How a Word From Biochemistry Became a Hashtag

It is a little after eleven at night and the video starts mid-sentence. A man in a hooded top sits in a parked car, phone propped on the dashboard, the interior light giving his face the colour of candle wax, and in the first twenty seconds he says the word four times. Peptides. He gets round to naming one only later, and when he does it is a string of letters and digits that sounds like a number plate. He does not say what it was tested in. He says what it did for him, and the counter beneath the clip says it has been watched a little under two million times. He is a composite. There are thousands of him. But the shape of the clip is exact.

So why are peptides trending? A word that spent a century as neutral chemistry, naming a size of molecule and a kind of chemical bond and nothing else, picked up borrowed prestige from the most talked-about medicines of the decade. It then landed in a medium that rewards exactly what the word does well: sounding precise while committing to nothing. By 2026 it had become a genre of content in its own right, and medicines regulators on three continents spent the spring responding to what that content was attached to. This piece is about the word's journey and those responses. Whether any particular compound does anything is a separate question, and the feed, for structural reasons, cannot answer it.

A hundred quiet years

For most of its life the word was furniture. It was coined at the start of the twentieth century to describe a chain of amino acids joined by the bond that holds proteins together, and for the next hundred years it did that job and no other. It turned up in textbooks, in the methods sections of papers, and on the order forms of laboratories that built short sequences to specification. It carried no promise. A peptide could be a hormone, a venom, a fragment of a digested meal or an inert scrap cut from a larger protein. The word said how a molecule was built and nothing about what it did.

That silence is what everything else in this story depends on. Most words that become marketing words arrive already loaded, as "detox", "boost" and "clean" did, and the copywriter's job is to aim them. "Peptide" arrived empty. It had the sound of expertise, with the hard consonants and Greek root of a term you would only use if you knew what you meant, but no content anyone could contradict. "This is a peptide" is a true statement about an enormous number of molecules, and most of them do nothing interesting whatsoever.

The medicines that taught the public the word

The prestige came from elsewhere, and quickly. Between roughly 2021 and 2024 a class of prescription medicines for type 2 diabetes and obesity became the most discussed drugs in the world, in newspapers, on earnings calls and in the queue at the chemist. Semaglutide and tirzepatide are peptides: modified amino-acid chains designed to imitate hormones released by the gut. For the first time the word was attached, in ordinary conversation, to medicines that visibly changed people's bodies and were famous enough to run short.

What passed to the public in that moment was not information. Hardly anyone who met the word through those headlines learned what it means. What they learned was an association: peptide, the kind of thing that works. The name of a whole category had borrowed the reputation of its two most famous members, much as "biotech" once borrowed the reputation of the few companies in it that made money. That reputation was then on offer to every other member of the category, including thousands of sequences that had never been given to a person under controlled conditions and, in most cases, never would be.

What the feed rewards

The medium those molecules met was not neutral either. One of the largest studies of how information travels on a social platform, published in 2018, followed around 126,000 stories shared by some three million people. It found that false claims reached more people, faster and further, than true ones, and that human sharing rather than automated accounts drove the difference. The authors' best explanation was novelty: false stories were more surprising, and surprise is what people pass on 2. Health is no exception. A systematic review of health misinformation on social media found it recurring across every topic the included studies examined, and especially common in posts about drugs and tobacco products and about vaccines 3.

Short-form video puts its own weight on top of that. A clip is ranked by whether people keep watching, and what keeps them watching is a face, a first person, a confession and a result. Each of those is the opposite of what a finding needs before it can be interpreted. A face is a sample of one. A first person has no control group. A confession is not blinded. A result reported by someone who wanted it cannot be separated from the wanting. None of this requires anyone to be dishonest. Most creators believe what they say, and that is why it works: on a small screen, sincerity reads as credibility far more easily than a confidence interval does.

And the word fits the format like a key fits a lock. A clip cannot spell out a sequence without losing its audience, but it can say "peptides". It cannot explain what an animal model is, but it can say "studies show". The noun survives the trip and the qualifiers fall off. A viewer who knows no chemistry comes away knowing one thing: there is a category, it is scientific, and the people in the videos are in on it.

What researchers found when they counted

In 2026 a group of orthopaedic researchers did the unglamorous thing and counted. They searched YouTube, Instagram, TikTok, Reddit and X for posts about peptides and musculoskeletal health and analysed 239 of them, which between them had been viewed 10,087,381 times 1. The most-mentioned compound was BPC-157, a sequence with a large animal literature and no completed controlled trial in people, and posts often paired it with TB-500 1.

The authors found that influencer content, especially on the two most visual platforms, made the most dramatic claims with the least support 1. Their figures are a snapshot rather than a census: one search, one period, one field of medicine, published as a conference abstract rather than a full paper. But the proportions tell the story. The sentence "this might hurt you" appeared in about one post in eight. The sentence "this fixed me" was so common it barely needed counting.

Research use only, worn as a costume

The trend had a supply problem, and the feed solved it. Most of the compounds in these videos are not licensed medicines in Britain, the European Union or the United States. They circulate as laboratory reagents under a label, for research use only, which was written for a different market and says in effect that the substance is for experiments and has not been assessed for anything else. The label is legitimate and necessary, because science needs a route to compounds that are not medicines. On a feed, though, it does a second job. It lets a seller say nothing about use while a creator one swipe away says everything.

Regulators noticed this division of labour. When the UK medicines regulator responded to press reporting on clinics offering experimental peptide injections, its spokesperson summed up the position in one sentence: "We disregard claims that products are for 'research purposes' if it is clear that such claims are being used as an attempt to avoid medicines regulations" 4. The legal reason is simple. In British law a medicinal product includes any substance presented as having properties for treating or preventing disease in people 7. Presentation is enough. A reagent stays a reagent until someone attaches a promise about the human body to it, and after that the law no longer cares what the label says.

Three regulators, one spring

In about seven weeks of 2026, medicines authorities in three countries acted on different parts of the same phenomenon. They were not coordinated and they were not doing the same thing, which is what makes them worth comparing.

The British action was about words. In April, after a newspaper investigation found clinics advertising experimental peptides with claims about ageing and recovery from injury, the MHRA said it was looking into them. It made clear that research-use framing would not shield a product sold with a medicinal promise 4. The feed's vocabulary of healing, repairing and protecting is, in British law, the vocabulary that turns a reagent into an unlicensed medicine 7.

The American action went the other way. On 22 April the FDA updated its page listing bulk substances that may present significant safety risks in pharmacy compounding, the list the trade calls Category 2, and a group of peptides, among them BPC-157 and TB-500, now sits in a withdrawn section 6. This kind of administrative change travels badly, because "off the FDA's risk list" compresses easily into "cleared by the FDA". The document says something much narrower. Category 2 concerns which substances compounding pharmacies may use while nominations are reviewed. Leaving it approves nothing, licenses nothing and says nothing about whether any of these molecules does what the videos claim. The same page still records the agency's stated concern about compounded peptides: a risk of immune reactions arising from aggregation and peptide-related impurities 6.

The South African action was about manufacture, and it involved the two most famous peptides of all. On 11 May inspectors from SAHPRA and the South African Pharmacy Council visited a pharmacy in Silverton, Pretoria, that was producing semaglutide and tirzepatide products under the heading of compounding. The regulators announced the action on 23 May. In South African law, compounding means preparing a medicine for an individual patient against a prescription; the regulators said this was manufacturing and marketing for broad commercial distribution, largely for weight management. The injectable stock was seized, a recall was ordered, and SAHPRA noted reports of adverse events including hospital admissions. The Pharmacy Council warned that pharmacy staff involved in such supply risked removal from the register 5.

Taken together, the three actions show where the trend's joints are. The British one targets the promise. The American one concerns the pharmacy counter through which a peptide might legitimately reach a patient. The South African one concerns the factory behind the counter. None was aimed at the feed itself, at the man in the parked car who is selling nothing, or says he is not. Regulators act on sellers, claims and premises. A hashtag has none of those, which is why it moves faster than they do.

Where the hashtag ends

It would be easy to end with a scold, and wrong. The people in these clips are mostly curious, often injured, sometimes desperate, and responding rationally to a medical system that had little to offer them for a torn tendon or a slow recovery. The word reached them because it seemed to offer a way round that system with the vocabulary of science attached. That is a cultural fact worth taking seriously rather than sneering at.

But the word's journey can be traced backwards, and doing so is instructive. On the feed it means a category of thing that works. In a pharmacology textbook it still means a chain of amino acids and nothing more. Between those two meanings sit the questions a hashtag cannot hold: which sequence, made by whom, at what purity, tested in what, compared against what, and for how long. Every one of them has an answer, and none fits in fifteen seconds. That is where a piece about culture has to stop, and where the chemistry begins: the vial, the certificate and the trial that was or was not run.