GLP-1: Volume Is Outrunning Price Cuts, and Lilly Is Capturing the Profit

Lilly’s quarterly filing reports worldwide volume growth of 53% and realized-price decline of 15%; the corresponding US figures are +41% and −7%, and outside the US +76% and −29%. These are company-wide contributions to revenue growth, including diabetes and other products. They are not counts of obesity patients, and they should not be compounded as an exact price-times-units calculation. The same filing attributes much of the international price pressure to China’s reimbursement listing for Mounjaro. Lilly Q3 2026 Form 10-Q¹.

Since our August study of the GLP-1Glucagon-like peptide-1, a hormone the gut releases after a meal. It prompts the pancreas to release insulin when blood sugar is high, slows the emptying of the stomach and signals fullness to the brain. The body breaks it down within minutes. The drugs in this report copy it in forms that last days, which is why they are called GLP-1 drugs; several also act on the receptors for related hormones. value chain (the companies paid along the way from the discovery of these drugs to the patient), the evidence has moved from price to volume. That study was right that in most US channels (the different ways patients get and pay for the drugs) the patent holder no longer sets the price. So far, though, lower prices have mostly brought in more patients rather than cutting revenue.

The gap between volume growth and price cuts is now narrowing. Lilly's worldwide revenue growth fell from 56% in the first quarter to 48% in the second and 39% in the third. The US price cut was also deeper in the third quarter than in the second. On the Q3 call², Lilly's chief financial officer said US price fell 10% in the quarter, leaving out changes to Lilly's estimates of the rebates it owes. In the second quarter the cut was roughly 9%, the figure Lilly gave in its Q2 2026 release³. For 2027, Lilly expects price to stay "a headwind, more than offset by volume growth".

Who profits depends on these prices. A US patient paying out of pocket now pays $149 to $449 a month. The government has fixed two prices inside that range. Manufacturers get $245 a month under the Medicare GLP-1 Bridge (a temporary programme through which Medicare pays for obesity drugs) and in Medicaid. From 2027, a law caps the Medicare price of semaglutide (Novo Nordisk's drug) at $274. For insured patients, the net price (what the drugmaker keeps after rebates) is under the same pressure, though its level is not disclosed. When the price is roughly the same whichever company makes the drug, the profit goes to whoever sells the most doses, at the lowest cost, of the drug that patients and doctors prefer. On current evidence that is Lilly. Its drugs account for about six in ten US obesity prescriptions, and it kept an 83.9% gross margin (the share of sales left after the cost of making the products) in a quarter when its prices fell 15%. Novo Nordisk is still adding patients but expects its US sales to fall this year. The middlemen are pharmacy benefit managers (companies that run drug coverage for insurers and employers), compounders (pharmacies that make their own versions of the drugs) and telehealth platforms. They are handling more volume and keeping thin margins, with one partial exception. The discoverer of Lilly's pill still collects the same royalty rate, on pill sales that have started slowly.

The August primer’s most useful distinction is between two questions. An efficacy estimand asks how well treatment works under the trial’s assumptions about continuing therapy. A treatment-regimen estimand includes outcomes under a different set of assumptions about discontinuation and other events. Read the trial’s definition before comparing results. Neither headline percentage alone gives the outcome an individual patient should expect.

Population, dose, duration, background treatment and comparator matter as well. An obesity trial without diabetes is not interchangeable with a trial in people with type 2 diabetes. A sponsor-funded indirect comparison adjusts results from separate trials; a randomized head-to-head assigns participants within one study. It is stronger evidence for that specific comparison, but does not answer comparisons against a different dose or drug. The pipeline tables below identify those limits.

The October retatrutide result illustrates the rule. Its 20.9% weight-loss result was against Ozempic 2 mg in adults with type 2 diabetes at 80 weeks, using the efficacy estimand. It is not a direct comparison against Wegovy’s obesity doses or tirzepatide. Discontinuation because of adverse events was higher on retatrutide. Lilly TRANSCEND-T2D-2 release, 28 October 2026⁸.

Finally, a registry’s primary-completion date marks completion of collection for its primary outcome; it is not a promised announcement date or approval date. A trial still running has created a funding obligation and the possibility of evidence. It has not yet demonstrated a commercial advantage.

The older study correctly connected manufacturing to pricing, but overstated the similarity of the production routes. Semaglutide’s peptide backbone is made by yeast fermentation and subsequently modified, according to the FDA’s chemistry review. It should not be described as simply the same repeated chemical-synthesis process as every other peptide. Peptide supply can involve fermentation, chemical synthesis, purification and later modification, depending on the molecule and process. FDA Wegovy chemistry review, 2026⁹.

Oral semaglutide remains a peptide even when pressed into a tablet. It needs an absorption-enhancing formulation because swallowing the molecule does not deliver all of it into the bloodstream. Orforglipron is a small molecule made by chemical synthesis. These are different production and formulation problems, even though both medicines are pills. Foundayo prescribing information¹⁰; FDA oral semaglutide labeling¹¹.

The dose arithmetic is informative but does not reveal factory costs: 25 mg of semaglutide each day is 175 mg a week, compared with a 2.4 mg weekly injection, or about 73 times as much active ingredient. Those are the doses being compared, not every approved injection dose. Absorption, exposure and formulation differ. Neither the 73-times ratio nor consolidated gross margins establishes a 73-times cost difference.

That is the manufacturing thesis in its defensible form: an efficient production route and full plants can help a company sustain margins at lower prices. The magnitude must be tested against disclosed results. Neither leader publishes cost per patient-month, and product mix, currencies, rebates and accounting charges also affect reported margins. The updated Lilly–Novo comparison below keeps those distinctions visible.

List price is the posted amount before rebates and discounts. Manufacturer net revenue is what the drugmaker retains after the relevant deductions and estimate adjustments. Cash price is an offer to a patient paying outside normal insurance, often subject to a dose, introductory period or programme condition. Patient out-of-pocket cost is what that patient pays under their particular coverage. These measures cannot be substituted for one another.

A pharmacy benefit manager, or PBM, manages drug coverage for a plan or employer. It negotiates terms, builds the formulary—the list of covered drugs—and helps process claims. Rebates can make the net price very different from list. Changes to estimates of rebates owed for earlier sales can also alter a quarter’s reported price change without changing the current offer at the pharmacy. That is why the October analysis distinguishes Lilly’s reported US price decline from management’s estimate excluding prior-period rebate adjustments.

Government purchasing creates additional channels. The Medicare GLP-1 Bridge buys eligible medicines at a $245 manufacturer net price per monthly supply; the beneficiary copay is $50. CMS says the demonstration operates outside the normal Part D payment flow and runs through 31 December 2027. The $50 is not the manufacturer’s total revenue. The separate $274 Medicare maximum fair price for semaglutide takes effect in 2027 under a different mechanism. CMS Bridge guidance⁵; CMS negotiated-price fact sheet¹².

These channels make government and payer decisions central to the business, without proving that all sales are government-priced. Counting how many channels are administered does not measure their share of prescriptions or revenue. Commercial net prices remain undisclosed. The updated price ladder below dates each observation and distinguishes dose-specific cash offers, statutory prices and estimates.

Discovery royalties and manufacturing margins pay for different risks. A discoverer can license a molecule and receive a percentage of the licensee’s net sales while the licensee funds later trials, plants and commercialization. A contractual royalty rate can remain intact through a price cut, but the royalty dollars still depend on the net-sales base. Attractive terms do not establish that today’s income is large. The Chugai discussion below makes the rate, sales base and undisclosed payment amount separate questions.

Likewise, growing demand does not guarantee pricing power for every supplier. Capacity ownership, competing plants, contract terms and the customer’s ability to make the ingredient itself matter. But a supplier’s acquisition is not proof that it lacked bargaining power, and management guidance is not an achieved margin. The Bachem reassessment below therefore preserves both the weak first-half actuals and the stronger full-year forecast.

Patents and evidence still matter alongside cost. A patent can protect an originator from a copy without protecting the price against a payer or a better rival. Regulatory exclusivity, patent challenges and settlements differ by country. A successful generic launch abroad does not establish the date of US generic entry. Clinical differentiation can also support demand or coverage even when a government channel fixes reimbursement. The updated competitive and returns sections test these mechanisms against the evidence.

On 28 October Lilly reported TRANSCEND-T2D-2⁸, the first Phase 3The last of the three stages of testing a new drug in people. Phase 1 checks safety in a small group. Phase 2 tests doses and early effect in a larger one; a Phase 2b is the later, larger dose-finding part of that stage. Phase 3 compares the drug with a placebo or a rival drug in a large group, and a company files for approval on its results. trial comparing retatrutide with another incretin (a drug that copies the gut hormones released after a meal). In adults with type 2 diabetes, retatrutide 12 mg cut body weight by 20.9% at 80 weeks against 9.4% for Ozempic 2 mg, the highest approved dose of Novo's diabetes brand of semaglutide. Blood sugar, measured as A1C, fell 2.2 points against 1.8. Discontinuation because of side effects was 7.9% at the 12 mg dose against 2.7% on Ozempic. Lilly had already reported two more obesity trials in July and set a US filing for the first quarter of 2027 in its 23 July release³⁹; the full TRIUMPH-2 results appeared in The Lancet on 29 September.

Two label and safety results strengthen the drugs Lilly already sells. On 28 August the FDA approved Mounjaro to lower the risk of heart attack, stroke and cardiovascular death in type 2 diabetes, based on SURPASS-CVOT¹⁷, which compared it with Lilly's older Trulicity rather than with placebo. On 30 September ACHIEVE-4²⁶ showed Foundayo did not raise cardiovascular risk against insulin glargine (hazard ratioA measure of how often an event, such as a heart attack or stroke, happens in one trial group compared with another over the same period. A ratio of 1 means no difference; below 1 means fewer events in the drug group, above 1 means more. 0.84). The same day Lilly reported a Phase 2b trial of EloraTZP²⁷, which adds Lilly's amylinA hormone released along with insulin that also reduces appetite. Drugs that copy it, such as Lilly's eloralintide, Novo's cagrilintide and Zealand's petrelintide, are being tested on their own and combined with GLP-1 drugs. drug eloralintide to tirzepatide: 23.3% weight loss at 48 weeks against 14.8% for tirzepatide 15 mg alone, in people with obesity and type 2 diabetes.

The Bridge has its first uptake number. The White House said on 18 September that 600,000 seniors⁶ had used it in its first two months, saving $216 million. Lilly said on its call that more than half a million people had filled a prescription for a Lilly medicine through the Bridge, that seven in ten new Bridge patients start on a Lilly drug, and that about 80% of Bridge patients had not been on an incretin before. The same fact sheet put all 50 state Medicaid programs on most-favoured-nation pricing (the lowest price paid in other rich countries) through a new "GENEROUS" payment model. That changes what Medicaid pays, not whether a state covers obesity; only 13 states did in January, according to KFF⁴⁰. States that want to add obesity coverage can join CMS's BALANCE model⁴¹, under which CMS negotiates GLP-1 prices on their behalf, until 1 January 2027; CMS has not published how many have joined.

In commercial insurance, CVS Caremark, the largest pharmacy benefit manager (PBM, the firm that runs drug benefits for employers and insurers and negotiates rebates with manufacturers), put Zepbound back on its main formulariesA formulary is an insurer's or PBM's list of the drugs it covers, often in tiers with different copays. A drug left off the list is not paid for, or costs the patient much more, so a place on it decides whether insured patients can get the drug at all. on 1 October, ending the exclusive Wegovy arrangement it began in July 2025 (FOX 13²⁸, reporting the May announcement). In the cash market, prices kept edging down. Lilly's LillyDirect³⁶ page now lists the 2.5 mg and 5.5 mg Foundayo doses at $149 and $199, "temporarily reduced through December 31, 2026". TrumpRx.gov⁴², the government's cash-price site, lists the Wegovy pill at $149, the Wegovy pen and Ozempic at $199, and Zepbound at $299.

Lilly broke ground on a $6.5 billion plant in Houston²¹ on 21 September to make Foundayo's active ingredient, and on a $5 billion site near Richmond⁴³ on 22 October for cancer drugs. Samsung Biologics' bid for PolyPeptide, the peptide contract manufacturer, was declared successful on 16 October with 87.9% of shares (EQS release³⁰). UnitedHealth's pharmacy benefit arm, Optum Rx, reported a third-quarter operating margin of 4.3% against 3.9% a year earlier, on flat revenue of $39.9 billion and fewer prescriptions (UnitedHealth Q3 release²⁹).

Lilly's Q3 10-Q also discloses that in July 2026 several generic companies filed applications to sell copies of Mounjaro and Zepbound before some or all of the listed patents expire, and that Lilly sued them in Delaware in August. Tirzepatide's compound patentThe patent on the drug molecule itself. Later patents can cover doses, devices or ways of making the drug, but the compound patent is usually the one whose expiry lets generic copies onto the market. runs to 2036 in the US, so this is the start of a long process. The generic filings are the first sign that tirzepatide is now valuable enough to attract the challenges semaglutide has faced for years.

Three things that could have happened did not. Novo has not reported its third quarter. The FDA has not finalised its May proposal to bar outsourcing pharmacies from compounding semaglutide and tirzepatide from bulk ingredient; the latest Federal Register³⁵ notice only extended the comment period. And tirzepatide has not returned to the FDA's shortage list; the agency's shortage data⁴⁴ has no tirzepatide entry, and its semaglutide entries⁴⁵ list only Rybelsus tablets as to be discontinued, after Novo launched its new Ozempic tablets in May.

The table gives each number with its own definition, because the numbers measure different things and should not be added up.

The US is already a mass market. About one adult in nine uses one of these drugs for weight loss. Gallup's measure of the adult obesity rate has fallen from 39.9% in 2022 to 36.8% so far in 2026. Gallup notes that the two have moved together, but it does not claim the drugs caused the fall, and neither do we. Novo reported about 575,000 US Wegovy prescriptions in a single week in July, pen and pill together. Lilly said US prescriptions of incretin drugs (the wider family of gut-hormone drugs that includes the GLP-1s) for obesity rose 82% in the third quarter from a year earlier, against 32% for all incretin prescriptions.

The US is also far from saturated. Lilly said on the call that it reaches about one in ten eligible patients in the US and about two in a hundred outside it. These are management estimates, and nobody measures the number of eligible patients precisely. Medicare shows the scale of the demand still to come. In 2024, 2,282,000 people enrolled in Part D, Medicare's prescription-drug coverage, were taking semaglutide (CMS¹²), mostly for diabetes. That is because Medicare could not pay for a drug used for weight loss alone until the Bridge. Some 600,000 seniors joined the Bridge in its first two months.

Outside the US, the only count published by a company is Novo's. Novo reached 4.9 million people with its obesity products in the twelve months to June 2026, up from 2.9 million a year earlier, according to its H1 2026 report⁴. Its diabetes and obesity products together reached 46.5 million, but that figure includes insulin. No primary source gives a worldwide total of GLP-1 users, and this study does not estimate one.

Revenue is units times price. If units grow faster than price falls, revenue grows. That identity is trivial; what matters is why units are growing this fast, and the evidence points to three reasons.

The first is that lower prices opened new buyers, not just cheaper sales to old ones. Medicare could not pay for a drug used for weight loss until the Bridge began on 1 July. About 600,000 seniors joined in two months, and Lilly says about 80% of them had never taken an incretin. The cash market opened the same way: Lilly's LillyDirect and Novo's NovoCare now sell directly to patients at $149–$449 a month, a price that a patient without insurance coverage can pay and could not at $1,000. In China, inclusion on the national reimbursement list cut the price and put Mounjaro within reach of the state insurance system. Each is a new pool of patients. None existed at the old price.

The second reason follows from the first, and it changes how the price number should be read. Lilly's "realized price" is an average across every channel. When a new patient arrives in Medicare at $245, or in China at the new list price, the average price falls even if no existing patient pays less. Part of the 15% decline is therefore mix: more of the volume is in cheaper channels. Lilly does not split mix from like-for-like price cuts, so the size of each is not known. But its own description supports the reading. It says the price decline outside the US came mainly from the China listing, a new channel, and in the US mainly from Zepbound, the brand growing fastest in Medicare and cash. This is an inference from Lilly's disclosure, not a figure Lilly reports. If it is right, the falling average price is partly a sign of the market widening, not only of pricing power being lost.

The third reason is supply. In 2023 and 2024 tirzepatide was in shortage, and Lilly could not sell more units however much demand there was. The FDA declared the shortage over in a declaratory order of 19 December 2024⁵⁶, and Lilly's capital spending reached $8.4 billion in the first nine months of 2026, against $5.3 billion a year earlier (Q3 10-Q¹, cash flow statement). Volume can only outrun price if the units exist. They now do.

Price cuts will reach existing patients, not just new channels. From 1 January 2027, Medicare will pay no more than its negotiated maximum fair priceThe legal name for the price Medicare sets for a drug it has chosen for price negotiation under the Inflation Reduction Act. Once it takes effect, Medicare's drug plans pay no more than that price. of $274 a month for semaglutide in Part D, 71% below the 2024 list price of $959 (CMS¹²). Novo will cut the US list price of Wegovy and Ozempic to $675 on the same day, by about half and a third respectively (Novo H1 2026⁴). Lilly's management said on the Q3 call that commercial access should be "relatively stable through 2027", with most new coverage coming from Medicare and Medicaid, which pay $245. More of Lilly's volume will be sold at government prices.

Generics will set the price abroad. Semaglutide's compound patent has expired in India, Canada, Brazil and, according to Yicai⁵⁷, China. Indian generics launched on 21 March 2026 at Rs 1,300–4,200 a month (ThePrint⁵⁸). Tirzepatide's patents run longer, to 2036 in the US, but it competes on price with whatever semaglutide costs in each market. That is the main reason Lilly's price outside the US is falling faster than at home.

None of these crosses the line on the evidence available today. Each narrows the gap, which is what the last three quarters show.

The top Wegovy pill dose is 25 mg a day, or 175 mg a week. The standard pen gives 2.4 mg a week. The pill therefore uses 73 times as much semaglutide, for a comparable weight loss in separate trials: 16.6% for the pill in OASIS 4 and 14.9% for the 2.4 mg pen in STEP 1. The new tablets absorb about twice as well as the old ones, which narrows the gap but does not close it. The August study used an absorption figure of under 1%, from a 2024 paper on oral semaglutide⁷¹. The label now gives the higher figure for the new tablets.

Foundayo is orforglipron, a small molecule, the kind of chemical most ordinary pills contain. It is made by chemical synthesis in conventional plants. It is absorbed well enough to be taken "any time of day without food or water restrictions", as Lilly's approval release⁷² puts it. Oral semaglutide, at least in its diabetes tablets, must be taken on an empty stomach with up to four ounces of water, and the patient must wait at least 30 minutes before eating, according to the Rybelsus and Ozempic tablets label. Lilly is building two plants for Foundayo's active ingredient: the $6.5 billion Houston site, where it broke ground in September, and the $6 billion-plus Huntsville⁷³ site announced in December 2025. Lilly says small molecules are "more readily scalable", meaning easier to make in larger quantities. According to the White House fact sheet⁶² on the November 2025 agreements, Novo committed an additional $10 billion of US investment, including producing the Wegovy tablet "end-to-end in the U.S."

Neither company discloses what a month of either pill costs to make. So Foundayo's cost advantage is inferred from how the two pills are made rather than measured. The inference rests on two facts that are disclosed: how much drug each pill needs, and what kind of plant makes it. The August study made this argument before either pill had a full quarter of sales. Since then, one new fact has weakened it slightly: Novo's new tablets absorb better than Rybelsus did. Another has strengthened it: Lilly is now building the plants.

For the patient, the two pills cost the same. Both start at $149 a month for self-pay patients, those paying cash. Novo charges $199 for the 4 mg Wegovy pill and $299 for the 9 mg and 25 mg doses (NovoCare⁶⁰). Lilly charges $149 for the 0.8 mg and 2.5 mg Foundayo doses, $199 for 5.5 mg and $299 for 9 mg and above, with the 2.5 mg and 5.5 mg prices "temporarily reduced through December 31, 2026" (LillyDirect³⁶). Insured patients pay as little as $25 with either company's savings card. Medicare Bridge patients pay $50, and the maker receives $245 for either pill.

What each company receives on average is different. Novo's Wegovy pill earned roughly $169 per US prescription in the second quarter (DKK 3,141 million ÷ 6.40 ÷ about 2.9 million prescriptions). Novo says the uptake was driven by patients new to GLP-1 drugs, "across all doses in the self-pay channel, particularly the 1.5 mg, 4 mg and 9 mg doses". That means much of the Wegovy pill's volume is self-pay at the $149–$299 end. Dividing Foundayo's $184 million by Lilly's count of "over half a million" prescriptions gives at most about $368 per prescription in the third quarter. The true figure is lower if the count was well above half a million, and the $184 million includes the first sales outside the US. The two averages are not like-for-like and should not be read as prices. They do suggest the Wegovy pill's lead is concentrated in the cheapest part of the market.

So far, pills are mostly bringing in new patients. Novo describes Wegovy pill uptake as "driven by GLP-1-naïve patient uptake", meaning patients who had not taken a GLP-1 drug before. In the second quarter, Novo's combined US Wegovy sales rose 4% at constant exchange rates even though pen sales fell 22%. Novo puts the pen's fall down to lower prices rather than lost patients: "partly countered by increased volumes". In the Medicare Bridge, Lilly said on the call, about 80% of patients had not been on an incretin before, and about a quarter chose a pill.

The opposite risk is cannibalisation: pills taking patients from the same company's injections. That risk is real, but it comes later. Lilly is positioning Foundayo for maintenance, the long phase after the main weight loss when a patient wants to keep the weight off at lower cost. Its May 2026 maintenance trials⁷⁴ (ATTAIN-MAINTAIN and SURMOUNT-MAINTAIN) showed patients holding their weight after switching from higher-dose injections. A patient who moves from Lilly's Zepbound injection at $449 in cash to Foundayo at $299 brings Lilly less revenue, though the pill probably costs Lilly less to make. Lilly has not updated the share of Foundayo patients who are new to this class of drug since its launch weeks, when it said 80%, and the next update to that share is the figure to watch.

The August study predicted that pills would make up more than a third of new US obesity starts by the end of 2027. On the evidence so far they are at about a quarter. That fits the Lilly base case, in which pills bring new patients into Lilly's obesity business before 2028 and take some share from injections after it.

Lilly has the two strongest late-stage results. Retatrutide's larger weight loss comes with more side effects than tirzepatide's. In TRIUMPH-1, 42.4% of patients on 12 mg reported nausea against 14.8% on placebo, and 11.3% stopped treatment because of side effects against 4.9%. Dysesthesia, an unpleasant skin sensation, was reported by 12.5% on 12 mg in TRIUMPH-1 and 6.5% in TRANSCEND-T2D-2. The 4 mg dose, reached in a single step, gave 19.0% weight loss, and a smaller share of patients stopped on it than on placebo. That suggests the drug may end up being used at more than one dose, for different patients. There is already demand for retatrutide. In August Lilly filed six lawsuits against sellers of black-market retatrutide and reported more than 14,000 illegal listings in over 100 countries (Lilly, 12 August¹⁵).

EloraTZP is the second. It combines tirzepatide with eloralintide, a drug that copies amylin. Amylin is a hormone released along with insulin, and it also reduces appetite. In a Phase 2b trial of 367 adults with obesity and type 2 diabetes, EloraTZP produced 23.3% weight loss at 48 weeks against 14.8% for tirzepatide 15 mg alone, and cut A1C, a measure of average blood sugar, by up to 2.9 points. Phase 3 begins in the fourth quarter of 2026, so it is a product for the end of the decade.

The new Novo obesity drug closest to approval is CagriSema, a weekly combination of semaglutide and cagrilintide, another drug that copies amylin. It produced 20.4% weight loss at 68 weeks in REDEFINE 1, counting everyone randomised (ACC summary of the NEJM paper⁷⁵), roughly level with tirzepatide. In the direct comparison, REDEFINE 4, it lost to tirzepatide: 23.0% against 25.5% at 84 weeks among patients who stayed on treatment, and 20.2% against 23.6% counting everyone, according to Novo's 6-K of 23 February 2026⁸¹. In REIMAGINE 4, in type 2 diabetes, it matched tirzepatide 15 mg on weight but not on blood sugar, according to Novo's H1 report⁴. Novo lists a US decision on CagriSema for obesity in the fourth quarter of 2026. Approval would give Novo a second strong injection, but not one better than tirzepatide.

Amycretin, now also named zenagamtide, is a single molecule that acts on both GLP-1 and amylin receptors. In a Phase 1b/2a study published in The Lancet in June 2025, the weekly injection produced up to about 24% weight loss at 36 weeks (ACC summary⁷⁶). That was a small trial of 125 people. Novo has begun the Phase 3 AMAZE programme. Its AMAZE 8 trial, which compares amycretin with semaglutide, reports results in the first half of 2029. In September Novo also licensed Hengrui's HRS-1596, a once-weekly GLP-1/GIP pill, for $300 million upfront and up to $2.6 billion in total (Novo, 29 September²⁴). It is ready for Phase 1.

Amgen's MariTide is built to be taken once a month or less. Its Phase 2 trial produced 12.3–16.2% weight loss at 52 weeks counting everyone randomised, and up to about 19.9% among those who stayed on it (TCTMD⁷⁷). When the dose was raised gradually, 8% of patients stopped because of side effects, against 12–27% when it was not. Amgen's Q2 release⁷⁸ lists nine Phase 3 studies running and says patients might stay on it "with as few as 4 or 6 doses per year". No Phase 3 result has been reported. MARITIME-1 and MARITIME-2 reach primary completion, the date the last patient is measured for the main result, on 21 January 2027, according to the registry⁷⁹.

Pfizer's monthly injection PF'3944, which it gained when it bought Metsera and has now named berobenatide, produced 12.3% placebo-adjusted weight loss (the loss beyond what the placebo group lost) at 28 weeks in Phase 2b (Pfizer, 3 February⁸²). Pfizer's Q2 2026 release⁸³ says ten Phase 3 studies of it will advance in 2026, within a programme of more than 20 obesity trials. VESPER-6, a Phase 3 trial of monthly dosing, is due to complete in May 2028, according to its registry record⁸⁰.

Boehringer Ingelheim's survodutide, partnered with Zealand, met its Phase 3 endpoints with 16.6% weight loss at 76 weeks in SYNCHRONIZE-1 (Zealand H1 2026¹⁶). Viking's VK2735 produced 14.7% after 13 weeks in Phase 2. Its two Phase 3 trials, with about 5,500 patients, complete in 2027. Structure's aleniglipron, a small-molecule pill like Foundayo, produced 16.2% at 72 weeks in an open-label extension, a follow-on study in which patients knew they were taking the drug. Its Phase 3 results are due in the second half of 2028. Roche's enicepatide produced 15.5% at 48 weeks in people with type 2 diabetes, and its amylin drug petrelintide, partnered with Zealand, produced 10.7% at 42 weeks with very few gastrointestinal side effects (Roche, 5 March⁸⁴). On 22 September Zealand started petrelintide's Phase 3 programme: three trials with about 7,000 participants, with the main result measured at 64 weeks (Zealand release²³).

In the old market, a better drug earned a higher price. The August study listed tests that would show its conclusions were wrong. One of them checks whether that still holds: whether retatrutide or high-dose CagriSema "commands a materially higher realised net price per patient-month", meaning the price actually collected after discounts for each month a patient is treated. Neither drug is approved, so the test cannot be run yet. One disclosed fact does bear on it: the commitment quoted above from the November 2025 agreements, as the White House fact sheet⁶² describes them. Lilly's Q3 10-Q¹ says it agreed "to launch new medicines with a more balanced pricing approach across developed nations". A drug launched on those terms cannot easily be priced in the US far above what it sells for in Europe.

If that holds, retatrutide's advantage will show up in sales volume and market share rather than in price. It would win the patients who need more than 20% weight loss, the patients near the surgery threshold, and the formulary placements (spots on insurers' lists of covered drugs) that go to the strongest drug. It would also keep those patients on a Lilly drug as tirzepatide approaches its 2036 patent expiry. That would still be worth a great deal. This is an inference from the agreements. Retatrutide's US price has not been announced.

Both companies cut cash prices and both signed the November 2025 agreements, so the US price reset itself does not explain the gap. The two companies differ in where their sales were exposed. Novo's US obesity sales relied on the insured channel, meaning patients whose insurance pays for the drug. There, its exclusive Wegovy deal with CVS Caremark ended on 1 October, when Zepbound returned to Caremark's formularies (its lists of covered drugs). Novo's US pen sales fell 22% in the second quarter because its realized prices, what it actually receives after discounts, were lower. Novo also faces generic semaglutide in Canada, India and Brazil. In August it said that in Canada "lower realised prices" had followed, while "volumes have remained broadly stable". Tirzepatide has no generic competition anywhere. Its US compound patent, the patent on the drug itself, runs to 2036, and the generic companies that challenged it in July are at the start of a long process.

Lilly also set prices differently in each market. In China it cut the price of Mounjaro, its other brand of tirzepatide, by about 80% to win a place on the national reimbursement list (the drugs China's national health insurance pays for), and sales volumes rose. In the UK it raised the list price by as much as 170% from September 2025. Novo cut its China Wegovy price by about 48% in January (Yicai⁵⁷) as the semaglutide patent expired there.

The growth in volume came from a one-time event and should not be read as a trend. In 2026 three large new pools of patients opened at once: Medicare through the Bridge, the cash market through LillyDirect, NovoCare and TrumpRx, and China through the reimbursement listing. Each pool fills once. Lilly's worldwide volume growth has already fallen from 65% to 53% in two quarters, and its revenue growth from 56% to 39%. If it keeps falling along that path, revenue growth reaches single digits within a year or two.

Meanwhile the price keeps falling, and the cuts are moving from the new channels into the business Lilly already has. Lilly's underlying US price decline was about 9% in the second quarter and 10% in the third. On 1 January 2027 two things happen together: Medicare's maximum fair price of $274 for semaglutide takes effect in Part D, and Novo cuts the list price of Wegovy and Ozempic to $675. For every insured patient whose coinsuranceA percentage of the drug's price that an insured patient pays at the pharmacy, instead of a fixed copay. When the percentage is taken of the list price, a lower list price lowers what the patient pays. or plan payment is tied to list price, the reference price for this class of drugs will halve. Lilly will have to respond. Its own management told investors that commercial access will be "relatively stable through 2027". That means any growth in the number of people covered will come from Medicare and Medicaid, at $245.

Insurance coverage is fragile, and employers are pulling back. In the Business Group on Health's survey of 127 employers, the share of large employers covering GLP-1s for obesity fell from 72% in 2025 to 60% in 2026. Some 14% have dropped it or plan to drop it by 2027 (SHRM²⁰). The consultancy Marsh found 6% of employers dropped coverage in 2026, 5% plan to in 2027, and 27% are adding restrictions such as minimum body-mass thresholds (AP, 16 September⁸⁵). The AP report describes the drugs as the single largest contributor to the record increase in employer health costs expected for 2027. Medicaid programmes covering obesity fell from 16 to 13 between October 2025 and January 2026 as California, New Hampshire, Pennsylvania and South Carolina dropped it, according to KFF⁴⁰. The Bridge is a demonstration programme that ends on 31 December 2027, and CMS told Part D plans on 21 April that it will delay the Part D part of its BALANCE model for 2027 (CMS memo to plans⁸⁶). If the Bridge lapses, the seniors who joined it have nowhere obvious to go.

Many patients stop taking the drugs. In a study of 125,474 adults who started a GLP-1 for overweight or obesity between 2018 and 2023, 64.8% of those without diabetes stopped within a year (Rodriguez et al., JAMA Network Open, 2025⁸⁷). In KFF's poll, 14% of users said they had stopped because of cost and 13% because of side effects. A market in which most new patients leave within a year has to keep signing up new ones just to stay the same size.

Cheap alternatives are still on sale. Compounded semaglutide is still advertised from $86 a month (Ozari Health, August 2026⁶³). The FDA's proposal to stop industrial-scale compounding is still only a proposal. Generic semaglutide is on sale in India for Rs 1,300–4,200 a month, and in Canada and Brazil, and the companies making it will look for any market they can enter. Novo is under pressure and has every reason to use its lower-cost pill dose ($149) and its 2027 list-price cut to win market share on price.

There are also risks the market has not priced in. In February the UK regulator added a warning of a "very rare" risk of NAION, a cause of sudden vision loss in one eye, with semaglutide, and said it is reviewing the evidence for other GLP-1 drugs (MHRA, 5 February 2026⁸⁸). In the TRIUMPH-1 trial, one in eight patients on retatrutide's top dose had dysesthesia, a skin side effect. Generic companies have begun challenging tirzepatide's patents. And tirzepatide itself will eventually be selected for Medicare price negotiation; Lilly's 10-Q says it expects "other significant products will be selected in future years".

Some of this case is right, and it changes how firmly the study holds its answer. The gap between volume growth and the price decline is narrowing, the underlying US price decline is deepening, and 2027 brings a new low reference price. Those are facts. So the study is most confident that volume growth outpaces the price decline over the next few quarters, and less confident beyond 2027.

Three parts of the case are weaker than they look. The first is the idea that each new pool fills once. Lilly says it reaches about one in ten eligible Americans and about two in a hundred eligible people abroad. Even allowing for some optimism in those estimates, the pools are large compared with the number of patients treated so far. Volume growth will keep slowing in percentage terms as the base grows. The question is whether it slows below the break-even rate, meaning the volume growth needed just to offset the fall in price. At Lilly's current US price decline, that rate is about 11%.

The second is how the price figures are read. Lilly's average realized price is falling partly because more of its units go to cheaper channels, and not only because existing patients pay less. If the average is falling because of that shift in mix, the market is getting wider. Lilly does not disclose the split, so this cannot be proven from the filings. It should be treated as the more likely of two readings, not as fact.

The third is the claim that lower prices must mean lower profit. Lilly's gross margin rose to 84.0% in the first nine months of 2026 from 83.3%, while its realized price fell 14%. Lilly attributes the improvement to "improved cost of production and favorable product mix". That is what happens when a plant with largely fixed costs makes more units: the cost of each unit falls. It also means that in a price war, the company best placed to survive is the one with the most volume to spread the cost of its plants over. Today that is Lilly.

The employer pullback is real, and it is concentrated among the largest employers. The Business Group on Health surveyed 127 of them. KFF's annual survey of firms with 200 or more workers found 18% covering GLP-1s for weight loss in 2026 against 19% in 2025, and 33% among firms with 5,000 or more workers; 79% of the firms that do not cover the drugs say they are not likely to start (KFF, 22 October³²). Taken together, the two surveys say employer coverage is flat to falling and is not where new patients will come from. The two surveys are still the strongest new evidence against the study's case on volume, and they are why the cash channel matters. A patient whose employer stops paying can still buy Zepbound at $299–$449 or Foundayo at $149–$299, so Lilly still makes the sale, at a lower price. Lilly's guidance that commercial access will be "relatively stable through 2027" fits KFF's broad survey better than the large-employer one. Lilly's 2027 reports will show which survey matches its own business.

Discontinuation is the other hard point, and it is already in the numbers. The volume Lilly reports is net of patients who stopped, and US obesity incretin prescriptions still rose 82% in the third quarter. Persistence, meaning how long patients stay on treatment, is also improving at the margin: Lilly's maintenance trials and lower-dose options give patients a cheaper way to stay on treatment, and the JAMA cohort found that discontinuation fell as weight loss increased, which favours stronger drugs.

Reassessment, rather than a new event. Bachem’s 35–40% growth guidance and low-thirties margin outlook were published on 30 July, before the August study. The change here is the weight assigned to that existing counter-evidence. The full-year result remains the test. Optum Rx’s later margin improvement is new evidence, but it concerns the entire segment and does not isolate obesity-drug economics.

The chain runs from the firms that discovered the drugs and collect royalties, through the drugmakers and the contract manufacturers that make ingredient for them, to the PBMs, telehealth firms and pharmacies that sell the drugs and the employers, insurers and government that pay for them. When a month of the drug sells for roughly the same $149–$449 whoever makes it, the profit on each sale is that price less the cost of making and delivering the dose. Profit in the chain then goes to whoever sells the most doses at the lowest cost, and to anyone who holds a claim on those sales that a price cut does not reach. The August study ranked the chain on that logic, grading each layer by whether it keeps a return above its costs once a competitor, a pill, new factories, a generic or a government price arrives. Twelve weeks of new evidence leave its top unchanged and move two layers in the middle.

Lilly is at the top, and its position strengthened. It sells about six in ten US obesity prescriptions and about 58% of incretins abroad, by its own account on the Q3 call. Its gross margin was 84.0% in the first nine months against 83.3% a year earlier, while its realized price fell 14%. Reported operating income rose less, 17% in the third quarter, because Lilly booked $4.2 billion of acquired research chargesCosts a drugmaker records at once when it buys a company or licenses a drug that is still in development. US accounting rules treat them as an immediate expense, so they cut operating income in the quarter the deal closes., mostly for AtaiBeckley and Curevo; before those charges, operating income rose about 59% ($12.8 billion against $8.0 billion, our arithmetic from the Q3 release¹³). It is building the cheapest route to a large-volume pill, and it has the strongest drug close to filing. The August study gave Lilly its top grade on cost; the volume evidence since then adds a second reason. A plant that runs fuller makes each dose cheaper, and the company with the most volume is the one best placed to run its plants full. Lilly's rising margin is consistent with that, though the company does not disclose utilization. A lower price therefore helps Lilly more than it hurts it, at least for now: the volume it brings lowers Lilly's cost per dose as well.

The discovery royalty keeps its place but not much money yet. Chugai, majority-owned by Roche, receives "tiered royalty payments on worldwide net sales ranging in percentages from mid single digits to low teens" on Foundayo, recognized in Lilly's cost of sales, according to Lilly's Q3 10-Q¹. Foundayo sold $281 million in the first nine months of 2026. At those rates, that is roughly $14–37 million of royalty, an illustrative range that Chugai does not confirm: its Q3 announcement³³ mentions royalty income from Hemlibra and Nemluvio and does not break out orforglipron. The point of the August grade was the rate, which nothing since has touched. The amount depends on how fast Foundayo grows, and so far it has grown more slowly than the Wegovy pill.

The August study graded peptide contract manufacturers as price-takersA price-taker is a seller with no power over its price, which accepts what its customers or the market set. Its profit then depends on its costs and on how much it sells., with Bachem "exposed". Bachem's already-published July guidance argues against that; its first-half results still fit it. Its first-half 2026 sales grew only 7.3% in local currencies and its EBITDA margin fell to 25.4% from 29.1%. Sales of commercial active ingredients, the line that supplies marketed drugs, fell 19.3% in local currencies, while clinical-development work rose 39.5%. It still guides to full-year growth of 35–40% and an EBITDA margin "in the low thirties", with its new Building K plant ramping and a large-volume site at Sisslerfeld planned with a partner, more than CHF 500 million of investment for production by 2030 (Bachem, 30 July⁸⁹). That is almost exactly one of the August study's falsifiersA falsifier is a result, named in advance, that would show a study's conclusion is wrong. The base studies listed theirs so that later evidence could be scored against them, which the Scorecard in Part VI does.: "Bachem delivers 35–40% FY growth with EBITDA margin above 30%". If the full-year results in early 2027 meet the guidance, the grade is wrong. There is a reason it might be. Lilly's next obesity drug, retatrutide, is a peptide, and Novo's pill needs 73 times as much semaglutide per patient-week as its pen. Both raise the demand for peptide ingredient. Samsung Biologics' successful bid for PolyPeptide (87.9% of shares tendered by 12 October) is a bet on the same demand. Thermo Fisher, the largest diversified supplier, still does not mention GLP-1 drugs in its Q3 2026 release⁹⁴; its laboratory products and biopharma services segment earned a 15.5% margin. We now hold the grade at "exposed, under review", with Moderate confidence: the reported half-year supports "exposed", and only the guidance argues against it.

The August study graded the pharmacy benefit managers as exposed, on the evidence that their segment margins were falling while their revenue grew. Two of the three still fit. CVS Health's Health Services segment, which contains Caremark, earned an adjusted operating margin of 3.22% in the first half against 3.53% a year earlier (CVS Q2 10-Q⁹⁰). Cigna's Pharmacy Benefit Services unit, inside Evernorth, saw second-quarter pre-tax adjusted income fall 27% to $609 million on 8% higher revenue, which Cigna attributes to "client- and customer-focused initiatives" (Cigna Q2 10-Q⁹¹). The third does not. Optum Rx's operating margin rose to 4.3% in the third quarter from 3.9%, on flat revenue and 5% fewer prescriptions, which UnitedHealth attributes to "continued adoption of biosimilars and operational improvements" (UnitedHealth Q3²⁹). None of the three breaks out GLP-1s, so the margins are segment facts, not obesity facts. What can be said is that a PBM can still raise its margin while obesity drugs move to cash and Medicare channels, if it earns enough elsewhere. The layer is thinner than it was, not empty.

The insurers and employers who pay for covered prescriptions are the one part of the chain where lower prices are good news, and they are responding by rationing rather than expanding. As the coverage figures in The Strongest Case Against show, the largest employers cut coverage of obesity GLP-1s from 72% to 60% in a year, while coverage across firms with 200 or more workers held at about 18%. Their case for paying depends on how long patients stay on the drug and how long the payer keeps the patient. A Lilly-sponsored real-world study of adults over 55 found non-drug healthcare costs up to $607 a month (38%) lower after 12 months on Zepbound than in similar untreated adults, driven by fewer hospital admissions (Lilly, 26 August⁹⁵). That is a company-sponsored study in an older group, and a KFF analyst told AP that some early studies show no short-term reduction in the use of other health care. Medicare, which keeps its members for life, has the strongest reason to pay; an employer whose staff change jobs every few years has the weakest. That asymmetry is one reason the volume is moving into government channels at $245.

Hims & Hers shows what the cash channel pays a middleman. Its second-quarter revenue rose 38% to $753 million. Most of the increase came from outside the US, where it has bought telehealth businesses: revenue there was $131 million against $8 million a year earlier, and US revenue rose 16%. Its subscribers rose 19% to 2.9 million. In March it moved its US weight-loss offer to branded GLP-1 drugs and kept compounded ones "on a limited scale". Its gross margin fell to 64% from 76%, mainly because of its weight-loss offerings, which it says have "shorter shipping cadences and increased fulfillment costs", and it lost $86.3 million (Hims Q2 10-Q¹⁴). It now books prescriptions made through its platform partners as an agent, taking a fee, rather than as the seller. Novo's subscription programme runs through telehealth firms including Ro, WeightWatchers and LifeMD. These firms are a sales channel for the brands, paid for the service, with no claim on the drug.

The government earns no profit in this chain; it sets the price the others work around: $245 in the Bridge and Medicaid, $274 in Part D from 2027. The challengers sit outside the profit pool until they launch, and when they do, they will sell into a $149–$449 market under most-favoured-nation launch pricing, against tirzepatide and, soon, retatrutide. That makes their economics a question of cost and share, not of price. Viking's raise in September and Structure's $1,342.7 million of cash and investments at 30 June (Structure Q2 10-Q⁹⁶) give both the money to reach their readouts; neither has yet shown a drug that beats tirzepatide in a long trial.

Peptides and small molecules are made in different kinds of factory. A peptide is a chain of amino acids that must be produced, purified and, for an injection, formulated and filled into sterile delivery devices. The production route varies: semaglutide’s backbone is produced by yeast fermentation before modification. Novo's capital spending in 2026, about DKK 55 billion, goes into "active pharmaceutical ingredients (API), additional aseptic production and finished production processes as well as packaging capacity", in its own words (Novo H1 2026⁴). A small molecule like orforglipron is made by ordinary chemical synthesis and pressed into tablets. Lilly's Houston plant will train its workforce in "advanced chemical synthesis manufacturing". Lilly says such molecules are "more readily scalable", meaning production is easier to expand.

Lilly has announced ten US sites since 2020, in a programme of more than $50 billion (Lilly, 21 September²¹). Five make active ingredients: Houston ($6.5 billion, for Foundayo), Huntsville (more than $6 billion, including orforglipron), Richmond ($5 billion, for cancer drugs) and two in Lebanon, Indiana. Four make injectable medicines and devices: Research Triangle Park and Concord in North Carolina, Kenosha in Wisconsin, and the Lehigh Valley in Pennsylvania. The Lehigh Valley site ($3.5 billion) is for next-generation obesity drugs including retatrutide and is due to be operational in 2031, per the January release⁹². The Houston site alone will employ more than 600 people and 4,000 construction workers.

None of the claims has failed. Most cannot be tested yet, because their test dates fall in 2027 or later. The evidence so far supports them. Novo's adjusted gross margin is moving further from the level that would disprove the August study's cost argument. Lilly's adjusted gross margin is above its FY2025 level. Lilly's US prices are falling while its US volume is rising, as the Lilly base study predicted for every year to 2029.

Four items are weaker than they were. Bachem's "exposed" grade is at risk. The August study named a level of growth and margin that would prove the grade wrong, and Bachem's own guidance now forecasts that growth and margin. Its FY2026 results, due in early 2027, will settle it. Viking's grade was "science credible, funding not", and the funding half no longer holds after a $575 million raise. Zealand had no grade in August because its own filings had not been opened; they now have been. The August argument that a peptide sold as a pill (an oral peptide) lowers margin is also weaker. Novo's newer tablets absorb about twice as well as Rybelsus, its earlier tablet, did. Even so, a week of the top-dose pill still needs 73 times as much drug as a week of the injection pen.

The August study graded CVS "exposed" in the PBM layer, meaning the pharmacy benefit managers (PBMs) that run prescription-drug plans for insurers and employers. CVS's own figures still support that grade. Optum Rx's margin rose, though, so the squeeze is not hitting every company in the layer equally. The Lilly base study's central verdict is that rising volume will more than make up for falling prices through 2028. Every number still supports it, but with less room than before: revenue grew 48% in the quarter the study was written about, and 39% in the quarter after.