composition of a market designed to be invisible — from laboratory submission logs, blockchain flows, seizure records and complaint filings — and how much confidence can any such estimate bear? Every figure in this piece is a partial answer to that question. A full one is where a technical article would begin.