The compounded GLP-1 crackdown.

The word people keep using is ban. It is the wrong word. What happened to compounded GLP-1 drugs across 2025 and into 2026 was not a ban in any normal sense. It was the expiration of a temporary legal exception that had been sitting, in plain text, inside a 1997 statute the entire time — an exception that switched off the moment the conditions that created it disappeared.

That is the story I want to tell precisely, because the coverage has been imprecise. There is a real timeline, with real dates. And there are real losses, distributed in a way almost nobody describes accurately. The losses did not fall mainly on the drugmakers, who in fact won. They fell on a layer of telehealth intermediaries that had quietly become a multibillion-dollar industry, and on a patient population — well over a million Americans by most estimates — who had been paying roughly a tenth of the brand price.

My thesis is simple: this was the most predictable regulatory event of the decade, and the fact that so many sophisticated companies were caught flat-footed by it tells you more about the era than the rule change itself does.

The mechanism first, because it is the whole thing. Under sections 503A and 503B of the Federal Food, Drug, and Cosmetic Act, compounding pharmacies and outsourcing facilities are generally barred from producing copies of commercially available, FDA-approved drugs. There is one major exception: when a drug sits on the FDA's official shortage list, compounders may make what is essentially a copy to fill the gap. Semaglutide (Ozempic, Wegovy) and tirzepatide (Mounjaro, Zepbound) both went onto that list — tirzepatide in December 2022 — as demand vaporized supply. The exception opened. An entire industry rushed through it.

The timeline, with the dates that actually matter

The crackdown is not one event. It is a cascade of four.

First, tirzepatide. The FDA declared its shortage resolved in late 2024 and, after a legal challenge, reaffirmed that decision in December 2024. That started two clocks. State-licensed 503A pharmacies had until February 18, 2025 to stop compounding tirzepatide. The larger 503B outsourcing facilities — the industrial-scale players — had until March 19, 2025.

Then semaglutide, the bigger market by far. The FDA declared that shortage resolved on February 21, 2025. Same structure, slightly longer runway: 503A pharmacies had until April 22, 2025; 503B facilities until May 22, 2025. By the end of spring 2025, the legal basis for mass-compounding either molecule was simply gone.

The fourth and most decisive step came later, and the headlines undersold it. On April 30, 2026, the FDA proposed removing semaglutide, tirzepatide, and liraglutide from the 503B bulk drug substances list entirely. That move matters most. Ending a shortage exception leaves a door that could theoretically reopen if a new shortage appeared. Removing the molecules from the bulks list closes the door structurally — so that even a future shortage would not automatically reauthorize industrial compounding of these specific drugs. The temporary became permanent.

The legal fight the compounders were always going to lose

The industry did not go quietly. The Outsourcing Facilities Association sued the FDA over both the tirzepatide and semaglutide determinations, seeking preliminary injunctions to keep the shortage exceptions alive. The courts declined, in each case. The deadlines held.

I think the litigation was less a real legal strategy than a stalling tactic, and a revealing one. The compounders' strongest argument was never that the law entitled them to keep going — the statute is clear that the exception is tied to the shortage. Their argument was effectively economic: a lot of patients and a lot of revenue now depend on this, so the disruption should be slowed. That is a policy plea dressed as a legal claim, and courts treat those exactly the way you would expect.

Who actually lost — and it wasn't the drugmakers

Here is where the "actual losses" part gets interesting, because the intuitive read is backwards.

Novo Nordisk and Eli Lilly did not lose. They reclaimed a market that compounders had been serving at a fraction of the price. The clearest evidence is what each manufacturer did next: cash-pay channels. Brand semaglutide had been running near $1,799 a month at list for Ozempic and close to $1,999 for Wegovy, with compounded versions undercutting them at around $199. Once the exception closed, the manufacturers moved to occupy the low end themselves with direct cash programs priced far below list. They did not lose the patients. They repriced to keep them.

The layer that actually took the hit was telehealth. The cleanest illustration is Hims & Hers, which had become the public face of compounded semaglutide — offered in more than 30 states, with the company spending nearly $145 million in a single quarter marketing its drugs. By the end of 2024 Hims had around 100,000 GLP-1 members generating roughly $240 million in annualized revenue, and it had told investors to expect a $725 million weight-loss business in 2025.

Then came the most instructive single day in the whole saga. In late April 2025, Novo Nordisk announced a partnership to sell brand Wegovy through Hims. Two months later, on June 23, 2025, Novo terminated it — accusing Hims of "illegal mass compounding" dressed up as "personalization" and of deceptive marketing. Hims stock fell 34.63% in a single session, to $41.98. The intermediary that had ridden the exception up rode the crackdown down, in public, in one day.

The patient cost nobody wants to total

The other set of losers is harder to put on a chart, which is probably why the coverage skips it. For well over a million Americans, compounded semaglutide at roughly $199 a month had been the only version they could actually afford. The crackdown did not invent a cheaper brand option for them. It removed the cheap option and left a manufacturer cash-pay price that, while far below list, still sits well above $199 for many.

I want to be careful here, because this magazine does not tell anyone what to take or where to get it, and I am not doing that now. I am making a narrower, factual point: a price step exists, the crackdown widened it for a real population, and a clear-eyed accounting of the policy has to put that cost on the ledger — alongside the genuine safety case for closing the loophole.

The counter-argument, taken seriously

Because the counter-argument is strong. The case for the crackdown is not regulatory pedantry. Compounded drugs do not go through the FDA's approval process; quality is uneven and oversight is thin. Novo Nordisk's own investigation reported that much of the active ingredient in mass-compounded semaglutide was sourced from suppliers in China, outside the inspection regime that governs approved manufacturing. "Personalization" — adding a vitamin, nudging a dose — was, in many cases, a thin legal costume worn to keep mass-producing copies after the exception had closed. A regulator that let that continue indefinitely would be abdicating its core job.

So I do not think the FDA was wrong to close the door. I think the honest position is that the closure was correct and the costs were real, and that pretending one of those two things away — in either direction — is the move to distrust. The people cheering a pure victory and the people mourning a pure tragedy are both editing the ledger.

Why the surprise is the real story

Which brings me back to the thesis. The most underrated fact in this entire episode is that none of it was unpredictable. The exception was statutory, conditional, and explicitly temporary. It was always going to close when the shortages ended, and the shortages were always going to end once Novo and Lilly finished scaling manufacturing. A first-year regulatory analyst could have drawn this timeline in 2023.

And yet a multibillion-dollar industry organized itself as if the gap were permanent — building marketing engines, revenue guidance, and investor narratives on top of a legal exception with a sunset clause baked in. That is the part worth sitting with. The GLP-1 compounding boom was not killed by a surprise crackdown. It was killed by the expiration of exactly the condition that created it, on roughly the schedule anyone could have forecast. The losses were real. The surprise was the part that should never have happened.

That is the difference between a ban and a sunset. A ban is something done to you. A sunset is something you signed up for, in writing, and then chose to forget.

Ozemback — June 2026