The Shortage List Is the Business Model

For roughly three years, one of the fastest-growing businesses in American healthcare rested on a single row in an FDA web page. Not on a patent, not on a clinical trial, not on a manufacturing breakthrough. On the presence of the word semaglutide in the agency's drug shortage database. When that row was deleted in February 2025, the business model did not adapt. It changed species. What has happened since, through the warning letters of March 2026 and the FDA's April clarification, is the cleanest case study in recent memory of a market mistaking a regulatory artifact for a durable franchise.

Start with the mechanics, because almost everyone in the debate gets them wrong. The FD&C Act gives compounders two legal lanes. Under section 503A, a state-licensed pharmacy may compound a drug for an individual patient on receipt of a prescription. Under section 503B, an outsourcing facility may compound in larger batches, without patient names, under current good manufacturing practice. Both lanes carry a restriction that matters more than any other: they may not regularly compound what are essentially copies of commercially available, FDA-approved drugs. Wegovy and Zepbound are commercially available and FDA-approved. On its face, that should have been the end of the compounded GLP-1 industry before it began.

The shortage list is the exception that swallowed the rule. When a drug appears on FDA's shortage database, the copying restriction lifts, because the statute presumes the commercially available product is not actually available. A compounding pharmacy filling a semaglutide prescription during the shortage was not exploiting a loophole in the colloquial sense. It was operating inside an explicit statutory carve-out, one that exists for good reason: patients should not go without a medicine because a manufacturer cannot make enough of it.

But a carve-out written for continuity of care turned out to be something else when paired with direct-to-consumer telehealth. The shortage made compounded semaglutide legal. Telehealth made it scalable. And the price gap made it irresistible: a branded pen at a four-figure list price against a compounded vial at a fraction of it. An entire industry, platforms, pharmacies, provider networks, marketing agencies, grew up inside that gap. Hims and Hers was still confident enough in the model to launch a compounded Wegovy at $49 a month in February of this year, nearly a year after the shortage ended, and was rewarded with Novo Nordisk's first US patent suit against a compounder within days.

Note what the industry was actually selling. Not semaglutide. Semaglutide was available from Novo the whole time, at a price. What the shortage list sold was permission. The list was the moat, the margin, and the business plan, and it belonged to the FDA.

The delisting played out with the inevitability of a tide table. Tirzepatide came off the list in late 2024, semaglutide followed in February 2025, and the Outsourcing Facilities Association sued to argue the shortages were not really over. The district court denied the association's preliminary injunction on tirzepatide in March 2025 and on semaglutide in April 2025. FDA wound down its enforcement discretion in stages, and by May 22, 2025, the grace period for the last category of compounders had closed. There was no ambiguity left to hide in.

What replaced the shortage is a doctrine, and it is worth reading in full because the entire surviving industry now stands on it. FDA's position, restated in April 2026, is that a compounded drug is essentially a copy of a commercial product if it contains the same active ingredient in the same, similar, or an easily substitutable strength by the same route of administration, unless, and this is the sentence that launched a thousand funnels, "a prescriber determines and documents" that the compounded product contains a change producing "a significant difference" for an identified individual patient.

That clause is what the market now calls personalized dosing. Read it again, though, and notice who it makes responsible. Not the platform. Not the marketing agency. Not the patient clicking through a three-minute intake quiz. The prescriber. A licensed clinician must determine, and must document, that this specific patient requires a difference from the approved product that is clinically significant. A different strength because the patient is titrating more slowly. A formulation without an excipient the patient cannot tolerate. A documented clinical reason, in the chart, per patient, defensible to a state board. The statute did not create a dosing menu. It created a standard of care, and it put a named professional's license behind every order.

The gap between that legal substance and the marketing claim is precisely where the enforcement has gone. In March 2026, FDA sent thirty warning letters to telehealth companies, and the violations it chose to cite are telling: not the existence of compounding, but the marketing of it, claims implying sameness with the approved products, and branding that obscured who actually made the drug, as if the telehealth company itself were the compounder. Commissioner Makary framed it as a new era of attention to direct-to-consumer claims. Novo, meanwhile, has moved the fight from the shortage list to patent and trademark law, where the shortage is no defense at all. The regulatory boundary did not disappear when the shortage ended. It moved from a database row to the advertising copy, the label, and the prescriber's chart note.

Three consequences follow, and none of them are widely priced in.

First, the shortage era is not coming back in its old form. Even if semaglutide or tirzepatide were to relist, the agency has now shown its enforcement playbook: defined wind-down windows, no tolerance for indefinite discretion, and a federal court that was unimpressed the first time. Any business plan that assumes the list reopens quietly and stays open is planning to lose.

Second, the surviving legal lane is narrower than its marketing suggests, and its width is set by documentation quality rather than by statute. A personalized-dosing program in which the significant-difference determination is real, individualized, and recorded is practicing medicine within the compounding exemptions. A program in which the determination is a checkbox generated by the same software that took the payment is manufacturing unapproved drugs with extra steps. The difference between those two programs is invisible in the advertisement and total in the medical record, which is why the medical record is where regulators and plaintiffs will look.

Third, the compliance surface has moved from the pharmacy to the clinical layer. During the shortage, the load-bearing artifact was the batch record at the outsourcing facility. Now it is the prescriber's note: who reviewed this patient, what difference was determined, where it is documented, and whether a thousand such notes would survive being read side by side. Platforms that were built as marketing machines with a prescription pad attached will find this a hostile environment. The ones built to make provider review real, auditable, and boring will find it a tailwind.

There is a broader lesson here, and it extends well past GLP-1s. Regulated markets periodically produce these windows, an artifact of the rules that temporarily permits what the rules generally forbid. The window always looks like an industry while it is open. The businesses that confuse the window for the industry tend to be very profitable right up to the day the window closes, and then they are case studies. The shortage list was never the product. It was never even the business. It was the boundary, and the businesses built to respect the boundary are the only ones still standing on the right side of it.