The Problem of Too Much Demand

Novo Nordisk spent a hundred years being admirably dull. Founded in Denmark in the 1920s to manufacture the newly discovered insulin, it grew into one of the world's two or three dominant diabetes-care companies — foundation-controlled, long-horizon, deeply unglamorous. Its scientists had spent decades on a class of molecules called GLP-1s, which mimic a gut hormone that regulates blood sugar and, as it happens, appetite. Out of that patient work came semaglutide: approved for diabetes in 2017 under the name Ozempic, and for weight management in 2021 as Wegovy.

What happened next had no precedent in pharmaceutical history. Clinical results showing weight loss in the range of fifteen per cent collided with social media, celebrity culture and an obesity epidemic affecting a billion people, and demand went vertical. "Ozempic" stopped being a brand and became a word — in monologues, in headlines, eventually in the dictionary conversations. Prescriptions ran to millions; surveys suggested enormous fractions of the adult population wanted the drug; and Novo's factories — built for the sober arithmetic of diabetes care — were suddenly being asked to supply a global consumer phenomenon.

The market did the rest. Novo's valuation multiplied until, in 2023, it passed LVMH to become Europe's most valuable company — a Danish insulin maker worth more than the entire annual output of Denmark. Its results began visibly moving national statistics: economists noted that much of Danish growth was attributable to one company, the central bank held interest rates lower partly against the currency effects of pharmaceutical export earnings, and commentators reached for an uncomfortable precedent — Finland, Nokia, and what happens to a small country wired to one corporate engine.

Then the blessing began behaving like a crisis. Demand outstripped manufacturing capacity almost immediately and stayed ahead of it for years; injectable biologics cannot be scaled like software, and a new sterile filling line takes years to build and validate. The company found itself doing things no marketer dreams of: deliberately restraining launches in new countries, holding back starter doses so existing patients could stay supplied, and asking doctors, in effect, to stop creating new customers. Diabetes patients complained of shortages caused by demand for weight loss; regulators opened questions; compounding pharmacies in America began legally producing unauthorised copies during the official shortage — competition conjured into existence by the company's own inability to supply.

Money answered part of it. Novo committed sums in the tens of billions to new plants across Denmark, France and the United States, and its foundation moved to buy one of the world's biggest contract manufacturers to lock up filling capacity — a deal large enough to draw regulatory scrutiny of its own. But capacity was only the first front. American politics opened a second: list prices around a thousand dollars a month for a drug most of the population arguably could use made Novo a standing exhibit in Senate hearings on drug pricing. And competition opened a third: Eli Lilly's rival tirzepatide arrived with trial results showing greater average weight loss, and the two companies settled into the classic duopoly race — next-generation molecules, oral versions, trial readouts traded like artillery.

Through it all ran the strangest strategic question in the case: how big is this market actually? Forecasts for the obesity-drug market by 2030 ranged from fifty billion dollars to a hundred and thirty — a spread so wide it amounted to a confession. Every capacity decision, every plant, every acquisition was a bet on a number nobody could know, taken under conditions where building too little meant handing the market to Lilly, and building too much meant owning the world's most expensive empty clean-rooms.

The descent, when it came, needed no scandal — only arithmetic and expectation. Through 2024 and 2025 the extraordinary premium deflated piece by piece: a next-generation readout that was merely good rather than miraculous knocked a fifth off the shares in a day; Lilly's data kept pressing; American pricing pressure hardened; supply normalised just as investors noticed that normal supply also meant normal growth. The share price roughly halved from its peak; the chief executive who had presided over the miracle years departed in 2025 amid board pressure; thousands of job cuts followed. Nothing about the underlying business had failed — revenues and profits remained enormous — but the company had been priced for a story, and the story had become a business again. Readers of the WeWork case will recognise the mechanism operating here without a single governance sin.

Ahead sits the industry's oldest cliff with the industry's newest numbers on it. Semaglutide's key patents expire in major markets in the early 2030s — in some large countries sooner — at which point biosimilar copies will do to Wegovy's price what generics always do. The defence is the eternal one, run at unprecedented scale: pour today's extraordinary earnings into the pipeline — oral versions, next-generation molecules, trials in heart disease, kidney disease, liver disease, even dementia — so that when the cliff arrives, something new is standing on the far side. Whether the pipeline delivers is a question currently being answered in trial data, one readout at a time.

The case remains genuinely open, which is why it closes this series. A hundred-year-old company built the decade's defining product, was crowned Europe's champion, distorted its home country's economics, met a demand it could not supply, and was then repriced by success becoming ordinary — all without a villain anywhere in the story. The question it leaves for any leader is the one the forecasting grammar exists to answer honestly: when your own plan assumes a number — for demand, for growth, for how long the miracle lasts — what is the widest range you would defend out loud, and what would you build differently if you said the range instead of the number?