For roughly three years, the specialty pharmacy channel had one problem with GLP-1s, and it was a simple one: it could not get any. Then, between October 2024 and May 2025, the shortage closed. Product arrived. And the channel discovered that the thing it had been waiting for had, in the meantime, stopped being the thing that mattered.
Here is the thesis. The end of the GLP-1 shortage is remembered as a compounding story — the regulatory event that shut down the 503A and 503B copies. That reading is too small. The more consequential thing that happened is that the manufacturers used the eighteen months around the shortage's resolution to build direct-to-patient distribution that routes around the specialty pharmacy stack entirely. Specialty pharmacy won the supply war and lost the distribution war, in the same window, and largely without noticing.
I think this is the most under-covered structural shift of the era. It gets missed because it looks like good news at every individual step: more supply, more manufacturing capacity, lower cash prices. Each of those is genuinely good. Together they describe a channel being disintermediated.
The minimum background. The FDA declared the tirzepatide shortage resolved on October 2, 2024, then reaffirmed it on December 19, 2024 after a challenge from the Outsourcing Facilities Association, giving 503A pharmacies until February 18, 2025 and 503B outsourcing facilities until March 19, 2025. Semaglutide followed on February 21, 2025, with 503A and 503B deadlines of April 22 and May 22, 2025. We covered the enforcement side of that in the compounded GLP-1 crackdown. This essay is about what happened to the legitimate channel while everyone was watching the illegitimate one die.
The shortage was never a chemistry problem. It was a filling problem
The first thing to get right is what was actually scarce.
It was not the peptide. It was sterile fill-finish capacity — the industrial step of putting a drug into a pen or a vial under aseptic conditions. That is why the shortage took years rather than quarters to clear. You cannot buy your way out of a fill-finish constraint with a purchase order. You buy your way out with plants.
And so both companies bought plants. Novo Holdings acquired Catalent in a $16.5 billion transaction announced in February 2024 and closed in December 2024, with Novo Nordisk taking three fill-finish sites — Anagni, Bloomington and Brussels — for $11 billion. Eli Lilly announced $27 billion for four new US manufacturing sites in February 2025, on top of a Lebanon, Indiana facility already carrying a $9 billion price tag.
Note the timing. Lilly's $27 billion announcement came six weeks before the semaglutide shortage was declared over. These were not reactions to scarcity. They were bets on a post-scarcity market — and a company that has committed $27 billion to capacity has an extremely strong incentive to control where the output goes.
A manufacturer that solves its own supply constraint stops being a supplier and starts being a distributor. That transition is the whole story.
The channel that solved cold chain lost to the channel that solved the credit card
Specialty pharmacy exists because some drugs are hard. High cost, refrigeration, limited distribution networks, REMS requirements, prior authorization labor, adherence monitoring, injection training. GLP-1s check nearly every box. On paper, this class should have been the specialty channel's defining category.
Watch what the manufacturers built instead.
LillyDirect launched in January 2024. Self-pay Zepbound single-dose vials arrived in August 2024 at $399 and $549 a month, then were cut to $349 and $499 in February 2025 — the same month the semaglutide shortage was closing. Novo answered in March 2025 with NovoCare Pharmacy, offering Wegovy at $499 cash across all doses. We took that pricing apart in the LillyDirect $499 strategy, decoded.
Line those dates up against the shortage resolutions and the sequence is unmistakable. The direct channels were stood up in the exact months supply normalized. That is not a coincidence, and it was never really a promotion. It was a land grab executed at the one moment a manufacturer had the inventory to sustain it.
The most underrated number here is $499, not any of the shortage dates
Here is why that figure matters more than the regulatory calendar.
Every function specialty pharmacy performs is downstream of a single fact: a third party is paying. Prior authorization exists because a payer demands it. Step therapy exists because a formulary imposes it. Benefit investigation, copay assistance coordination, appeals management — all of it is labor generated by the presence of an intermediary. Remove the payer and you remove the reason the channel exists.
A $499 cash price does not make coverage irrelevant. It makes coverage optional for a large enough slice of patients to change the negotiating table. And it does so at a number close enough to plausible net-of-rebate pricing that the entire benefit-design apparatus starts to look like friction rather than value.
The system pushed back, as systems do. CVS Caremark made Wegovy the preferred GLP-1 on its standard commercial template formulary from July 1, 2025, excluding Zepbound — a demonstration that a PBM can still move enormous volume by fiat. And Novo's partnership with Hims & Hers, announced in April 2025, was terminated by Novo on June 23, 2025, which read to me as a manufacturer discovering it did not want a distribution partner so much as a distribution channel it owned.
Both events point the same way: everyone in the middle is now negotiating for the right to remain in the middle.
What the specialty channel actually kept
I want to avoid overstating the collapse, because specialty pharmacy did not lose everything, and what it retained is instructive.
It kept the medically complex. Semaglutide's cardiovascular indication, approved in March 2024 on the strength of the SELECT trial's 20% relative reduction in major adverse cardiovascular events, and tirzepatide's obstructive sleep apnea indication approved in December 2024, both create populations where the prescription sits inside a broader clinical picture that a cash checkout page cannot manage. It kept anything running through commercial insurance at scale, which is still most of the volume. And it kept the institutional relationships — health systems, employer plans, and the newer Medicare structures we examined at the dispensing window.
But look at that list and the common thread is unflattering. Specialty pharmacy retained the business that is complicated because of administration, not because of the drug. That is a real business. It is not a defensible one, because administration is precisely what a manufacturer with its own pharmacy and its own logistics can simplify away.
Meanwhile the technical moat is thinning. The oral era — oral semaglutide, and whatever orforglipron becomes — removes refrigeration, removes injection counseling, and removes the cold-chain logistics that were the most concrete argument for a specialized channel in the first place. A pill in a bottle is not a specialty product. It is a product.
Let me be explicit about scope. This is market and field analysis. Ozemback does not recommend, rank, or refer anyone to any pharmacy, platform, manufacturer program, clinic, or prescriber, and nothing here is guidance about any individual's treatment — that belongs to a person and their licensed clinician. The subject is a distribution channel, not a decision.
The strongest case that I am wrong
Three objections, and the second is the one that could sink the argument.
One: direct cash channels remain small next to insured volume. Most GLP-1 prescriptions in the United States still adjudicate through a plan, a PBM and a network pharmacy. Declaring the specialty channel disintermediated on the strength of a self-pay program is reading a trend line as a finish line.
Two: the direct channels are defensive pricing, not a business model. $349 and $499 exist to blunt compounded competition and to give the manufacturers a floor in formulary negotiations. Once compounding is fully suppressed and coverage broadens, the rational move is to let those channels quietly wither and return to the profitable insured book.
Three: specialty pharmacy is not a channel at all. It is a bundle of capabilities — cold chain, adherence, prior authorization labor, patient support. Capabilities do not get disintermediated. They get re-hired by whoever wins.
My read: the first objection is factually correct and rhetorically weak, because the question is direction, not present share. The second is genuinely strong, and I would take it more seriously if the direct channels were marketing spend. They are not. Owned pharmacy operations, owned telehealth relationships and owned fulfillment are capital and organizational commitments, and companies do not build those in order to dismantle them three years later. The third objection is right — and it is also the concession I am making. The capabilities survive. What does not survive is the independent economics of the firms that currently sell them.
What actually ended in May 2025
So here is the plain version.
The end of the GLP-1 shortage is filed in most coverage as a return to normal: supply restored, gray market closed, system functioning again. I think that is exactly backwards. Scarcity was the condition under which the old distribution map survived, because when nobody can get product, nobody questions who is standing in the middle of the pipe. Abundance is what forced the question.
What ended in May 2025 was not a shortage. It was the last period in which a manufacturer needed the existing channel more than the channel needed the manufacturer. Everything since — the cash pricing, the owned pharmacies, the terminated partnerships, the $27 billion in plants — is the working-out of that inversion.
And the class is still growing. Which means the interesting question for the next two years is not whether patients can get these drugs. It is who they get them from, and how many participants are left standing between the plant and the person.
Ozemback — August 2026