Playbook › Themes
GLP-1, obesity & metabolic pharma
The field-level conclusion
Separate two questions that the field's own promoters keep fused: is obesity a forever business, and is any given incumbent a forever winner in it? The first is a confident yes — ~1bn addressable patients, chronic lifelong dosing, expanding indications (CV, MASH, sleep apnea, CKD), and as of 2026-07-01 the first-ever Medicare obesity coverage via the CMS BALANCE / GLP-1 Bridge pilot. The second is a separate bet entirely, and on 2026-07-31 the evidence said the answer for Novo Nordisk is no.
The field's defining dynamic: volume expands while value per patient collapses. Every pricing vector points the same way and most are dated, not speculative:
| Vector | Status as of 2026-07-31 |
|---|---|
| MFN / TrumpRx | Live since Jan 2026. Ozempic ~$1,000 → $350/mo; oral Wegovy $149 |
| IRA Medicare negotiation | Effective 2027-01-01. Ozempic $274/mo, −71% off list; Wegovy $385 |
| Amazon Pharmacy | Carries oral Wegovy from $149 cash to ~200m Prime members |
| Canada | Patent lapsed on an unpaid fee; Dr. Reddy's generic approved April 2026 — first G7 generic. At 3 generics, statutory price falls to 35% of brand |
| China | Core patent expired 2026-03-20; >10 generic filings |
| US / EU | Intact — US June 2032, EU March 2031 |
Net: a category whose dollar value probably still grows at the field level, and whose incumbent-specific economics are being competed and legislated away simultaneously.
Why the duopoly stopped being a duopoly
The historical entry barrier was "$3–5B and ~10 years to get a credible incretin into Phase 3." Chinese biotech out-licensing collapsed it to roughly $300–500m upfront and 3–4 years, because the asset arrives already de-risked through Ph1/2 at a fraction of Western cost. Q1 2026 alone saw ~$60B of Chinese biotech out-licensing: AstraZeneca–CSPC (up to ~$18.5B, as many as 8 incretin assets), GSK–Hengrui (~$500m upfront), Lilly–Innovent, Pfizer–Metsera ($4.9B).
Any top-15 pharma can now buy a seat at this table for less than one quarter of Novo's operating profit. Whatever "R&D scale" meant as a moat in this field in 2023, it does not mean it now. Also in the queue: Amgen MariTide (monthly dosing), Roche (Carmot CT-388 + Zealand petrelintide), Viking VK2735 (Ph3 fully enrolled), Structure aleniglipron.
The structural shift that matters most — injectable → oral small molecule
This is the under-priced change in the field. Peptide manufacturing requires ~13,000 kg of input per kg of API versus 168–308 kg for a small molecule, plus aseptic fill-finish, cartridges and devices. Small-molecule orals can be made by any competent CDMO at a fraction of the COGS. Lilly's orforglipron (Foundayo, approved 2026-04-01) is a small molecule; Novo's oral semaglutide is a peptide needing an absorption enhancer and enormous API tonnage — the highest-cost formulation in the portfolio competing against the lowest.
Consequence for the field: manufacturing capacity stops being a moat and becomes a depreciation schedule. Any thesis in this space resting on "they have the capacity" was a 2023–25 thesis. Check what the capacity is for.
Efficacy leaderboard as of 2026-07-31
Lilly leads on every published head-to-head. This is the single cleanest fact in the field.
| Asset | Company | Weight loss | Status |
|---|---|---|---|
| Retatrutide (triple agonist) | LLY | ~30% | BLA Q1 2027 |
| Tirzepatide / Zepbound | LLY | 23.6–25.5% | Marketed, outsells Wegovy |
| CagriSema | NVO | 20.2–23.0% | Failed non-inferiority vs tirzepatide, Feb 2026 |
| Oral semaglutide | NVO | ~17% | Marketed, launch beating expectations |
| Orforglipron (oral, small molecule) | LLY | ~12% | Approved 2026-04-01 |
Lilly holds ~60% of the US GLP-1 market and overtook Novo ex-US in May 2026. Lilly guides +20–25% revenue; Novo guides −4 to −12%. The ~3x forward-multiple gap between the two is the market's verdict on who won, and it reads as rational rather than as an arbitrage.
The best analogue for the post-cliff arc — Novo's own insulin history
Rejected analogues first: Gilead/Sovaldi is the wrong decline mechanism (hep-C cured its patients away; obesity is chronic). Humira/AbbVie fits the cliff event — AbbVie survived only because Skyrizi and Rinvoq were already >$10B franchises when it hit.
The best fit is Novo's own insulin franchise. Same company, same payers, same PBM rebate mechanics: volumes held, US net price fell ~80% over a decade, and the segment became a low-growth annuity (FY2025: −1% CER). Novo discontinued Levemir rather than sell at the capped price. Insulin is the base case for what GLP-1 becomes after 2032, and it is a worked example sitting inside the same income statement.
What to check first on any name in this field
- Decompose the yield/valuation before believing it — see [[pitfall-dyt-inverts-when-price-caused-the-yield]], which was found on NVO in this field.
- Label the revenue basis — script data is gross, filings are net, and the spread ran ~70% in GLP-1. See [[pitfall-drug-sales-gross-vs-net-basis]].
- Recompute every multiple for non-USD reporters — see [[pitfall-yahoo-mixes-cad-usd-on-cross-listed-issuers]] (NVO's EV/EBITDA printed as 1.92 against a true 9.7x).
- ROIC trend, not ROIC level. NVO's fall from 70.6% → 55.5% → 36.0% across FY23–25, against invested capital tripling with flat NOPAT, was the earliest clean quantitative signal of moat erosion — visible before the narrative turned.
- Gross margin is the net-price tell. NVO's 84.7% → 81.0% in one year showed the rebate spread widening before any of it reached commentary.
Verdict for the watchlist
No entries added. NVO held at conviction 5.0 (HOLD, do not add) — the incumbent is at a
decade-low multiple and deteriorating on every trend line, which is the "good-and-fading
at a fair price" cell of the §0 grid, not the value cell. LLY was not analysed and is the
obvious gap: it has won every head-to-head in the field but trades ~40x forward, so the
question there is entirely price, not quality. Worth a /analyze before any conclusion
about the field's investability is treated as complete.
What would change this
- NVO: zenagamtide/amycretin Ph3 AMAZE data (~2028) is the whole long-term thesis; oral Wegovy's script trajectory holding as orforglipron scales is the near-term one.
- Field: whether volume growth can outrun a 5–10x price decline. Answered on FY2026 results in early 2027, not before.
- A pure-play at a genuinely distressed price, or LLY de-rating to a level where winning the field is available at a fair multiple.
History
- 2026-07-31 — field swept for the first time via the NVO analysis. Three playbook notes
came out of it. Overdue at ~100 days (
cycle) or on any of the dated catalysts above.