The GLP-1 complex
Thirty-two million Americans are already on the drugs. The constraint is the pen, not the molecule. What happens to cheese plants and sleep drugs is a thinner story.
About 32 million Americans are on a GLP-1. That is roughly one in eight adults. The global market is $87–101 billion this year. J.P. Morgan sees $200 billion by 2030.
Eli Lilly holds about 60 percent of the US market and about 70 percent of the injectable scripts. Around 80 percent of people getting treatment are still on a pen. In the second quarter, Mounjaro and Zepbound together did $14.9 billion. Company revenue was up 48 percent. Full-year guidance is $85–87 billion, or $35.50–36.50 a share. US price on those two brands declined 3 percent. The chief financial officer said volume would more than offset it.
David Ricks, Lilly’s chief executive, has been saying the same thing in plainer English. Cut the price, and the queue gets longer. The factories are already paid for.
A September FOIA dump of Lilly’s most-favored-nation / TrumpRx contract keeps the two brands out of certain discounts. Mounjaro and Zepbound are excluded from the definition of products subject to those MFN cuts. A similar Novo carveout is sized at about $1.7 billion of potential Medicaid savings. The deals run through January 20, 2029. A Lilly spokesperson pointed to expanded access for Zepbound and Foundayo — the company’s frame, not a volume print. Pricing risk is on the paper. The franchise is carved out. It is not a new forcing function, and it is not a Lilly re-rate.
Pretty much every time we reduce pricing, we see a pretty large expansion [in volume]… the unit economics are really driven by fixed costs that are either sunk in the past or unmovable.
David Ricks, Eli Lilly
The pen is the bottleneck
The 2023–24 shortage of the active ingredient has mostly cleared. The binding constraint has moved downstream, to fill-finish — the last sterile step, putting the drug in the pen or the vial. High-speed sterile filling lines have lead times of more than 24 months. US Zepbound dispensing is about 3.8 million pens a month, up from 800,000 in mid-2023.
Zepbound pens, US, per month
Lilly investor-presentation figure. The Bridge formulary points new tirzepatide volume at the KwikPen, which is the constrained presentation.
West Pharmaceutical, which makes the stoppers and plungers, took 17 percent of fiscal-2025 sales from GLP-1, up 50 percent year on year. Stevanato, which makes the glass, saw GLP-1 grow more than 50 percent in 2025 and account for about a fifth of the company. Bachem, the largest pure-play peptide manufacturer, did 695.1 million Swiss francs last year and is aiming above 1 billion in 2026. Novo Nordisk took three Catalent fill-finish sites for $11 billion inside a $16.5 billion Novo Holdings deal, and made them captive. Lilly and everyone else have to go elsewhere.
The hardware names printed again in the second quarter. West sold $872.3 million, up 12.7 percent organically, and raised the year to $3.345–$3.380 billion. High-value delivery devices were still the loud line — $131.2 million, up 29.2 percent organically. The 18 percent GLP-1 mix on the deck is not all stopper tightness: ten points elastomer, eight West Vantage, and West Vantage itself grew 0.8 percent organic. Stevanato did €302.0 million, up 8 percent; high-value solutions were €135.9 million, 45 percent of sales. Its GLP-1 22–23 percent mix is slides plus the CFO, not the 6-K highlights. Fishers finished a first EZ-fill vial process qualification; the full ramp is end-2028. Injectable unit volumes are still undisclosed. Whether either company is collecting a capacity-constraint rent, versus just riding mix, is still inferred.
The oral derailer is live. Lilly’s Foundayo — orforglipron, a pill, not a peptide — was approved on April 1 and shipping on April 6. On the second-quarter call, Lucas Yuffa put about one in four new starts on Foundayo, and about 80 percent of people getting treatment still on an injectable. Those are different bases. Do not write that a quarter of the franchise is oral. The 8-K printed Foundayo at $98 million, not the mix. Novo’s Wegovy pill has done three million prescriptions since January 5; more than 80 percent of those new scripts were new to GLP-1, expanding the pool rather than replacing the pen. Injectable Wegovy HD is still strong on the same tape. Class-level oral share above 30 percent is not established. West and Stevanato’s mix is up, not down. The June honest gap stays open.
The compounding valve is still a proposal. A May 1 Federal Register notice would keep semaglutide, tirzepatide, and liraglutide off the 503B bulks list. Comments closed July 30. No final determination has been fetched. That is a close, not a reopen, and not a ban.
The Medicare GLP-1 Bridge opened on July 1 at $50 a month, through December 31, 2027, outside Part D. Ricks called it coverage for 20 million eligible Americans — 35 percent more people with coverage. That is a Lilly earnings-call print of a coverage-expansion denominator. The 3.8 million figure the wiki keeps is KFF 2023 claims: 65 or disabled, a high enough BMI, and carved out if they already have type 2 diabetes, moderate-to-severe sleep apnea, MASH, or a prior GLP-1. CMS pages retrieved at the end of August publish no eligible-headcount. Klomp’s “single-digit millions” is quoted by KFF, not independently retrieved as a CMS transcript. The two numbers differ by about five times. The wiki’s operative figure stays 3.8 million because it is the narrower Bridge-eligible print and it was here first. Ricks’s 20 million stays as management’s framing. Actual Bridge scripts are still none-found — no CMS utilization table, no Lilly 8-K script count, no IQVIA Bridge-tagged print.
Lilly has also introduced a multi-dose KwikPen — one device for a month’s supply. That cuts pens per patient. More covered patients raise pen demand; fewer pens per patient lower it. No one has netted the two.
Less grocery, more protein, fewer operations
GLP-1 users consume 21 percent fewer calories and spend about 33 percent less on groceries. Protein consumption is up 65 percent. Whey protein isolate sells into food at a 25 to 30 percent premium. To make whey you have to make cheese. Lorcan Roche Kelly’s line is that four or five massive US cheese plants are going up, and “you’re going to be swimming in cheese.” Most of the dairy book is private. There is not much to own.
The surgical substitution is firmer. Metabolic bariatric surgery fell about 34.1 percent from 2022 to 2024 as GLP-1 use rose 140.4 percent. Weight-loss operations dropped from a peak of more than 230,000 in 2022 to about 177,000 in 2024. Intuitive Surgical said US da Vinci bariatric cases declined by a high-single-digit percentage in the second quarter, which management tied to rising GLP-1 use. Practices are pivoting to combined medical-and-surgical obesity care, not exiting. Obesity is 90 to 95 percent untreated, so a mix-shift inside a growing pool can look like cannibalization.
Retatrutide, Lilly’s next injectable, posted 28.3 percent weight loss in trials — “approaching bariatric surgery levels,” in the chief scientific officer’s words. A US biologics filing in the first quarter of 2027 is still Lilly paper — the 8-K, a July 23 press release, and Skovronsky. A Southern District of Indiana order (1:24-cv-01503, September 30, 2025) upheld protein-as-drug and vacated the analogous-to-protein piece back to the agency. A Seventh Circuit date of September 24 has not been confirmed from PACER, CourtListener, or the circuit calendar. Classification delay is the break. Do not treat the filing as in. The drug is not approved and not launched. No source sizes its effect on surgical volumes.
The next dated look is the third-quarter call: Lilly’s IR page lists October 29, 2026, at 10:00 AM Eastern. The confirm into that print is volume still offsetting a guided US price decline, Foundayo mix rising on new patients rather than a closed injectable pool, and the biologics path intact. Hardware last-prices stay on West and Stevanato, not on Lilly.
On September 21 Lilly broke ground on a $6.5 billion Houston / Generation Park API plant for Foundayo and other small-molecule medicines — one of ten US manufacturing sites announced since 2020. The capex itself was announced a year earlier, on September 23, 2025. It is a dated construction start. It is not a 2026 oral-mix print, and it is not a reason to re-rate Lilly or the hardware names.
A named option in sleep
Lilly paid $6.3 billion for Centessa Pharmaceuticals and an orexin-targeting drug. On a podcast, three technologists called it the GLP-1 playbook a second time: buy the receptor class before the category exists, then turn a narcolepsy compound into a lifestyle market. There is no trial data in the record. There is no Lilly statement of intent. The mechanism on the wiki is filed as hypothesis, not active. If the indication stays narcolepsy, it is a bolt-on, not a franchise. Treat it as a named option. It is not a second GLP-1.