The Cheap Price Is the Volume: Lilly Grew 56% While Cutting Cash Prices, Novo Fell 4% — Same Channel, Opposite Outcomes, and Mix Explains Neither

Same Channel, Opposite Outcomes

In the first quarter of 2026, two companies sold the same class of drug, into the same country, under the same deal with the same government, at cash prices they had both just cut. One of them grew revenue 56% to $19.8 billion and raised its full-year guidance by two billion dollars. The other watched adjusted sales fall 4% at constant exchange rates and guided to a full-year range that is negative at both ends. [1] [5]

That is the finding. Everything else in this report is an attempt to explain it.

The hypothesis we set out to test was the consensus one: that the collapse in GLP-1 cash-pay prices is a channel-mix tax — that as patients migrate from insured prescriptions into direct-to-consumer self-pay, revenue per script falls faster than units rise, and volume growth does not offset it. It is an elegant thesis. It is also, on the evidence of the primary filings, wrong in both of its halves — and wrong in the direction of the opposite call.

Start with the thing the hypothesis most needs to be true: that the mix shift has not yet been tested at scale. It has. Lilly's CFO Lucas Montarce told the Q1 2026 call that self-pay accounts for approximately 45% of total Zepbound prescriptions and 55% of new prescriptions. [15] The mix shift is not coming. It is half-done, and the marginal channel — the one that determines where the book converges — is already cash. Novo's disclosure is the mirror: around 100,000 of roughly 270,000 weekly US injectable Wegovy prescriptions were filled in the self-pay channel in the week ending 17 April 2026. [5] On our arithmetic that is about 37% — meaning Lilly, the company that grew 56%, is more cash-exposed than the company that shrank.

Same MFN agreement. Same TrumpRx portal. Same $245 Medicare price with the same $50 copay. Comparable self-pay penetration, tilted against the winner. And a divergence of roughly thirty-five points between the two companies' guided growth rates: Lilly at $82.0 billion to $85.0 billion, Novo at -4% to -12% at CER. [1] [5]

A variable that both companies share cannot explain an outcome that diverges that far. Whatever is killing Novo, it is not the existence of a cash channel, because Lilly has more of one and is thriving. The operative variable is share — and, underneath share, the price point each company chose to set inside the cash channel. Lilly capped LillyDirect refills at $449, which is essentially level with what Leerink Partners estimates the average US commercial plan already pays for Zepbound. Novo set its standard cash price at $349 — below its own implied insured net. [11] [16] [12] Dilution is not a property of the channel. It is a property of the price.

One honest caveat before we argue any of it. The single genuinely untested leg of this story is government pricing, not DTC pricing. The Medicare GLP-1 Bridge — $245 a month, $50 copay — went live on 1 July 2026, thirteen days before this report's date, and has touched zero reported quarters. [9] We are not claiming the story is over. We are claiming that the part of it that has already happened says the opposite of what the hypothesis predicted.

Lilly's Bridge: Volume Beats Price Five to One

Lilly publishes exactly the decomposition the mix hypothesis requires, and it goes the wrong way for the hypothesis. The Q1 2026 10-Q gives the price/volume/FX bridge across three columns — US, outside US, consolidated:

Volume 49 % 95 % 65 % / Price (7) (25) (13) / Foreign exchange rates — 11 4 / Percent change 43 % 81 % 56 % [2]

Worldwide, volume of 65% outruns price erosion of 13% by roughly five to one. In the United States — the only geography where LillyDirect, TrumpRx and the MFN deal operate at all — volume of 49% outruns price of 7% by seven to one. [2] [1]

Source: Eli Lilly Form 10-Q, quarter ended 31 March 2026

The obvious rejoinder is that this is a pre-mix-shift snapshot: that we are looking at a book still dominated by insured scripts, and the cliff arrives once cash-pay scales. That rejoinder is unavailable. The natural experiment the hypothesis proposes has already run. Montarce's 45%/55% disclosure means nearly half of the Zepbound book is already cash and a majority of new starts are cash — so the stock of scripts is converging upward toward a flow that is majority self-pay. [15] The mix shift is not a future event to be modelled. It is the condition under which these numbers were printed.

And what did it print? US Zepbound revenue increased 79% to $4.1 billion, in Lilly's own words "primarily driven by strong demand, partially offset by lower realized prices, including previously announced reductions in cash pay prices." [1] Worldwide Zepbound grew 80%. [2] Mounjaro grew 125%. [2] Between them, Mounjaro at $8,662 million and Zepbound at $4,160 million account for the majority of a $19,799 million quarter. [2] Whatever happens to GLP-1 realised price happens to Lilly — there is nowhere for it to hide. It grew 56% anyway. [1]

The complication we are not going to bury

Two things cut against this, and they belong here rather than in a footnote.

First, Lilly's -7% US price line is flattered by a one-off. The 10-Q says, of both incretins, that "Lower realized prices in the U.S. were partially offset by a favorable one-time adjustment to estimates for rebates and discounts during the three months ended March 31, 2026," and it sizes the category: adjustments from changes in estimates for US sales returns, rebates and discounts were "less than 3 percent" of US revenue in Q1 2026, against 1% a year earlier. [2] On our arithmetic — 3% of $12.1 billion of US revenue — that is up to roughly $360 million, or up to about two percentage points of year-on-year swing in the US price line. [1] This is researcher arithmetic on Lilly's disclosed inputs, not a Lilly figure, and the honest conclusion is that underlying US price erosion is worse than the reported -7%. We will hold that thought and come back to it, because it matters for the model in section five — and, awkwardly for the hypothesis, it does not change the direction of anything.

Second, and far more seriously: Novo exists. Its adjusted sales, excluding the 340B provision reversal, decreased by 4% at CER, "driven by lower realised prices, partly offset by GLP-1 volume growth across geographies." [5] That is the only place in this entire dataset where the revenue line actually breaks. It is therefore the only evidence a mix thesis can lean on, and it deserves a fair hearing. We give it one — and then we show that Novo's own filing refuses to support the reading placed upon it.

The question that organises the rest of this report is not "does volume beat price?" It is: why does the same channel produce +56% at one company and -4% at the other?

The Scariest Number Is Made in China, Not in LillyDirect

Before we get to Novo, we have to dispose of a number that has been doing a great deal of unearned work in the bearish case.

Lilly's consolidated realised price line looks like it is falling off a cliff. In Q4 2025, worldwide revenue rose 43% "driven by a 46% increase in volume, partially offset by a 5% decrease due to lower realized prices." [4] One quarter later it is -13%. [1] A price line that deteriorates from -5% to -13% in ninety days, in exactly the ninety days when cash-pay scaled — that looks like the mix cliff arriving on schedule.

It is not. Look at the US column.

Q4 2025: "Revenue in the U.S. increased 43% to $12.9 billion, driven by a 50% increase in volume, partially offset by a 7% decrease due to lower realized prices." [4] Q1 2026: "Revenue in the U.S. increased 43% to $12.1 billion, driven by a 49% increase in volume, partially offset by a 7% decrease due to lower realized prices." [1] The same 7%. Zero acceleration — across precisely the window in which self-pay scaled to 45% of total Zepbound scripts and 55% of new ones. [15]

The entire consolidated deterioration lives outside the United States, in a price line of (25) against volume of 95 %. [2] And Lilly tells you, in one sentence, what it is: the lower realised prices outside the US "were driven primarily by the addition of Mounjaro to the National Reimbursed Drug List (NRDL) in China." [2] [1]

A Chinese government formulary. Not LillyDirect. Not NovoCare. Not TrumpRx. Any analysis that folds the -25% ex-US number into a US cash-pay narrative is making a factual error that a reader can catch by reading a single sentence of the 8-K.

There is a bonus in the ex-US column that is worth pausing on, because it is the closest thing this dataset contains to a controlled experiment on the elasticity question. Outside the US is also where semaglutide has actually lost exclusivity. IQVIA reports the loss of exclusivity beginning in early 2026 across India, China, Brazil, Turkey and Canada — markets that "collectively account for 38% of adults living with obesity globally." [19] So the ex-US book is simultaneously absorbing a government reimbursement price cut and the arrival of generic competition on the rival molecule.

And in that book, Lilly's realised price fell hardest — (25) — and its revenue grew fastest: 81%, on volume of 95 %. [2] The market where the price shock is most severe is the market where the growth is most violent. If you want a single fact that inverts the hypothesis, it is that one.

Two housekeeping points that the rest of the report depends on, and that belong here.

The first is a limitation that constrains us as much as anyone. Neither company decomposes its reported "price" line into rate cuts within a channel versus channel-mix shift. Lilly publishes one number — Price (7) (25) (13) — and no split. [2] Novo publishes no price/volume decomposition at all, only the repeated qualitative phrase "lower realised prices." [5] Every mix claim in this space is therefore an inference, not a disclosure — including the hypothesis we are killing, and including our own counter-argument. We will be explicit throughout about which numbers are disclosed and which are ours.

The second is a small, telling absence. The words LillyDirect, TrumpRx, Medicare and Medicaid appear zero times in Lilly's Q1 2026 earnings release. The only phrase Lilly uses is the euphemism "cash pay prices." [1] A company that believed its channel was being structurally repriced would say more, not less.

Finally, for the avoidance of doubt: no US generic of semaglutide or tirzepatide exists today, and none features in the 2026-2027 price curve. [19] The American price path through 2027 is set by policy, by the Lilly-versus-Novo cash-price war, and by orals priced below injectables — not by generic entry.

Gross Margin Is Where a Mix Cliff Must Show First — and It Didn't

If a destructive channel-mix shift is under way, there is exactly one place it must appear before it appears anywhere else, and that place is gross margin.

The logic is mechanical. A cash-pay script carries no PBM rebate, no channel fee, no 340B or Medicaid best-price leakage. It is a different revenue-per-unit at the same cost-per-unit. Migrating scripts out of the insured channel and into the cash channel therefore changes realised revenue per unit sold before it changes demand, before it changes guidance, before it changes anything else on the income statement. Gross margin is the seismograph.

Lilly's Q1 2026 gross margin was 81.9%, "a decrease of 0.6 percentage points versus the same quarter last year. The change was primarily driven by lower realized prices." Non-GAAP: 82.6%, a decrease of 0.9 percentage points. [1] [2]

Sixty basis points. That is the scratch left by a mix shift that is 45% complete on the stock and 55% complete on the flow. [15]

And it gets worse for the hypothesis, because the full-year number moves the other way. In fiscal 2025 — the year Lilly cut cash prices and signed the MFN agreement — its gross margin rose: "Gross margin as a percent of revenue 83.0 % 81.3 %," 2025 against 2024. [3] That is not a year in which price was spared, either. The FY2025 bridge reads "Volume 53 % 46 % 50 % / Price (10) % — % (6) % / Foreign exchange rates — % 2 % 1 % / Percent change 43 % 48 % 45 %" — US realised price down a full 10%, a deeper cut than Q1 2026's -7%, with revenue up 45% to $65,179m and margin expanding. [3]

Source: Lilly FY2025 10-K and Q1 2026 8-K; Novo Nordisk FY2025 and Q1 2026 6-Ks

Now Novo, where the margin damage is real and an order of magnitude larger. Adjusted gross margin in Q1 2026 was 80.6%. [5] Full-year 2025 gross margin was 81.0%, against 84.7% in 2024 — a 3.7 percentage point fall, and note the timing: that happened before MFN pricing bit. [7]

Here is where the mix thesis's best piece of evidence turns in its hand. If you want to argue that channel mix destroyed Novo's margin, you have to get past Novo's own explanation of its margin, which names mix as a tailwind. Its 6-K attributes the gross margin move to lower realised prices, one-time costs and a negative currency impact — "partially countered by a positive product mix from increased GLP-1 sales." [5]

The one company in this story whose margin actually broke says that mix helped it. That sentence is the single most inconvenient fact for the hypothesis in the entire ledger, and it comes from the hypothesis's own star witness.

Why "cash price is below list price" is the wrong comparison entirely

Underneath all of this sits a piece of arithmetic that most cash-versus-insured comparisons get backwards. The naive mix model compares a $349 or $449 cash price against a four-figure list price and concludes catastrophe. But nobody pays list, and nothing like list ever reaches the manufacturer.

Novo's 20-F discloses the size of the gap with unusual candour: "total US rebates, discounts and sales returns amount to DKK 394,631 million, corresponding to 70 % of gross sales in the US ( 69 % in 2024 and 74 % in 2023 )," on gross sales of DKK 729,423 million. [6] Seventy per cent. The insured channel already nets roughly thirty cents on the list dollar.

Lilly's disclosure is less direct but no less startling. Its US rebate, discount and return roll-forward reads: "Sales rebate, discount, and return liabilities, beginning of year $ 10,313 ... Reduction of net sales (1) 62,135 41,452 ... Cash payments (57,299) (41,807) ... end of year $ 15,148" — against US net revenue of $43,481 million. [3] Lilly wrote off more in US rebates and discounts than it booked in US net revenue. On our arithmetic that implies roughly 58.8% of US gross sales deducted in 2025, up from about 57.7% in 2024 — a reconstruction, since Lilly does not publish a gross-sales line, and one that shows the rebate wedge widening by only about a point a year.

Once you see the gross-to-net machine, the entire framing of the mix hypothesis changes. A cash script does not fall from list to $349. It falls from net to $349. And whether that is a fall at all depends entirely on where net was — which is the subject of the next section, and the argument of this report.

Dilution Is a Price You Chose, Not a Channel You Entered

The mix hypothesis assumes the cash channel is intrinsically dilutive. It is not. Dilution is a property of the price you set inside it, measured against the net price you were already realising. Two companies entered the same channel. One priced at parity with its own net. The other priced below. That single decision — not the channel, not the deal, not the portal — is what separates +56% from -4%.

Lilly priced at its net

From Lilly's own announcement of the MFN agreement: "Zepbound multi-dose pen will be available at the lowest dose at $299, with additional doses up to $449 representing a $50 discount to current direct-to-patient prices... When patients refill their multi-dose pen prescription on the LillyDirect digital health platform, they will pay no more than $449." [11]

Now the comparison that matters. Leerink Partners' team, led by David Risinger — this is a third-party estimate, not a company disclosure — "estimates that the average U.S. commercial plan pays about $450 to $500 per month for Zepbound." [16]

Read those two numbers side by side. The cash refill cap is $449. The estimated commercial net is $450 to $500. Moving a Zepbound script from insurance to cash costs Lilly approximately nothing per script. The channel that was supposed to be a tax is, at Lilly's chosen price, roughly revenue-neutral — and it comes with no rebate administration, no PBM intermediation, and a direct relationship with the patient.

Novo priced below its net

Novo did something different, and it did it voluntarily. In November 2025 it announced that Wegovy 0.25 mg and 0.5 mg and Ozempic were available "at a limited time price of $199 per month to new self-pay patients," and that after two months "patients will move to the new standard monthly self-pay price which Novo Nordisk is lowering from $499 to $349 per month." [12] It did this, in its own framing, months in advance of its MFN commitment — a voluntary acceleration of its own price erosion. The live NovoCare grid confirms the standard: Wegovy pen at $349/mo across 1 mg, 1.7 mg and 2.4 mg, the HD pen at $399, and the pill at $149 to $299 by dose. [13] Novo has since added a subscription option through telehealth partners, at a price it does not disclose. [5]

What was Novo's insured net? Novo does not disclose it, so we have to back it out — and we flag clearly that what follows is researcher arithmetic, order-of-magnitude only. Take Novo's US Operations Wegovy injectable sales of DKK 9,493 million for Q1 2026 [5], convert at roughly 6.38 kroner to the dollar, and divide by the disclosed script counts — around 230,000 weekly as of January 2026 rising to around 270,000 by mid-April [7] [5] — and you get a blended realised price somewhere around $424 to $498 per monthly script. That blend is heavily caveated: revenue is sell-in to wholesalers while prescriptions are sell-out, so inventory swings distort it, and the FX rate is approximate.

But it cross-checks. Two companies, two entirely different methods — Leerink's bottom-up estimate for Zepbound and our top-down back-out for Wegovy — land in the same neighbourhood: the insured channel nets roughly $450 to $500 a month, not $1,000-plus. [16]

Now strip out the self-pay portion. If roughly 37% of Novo's US injectable scripts clear in the self-pay channel [5] — many of them at the $199 introductory price — and the overall blend is $424 to $498, then the remaining insured majority must be clearing at something like $460 to $525. That is our estimate, not Novo's, and it carries the further caveat that Novo's 70% gross-to-net is franchise-wide and includes insulin, whose rebate profile is worse than Wegovy's. [6]

Set the two decisions against each other and the picture is stark:

• Lilly: cash at $449, insured net estimated at $450-$500. Migration is roughly neutral. [11] [16]
• Novo: cash at $349 standard, $199 introductory — against an implied insured net of roughly $460-$525. Every migrating script surrenders something like a quarter of its realised price. [12] [13] [5]

	Eli Lilly — Zepbound	Novo Nordisk — Wegovy injectable
Standard cash / self-pay price[11][12][13]	$449 refill cap on LillyDirect (doses $299-$449)	$349/month standard (all doses 1mg-2.4mg); HD pen $399
Promotional / subscription price[11][12][5]	Lowest dose $299	$199/month introductory (first 2 months); a telehealth subscription option at an undisclosed price
Estimated insured net price[16][5]	$450-$500 (Leerink Partners estimate for the average US commercial plan)	~$460-$525 (our back-out from disclosed sales and script counts; order-of-magnitude only)
Cash price vs insured net[16][11][13][5]	At parity — migration is roughly revenue-neutral per script	Below net — each migrating script surrenders ~25% of realised price
Self-pay share of scripts[15][5]	~45% of total Zepbound scripts; 55% of NEW scripts (CFO, Q1 call)	~100,000 of ~270,000 weekly US injectable scripts = ~37% (our arithmetic on disclosed counts)
Q1 2026 outcome[1][2][5]	Revenue +56% to $19.8bn; US price line (7); gross margin 81.9% (-0.6pp)	Adjusted sales (4%) CER; US Operations (11%) CER; adjusted gross margin 80.6%
FY2026 guidance[1][5]	RAISED $2bn to $82.0-85.0bn (implies ~26-30% growth)	Adjusted sales growth -4% to -12% at CER

The model, in prose

Run the arithmetic and it closes on itself in a way that is hard to dismiss as coincidence.

Take Leerink's commercial-net range and use its midpoint. Take Lilly's disclosed 45% self-pay share of Zepbound scripts. Assume — and this is an assumption, not a disclosure — an average self-pay realisation of about $400, which sits inside the $299-$449 LillyDirect grid. [11] [15] [16] Blend them, and the resulting realised price is about 7% below an all-commercial book.

Lilly's disclosed US price line is -7%. [2] The channel-mix arithmetic and the reported price line reconcile.

Now change exactly one input. Leave the mix at 45%, leave the volume alone, leave the commercial net alone — and re-price the cash channel at Novo's $349. The same book now prints roughly -12%. Novo's US Operations adjusted sales fell 11% at CER, "driven by lower realised prices." [5]

Lilly, at Novo's price point, prints Novo's price line. Same mix. Same volume. One different decision.

Lilly today ($400 cash, 45% self-pay)Lilly at Novo's $349 cash priceAdd a 15% Medicare Bridge tranche at $245Reset

Insured commercial net price per monthestimate475$

Average realised self-pay price per monthassumption400$

Self-pay share of scriptsfact45%

Medicare Bridge share of scriptsassumption0%

Medicare / MFN price per monthfact245$

Commercial share of scripts55%

Blended realised price per script$441$

Realised-price change vs an all-commercial book-7.1%

Researcher scenario arithmetic on third-party and company inputs. The $400 average self-pay realisation is an ASSUMPTION sitting inside LillyDirect's disclosed $299-$449 grid, not a disclosure. The default scenario returns roughly -7%, which lands exactly on Lilly's disclosed US price line of (7) — the channel-mix arithmetic and the reported price line reconcile. Re-price the cash channel at Novo's $349 and the same book prints roughly -12%; Novo's US Operations adjusted sales fell 11% at CER. Same mix, same volume, one different decision. Neither company decomposes its price line into rate cuts versus channel mix, so this model is an inference, not a reconciliation of disclosures.

This is why we adjudicate the central conflict — is the cash channel dilutive? — as company-specific rather than class-wide. Both positions in that argument are true. They are simply true about different companies. AMCP, incidentally, gives the cleanest measure of how modest the real DTC cut actually was: the manufacturers "sell most doses of Wegovy and Zepbound directly to eligible cash-paying patients via their own direct-to-consumer platforms for $499 per month, meaning the MFN price would represent a 30% discount." [17] The headline framing — four-figure list price collapsing to a few hundred dollars — is a 74%-sounding cut against a price almost nobody paid. The real cash-channel cut is about 30%.

A convention on list prices, since they will keep intruding: we use approximately $1,000 to $1,100 a month for Zepbound and Ozempic and approximately $1,350 for Wegovy, and we say "approximately." [10] [16] The sources differ at the margins because of dose mix and reference year, not because they disagree; CMS's $959 for a 30-day supply of the semaglutide selected drug is a CY2024 figure across three brands. [8] No list-price figure in this report is quoted to greater precision than that, because none can be.

And one last inversion to close the section, because it demolishes the intuition underneath the whole mix story. The NovoCare price guide offers commercially insured patients "Pay as little as $25 (subject to a max savings of $100/mo)" across all Wegovy doses — against $349 a month cash. [13] The insured patient still pays far less than the cash patient. Cash-pay is not a cheaper deal that lures insured patients out of their coverage. It is the only deal available to people who have no coverage at all.

The Cheap Price Is the Volume — and Novo Is Losing the Units

The hypothesis contains a hidden ceteris paribus. It says the volume growth does not offset the mix — as though volume were an exogenous quantity, arriving on its own schedule, from which one then subtracts the price cut. But the cheap cash price is what buys the volume. You cannot hold volume constant and subtract the price reduction, because they are the same lever pulled once.

Lilly's management says so on the record. Dave Ricks on the Q1 2026 call: "75% of ex-US business for Mounjaro is out of pocket, and US is a meaningful portion as well — we see quite expansionary volume, perhaps nonlinear, to price reductions." And, more bluntly: "pretty much every time we reduce pricing, we see a pretty large expansion." [15] That is a company describing elasticity, not describing a tax. And the structure of Lilly's own mix disclosure — 55% of new Zepbound prescriptions are self-pay, against 45% of the total — tells you the cash channel is where the growth is being recruited, not where the base is leaking. [15]

The most spectacular evidence, though, comes from Novo, which is why it is so awkward for the case against the cash channel. Novo's Wegovy pill launched on 5 January 2026 at self-pay prices ranging from USD 149-299 per month by dose. By the week ending 17 April 2026, "total weekly prescriptions exceeded 200,000, mainly driven by GLP-1-naïve patient uptake and the 1.5 mg and 4 mg doses in the self-pay channel. Total prescriptions for Wegovy pill in Q1 2026 reached around 1.3 million, and total prescriptions since launch now exceed 2 million, making it the strongest-ever GLP-1 volume launch in the US." [5]

Read that again with the hypothesis in mind. The strongest volume launch in the history of the class is a cheap, self-pay, oral product recruiting patients who were never on a GLP-1 at all. That is not cannibalisation. That is market creation, and the low price is the mechanism.

(One discount to apply: Wegovy pill Q1 2026 sales were DKK 2,256 million, and Novo says the figure was "positively impacted by pre-launch pipeline fill with wholesalers and telehealth partners." [5] The revenue number is inflated by channel fill. The prescription counts are not.)

The third leg of the volume argument is what happened to the cheap-supply competition. Hims & Hers — the largest compounded-GLP-1 distributor — announced on 9 March 2026 that "the company will no longer advertise compounded GLP-1 offerings on its platform or in its marketing, and existing patients will have the opportunity to transition to FDA-approved medicines." Its stated reason: "Progress in the industry has made a broader range of FDA-approved GLP-1 treatments available at more affordable prices." [14] It simultaneously became a branded distribution channel for Novo. [14]

The brands' own price cuts killed the compounding arbitrage, and the largest arbitrageur voluntarily changed sides and became a volume conduit. That is a competitive weapon working exactly as intended — and it is incremental branded volume that is entirely invisible to a model which assumes cash scripts merely cannibalise insured ones.

The fact that finishes the mix thesis

Here is the structural problem. The mix hypothesis requires units growing while revenue per unit falls. That is what a mix shift is. If the units are not growing, mix is not the diagnosis — competition is.

On its diabetes book, Novo is losing the units. Ozempic came in at DKK 27,825m, down 15% as reported and (8%) at CER. Rybelsus was DKK 4,572m, down 20% and (15%) at CER. [5] Novo's own outlook cites "intensifying competition" in US Operations. [5] These are not scripts being repriced. These are scripts being lost — to Lilly, whose Mounjaro grew 125%. [2]

And on obesity, the pairing is brutal. "Sales of Wegovy injectable in Q1 2026 decreased by 11% at CER, driven by lower realised prices, partially countered by increased volumes." Immediately followed by: "The volume growth of the branded obesity market in the US was 85%." [5] Novo's US Operations obesity care sales rose 9% at CER, to DKK 11,752m, against an 85% market. [5]

A company growing 9% into a market growing 85% is not suffering from mix. It is suffering from share. And this is not a one-quarter artefact: the same gap opened in FY2025, when Wegovy sales in the US grew 16% at CER "driven by increased volumes, partially countered by lower realised prices" while "the volume growth of the branded obesity market in the US was 108%." [7]

Two years running, Novo has grown a fraction as fast as the market it invented. Mix is the mechanism by which the share-loser gets hurt. It is not a class-wide tax.

The best evidence against us

We owe the hypothesis its strongest card, and Novo deals it. In the outlook section of its Q1 report, Novo writes that uptake of the Wegovy pill "is reflected in the outlook, based on a range of assumptions related hereto such as market penetration, potential negative impact on the growth of the injectable obesity medication category as well as channel mix." [5] Its guidance commentary also explicitly names "lower realised prices linked to investments in market access amplified by the MFN agreement." [5]

That is a company telling you, in writing, that its cheap oral may cannibalise its expensive injectable, and that channel mix is in its guidance assumptions. It is the closest either company comes to endorsing the mix hypothesis on the record, and we do not wave it away. It means the mix effect is real at Novo — which is precisely our adjudication. What it does not do is generalise. Lilly, facing the same channel with more cash penetration, raised guidance.

And note the size of the prize that the cheap price is buying. J.P. Morgan Global Research — an estimate, and we name it as such — forecasts approximately 25 million Americans on GLP-1 treatment by 2030, up from around 10 million in 2025, and attributes 2026 growth directly to the thing the hypothesis treats as the problem: "In 2026, the size of the GLP-1 market is expected to grow significantly thanks to reduced prices, seniors getting access to obesity drugs and the approval of oral GLP-1s." [20]

The Steelman: The Cliff That Hasn't Happened Yet

Everything above is an argument about quarters that have been reported. It would be a poor report that stopped there, because the parts of this price reset that genuinely cut net realised prices are, almost without exception, still in front of us. What survives of the hypothesis is real. It is simply a price story and a forward story, not a mix story.

Here is what survives, in ascending order of danger.

One: per-script, cash really is dilutive at Novo. We have made that case ourselves. If you own Novo, the mix thesis is not wrong — it is your central problem, compounded by a share problem.

Two: the list-price scaffolding comes down on 1 January 2027. Novo will lower the WAC of Wegovy and Ozempic to USD 675, "representing reductions from the current list price of approximately 50% and 35% for Wegovy and Ozempic, respectively," and warns that "The change in list prices is expected to impact Novo Nordisk's cash flow in 2027." [5] [6] Once WAC is $675, the 70%-rebate machine has far less gross to give back, and cash and net prices converge from above. Nothing in any reported quarter reflects this.

Three: the orals reset the entry price of the class, and the government set the floor. NovoCare lists Wegovy pill at $149/mo for both 1.5 mg and 4.0 mg; Lilly's agreement states "Orforglipron will be available with the lowest dose starting at $149 with additional doses up to $399." [13] [11] Even at top dose the orals undercut the injectables — $299/mo for Wegovy pill 9 mg and 25 mg against a $349/mo pen and Lilly's $449 refill cap. [13] [11] And this is not a promotion that can be withdrawn: the White House states that "the initial dose of those drugs will be priced at $150 per month through TrumpRx," and AMCP confirms "the manufacturers agreed to price starting doses of these products at $150 per month." [10] [17] The fastest-growing format in the class is a price-down product with a government-fixed floor. That is a genuine, structural, forward headwind to revenue per patient — and it is the format both companies are pushing hardest.

Four, and this is the one that matters: the Medicare GLP-1 Bridge.

CMS describes it plainly: "The Medicare GLP-1 Bridge is a short-term demonstration run by CMS that will provide eligible Medicare Part D beneficiaries with access to certain GLP-1 drugs between July 1, 2026, and December 31, 2027... Eligible beneficiaries will have a $50 copay." It operates "outside of the Medicare Part D benefit's coverage and payment flow." [9] The manufacturer price behind that $50 copay is $245: "The Medicare prices of Ozempic, Wegovy, Mounjaro, and Zepbound will be $245," applying, per AMCP, to all doses and indications. [10] [17]

This went live on 1 July 2026. It appears in zero reported quarters. Everything in this report — Lilly's +56%, its 81.9% margin, its raised guidance, Novo's -4% — was printed before a single Bridge script was filled. [9]

And unlike the cash channel, $245 is a real cut to a real net price. Leerink's team — again, a third-party estimate — "estimates that the future $245 monthly price for Zepbound is about 55% of current Medicaid and Medicare net prices." [16] That is a genuine reduction of roughly 45% in realised price on every script that moves through it. Compare that to the roughly net-neutral cash channel and you can see immediately where the actual pricing risk lives: not in LillyDirect, but in Washington.

The volume base is enormous. CMS discloses 2,282,000 Medicare Part D enrollees who used the semaglutide franchise in CY2024, against $15,161,908,000 of Part D gross covered prescription drug costs. [8] Lilly is unambiguously excited: Ilya Yuffa on the Q1 call — "We are excited about having Part D access starting to activate for obesity medicines in July...The $50 per month copay is an important element of affordability for seniors." [15] That excitement is a volume bet against a price cut, which is the entire thesis of this report — but it is a bet, and it has not settled.

Layered on top, from CY2027, is the IRA's statutory instrument. CMS's negotiated maximum fair price for the Ozempic/Rybelsus/Wegovy selected drug is $274 for a 30-day supply, a 71% discount to the $959 CY2024 list. [8] And because CMS negotiated the three brands as a single selected drug — its own file list confirms it: "File for the MFP Explanation for Ozempic; Rybelsus; Wegovy" — that $274 cannot be ring-fenced to diabetes. [22] It hits Wegovy too.

The government has, in short, made itself the cheapest channel in America: $245 to $274 on Medicare, against $346 to $350 on the government's own TrumpRx cash portal, which is live, against $349 to $449 on the manufacturers' own platforms. [10] [17] [21] [13] [11] That is a strange inversion, and it is where the next chapter of this story will be written.

There is a counter-current worth naming. The Medicaid half of the MFN deal is a price cut with very little volume behind it. Both companies agreed to "provide every State Medicaid program in the country access to MFN drug prices" [10], yet KFF reports that "Obesity drug coverage in Medicaid remains limited, with 13 state Medicaid programs covering GLP-1s for obesity treatment under fee-for-service (FFS) as of January 2026," and flags that "how the new lower costs compare to the net prices state Medicaid programs currently pay and how states will respond amid tightening budget conditions remain unclear." [18] Novo's own outlook cites "negative impact from reduced obesity medication coverage in Medicaid." [5] The government channel giveth and it taketh.

1. 6 Nov 2025

   MFN deal signed: Medicare at $245, TrumpRx at ~$350, orals at $150[10][11][17]

   Both companies agree Medicare/Medicaid pricing of $245/month with a $50 copay, an average ~$350 on TrumpRx, and $150 starting doses for orals. Lilly's release carries the load-bearing caveat: "The scope of the agreement does not include pricing obligations in the commercial channel."

2. 17 Nov 2025

   Novo cuts its standard cash price from $499 to $349 — voluntarily, and early[12][11]

   $199/month introductory for new self-pay patients on the two lowest doses. Novo brings these prices forward "months in advance" of its MFN commitment. Lilly's comparable refill cap is $449.

3. 5 Jan 2026

   Wegovy pill launches at $149-299/month self-pay[5][13]

   By mid-April: 200,000+ weekly scripts, 2 million+ since launch, "the strongest-ever GLP-1 volume launch in the US" — mainly GLP-1-naive patients in the self-pay channel.

4. 9 Mar 2026

   Hims & Hers exits compounded GLP-1s and becomes a branded channel[14]

   The largest compounded-GLP-1 distributor stops advertising compounded offerings because "a broader range of FDA-approved GLP-1 treatments [are] available at more affordable prices." The brands' own cash cuts killed the arbitrage.

5. 30 Apr 2026

   Lilly raises FY2026 guidance by $2bn — with the mix shift half-done[1][15]

   Q1 revenue +56% to $19.8bn; gross margin 81.9%; self-pay at 45% of total and 55% of new Zepbound scripts. Guidance goes from $80-83bn to $82.0-85.0bn.

6. 1 Jul 2026

   MEDICARE GLP-1 BRIDGE GOES LIVE — $245/month, $50 copay[9][10][16]

   A CMS demonstration running to 31 Dec 2027, outside the Part D coverage and payment flow. Leerink estimates $245 is only ~55% of current Medicare/Medicaid net — a genuine ~45% cut, unlike the roughly net-neutral cash channel. It has touched ZERO reported quarters.

7. Oct/Nov 2026

   Q3 2026 results — the first quarter containing Bridge scripts[9]

   The steelman of the hypothesis. If either company prints a realised-price shock here, the mix effect was not second-order — but it would be a GOVERNMENT-pricing story, not a DTC one.

8. 1 Jan 2027

   Novo's WAC falls to $675 — list price cut of ~50% (Wegovy) and ~35% (Ozempic)[5][6]

   Once WAC is $675, the 70%-rebate machine has far less gross to give back, and cash and net prices converge from above. Novo warns it "is expected to impact Novo Nordisk's cash flow in 2027." Nothing in any reported quarter reflects this.

9. CY2027

   IRA maximum fair price of $274 binds on the Ozempic/Rybelsus/Wegovy selected drug[8][22]

   A 71% discount to the $959 CY2024 list. Negotiated as a single selected drug — so it hits Wegovy, not just the diabetes book.

10. Late 2027

    THE ONE SCENARIO THAT MAKES THE HYPOTHESIS RIGHT: Klomp's 24-month commercial glide path[17][11]

    CMS Medicare director Chris Klomp said manufacturers agreed to take injectable commercial prices to the Medicare/Medicaid level within 24 months. If honoured, there is no $450-$500 commercial net left for Lilly's $449 to sit at parity with, and the cash channel flips from neutral to sharply dilutive. It is a verbal statement, directly contradicted in writing by Lilly.

The one scenario that makes the hypothesis right for the whole class

At the November 2025 announcement, Chris Klomp, Director of Medicare at CMS, "stated that manufacturers have agreed to lower the price of injectable GLP-1s to the same level as the price offered to Medicare and Medicaid over the next 24 months." [17]

If that is honoured, the entire US injectable book converges on approximately $245 by late 2027. At that point Lilly's $449 cash price is no longer at parity with a $450-$500 commercial net, because there is no $450-$500 commercial net. The cash channel flips from roughly neutral to sharply dilutive, and the mix hypothesis becomes right — not because mix was ever the mechanism, but because the price level collapsed underneath it.

Against that stands Lilly's own written language: "The scope of the agreement does not include pricing obligations in the commercial channel." [11] BioPharma Dive reports that Lilly "does not have any obligations" to cut commercial prices, though it does expect them to "come down over time, but also expects significant volume upside." [16]

We adjudicate this the way any careful reader should: the written company disclosure outranks the verbal official. There is no commercial pricing obligation today, and no filing corroborates one. But Klomp's statement is the single largest identified forward risk in this entire analysis, and we flag it as such.

What would change our mind

We would rather be specific than safe. Here is what would falsify this report.

• A Lilly guidance cut that management explicitly pins on channel mix. Today the mix shift is roughly half-executed — 45% of total Zepbound scripts, 55% of new — and it produced a $2 billion guidance raise, US Zepbound +79% to $4.1 billion, and a gross margin of 81.9%. [15] [1] If self-pay climbs past 55-60% of total scripts and guidance is still going up, the mix thesis is unrecoverable. It needs a cut, and it needs management to name mix as the cause. Nothing less counts.
• A break in Lilly's gross margin below roughly 78-79%. That is where a destructive mix shift must first appear. Q1 2026 was 81.9%, and FY2025 rose to 83.0% from 81.3% in the very year the cash prices were cut. [1] [3] Until margin breaks, there is a sixty-basis-point scratch and no cliff.
• Actual evidence of cannibalisation — commercial script counts declining as self-pay rises. Right now the cheap price is recruiting the naive, not converting the insured, and the insured patient still pays $25 with a copay card against $349 cash. [5] [13] If commercial scripts start falling as self-pay rises, mix acquires a mechanism. Until then it has none.
• The Medicare Bridge quarter — Q3 2026, reported in October or November. This is the steelman. If either company prints a realised-price shock in the first quarter that contains Bridge scripts, we were wrong to call the mix effect second-order. But note carefully: that would be a government-pricing story, not a DTC story, and it cannot be written in the past tense today. [9]
• Novo's US volume line, not its price line. If Novo's units keep shrinking — Ozempic (8%) at CER, its diabetes book weak — its decline is a share loss and the word "mix" is doing no analytical work. [5] Conversely, if Novo's US units start growing strongly while its US revenue still falls, the hypothesis is vindicated in its pure form and our share adjudication fails.

Housekeeping, because the numbers in this sector are treacherous

Three disciplines the reader should demand of anyone writing about this market, including us.

There is no "$150 billion market" in the present tense. The two companies who are this market run at roughly $85 billion annualised today — Lilly's incretins at about $51 billion a year on Q1 run-rate, Novo's GLP-1 franchise at about $33 billion, on our arithmetic from the two product tables. [2] [5] The nine- and twelve-figure market projections you have read are 2030 forecasts, and they only materialise if cheap pricing drives the volume the hypothesis says is not offsetting anything.

IQVIA's market figures are list-price figures and must be labelled as such, every time. IQVIA's ForecastLink — an estimate — has "The global Obesity Medicines market reached $66 billion in list price sales value in 2025... and is forecast to rise to $92 billion in 2026, before reaching between $105 billion and $200 billion by 2027 and beyond." [19] Against Novo's disclosed 70% US gross-to-net, a list-priced market number overstates realised US revenue by roughly threefold. [6] Worse: Novo's WAC cut to $675 will shrink a list-priced market with zero change in the number of patients treated. [5] And the $105 billion to $200 billion spread for 2027 is IQVIA quietly conceding that price, not volume, is the swing factor.

Do not quote Novo's headline growth. Reported net sales of DKK 96,823m and 32 % growth at CER are an accounting artefact — the quarter was "positively impacted by a provision reversal related to the 340B Drug Pricing Program in the US," against a provision that stood at USD 4.2 billion at 31 December 2025. [5] [7] The number that describes the business is the adjusted -4% at CER.

The call

The cash-pay price cut is not a mix tax that volume fails to offset. It is the volume engine. Lilly, with self-pay already at 45% of total Zepbound scripts and 55% of new ones, grew revenue 56% on volume of 65% against price of (13), held gross margin at 81.9%, and raised full-year guidance by two billion dollars to $82.0 billion to $85.0 billion. [15] [2] [1] Novo, which set its cash price below its own insured net rather than at parity, fell 4% at CER and is losing the units themselves. [5]

The same channel produced a roughly thirty-five-point divergence in guided growth. A variable both companies share cannot do that. The operative variable is share, and the lever is the price you choose inside the channel — not the channel itself.

The cliff exists. It is just not where everyone is looking. It is at $245, it is administered by CMS, it went live thirteen days ago, and it has not appeared in a single reported quarter. [9] Publish the divergence, not the cliff.