Who keeps the profit as peptide medicines move from GLP-1 hype to long-term care?

Peptide medicines are engineered chains of amino acids: the body’s signalling language, adapted so they can be manufactured, injected or swallowed, and used for years. The theme brings together industries that developed separately: century-old insulin plants, academic hormone biology, industrial chemistry, injection-device engineering, the difficult history of obesity treatment, US pharmacy-benefit negotiations, and newer manufacturers in China and India.

GLP-1 drugs exposed those connections because demand briefly exceeded capacity to make the molecule, fill it safely and package it in pens. Obesity is the largest current chapter in a far older therapeutic category. The market is now shifting from scarcity towards pricing, long-term adherence and competition. Protected drug owners still capture the largest share of profit; qualified device and formulation specialists can retain a meaningful share. Ingredient suppliers, compounders and generic entrants are more likely to lose the scarcity rents of the shortage period.

The first chain: insulin teaches medicine to make a hormone

In the summer of 1921, at the University of Toronto’s physiology department, a young surgeon and a medical student were trying to isolate a substance no one had yet captured. Frederick Banting, who had served in the First World War and returned to a poorly paid medical practice in London, Ontario, had become convinced that the pancreas produced a substance that controlled blood sugar, but that digestion destroyed it before extraction. The department head, the Scottish physiologist J. J. R. Macleod, gave him laboratory space, dogs and an assistant, Charles Best.¹ Their extracts lowered blood sugar in diabetic dogs.

Turning that result into a medicine required a third researcher: James Collip, a University of Alberta biochemist visiting Toronto, who purified the extract sufficiently for use in people. In January 1922, Leonard Thompson, a fourteen-year-old with diabetes, became the first patient to receive the treatment and survived.¹

Before insulin, a child diagnosed with type 1 diabetes was expected to die, often after months on starvation diets. Insulin turned the disease into one managed daily for a lifetime. That shift—from death sentence to chronic condition—became the template for much of peptide medicine.

A sentence written in amino acids

Insulin is a peptide hormone. Amino acids are its building blocks: arrange them in a precise sequence and the resulting chain can carry a signal to cells. Insulin’s signal is to take in glucose. It is like a sentence assembled from an alphabet of roughly 20 letters, in which one changed letter can alter the meaning. The analogy has limits: a peptide also folds into a three-dimensional shape, and that shape, not simply its sequence, determines whether it fits a cell receptor.

That sequence was unknown in 1922. The Cambridge biochemist Frederick Sanger spent roughly a decade breaking insulin down piece by piece. In 1955, he reported its complete amino-acid sequence—the first demonstration that a protein has a single, exact chemical order.² The work earned him the first of two Nobel Prizes. Its commercial importance emerged later: once a molecule’s sequence is known, it can in principle be copied, altered and improved.

Building the chain on a bead

Making those copies was harder. Chemists could join amino acids in solution, but every step required isolation before the next link could be added, compounding losses. In 1959, Bruce Merrifield, a biochemist at the Rockefeller Institute in New York, sketched a different approach: attach the first amino acid to a small insoluble bead, add the next, wash away material that did not attach, and repeat.³ He published the method, solid-phase peptide synthesis, in 1963.⁴ Within several years, his laboratory had automated it; by 1969, it had used the method to build the 124-amino-acid enzyme ribonuclease.⁵ He received the 1984 Nobel Prize in Chemistry.³

The process resembles assembling a necklace on a fixed clasp: the clasp holds the growing chain while successive beads are threaded on, and loose beads are washed away. But the limits of the analogy explain much of the industry’s economics. Even if each coupling works 99 per cent of the time, a 30-step chain is fully correct only about three-quarters of the time. The remaining near-miss molecules must be removed, and a near miss in a drug can become a regulatory problem. Folding, purity, stability and scale turn a laboratory technique into an industrial discipline. The gap between making a molecule and making tonnes of it consistently is where much of the manufacturing risk—and profit—remains.

The first industrial franchises

Insulin also taught companies how to commercialise a peptide. Toronto’s discoverers sold their patent to the university for a nominal sum, and the university needed partners able to manufacture the drug at scale. Eli Lilly $LLY, based in Indianapolis, reached an agreement and was selling insulin within two years.¹ In Denmark, the Nobel-winning physiologist August Krogh brought the Toronto method home. The Danish insulin laboratories founded in the 1920s eventually merged into Novo Nordisk in 1989.¹ The two companies that now dominate obesity medicine were therefore shaped by a century of manufacturing and selling injectable hormones.

The inheritance extended beyond the molecule. Diabetes created endocrinology practices, clinics and nurses able to teach self-injection; cold-chain distribution for fragile proteins; reusable and later disposable injection devices; and insurance arrangements for medicines taken indefinitely. That infrastructure was in place long before Wegovy.

Peptide medicines also expanded well beyond diabetes. Novartis $NOVN.SW sells Sandostatin, a long-lived version of somatostatin for hormone-secreting tumours, and Lutathera, a peptide that delivers a radioactive payload to the same tumours. Together, the drugs generated about $527 million in the second quarter of 2026, less than 4 per cent of Novartis’s sales.⁶ Ipsen $IPN.PA sells lanreotide as Somatuline and the hormone therapy Decapeptyl; together, they generated roughly €498 million in the same quarter.⁶ Takeda $4502.T sells a GLP-2 treatment for short-bowel syndrome; Recordati $REC.MI sells pasireotide for rare hormonal disorders; Rhythm Pharmaceuticals $RYTM sells setmelanotide for obesity caused by rare genetic defects; and Ironwood Pharmaceuticals sells linaclotide, marketed as Linzess, for chronic constipation.⁶

Is a peptide automatically a good business?

It is tempting to treat peptides as natural franchises: hormones the body already uses, made more durable and sold for life. The record is less forgiving. Early peptide drugs often required several injections a day because the body cleared them within minutes or hours. Many were expensive to manufacture, and some could not be taken orally because the gut digests peptides as it digests food.

Even established franchises can contract. Ironwood’s revenue fell 21 per cent in 2024 and a further 16 per cent in 2025; its market value is roughly one-third of its 2021 level.⁶ Peptide science is necessary but insufficient. A molecule becomes a business only if it lasts long enough in the body, can be made consistently, can be delivered by patients and is accepted by the payer.

By the 1980s, the scientific challenge had narrowed: the body produced many useful hormonal signals, but broke several of them down before they could become practical medicines.

↑ Back to Top

A gut signal with a two-minute life

In the early 1980s, molecular biologists studying the gene for glucagon, a pancreatic hormone that raises blood sugar, found that its instructions encoded more than glucagon. The same stretch of DNA also coded for two related peptides, later named glucagon-like peptide-1 and glucagon-like peptide-2.⁷ Their role was initially unclear.

Svetlana Mojsov helped establish it. Raised in Skopje, then part of Yugoslavia, she trained as a peptide chemist at Rockefeller University under Bruce Merrifield; in 1984, the two published the chemical synthesis of glucagon.⁸ After moving to Massachusetts General Hospital in Boston to run a peptide-synthesis facility, she began working with the endocrinologist Joel Habener.⁹ Mojsov proposed that the active form of GLP-1 was a shortened version of the peptide implied by the gene, and synthesised it by hand. In 1987, the Boston group and, at about the same time, Jens Juul Holst’s Copenhagen group showed that truncated GLP-1 strongly stimulated insulin release.⁷

The dinner reservation

The finding helped explain an observation physicians had made in the 1960s: glucose taken by mouth prompted much more insulin release than the same amount infused into a vein. The gut was signalling the pancreas ahead of a meal. Hormones that carry this signal are called incretins, and GLP-1 proved to be one of the most important.⁷

It works rather like a dinner reservation: the gut calls ahead so insulin is available when glucose arrives. Because its effect depends on glucose being present, GLP-1 lowers blood sugar with relatively little risk of driving it dangerously low. The metaphor has limits. GLP-1 also slows stomach emptying and affects appetite centres in the brain; researchers are still mapping those effects. It is one signal among several governing hunger and fullness, not a switch for willpower.

The catch

Native GLP-1 lasts in the bloodstream for only a few minutes before DPP-4, an enzyme, cuts it apart and the kidneys clear it.⁷ The biology was compelling, but a patient would have needed continuous infusion. In its natural form, GLP-1 was not a practical medicine.

The workaround came from the Gila monster. John Eng, an endocrinologist at the Veterans Affairs medical centre in the Bronx, had read that certain venoms inflamed the pancreas and wondered whether they contained hormone-like peptides. In 1992, he isolated one from Gila-monster venom and named it exendin-4.⁹ It resembled GLP-1 closely enough to activate the same receptor, but DPP-4 could not cut it, giving it a lifespan of hours rather than minutes. After the VA declined to patent the discovery, Eng patented it and licensed it to Amylin Pharmaceuticals, a small San Diego company.¹⁰ In April 2005, Amylin and Lilly received FDA approval for synthetic exenatide, sold as Byetta: the first GLP-1 receptor agonist, a drug that activates the GLP-1 receptor.¹⁰

Two lineages

Novo Nordisk took a different route, modifying the human peptide rather than borrowing the lizard’s. It changed one amino acid and attached a fatty-acid chain, allowing the molecule to bind albumin, the most abundant protein in blood. That protected it from DPP-4 and rapid kidney clearance. The resulting drug, liraglutide, lasted long enough for once-daily injection.¹¹ Semaglutide, a later refinement, extended the elimination half-life to about a week.¹² Novo’s experience modifying and manufacturing insulin helped turn these changes into products patients could use more easily.

Lilly entered the field through its Amylin partnership and internal metabolic research. It later chose a more ambitious design: one molecule that activates two gut-hormone receptors. That decision would shape the competitive contest discussed later in this story.

The new designers

A newer group of companies has tried to make peptide design more systematic. ペプチドリーム PeptiDream in Tokyo screens large libraries of ring-shaped peptides for compounds that bind difficult targets. Bicycle Therapeutics, in Cambridge, England, develops peptides constrained into two loops. Protagonist Therapeutics $PTGX in California designs peptides intended to remain stable when swallowed or to require less frequent injection. Gubra in Copenhagen designs metabolic peptides for partners; MBX Biosciences develops long-acting endocrine peptides; Entera Bio works on oral delivery; and Altimmune and NeuroBo have metabolic candidates.⁶

In this context, design is more than finding a molecule that binds a target. Binding is only the first test. The molecule must survive in circulation, avoid unintended targets, reach the relevant tissue, remain stable in an injection pen for months and be manufactured consistently at kilogram scale. Much of the commercial challenge begins after a promising peptide is found.

The financial results show why that work is difficult to monetise. Revenue at these companies often comes through licence fees and milestone payments, creating volatile results. PeptiDream’s revenue fell 60 per cent in 2025 after rising the previous year. Protagonist’s fell from $434 million in 2024 to $46 million in 2025 as milestone income receded. Gubra’s revenue rose almost tenfold in 2025 after a partnership deal.⁶ Bicycle, valued at about $250 million, had enough cash for roughly nine quarters at its first-half 2026 spending rate.⁶ Across the eight companies Empor tracks in this layer, the median operating margin has been deeply negative in each of the past five years.⁶

Early GLP-1 medicines illustrated the same distinction between scientific validation and a durable market. Byetta required twice-daily injections and caused nausea. It established the biology, but did not create a large obesity business and never approached the scale of later drugs. Showing that a signal works is one task; making a treatment that millions of people will take for years, and that payers will cover, is another.

By the late 2000s, Novo had the diabetes infrastructure to test that proposition. The remaining question was whether a diabetes drug could overcome obesity medicine’s long and damaged history.

↑ Back to Top

The warning label hanging over every obesity drug

On 15 September 1997, the US Food and Drug Administration asked the makers of fenfluramine and dexfenfluramine to withdraw the drugs; they agreed.¹³ Millions of prescriptions had been written. Fenfluramine, approved in June 1973 as an appetite suppressant, was widely paired with phentermine in the 1990s, when “fen-phen” clinics proliferated.¹⁴ Dexfenfluramine, a related compound, was approved in 1996. Within a year, Mayo Clinic physicians reported heart-valve damage in women taking the combination; further cases followed. The drugs were also associated with pulmonary hypertension, a rare and often fatal disease of the lungs’ blood vessels.¹³ A CDC follow-up report found abnormal heart valves among screened patients who had been exposed to the drugs.¹³

The episode was brief, but its effects endured. Fenfluramine remained on lists of drugs withdrawn for safety reasons decades later.¹⁴ For many physicians, regulators and insurers, it became evidence that strong initial demand for an obesity drug could obscure serious long-term risks.

What the scandal left behind

Three habits hardened after 1997, and they still shape who profits from obesity medicines.

Regulators sought more extensive safety evidence for drugs intended for large populations and long-term use. That increased the cost of entry, favouring companies with capital and established clinical-development capabilities.

Many physicians became reluctant to prescribe weight-loss medicines, while obesity medicine remained a small specialty.

Payers often treated weight loss as a lifestyle issue rather than a disease with measurable medical costs. Coverage remained limited, leaving even effective drugs dependent on patients able to pay themselves.⁶

The implication for today’s profit pool is direct. Prescription demand alone does not secure durable economics if safety concerns return, patients discontinue treatment, insurers withhold coverage or net prices fall. Evidence that a drug reduces heart attacks, strokes or liver complications matters more commercially than celebrity demand because it gives payers a basis to fund treatment.

The safety question has not disappeared. In 2023 and 2024, the FDA examined reports of suicidal thoughts among patients taking GLP-1 drugs. In January 2024, it said its preliminary review had found no evidence that the drugs caused them, while continuing its review.¹⁵ The review later ended without a finding against the class, illustrating how quickly concerns associated with earlier obesity drugs can return.

The size of the need

The underlying need has grown. The World Health Organization estimates that 890 million adults were living with obesity in 2022—about 16 per cent of the world’s adults, more than double the share in 1990. It attributes 3.7 million deaths in 2021 to higher-than-optimal body weight.¹⁶ Obesity is associated with type 2 diabetes, heart disease and fatty-liver disease, each carrying separate treatment costs.

Those figures establish potential need, not a market forecast. They do not determine how many people begin treatment, remain on it, receive coverage or use any one company’s drug. A large eligible population does not itself set price, persistence, coverage or market share.⁶

The gatekeepers

In the United States, many of those decisions are made by companies that do not manufacture drugs. CVS Health $CVS combines the Aetna insurer, the Caremark pharmacy-benefit manager and a pharmacy chain. UnitedHealth Group $UNH owns UnitedHealthcare and the Optum Rx benefit manager. The Cigna Group $CI operates Express Scripts through Evernorth. Elevance Health $ELV runs CarelonRx; Humana $HUM, Centene $CNC and Molina Healthcare $MOH insure millions of people, many through government programmes.⁶

A pharmacy-benefit manager, or PBM, sets formularies—the drugs a plan covers and the terms of coverage—negotiates manufacturer rebates for favourable placement and can require prior authorisation, requiring a doctor to justify a prescription before the plan pays. Insurers bear their members’ medical costs. Together, they determine whether a prescription becomes a reimbursed year of treatment.

Their incentives are mixed. An insurer pays for an expensive weight-loss drug immediately but may not realise any savings for years, potentially after the member has changed plans. Molina, whose membership is predominantly government-sponsored, spent 92 per cent of premium revenue on medical care in the second quarter of 2026, leaving limited room for costly new therapies.⁶

The central commercial objection to the obesity-drug boom is therefore discontinuation and affordability. Patients stop weight-loss drugs for reasons including side effects, cost and changes in insurance. Yet public global data do not reliably measure how many discontinue or how quickly.⁶

Novo’s diabetes business had already addressed several older constraints: weekly dosing, large-scale manufacturing and physician familiarity. In 2021, a higher dose of semaglutide crossed the regulatory line into chronic weight management.

↑ Back to Top

Novo turns diabetes infrastructure into Wegovy

In January 2010, the FDA authorised Novo Nordisk's Victoza, liraglutide, as a once-daily treatment for type 2 diabetes: the first human GLP-1 analogue designed for daily use.¹¹ At first, it extended Novo’s diabetes franchise, drawing on the same endocrinologists, sales force and pen-based injection habits.

Patients’ weight loss was difficult to ignore. Novo ran separate trials at a higher dose, and in December 2014 the FDA approved liraglutide for chronic weight management as Saxenda.¹⁷ Saxenda required daily injections and delivered more modest results, limiting it to a niche. The next molecule changed the commercial opportunity.

Semaglutide, twice

Semaglutide was approved for type 2 diabetes in 2017 as Ozempic, a once-weekly injection.¹² Novo then tested a higher dose in people with obesity. On 4 June 2021, the FDA approved it as Wegovy for chronic weight management, the first such approval for adults with obesity or overweight since 2014. In the trial highlighted by the agency, patients taking Wegovy lost, on average, 12.4 per cent more body weight than those taking placebo.¹⁸

The result changed expectations. Earlier drugs had generally produced single-digit percentage weight loss; Wegovy approached the range doctors associated with surgery. Its prefilled, once-weekly pen also made treatment more manageable.

The pen helped turn a peptide medicine into a plausible routine treatment. A vial and syringe require patients to measure a dose, handle a needle and store a fragile liquid correctly. A prefilled pen sets the dose and conceals the needle. But the device is part of the approved product: changing its design or supplier can require fresh regulatory evidence. Weekly dosing reduced friction; it did not resolve side effects, affordability or long-term adherence.

Why Novo was first

Novo’s advantage combined diabetes sales teams already calling on relevant doctors, regulatory experience with peptide drugs, decades of insulin manufacturing and an established supply of injection devices. Obesity-care sales rose from about DKK 6 billion in 2019 to DKK 82.3 billion in 2025, roughly a fourteen-fold increase over six years.¹⁹ In 2025, it still held 59.6 per cent of global branded obesity-treatment volume.²⁰

That scale created a lead, but not effortless economics. Group sales grew 6.4 per cent in 2025, after expanding by at least a quarter in each of the previous three years.⁶ Gross margin fell from 84.7 per cent to 81.0 per cent, partly reflecting depreciation on acquired manufacturing sites, restructuring costs and capacity investment.²⁰ Free cash flow fell from about 24 per cent of sales in 2024 to about 9 per cent.⁶ Novo’s market value declined from about $463 billion at the end of 2023 to $167 billion on 24 September 2026, and it traded at about nine times trailing twelve-month earnings.⁶ Analysts expected 2026 revenue to be slightly below 2025’s.⁶ Market leadership, in other words, did not settle the question of how durable its earnings would be.

The supplier behind the pen

The pens brought a new group of suppliers into the value chain. West Pharmaceutical Services $WST, based in Pennsylvania, makes stoppers, seals and containment components that touch injectable drugs, and manufactures self-injection devices under contract; its 2025 results cited obesity- and diabetes-led demand for those devices.²¹ Stevanato Group in Italy makes glass cartridges, pens and inspection equipment. Gerresheimer in Germany supplies glass and plastic containers and drug-delivery systems. Ypsomed, in Switzerland, designs and manufactures self-injection pens and autoinjectors used by many drug companies.⁶

These suppliers are difficult to replace because the pen is validated alongside the medicine. Engineers test patient use, regulators review the evidence and production lines are qualified together. A supplier change can require much of that work again, creating switching friction and bargaining power. Neither West nor its peers disclose how much revenue comes from GLP-1 programmes, so public filings do not show the theme’s precise contribution to their earnings.⁶

The case against the durable lead

The case for a permanent Novo advantage is incomplete. Margins have fallen as it invests in capacity, rivals are gaining ground, and initial prescriptions reveal little about persistence. The stronger test is whether patient-years of treatment and net sales hold up once supply has normalised and competition has fully arrived.

Wegovy’s opening was therefore not merely medical. It gave Lilly an opportunity to test whether activating two gut-hormone signals could outperform one.

↑ Back to Top

Lilly makes the market a contest rather than a franchise

On 13 May 2022, the FDA approved Lilly's tirzepatide, sold as Mounjaro, for type 2 diabetes.²² The molecule activates both the GLP-1 receptor and the receptor for GIP, another gut hormone that signals the pancreas before a meal. The comparison with Novo's single-receptor drug is useful only to a point. These pathways interact in ways researchers are still studying, and a dual agonist is a distinct drug, with its own benefits and side-effect profile, rather than a simple upgrade.

The difference was commercially important in clinical trials. On 8 November 2023, the FDA approved tirzepatide for chronic weight management as Zepbound. In the agency's summary, patients on the highest dose lost on average about 18 per cent more body weight than those on placebo.²³ Lilly had become a direct challenger to the market Novo had opened.

Two strategies

The companies made different bets. Novo relied on semaglutide's installed base, millions of Ozempic patients, global manufacturing and the first obesity label. Lilly paired stronger trial efficacy with rapid capacity investment, aiming to meet demand sooner. Both were negotiating net prices with payers while supply remained tight, a setting that favoured the manufacturer able to provide volume.

Lilly's sales showed the scale of that execution. Mounjaro generated $22.97 billion and Zepbound $13.54 billion in 2025, or about $36.5 billion combined.²⁴ Total revenue rose 44.7 per cent to $65.2 billion, while operating margin reached 45.6 per cent, up from 28 per cent in 2021.⁶ In the second quarter of 2026 alone, the two brands generated about $14.9 billion in sales.⁶ Lilly's market value rose from about $263 billion at the end of 2021 to about $1.1 trillion on 24 September 2026, when it traded at roughly 39 times trailing earnings.⁶

Its annual filing adds the qualification central to the investment case: higher US volume in 2025 was partly offset by lower realised prices.²⁴ Demand and pricing power are separate. A company can sell far more units while earning less per unit as discounts secure formulary access.

Lilly has also sought to control more of the delivery challenge. It licensed Camurus's FluidCrystal technology, which forms a slow-release depot under the skin, for up to four cardiometabolic compounds. In 2026, the companies expanded the agreement to include an option on an amylin-based compound.²⁵ Amylin is a hormone released alongside insulin that signals fullness. A monthly depot, rather than a weekly injection, could differentiate Lilly's medicines if weekly pens become standard.

Who leads what

“Leader” needs precision. Novo leads branded obesity-treatment volume share. Lilly leads in growth, operating margin and investor expectations. Investors value Lilly at more than fourteen times sales, compared with under four times sales for Novo.⁶ The market remains contested; the answer depends on the measure.

The challengers

A second contest is developing behind the two leaders. Amgen $AMGN, a $219 billion biotechnology company, is testing an obesity candidate but has no obesity revenue. Roche is building an obesity pipeline within a business whose $74 billion in revenue comes mainly from cancer medicines and diagnostics.⁶ Viking Therapeutics has no revenue, and its roughly $4.9 billion market value depends largely on VK2735, which is in two late-stage trials in injected and oral forms.⁶ Zealand Pharma $ZEAL.CO in Copenhagen is pursuing amylin with petrelintide; partnership income lifted revenue to DKK 9.2 billion in 2025, but its market value has since fallen by more than half from its 2024 peak.⁶

Pfizer $PFE has rebuilt an obesity pipeline through acquisition; Merck & Co. $MRK is pursuing liver disease; and Sanofi $SAN.PA retains substantial diabetes infrastructure but has limited obesity-led growth.⁶ 中外製薬 Chugai Pharmaceutical $4519.T conducts metabolic-disease research alongside its antibody business and earns an operating margin of about 47 per cent.⁶ In China, 江苏恒瑞医药 Jiangsu Hengrui Pharmaceuticals $600276.SS, 翰森制药 Hansoh Pharmaceutical $3692.HK, 联邦制药 The United Laboratories, 石药集团 CSPC Pharmaceutical $1093.HK and the insulin specialist 甘李药业 Gan & Lee Pharmaceuticals combine metabolic pipelines with established domestic sales forces.⁶

No pipeline candidate is economically equivalent to an approved medicine. The outcome depends on head-to-head clinical evidence, safety and tolerability, manufacturing capacity, payer coverage and realised net price. The volume of GLP-1 press releases does not answer those questions.

Is it winner-take-all?

A common view is that obesity medicine will resemble search, with one or two dominant suppliers. The record does not yet support that conclusion. Chinese competitors, new mechanisms such as amylin, oral medicines, long-acting depots and payers' ability to set manufacturers against one another could divide the market into segments. But the same crowded pipeline implies that many entrants will not achieve Lilly's margins. Leadership is contested; profit has so far been concentrated.

Demand had made the drug only one part of the bottleneck. Manufacturers still had to assemble amino acids, purify the active ingredient, fill sterile cartridges and produce hundreds of millions of reliable pens.

↑ Back to Top

The factories, pens and depots behind the headline drug

Inside a peptide plant, the drug begins on a bead. Large reactors hold resin, the modern descendant of Merrifield's solid support. Amino acids are added one at a time, with washing after each coupling, until a chain dozens of units long hangs from each grain. The chain is then cut free. The harder work follows: separating the correct molecule from near-misses, testing each batch and converting the purified powder into a sterile finished medicine. The process is repetitive, like an assembly line, but its arithmetic is unforgiving. Each impurity can become a regulatory problem, and a failed batch can write off kilograms of costly material.

Walking the chain

The supply chain runs from basic inputs to the patient. At the start are amino acids and specialised chemicals. 味の素 Ajinomoto $2802.T, the Japanese food and chemicals group, has a century of amino-acid expertise and supplies materials and process technology to peptide makers, though it does not disclose how much revenue comes from peptide drugs.⁶

Next come makers of the active pharmaceutical ingredient, or API: the drug substance itself. Specialists include Bachem in Switzerland, CordenPharma, a privately owned European manufacturer, and PolyPeptide Group, listed in Switzerland. Larger contract developers and manufacturers, known as CDMOs, have added peptide capability, including Piramal Pharma $PPLPHARMA, 药明康德 WuXi AppTec $2359.HK, Lonza $LONN.SW, Divi's Laboratories $DIVISLAB, Granules India through its acquired Senn Chemicals unit, 普洛药业 Apeloa Pharmaceutical, 翰宇药业 Hybio Pharmaceutical, 诺泰生物 Sinopep-Allsino Biopharmaceutical and Neuland Laboratories $NEULANDLAB.⁶ Novo and Lilly also manufacture much of their API internally, though public disclosures do not quantify it.

The purified drug is then dissolved and filled into cartridges or syringes under aseptic conditions—another regulated bottleneck. The cartridge finally goes into a pen or autoinjector. Suppliers include West, Stevanato, Gerresheimer and Ypsomed, as well as SHL Medical, a private autoinjector maker; Becton, Dickinson $BDX, whose prefillable syringes are widely used; AptarGroup; ニプロ Nipro; and テルモ Terumo $4543.T.⁶ Camurus, a Swedish company, occupies a related but distinct position: its formulation platform turns a drug into a long-acting depot.

Money flows in the opposite direction. Patients, employers and governments pay insurers and PBMs; those organisations pay pharmacies and, through negotiated net prices, brand owners. Brand owners pay API makers, fillers and device suppliers, while funding trials and sales forces and retaining what remains. Bargaining power tends to sit with the supplier that is hardest to replace.

Qualification is a moat; scarcity is a season

A supplier's production line is not interchangeable capacity. Before a drug company can use it, the process must be validated and regulators must accept evidence that each batch is made consistently. That creates genuine switching costs, particularly in sterile filling and devices, where human-factors testing adds another requirement.

Scarcity is different. During the shortage years, qualified peptide capacity could command favourable terms. That encouraged investment, and the new capacity is now arriving. Bachem's sales rose 14.8 per cent in 2025 to CHF 695.1 million, with an operating margin of 24.1 per cent. But it invested CHF 332.6 million in capacity that year—almost half of sales.²⁶ Its free cash flow has been negative in each of the past five years.⁶ In March 2025, CordenPharma announced a greenfield facility near Basel costing more than €500 million, with more than 5,000 litres of solid-phase synthesis reactor capacity.²⁷ PolyPeptide's revenue grew about 16 per cent in 2025, but it still lost money; its operating margin had fallen from 22.7 per cent in 2021 to 1.9 per cent.⁶

Those figures describe a capital cycle: rising demand, heavy investment and a period in which new capacity must find customers. They do not show that every supplier can preserve shortage-era pricing.

The market values Bachem at nine times sales, against 3.5 times for PolyPeptide.⁶ The gap implies that investors expect Bachem's process expertise and long customer qualifications, built since its founding as a peptide specialist, to endure through the construction boom. That may prove correct, but industrial capital cycles place the burden of proof on companies adding capacity.

Who executes better

Among listed specialists that disclose enough for comparison, Bachem leads in profitability and scale. WuXi AppTec's growth and margins are larger, but it does not report peptide revenue separately, so its results do not establish peptide-specific performance.⁶ Sinopep-Allsino reports a gross margin of about 63 per cent, but its free cash flow has been deeply negative for five years and its customer mix is less transparent.⁶

Among delivery suppliers, West is the largest high-quality listed company. Its operating margin nevertheless fell from 26.8 per cent in 2021 to 20.1 per cent in 2025 as its wider business normalised after the pandemic.⁶ Ypsomed moved in the other direction: its operating margin rose from about 6 per cent in the year to March 2022 to about 34 per cent in the year to March 2026. Delivery systems accounted for roughly 82 per cent of revenue, although total annual revenue fell 2.4 per cent in its latest year.⁶ Gerresheimer illustrates the limits of the broader injectable-demand thesis. It posted an operating loss and a net margin of about minus 14 per cent in its latest year, while its market value fell to about $1 billion from $3.2 billion at the end of 2023.⁶ Rising demand for injections does not protect every supplier.

Camurus shows the economics of a differentiated delivery asset. Its revenue grew 21 per cent in 2025 and its operating margin was 39 per cent; investors value it at almost fifteen times sales.⁶ Its formulation platform resembles intellectual property more closely than a component-production line. Its Lilly partnership, however, remained developmental rather than a source of booked product revenue.

What the data refuses to confirm

If suppliers were straightforward GLP-1 proxies, their revenue should rise a quarter or two after branded-drug sales. Empor tested that relationship across as many as seven years of results. API makers showed no consistent link at the expected two-quarter delay; delivery-company revenue growth tended to move against branded sales in the same quarter.⁶ Multi-year contracts, inventory reservations, customer-specific qualification, currency movements and diversified businesses obscure the connection.

A few years of patterns are evidence, not proof. The absence of a clean relationship does not make GLP-1 demand irrelevant, but it does undermine the simpler claim that every supplier offers a direct investment proxy for branded-drug sales.

The shortages did not merely enrich manufacturers. They temporarily changed who could supply patients at all.

↑ Back to Top

The shortage creates a side door, then Washington closes it

In March 2022, the FDA added Wegovy to its drug-shortage list. Ozempic followed in August.²⁸ Demand had outrun production, and a regulatory exception gave a new group of businesses a temporary role.

The bridge during the road closure

The exception involved compounding: pharmacy preparation of a medicine for an individual patient, such as a liquid form for someone unable to swallow a tablet. It is not a second approval system for copies of branded drugs. US law generally bars compounders from making what are essentially copies of commercially available approved drugs, but a shortage relaxes that restriction.²⁹ While semaglutide and tirzepatide remained on the shortage list, state-licensed pharmacies and larger outsourcing facilities could make and sell their own versions.

The arrangement was a bridge during a road closure, not a permanent competing highway. The FDA repeatedly warned about the products being sold through it. It received reports of adverse events, including dosing errors when patients measured compounded drugs themselves, and of products using forms of semaglutide that differed from the approved drug.³⁰

A channel, not a maker

Hims & Hers Health $HIMS became the best-known consumer-facing business using that route. Patients sign up online, a clinician reviews their case, and a prescription is filled and shipped. The company does not make peptides. During the shortage, it offered compounded GLP-1 treatment alongside other services, and revenue rose from $272 million in 2021 to $2.3 billion in 2025.⁶ Its role was as a care channel between patient and prescription. Novo and Lilly owned their molecules; Hims owned the customer relationship and benefited from a regulatory permission that was temporary.

The door closes

The FDA declared the tirzepatide shortage over in December 2024. On 21 February 2025, it determined that the semaglutide-injection shortage had also been resolved, saying manufacturers could meet current and projected national demand, while warning that local disruptions could persist.³¹ It allowed compounders a wind-down period: until 22 April 2025 for state-licensed pharmacies and until 22 May 2025 for outsourcing facilities.²⁹

That returned leverage to regulated brands, but it did not make treatment affordable or broadly accessible. List price, net price, insurance coverage and patients’ ability to pay remained separate issues. Hims continued to grow, reporting revenue growth of 59 per cent in 2025, but analysts expected a loss per share in 2026 and its latest quarterly earnings fell well short of consensus.⁶ The channel remained; its sustainable margin was less certain.

The final gatekeepers

With branded supply more plentiful, PBMs and insurers again became the decisive gatekeepers. CVS's Caremark, Cigna's Evernorth and UnitedHealth's Optum can select a preferred weight-loss drug for a plan, require prior authorisation and demand rebates for favourable placement.⁶ Distributors such as McKesson $MCK, Cencora $COR and Cardinal Health $CAH move products from factory to pharmacy, but retain little of the therapeutic economics; Cencora's net margin was about half of one per cent in its latest fiscal year.⁶

Two myths

Two common claims require qualification.

The first is that PBMs retain all the profit. PBMs influence net prices and collect fees and rebates, but insurers that often own them bear the medical cost of expensive drugs, while the broader access layer operates on thin margins. The companies Empor tracks generated about $1.1 trillion of revenue with a combined net margin below 2 per cent.⁶ UnitedHealth's operating margin fell from 8.7 per cent in 2023 to 4.2 per cent in 2025, while its market value had fallen by about a third since 2023. Those changes reflected pressures across US healthcare costs, not weight-loss drugs alone.⁶ Empor's tests found no consistent same-quarter relationship between branded peptide sales and payer margins.⁶ Processing a vast volume of claims is not the same as retaining the medicine’s profit.

The second is that the end of compounding guaranteed branded profits. A resolved shortage guarantees none of the conditions that make a chronic treatment durable: payer coverage, patient persistence and pricing discipline. Lilly’s own filing, which showed rising volume alongside lower realised prices, provides the counterexample.

As branded supply improved, the contest shifted from scarcity to market structure: whether long-term care would support an expanding protected market, a price-compressed one or a patchwork of local generic markets.

↑ Back to Top

The market spreads east, and the old molecule starts to become a commodity

In March 2026, semaglutide’s patent protection in India expired, and launches followed.⁶ In the United States, the previous four years had been defined by shortages, compounding and access disputes. In India, the commercial question shifted within weeks to price, local production and distribution, because multiple companies could sell the same molecule.

From franchise to commodity

A protected peptide medicine earns high margins because patents bar copies, clinical evidence supports its label, and a branded sales force and payer contracts drive volume. When comparable versions enter a market, patent protection disappears and the latter advantages weaken. Competition shifts towards manufacturing reliability, quality and commercial reach.

India’s pharmaceutical industry is built for that contest. Dr. Reddy's Laboratories $DRREDDY, Sun Pharmaceutical Industries $SUNPHARMA.BO, Cipla $CIPLA.NS, Zydus Lifesciences $ZYDUSLIFE, Lupin $LUPIN, Torrent Pharmaceuticals $TORNTPHARM.NS, Alkem Laboratories $ALKEM and Natco Pharma $NATCOPHARM are all generic semaglutide entrants; Biocon brings insulin and biosimilar expertise to follow-on GLP-1 products.⁶ Lower prices could make treatment available to patients unable to afford Novo’s brand. But eight or nine capable manufacturers selling the same molecule are unlikely to sustain exceptional margins. None discloses semaglutide revenue separately, so the prize for any individual company remains unclear.⁶

The same dynamic had reached the United States earlier with an older molecule. In December 2024, the FDA approved the first generic once-daily GLP-1 injection, a copy of Novo’s Victoza, liraglutide.³² Sandoz $SDZ.SW, the Swiss generics group, includes GLP-1 generics among its future products, a business model based on volume and price erosion rather than exclusivity.⁶

China beside it

China is both a branded-competition market and a manufacturing base. 信达生物 Innovent Biologics $1801.HK sells mazdutide, a domestic dual agonist acting on GLP-1 and glucagon receptors.⁶ 华东医药 Huadong Medicine already commercialises liraglutide in China, while Hengrui, Hansoh, CSPC, Gan & Lee and United Laboratories pair local pipelines with national sales forces.⁶ Hybio, Apeloa and Sinopep-Allsino make peptide ingredients for domestic and foreign customers. Their commercial route runs through China’s drug regulator, national medical-insurance administration and provincial procurement systems, which can provide approval and reimbursement while pressing prices lower.⁶

Regional, not global

India’s launches did not end semaglutide’s protection everywhere. Patent estates, local approvals, procurement rules and reimbursement vary by country, and Novo’s filing treats the timing of exclusivity loss as market-specific.³³ The molecule is becoming a commodity in some large markets while remaining a protected brand in others. Where competition takes hold, profit initially shifts from the owner of the evidence to the owner of the lowest-cost reliable factory; as competition deepens, more of the benefit reaches patients.

The franchises that stand apart

The wider peptide category offers a contrast. On 28 August 2026, the FDA approved Protagonist’s rusfertide, sold as Mimrylo, for polycythemia vera, a rare blood disorder in which the body makes too many red blood cells. The agency described it as the first drug of its kind: a synthetic mimic of hepcidin, the peptide hormone that regulates iron.³⁴ Protagonist licensed rusfertide to Takeda, and a second peptide, icotrokinra, an oral treatment for inflammatory disease, to Johnson & Johnson $JNJ.³⁵ Investors have valued the approvals and partnerships more highly than Protagonist’s uneven revenue: its market value rose from about $535 million at the end of 2022 to $8.9 billion on 24 September 2026.⁶ The company illustrates how partnered discovery can produce approved medicines outside obesity.

Ascendis Pharma, a Danish company, attaches drugs to a carrier that releases them slowly, turning daily injections into weekly ones. Its three endocrine products, Yorvipath, Skytrofa and Yuviwel, accounted for about 93 per cent of second-quarter 2026 revenue. Revenue rose 90 per cent in 2025, although the company still reported a net loss.⁶ Zealand’s amylin strategy, Camurus’s depots, Rhythm’s rare-obesity franchise, Ipsen’s lanreotide and Takeda’s GLP-2 therapy share a feature: each relies on a mechanism, patient group or delivery format that cannot simply be copied by buying semaglutide from an Indian plant.

These franchises are smaller and, on the evidence so far, less exposed to the incretin price war. Their protection is still limited. Ironwood’s falling revenue shows that differentiation buys time, not immunity.

Two chains of consequences

The bullish and bearish cases begin with the same development: lower prices.

In the bullish case, lower prices increase patient-years of treatment. New evidence on heart and liver outcomes broadens payer coverage, while oral pills or long-acting depots improve persistence. Volume absorbs newly built capacity, allowing total profit to rise even as revenue per patient falls.

In the bearish case, price concessions outpace volume. New plants operate below capacity, and a crowded pipeline produces more clinical assets than commercial winners. Brands retain their labels but earn less on each prescription; suppliers built for scarcity carry under-used facilities; and biotechnology companies valued on a share of a vast market find financing harder to secure.

The evidence so far does not settle the outcome. The next test is to focus on the signals that change before revenue does.

↑ Back to Top

The clues that reveal whether this is care or a craze

A prescription is written, approved by an insurer and filled. The economics depend on what follows. A patient may refill it for years, or stop after months because of nausea, cost or a change in coverage. For the manufacturer, the difference is between a chronic-care franchise and expensive, short-lived demand.

The relevant unit is therefore a reimbursed patient-year that persists. Prescription counts, viral demand and total-addressable-market estimates are noisier. Four signals provide a better early read on that unit.

Branded obesity-treatment volume

The first is global branded obesity-drug volume: doses actually dispensed. It reflects prescribing and refills before price changes and rebates reach reported revenue, helping distinguish new adoption from post-shortage catch-up. Novo publishes a market measure based on IQVIA prescription data; its latest annual report showed global branded obesity-GLP-1 volume growth of 104 per cent in 2025.²⁰

Continued volume growth alongside improved persistence would support the long-term-care case. A stall despite ample supply and lower net prices would suggest that affordability was no longer the main constraint.

How long patients stay on

The second measure—persistence by indication and payer—is the most important and least visible. No reliable global measure exists; evidence is scattered among insurers’ claims analyses and company-sponsored studies published irregularly.⁶ Patients stop treatment before the resulting decline appears in sales.

This measure tests whether clinical benefit becomes chronic revenue. Longer treatment duration among covered patients would support the franchise case. Persistence below the assumptions used to justify coverage would weaken the argument for broader access, regardless of trial results.

Volume against realised price

The third measure is the relationship between volume and realised price at brand owners. Lilly reports the split quarterly. In the quarter ended June 2026, Mounjaro and Zepbound generated nearly $14.9 billion combined; Lilly was due to report next on 29 October and Novo on 4 November.⁶

Rebates and discounts negotiated for the following year can affect price before they affect volume. If volume growth continues to exceed price erosion, broader use may still translate into profit growth. If prescriptions rise while net revenue flattens, more of the economic benefit is moving to payers and patients.

Coverage after new evidence

The fourth signal is whether payers expand coverage after new outcomes evidence or label extensions. When a trial shows fewer heart attacks or less liver damage, the commercial test is whether formularies change and prior-authorisation requirements ease. Those decisions are event-driven and disclosed through the FDA, insurers’ formulary updates and company statements.⁶

Positive trials that repeatedly fail to alter coverage would indicate that payers still treat obesity treatment primarily as a lifestyle cost.

A fifth measure would be useful but is unavailable: comparable utilisation of qualified peptide-ingredient, filling and device capacity. Companies report capacity in incompatible units, from reactor litres to device counts, and do not publish comparable utilisation data.⁶ Bachem’s half-year reports and device-makers’ commentary are imperfect substitutes. Falling utilisation and pricing after new plants open would suggest that scarcity, rather than durable demand, had supported upstream economics.

How the layers move together

One quarter of Wegovy sales does not reliably predict the next quarter at Bachem or West. Over the periods Empor examined, branded sales and supplier revenue showed no stable relationship; the observed patterns are evidence, not proof.⁶ Contracts, inventory decisions and diversified customer bases intervene between the layers. The supply chain is linked, but its quarterly results should be read through the earlier signals rather than treated as direct GLP-1 proxies.

Who keeps the profit

Protected branded medicines retain the largest share of profit because their owners control clinical evidence, labels, patents and payer contracts. Within that group, economics favour companies that combine differentiated efficacy with sufficient supply, although payers are capturing a larger share through lower realised prices.

The more durable supporting positions are validated formulation and device assets that are difficult to replace: depots, pens and sterile systems developed and approved alongside a drug. Ingredient suppliers can retain attractive returns where process quality and customer qualification outweigh the capacity now entering the market. Generic manufacturers, distributors and telehealth channels may gain volume, but generally have less control over price.

A sound view of the medicine does not guarantee a sound investment. Novo’s recent history illustrates the distinction: its drugs are more widely used than ever, while its market value has fallen by almost two-thirds. Returns depend on the layer of the value chain, the company and the price paid.

Insulin showed that a hormone can support a chronic-care franchise for generations. Fen-phen showed that an appetite drug can fail medically and socially, with consequences for payer behaviour that last decades. GLP-1 drugs have demonstrated exceptional demand. The next phase of profit will depend on affordability, persistence and the ability to sell a differentiated treatment after the shortage is forgotten.

↑ Back to Top

Glossary

• Amino acid: A chemical building block joined with others to form a peptide.
• API: Active pharmaceutical ingredient—the drug substance before it becomes a finished dose.
• Compounding: Pharmacy preparation of a medicine for an individual patient; it is not regulatory approval of a branded copy.
• Depot: A formulation that releases medicine gradually after administration.
• Fill-finish: Sterile filling of a drug into a vial, syringe, cartridge or pen.
• GLP-1: A gut-hormone pathway that affects insulin secretion, appetite and stomach emptying.
• GIP: Another gut-hormone pathway, paired with GLP-1 in tirzepatide.
• Gross-to-net: The gap between a medicine’s list price and revenue retained after rebates and discounts.
• Incretin: A hormone signal released after eating that helps the body manage glucose.
• PBM: Pharmacy-benefit manager—an organisation that negotiates formularies, rebates and pharmacy benefits.
• Peptide: A short chain of amino acids that can act as a biological signal or medicine.
• Persistence: The length of time a patient remains on treatment.
• Solid-phase peptide synthesis: Building a peptide step by step while it remains chemically attached to a solid support.
• SPPS: Abbreviation for solid-phase peptide synthesis.

References

1. The discovery of insulin: history review — PubMed ↩↩↩↩

2. The Nobel Prize in Chemistry: the development of modern chemistry — Nobel Prize Outreach ↩

3. R. Bruce Merrifield, Nobel lecture: Solid Phase Synthesis — Nobel Prize Outreach, 1984 ↩↩

4. Solid-phase peptide synthesis: a silver anniversary report — PubMed ↩

5. History of Merrifield solid-phase synthesis — PubMed ↩

6. Peptide Therapeutics research dossier and computed tables: scorecard, trends, links, results calendar and consensus — Empor, 24 September 2026 ↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩

7. History of the incretin concept — PubMed ↩↩↩↩

8. Mojsov and Merrifield, synthesis of glucagon — PubMed, 1984 ↩

9. The story of discovery: medications for diabetes and obesity emerged from research on a pancreatic hormone — NIDDK, 2021 ↩↩

10. Exendin-4: from lizard to laboratory and beyond — National Institute on Aging ↩↩

11. Victoza approved in the US — Novo Nordisk, January 2010 ↩↩

12. Ozempic (semaglutide) prescribing information — US FDA, 2017 ↩↩

13. Cardiac valvulopathy associated with exposure to fenfluramine or dexfenfluramine — CDC MMWR, 1997 ↩↩↩

14. Additions and modifications to the list of drug products withdrawn or removed from the market for reasons of safety or effectiveness — Federal Register, 19 July 2022 ↩↩

15. Update on FDA's ongoing evaluation of reports of suicidal thoughts or actions in patients taking GLP-1 receptor agonists — US FDA, January 2024 ↩

16. Obesity and overweight fact sheet — World Health Organization ↩

17. Saxenda (liraglutide) label summary — US FDA FDALabel ↩

18. FDA approves new drug treatment for chronic weight management, first since 2014 — US FDA, 4 June 2021 ↩

19. Strategic aspirations 2025 — Novo Nordisk Annual Report 2025 ↩

20. Financial performance — Novo Nordisk Annual Report 2025 ↩↩↩

21. West reports fourth-quarter and full-year 2025 results — West Pharmaceutical Services, 2026 ↩

22. Drug Trials Snapshots: Mounjaro — US FDA, 2022 ↩

23. FDA approves new medication for chronic weight management — US FDA, 8 November 2023 ↩

24. Eli Lilly and Company Form 10-K for 2025 — Eli Lilly, 2026 ↩↩

25. Camurus announces expansion of Lilly collaboration — Camurus, 2026 ↩

26. Bachem reports strong sales growth and sustained high profitability — Bachem, 2026 ↩

27. CordenPharma expands peptide platform with more-than-€500 million greenfield facility near Basel — CordenPharma, March 2025 ↩

28. Semaglutide shortage resolution decision memorandum — US FDA, 2025 ↩

29. FDA clarifies policies for compounders as national GLP-1 supply begins to stabilize — US FDA, 2025 ↩↩

30. FDA's concerns with unapproved GLP-1 drugs used for weight loss — US FDA ↩

31. FDA determines semaglutide injection shortage is resolved — US FDA, 21 February 2025 ↩

32. FDA approves first generic of once-daily GLP-1 injection to lower blood sugar in patients with type 2 diabetes — US FDA, December 2024 ↩

33. Novo Nordisk Form 20-F 2025 — Novo Nordisk, 2026 ↩

34. FDA approves first drug of its kind for polycythemia vera, a rare blood disorder — US FDA, 28 August 2026 ↩

35. Protagonist Therapeutics Form 10-Q for the quarter ended 30 June 2026 — SEC ↩