The Peptide Era: How Engineered Hormones Are Remaking Medicine

The Molecules That Shouldn’t Work

For most of the history of the modern drug industry, you did not build a company around a peptide.

Proteins are the molecules that do most of the work in your body, and they’re nothing more than long chains of smaller building blocks called amino acids. A peptide is just a short version of that chain: a handful of amino acids strung together, rather than the hundreds folded up into a full protein. They’re everywhere in human biology, ferrying signals from one organ to the next. As medicines, though, they were written off as a dead end. The reasons were plain and practical.

The body treats a peptide the way it treats food. Swallow one and your gut digests it. The same machinery that turns a steak into absorbable scraps shreds the drug before it ever reaches your blood. Inject it instead, and the bloodstream is barely friendlier: peptides get chopped apart or filtered out within minutes, far too fast to act like a daily medicine. And since they fall apart in the stomach, you generally couldn’t put them in a pill at all. A drug that a patient can neither swallow nor keep in circulation is, commercially, a very hard sell.

There was a strategic squeeze too. The industry had settled into two comfortable extremes. At one end sat small molecules, tiny chemical compounds like aspirin, cheap to manufacture and rugged enough to survive a trip through your gut. At the other, starting in the 1980s, came antibodies, large proteins borrowed from the immune system and turned into very precise, and very expensive, injected drugs. Peptides were stuck awkwardly in the middle. Bigger and frailer than a pill. Smaller and blunter than an antibody. “Real” drugs, the thinking went, were one or the other. That consensus has collapsed, and the speed of the collapse is the real story.

The drugs that broke it are a family of injections now sold under names that have become household words. Ozempic and Wegovy, two formulations of a molecule called semaglutide, from the Danish company Novo Nordisk. Mounjaro and Zepbound, two versions of a molecule called tirzepatide, from Eli Lilly. In barely five years they’ve become the fastest-growing class of medicines in the history of the pharmaceutical industry.

The numbers are absurd. Ozempic alone did around $17 billion in 2024. Mounjaro, about $11.5 billion. Wegovy roughly $8 billion, Zepbound nearly $5 billion, and both still climbing steeply. The wealth created here is hard to overstate. Novo Nordisk’s stock-market value briefly grew larger than the entire annual economic output of Denmark — as if a single company outweighed the country it lives in. And in November 2025, Eli Lilly became the first health-care company in history worth a trillion dollars, a milestone that until then belonged almost exclusively to Big Tech.

Here’s the part most patients never hear. The molecule that cracked the whole field open wasn’t dreamed up in a pharmaceutical lab. It was reverse-engineered from the venom of a lizard.

In the early 1990s, an endocrinologist named John Eng, working out of a Veterans Affairs hospital in the Bronx, was chasing an odd clue. Venom from the Gila monster — a heavy, slow desert lizard — seemed to enlarge the pancreas, the organ that makes insulin. The Gila monster eats only a few times a year, gorging and then fasting for months, all while holding its blood sugar remarkably steady. In 1992, Eng isolated the component of the venom responsible and named it exendin-4. Then he lined it up against the known human hormones, and found something startling. Exendin-4 was a near-twin of a gut hormone we make ourselves, called GLP-1. It shared more than half its structure. The difference was that the lizard’s version lasted in the bloodstream for hours instead of minutes. A synthetic copy of it became the very first drug of this class, approved in 2005 for type 2 diabetes.

To see why a lizard hormone mattered, you have to understand what GLP-1 actually does inside your own body. It’s made in the gut and released after you eat. Think of a hormone as a key and a receptor as a lock. GLP-1 fits a particular receptor on the cells of your pancreas, and when it docks, it tells those cells to release insulin — but only when blood sugar is genuinely high. That makes it a cautious, self-limiting regulator. It also acts on the brain, where it dials down appetite and signals fullness. In other words, your body already owns an elegant machine for steadying blood sugar and curbing hunger.

The catch is that it owns it for about two minutes. An enzyme in the blood — a pair of molecular scissors called DPP-4 — snips natural GLP-1 apart almost as fast as the gut can put it out.

So the entire trick of these drugs is to defeat those scissors.

Chemists rebuilt the hormone so the body can’t destroy it. They altered the exact spot where DPP-4 grips, so the blades slip. And on the longer-acting versions, they bolted on a fatty tail that anchors the drug to proteins already drifting around in the blood, keeping it in circulation for days. The result turns a hormone that lasts minutes into a medicine that lasts a week — one injection. Even the oldest objection, that peptides can’t be swallowed, is finally giving way, with the first pill versions reaching patients only in the last few years. Lilly’s tirzepatide goes a step further. It imitates not one gut hormone but two. The second, a partner signal called GIP, appears to deepen the effect on both blood sugar and weight.

Strip away the branding and the dollar figures, and that’s all these blockbusters are: a hormone your own gut already makes, re-engineered so the body can no longer tear it down.

The exciting, and slightly unsettling, part is that the same playbook is now being aimed well beyond diabetes and weight.

The most consequential surprise has been the heart. In a trial of more than 17,000 people who were overweight and had heart disease but not diabetes, semaglutide cut the rate of heart attacks, strokes, and cardiovascular death by about 20 percent. And the benefit started showing up before participants had lost much weight — a hint that the drug protects the heart directly, not just by slimming people down. Regulators have since approved it for exactly that. Researchers are also finding that these molecules quiet cravings. In early trials, people on semaglutide drank less alcohol and smoked less, apparently because the drug turns down the brain’s reward signal — the very circuitry that drives addiction.

What’s still unsettled is whether this is a durable new pillar of medicine, or a bubble inflated by demand the system can’t sustain. List prices run past $1,000 a month, and even though the manufacturers have started cutting cash prices sharply, the bill has strained insurers and national health budgets alike. Surging demand emptied pharmacy shelves for years — the shortages of both semaglutide and tirzepatide weren’t declared over until late 2024 and early 2025.

And there’s a deeper catch. These drugs manage a condition; they don’t cure it. Stop taking them, and appetite rebounds and most of the lost weight comes back. For a medicine pointed at tens of millions of people, that raises the question the field is only now starting to confront — whether the answer to obesity is a drug most patients have to take for the rest of their lives.

The Forty-Year Overnight Success

For most people, the drugs showed up all at once. Sometime around 2021 a weight-loss injection started appearing in the tabloids and on earnings calls in the same week. But the molecule inside the pen was forty years in the making, and almost none of that work, at the time, looked like the foundation of a fortune.

The trail runs from a gene fished out of a fish, through the venom of a desert lizard, to three decades of grinding chemistry. And past a row of scientists, more than one of whom watched other people collect the credit.

This is the runway, not the destination. But it’s worth walking, because nearly everything surprising about these drugs was already hidden in the way they were found.

Doctors had long noticed something strange about digestion. Swallow a spoonful of sugar and the pancreas dumps far more insulin than if you drip the identical dose straight into a vein. The food itself somehow sends word ahead — the gut warns the pancreas that sugar is coming, so the organ is primed before any of it arrives.

Biologists had a name for whatever carried that warning — an incretin — long before they could say what it actually was. One such messenger turned up in the 1970s. But it was too weak to explain the whole effect, so the hunt was on for a second, stronger one.

The break came in the early 1980s, in a lab at Massachusetts General Hospital run by an endocrinologist named Joel Habener. Habener wasn’t chasing a diet drug. He was studying glucagon — the hormone that raises blood sugar, insulin’s mirror image. To find the gene behind it, his team went fishing, almost literally, and pulled the DNA out of the pancreas of the anglerfish.

What they and others found was a surprise. The gene codes for a long precursor that the body later chops into pieces, and tucked inside it, right next to glucagon, sat two more segments nobody had described. They got christened glucagon-like peptides — GLP-1 and GLP-2. At first no one knew what either of them did.

Pinning down GLP-1’s job ran straight into a wall. Build the peptide exactly as the gene spells it out, test it, and it barely works.

The answer came from a chemist in the same building. Svetlana Mojsov had trained under Bruce Merrifield, who won a Nobel Prize for figuring out how to build peptides to order, one amino acid at a time. That skill let her ask a question others couldn’t easily test: what if the body trims GLP-1 after manufacturing it? What if the gene’s version is a rough draft, and the real hormone is a shorter piece cut out of it?

So she synthesized that trimmed form — just thirty-one amino acids long — made antibodies to hunt it down inside tissue, and showed that this fragment, not the full-length molecule, was the potent signal telling the pancreas to release insulin after a meal. Her paper, published in 1987, nailed down what GLP-1 actually is and which exact version of it matters.

Without that, there’s nothing to turn into a drug.

Yet for decades Mojsov’s name was largely missing from the story. The early patents on GLP-1 went to Habener and the hospital. She wasn’t on them — and patents are where the royalties live.

She spent more than ten years grinding through a legal fight to be named as an inventor on the GLP-1 patents. She eventually won.

Only recently has the record started to get set right. In late 2023 she was named — alongside Habener, the Toronto endocrinologist Daniel Drucker (who as a young researcher in Habener’s lab helped work out what the glucagon-like peptides do), and the Danish physiologist Jens Juul Holst — for a major innovation prize. In 2024 she shared the Lasker Award, often called America’s Nobel. In 2025, a Breakthrough Prize.

For a discovery underpinning a market worth hundreds of billions of dollars, the acknowledgment showed up almost forty years late.

Knowing what GLP-1 did was still a long way from having a medicine. The hormone had a flaw that looked fatal for any drug. The body destroys it in about two minutes, chopped apart by blood-borne molecular scissors almost as fast as the gut puts it out.

Researchers could prove the hormone worked. Drip it continuously into a vein in a hospital study and patients’ blood sugar comes down beautifully. But nobody is going to live tethered to an IV pole.

So through the 1990s GLP-1 sat in exactly this maddening spot: plainly powerful, practically useless. What the field needed was a version of the signal the body couldn’t immediately tear apart.

It arrived from an improbable direction: a lizard.

The Gila monster is a heavy desert reptile, and its venom carries a near-twin of human GLP-1 that lingers in the blood for hours instead of minutes. Turning that oddity into a product was its own long haul. After the finding was presented at a scientific meeting in 1996, a small San Diego biotech called Amylin Pharmaceuticals licensed the rights and set out to make a drug from the lizard peptide. In 2002 the far larger Eli Lilly came aboard as a partner and committed more than $300 million. The synthetic version, exenatide, sold as Byetta, won approval in April 2005 — the first medicine of its kind ever to reach patients, marketed as a “mimic” of the body’s own incretin.

It worked. But it was a blunt first attempt: a shot twice a day, every day, that often left people queasy.

From there the field climbed a ladder, and every rung was the same idea: take the hormone and re-engineer it to last longer in the body, so the shots could get rarer and the effect steadier. The decisive trick came from Lotte Bjerre Knudsen, a scientist at the Danish company Novo Nordisk. She attached a fatty tail to the peptide so it would cling to a plentiful protein already floating in the blood and ride along, shielded from the enzymes that would otherwise shred it. Her molecule, liraglutide, sold as Victoza from 2010, cut the dosing from twice a day to once. A deliberately refined successor, semaglutide — Ozempic — stretched a single injection to a full week when it landed in 2017. Tirzepatide, the two-hormone molecule, followed in 2022.

Each version lasted longer than the last. And, as it turned out, each one worked better.

And here the story turns from science to nerve. Every one of these drugs was approved, at first, only for type 2 diabetes. The weight loss was a side effect — a striking one that everybody noticed and almost nobody dared to chase. Because the market for obesity drugs was a graveyard.

For most of a century, diet pills carried a body count. It started with the amphetamines handed out for slimming in the mid-1900s, which turned out to be addictive. The worst reckoning came later. In the mid-1990s the most popular weight-loss treatment in America was a two-drug combination nicknamed fen-phen, and by 1996 doctors were writing more than eighteen million prescriptions a year for its ingredients. Then the reports landed: the drug was scarring patients’ heart valves, with damage showing up in roughly one user in three. It got pulled from the U.S. market in September 1997, trailing lawsuits that eventually ran past thirteen billion dollars.

The drugs that followed did no better. One was withdrawn in 2008 after it was found to double the rate of psychiatric problems, including suicidal thoughts. Another was pulled in 2010 for raising the risk of heart attacks and strokes.

So by around 2010, an obesity drug was exactly the kind of thing a serious pharmaceutical company stayed far away from: modest benefits, frightening risks, ruinous liability, and a condition many doctors still didn’t even regard as a disease worth treating for life.

That two firms — Novo Nordisk and Eli Lilly — walked straight back into that market is the gamble at the center of the whole bet. Both were diabetes companies. Both could see, right there in their own trial data, exactly what GLP-1 was doing to their patients’ appetites and waistlines.

Novo moved first, repackaging liraglutide at a higher dose as a dedicated obesity drug, Saxenda, in 2014. The weight loss was real, but slight — only a handful of percent. The vindication came in June 2021, when its once-weekly semaglutide was approved for obesity as Wegovy, on the strength of trials showing average weight loss around 15 percent — a number no remotely safe drug had ever come close to. Then Lilly went further. Its tirzepatide, cleared for weight loss as Zepbound in November 2023, drove losses above 20 percent — about forty-eight pounds at the top dose.

The market everyone else had fled became the richest prize in the history of the drug business.

Four decades of patient, unglamorous work — a gene pulled out of a fish, a fragment fought over in court, a lizard, a fatty tail — had finally compounded into an overnight success.

The Hormone Behind the Hype

The whole story turns on a strange fact about eating. Swallow a spoonful of sugar and your body fires off far more insulin than if you dripped the identical dose straight into a vein.

That gap isn’t a footnote. It has a name — the incretin effect — and, more importantly, a size.

Match the two routes exactly, holding blood sugar identical whether the glucose is eaten or infused, and you can measure how much extra insulin the eating itself buys. The answer is startling. In a healthy person, somewhere between 50 and 70 percent of the insulin released after a meal is triggered not by the sugar in the blood, but by chemical messengers the gut sends ahead of it.

That figure inverts the obvious picture. The pancreas doesn’t mainly take its orders from the sugar arriving in the bloodstream. It mostly takes them from a heads-up telegraphed by the gut.

The sugar is the cargo. The incretins are the phone call warning the loading dock that a delivery is on its way — so the crew is already in position before the truck backs in.

That phone call is made by hormones, and the one that matters here is GLP-1. It’s manufactured by specialized cells dotted along the inner lining of the intestine — L-cells — which sit in the gut wall like sensors, sampling whatever passes by. Within about fifteen minutes of the first bite, they start spilling GLP-1 into the blood.

The timing is the surprising part. That first surge goes out so fast that it arrives before much of the meal has traveled far enough down the gut to reach the bulk of the cells doing the releasing, which cluster lower in the intestine. The gut, in other words, anticipates. It starts priming the pancreas almost the moment food lands, instead of waiting to tally up the sugar afterward.

This is what lets a healthy pancreas mount a prompt early burst of insulin the instant blood sugar starts to rise. And that fast, well-timed burst is one of the first things to fail as diabetes sets in.

What GLP-1 does once it reaches the pancreas is the elegant part, because it does several things at once, and every one of them pushes in the same direction.

The first job is the one already familiar — it coaxes insulin out of the pancreas. But the crucial detail is the condition attached. GLP-1 only spurs strong insulin release when blood sugar is genuinely high. When sugar is normal or low, it goes quiet.

So it behaves like a thermostat, not a stuck-open valve. Older diabetes treatments simply whip the pancreas into making more insulin, or inject insulin directly regardless of need. This natural signal almost never drives blood sugar dangerously low, because it stops pushing the moment the job is done. That single property — a built-in floor — is the most important safety feature the eventual drugs would inherit.

The second job works the opposite end of the same problem. The body carries a hormone that does the reverse of insulin — glucagon, insulin’s mirror image — whose task is to order the liver to release its stored sugar into the blood. Between meals, when blood sugar is sagging, that’s exactly right. Just after you eat, it’s precisely the wrong instruction: the last thing a sugar-laden bloodstream needs is the liver pouring in more. GLP-1 quiets glucagon, and with that one move tells the liver to stop dumping.

So a single hormone clamps two brakes on the post-meal sugar surge at once — turning insulin up to clear the sugar, and turning the liver’s contribution down.

The remaining jobs slow the flood rather than clear it. GLP-1 puts a brake on the stomach itself, slowing how fast it empties into the intestine, so sugar trickles into the blood instead of arriving all at once — a spike flattened into a gentle slope. And it carries the message onward to the brain. Some of GLP-1’s signal travels in the blood, some is relayed along the vagus nerve — the main nerve trunk wiring the gut directly to the brain — and either way it reaches the brain regions that govern hunger. The result is the feeling of fullness. The gut tells the brain the meal has landed, and appetite winds down.

Count them up. Insulin up, liver hushed, stomach slowed, hunger quieted. Four levers, all worked by one short peptide, all aimed at the same end: absorb an incoming meal smoothly, and know when to stop eating.

That coordinated gut-to-brain conversation is the natural system everything that follows is built on.

This is where the link to diabetes gets sharp. And it’s not the obvious tale of a pancreas simply running out of insulin.

In people with type 2 diabetes, the incretin effect is broken. The 50-to-70-percent boost the gut normally supplies shrinks to roughly a third or less. The pancreas has stopped getting — or stopped heeding — the gut’s heads-up.

Careful experiments pinned down a revealing split between the two incretins. The older one, GIP — the weaker messenger found back in the 1970s — largely stops working in people with diabetes. Their cells have effectively gone deaf to it. But GLP-1 keeps working. In a landmark 1993 study, the German researcher Michael Nauck and colleagues infused extra GLP-1 into patients with type 2 diabetes and pulled their insulin response back toward normal, while the same trick with GIP did almost nothing.

Whether the weakened incretin effect causes the disease or results from it is still debated. But the practical upshot is the same, and it’s profound. Supplying more GLP-1 doesn’t merely flog the pancreas into working harder. It restores a specific signal that has gone quiet — repairing a genuine defect instead of brute-forcing the system.

That’s the difference between fixing a broken thermostat and just lighting more fires in the room.

All of these effects trace back to a single piece of hardware: the GLP-1 receptor — the lock the hormone fits into.

The reason one gut hormone can reach so many organs at once is that the same lock is installed far more widely than anyone first suspected. It studs the insulin-making cells of the pancreas, as expected. But copies of it also sit in the hunger and reward circuits of the brain, in the lining of blood vessels, in the kidney, and throughout the stomach and gut. And, strikingly, in the heart — including the small cluster of cells that serves as the heart’s natural pacemaker.

A hormone can act only where its lock exists. GLP-1’s lock turns out to be scattered across nearly every organ that bears on metabolism, and across several that seem, at first glance, to have nothing to do with sugar at all.

That one fact — a single key fitting many locks, planted in many places — is what makes the molecule so powerful and so unexpectedly far-reaching.

It’s also the opening the drug designers would walk straight through.

Two Minutes to Seven Days

This whole story comes down to one ratio: roughly two minutes against seven days.

Natural GLP-1 vanishes from the blood in about a minute and a half. The drugs built from it last a week. Closing that gap — making a fragile signal endure thousands of times longer — is the central feat of the entire field. And it wasn’t pulled off by inventing some sturdier molecule from scratch. It was pulled off by outwitting, one at a time, the two separate systems the body uses to clear a molecule like this away.

So to understand the drugs, you have to understand both disposal systems. Each one had to be beaten by a different trick.

The first is the pair of molecular scissors, an enzyme called DPP-4, that patrols the bloodstream. The piece worth understanding now is exactly where those scissors cut.

GLP-1 is a chain of building blocks strung end to end. DPP-4 grabs one end and snips off just the first two — a tiny clip that leaves most of the molecule intact but lops off the precise tip the hormone needs to dock into its receptor. The remnant is useless. It can no longer ring the doorbell. And all this happens within a minute or two of the hormone hitting the blood, which is why the body’s own GLP-1 is gone almost as fast as the gut can make it.

The second defense is the kidneys, and in some ways it’s the harder one to beat.

The kidney’s job is to scrub the blood, and it does so with what amounts to an extraordinarily fine sieve. Blood gets forced through dense tangles of tiny vessels whose walls are perforated like a strainer. Small molecules slip through the holes and drain into the urine. Large ones — above all the body’s own proteins — are too bulky to fit and stay behind in circulation. And the apparatus is relentless: a healthy pair of kidneys pushes something like 180 liters of fluid through that sieve every single day.

The cutoff between what passes and what stays sits at a particular size. GLP-1, a short chain weighing well under it, slips straight through. So even a version of the hormone the scissors couldn’t touch would still get rinsed out within hours.

In other words, two clocks run against the drug at once — one chemical, one mechanical. Beat only one and the molecule still fails.

Defeating the scissors turned out to be the smaller problem, and the fix is almost startlingly minimal.

The scissors work by recognizing one specific building block near the end of the chain — the second one in — and clamping onto it to make the cut. So chemists reasoned: replace that single block with a slightly altered version the enzyme can’t grip, and the blades simply slide off.

So they swapped it. In place of the natural building block at that spot they installed a man-made one called Aib — short for alpha-aminoisobutyric acid — that carries a small extra cluster of atoms exactly where the enzyme needs a clean handhold. To the scissors it’s greasy and ungrippable. They slip right off without cutting.

That one substitution — a single block out of about thirty — is most of what separates a hormone that survives two minutes from one that survives several hours.

The Gila monster’s exendin-4 pulled off the exact same feat by accident. It happens to carry a different block at that precise position, which is why its venom version outlasted the human one all along.

But hours are not a week, and the kidney’s sieve was still waiting.

The obvious idea — just build a bigger molecule, too large to fit through the holes — doesn’t work, because enlarging the hormone destroys the exact shape it needs to fit its receptor. So the elegant alternative, the trick that defines this entire class of drugs, was the opposite: leave the hormone small, but give it a way to grab onto something that’s already big.

That something is albumin, the single most abundant protein in human blood. The body churns out about ten grams of it a day, and it makes up roughly half of all the protein floating in your plasma. Its natural job is to be a courier, ferrying fatty molecules and other cargo around the bloodstream. Which means it’s built to bind grease.

So chemists bolted a short fatty-acid tail onto the drug — the “protraction” tail pioneered by Lotte Bjerre Knudsen — that works like a grappling hook. The tail catches a passing albumin molecule and clamps the drug to its host. The drug then rides along as a hidden passenger.

And the consequences are exactly what the designers were after. Albumin is a big protein — more than fifteen times the weight of the little drug clinging to it — and far too bulky to squeeze through the kidney’s sieve. So as long as the drug keeps its grip, it can’t be filtered out. And because that grip is deliberately loose, with the tail constantly letting go and re-attaching, the bloodstream ends up holding a large, slowly released reservoir of the drug — doled out a sip at a time instead of dumped and cleared all at once.

Notice that the length and stickiness of the fatty hook are what tune how long the whole thing lasts. Liraglutide used a shorter tail and a looser grip. It bought about thirteen hours — enough for one shot a day. Semaglutide uses a longer, stickier tail, paired with the ungrippable Aib block at the other end, and stretches a single dose to roughly a week. The same logic produced tirzepatide, the two-hormone molecule, whose own long tail likewise yields a once-weekly shot.

A hormone that lasted ninety seconds now lasts seven days. And it took just two edits to a thirty-link chain to get there: one to blind the scissors, one to dodge the sieve.

That once-a-week rhythm usually gets sold as a convenience. But its real importance runs much deeper, and it cuts two ways.

The first is steadiness. A drug taken twice a day spikes after each dose and sags before the next. And with these molecules, the spikes are exactly when the worst side effects — the nausea, the queasiness — tend to hit, while the sags are when the effect fades. A drug with a week-long life behaves completely differently. Its level in the blood rises and falls only gently, hovering near a plateau, which blunts both the side-effect peaks and the efficacy valleys.

The second benefit is more human. People are simply far better at sticking with a medicine they inject once a week than one they have to remember every single day. In real-world tracking, about two-thirds of patients were still on once-weekly semaglutide a full year after starting, versus only around forty percent for the once-daily liraglutide. By the same measure, the typical patient stayed on the weekly drug for nine months, and on a daily one for just four.

A medicine only works if people actually take it. And the molecule engineered to linger in the blood turns out to be — not coincidentally — the one patients actually stay on.

The hardest frontier of all is the one the body defends best: getting a peptide to work as a pill.

Everything that makes the injection necessary traces back to a single fact from the very beginning — the gut is, fundamentally, a machine for taking proteins apart. Swallow a peptide and the body can’t tell it from food. Stomach acid starts to unravel it, and a whole battery of digestive enzymes — the gut’s own scissors, far more numerous and varied than the lone one in the blood — chop it into scraps long before it could ever reach the bloodstream. And even a peptide that somehow survived the chemical assault hits a second wall: the lining of the gut is a tight barrier, evolved precisely to let small nutrients cross while keeping large molecules like intact peptides out.

By every normal rule, a swallowed peptide drug is just a strangely expensive snack.

So how do you sneak one through? You give the peptide a tiny, temporary bubble of protection at the exact moment it lands. The daily pill version of semaglutide — sold as Rybelsus, approved in 2019 — is pressed together with an absorption-boosting compound called SNAC. In the small patch of stomach lining where the tablet dissolves, SNAC does two jobs at once. It softens the acidity right there, so the acid and enzymes can’t shred the drug. And it loosens the stomach wall just enough to let some of the peptide slip across into the blood.

“Some” is the operative word. The method is barely good enough to function — only around one percent of each swallowed dose actually reaches circulation. To make up for everything lost, the pill has to be packed with more than ten times the drug of the injection. And it has to be taken on an empty stomach, with no more than a few sips of water, followed by a thirty-minute wait before eating — otherwise even that one percent collapses.

The trick was valuable enough that Novo Nordisk paid $1.8 billion to buy the small company that owned it. For now, that narrow doorway is the only way an oral peptide gets through.

One last property shadows every drug in this class: they are genuinely hard and expensive to manufacture. And the reason is rooted in how peptides are built.

A simple drug like aspirin is a single small molecule you can mix up by the ton. By contrast, a peptide has to be assembled link by link, in the slow one-amino-acid-at-a-time method Bruce Merrifield invented. Each new link demands its own full cycle of chemical steps — attach the block, wash, unmask the next attachment point, wash again. And because no step is ever perfectly efficient, a little material is lost at every turn. Chain together thirty-odd links and those small losses compound into a serious drag on the final yield.

Worse, the man-made Aib block can’t be produced by living cells at all, and the fatty grappling hook has to be attached separately. So neither one can simply be grown. The big manufacturers therefore split the job in two. They grow the plain middle stretch of the chain cheaply, by brewing it in giant vats of genetically engineered yeast — the same fermentation logic as making beer — and then build the difficult ends by hand-chemistry, before stitching all the pieces together and purifying the result.

Even with that division of labor, one detailed engineering estimate puts the cost of a single kilogram of finished semaglutide at around a hundred thousand dollars. And the sheer intricacy of the process is a big part of why supply ran short for years, and why prices have stayed stubbornly high.

Turning Down the Food Noise

Ask people what changed first on these drugs, and a striking number reach for the same words: the noise stopped.

They even have a name for it now — food noise. The low, constant chatter about eating that runs in the background of your waking hours. What’s for lunch. There are cookies in the kitchen. You shouldn’t, but maybe just one. You’ll start again Monday. For many people with obesity this loop is relentless, a running negotiation that never fully quiets, not even right after a meal. Then they take the first injection, and within days a large share report that it simply switches off. The kitchen stops calling. A plate can sit half-eaten without protest.

Notice that the word didn’t come from scientists. It bubbled up from patients trading notes online, and researchers only noticed because so many people, independently, reached for the exact same words to describe the exact same silence.

That detail carries a heavy implication: the chatter has a volume knob, and a weekly shot can turn it down.

For most of a century, obesity was treated as a problem of character. Eat less, move more, and if you can’t, that’s a failing of will. But if a molecule that acts on brain signaling can quiet the urge to eat, then the urge was a signal all along — something the brain generates and broadcasts, not a readout of a person’s discipline. Hunger, on this view, isn’t one simple feeling. It’s the output of a control system, and in obesity that system is miscalibrated, pushing harder and longer than the body needs.

The drugs don’t hand patients willpower. They lower the signal that willpower had been straining against.

This reframing is among the most consequential things to come out of the whole episode. And it began with a word patients coined themselves.

To turn down that signal, the drug has to reach the place the signal is made. And the brain is the hardest organ in the body to reach.

It sits behind a security perimeter called the blood-brain barrier. The vessels feeding brain tissue are sealed unusually tight, on purpose, to keep most of what circulates in the blood — large molecules above all — from leaking in among the delicate wiring. A drug clamped to a big blood protein is exactly the bulky passenger that barrier is built to turn away.

But the body left a few doors unlocked. A handful of small control stations sit on the brain’s side of the wall yet draw their blood supply from outside it, through vessels riddled with tiny windows, precisely so they can taste the bloodstream and report what’s in it. One of these, a node in the brainstem called the area postrema, is studded with GLP-1’s lock. It’s a chemical-sampling port wired straight into the circuits that bring a meal to an end — and it sits exactly where a blood-borne drug can get at it.

Here’s where the week-long lifespan does its real work.

The body’s own GLP-1 is gone in ninety seconds. It can only ever brush these sensors with a brief flicker after a meal. The drug, by contrast, held at a steady level for days, bathes them without pause — leaning on a button the natural hormone could merely tap. And deeper in, in the hunger-setting region called the hypothalamus, the same key reaches two opposing crews of neurons: one that drives the urge to eat, and one that signals enough. The drug works both at once. It eases off the accelerator while pressing the brake.

The felt result is the quiet patients describe. Not the white-knuckled restraint of a diet, but a real absence of the prompt to eat in the first place.

A second, cruder mechanism is stacked on top of the first, and it works down in the gut. By slowing how fast the stomach empties, the drug keeps food sitting longer, so a meal goes on feeling filling well after it’s finished. For the first few weeks, this physical fullness does much of the heavy lifting.

But the body adapts to the stalled stomach, and that gastric effect fades over the following months — which would seem to predict the whole drug petering out.

It doesn’t.

The appetite-lowering signal in the brain persists, and it’s this central effect, not the slowed stomach, that carries weight loss out over years. The two actions aren’t equal partners. The gut effect is a fast-starting primer. The brain effect is the durable engine.

That engine, though, runs only while the drug is present — which leads to the most sobering fact about the whole class. The numbers are worth seeing plainly.

In the original obesity trial, participants lost about 17% of their body weight over 68 weeks. Then they stopped the injections. The food noise returned, appetite rebounded, and within a year they had regained roughly two-thirds of what they’d lost, settling around 6% below where they began. The improvements in blood sugar and blood pressure drifted back toward the starting line right alongside the weight.

This isn’t the drug failing. It’s what the drug is. It manages a biological signal rather than resetting it — the same way a blood-pressure pill lowers your pressure only as long as you keep swallowing it, and the pressure climbs the moment you stop.

The lasting shift is conceptual. Obesity is being recast, like high blood pressure, as a chronic condition you hold in check rather than a state you fix and walk away from. For a medicine aimed at hundreds of millions, that implies treatment measured in decades, not months — the unresolved question hanging over the entire field.

The most intriguing turn is that the brain circuitry these drugs reach into governs far more than food.

The same reward machinery that makes a meal feel worth wanting — the brain’s dopamine-driven wanting system — also underlies the pull of alcohol, nicotine, and harder drugs, and GLP-1’s lock turns out to be planted right there too. Patients and their doctors stumbled onto it by accident: people taking the drugs for weight or diabetes found they’d also lost interest in their evening drink, or their cigarettes. Those anecdotes are now being put to the test. In a randomized trial published in early 2025, a team led by Christian Hendershot gave 48 adults with alcohol use disorder either low-dose semaglutide or a placebo for nine weeks. The treated group drank less on the days they drank, had fewer heavy-drinking days, and reported weaker cravings, with the heaviest drinkers benefiting most — and the smokers among them cut back on cigarettes as well.

The effects were modest and the trial was small. But the direction was unmistakable, and far larger studies are now running.

If they hold, it would mean these drugs stumbled onto something larger than appetite: a way to turn down the brain’s wanting itself. And with it, one of the most active new frontiers in the science of addiction.

Pulling More Than One Lever

A drug like semaglutide is one key turning one lock. GLP-1 fits a single receptor, and from that one connection yanks several levers at once: insulin, the liver, the stomach, appetite. A remarkably versatile signal. But it’s still one signal, and one signal has a ceiling.

The newest drugs make a different bet. Not that you perfect the single key, but that you carry several at once — one molecule that fits two locks, or three, each belonging to a different hormone, each wired into a different corner of the body’s accounting for food and fuel. Tirzepatide, the two-hormone molecule named at the very start, is the proof of concept. And the logic behind it is what’s now dragging the whole field toward drugs that speak three hormonal languages at the same time.

Tirzepatide is one chain that fits two locks. The first is GLP-1’s. The second belongs to GIP — the weaker incretin, the one diabetics had largely gone deaf to, the one the field had almost written off as the junior partner. Deliberately drafting a drug to switch that neglected lock back on was a gamble. It paid off.

The cleanest proof came when the two best drugs were finally put head to head — tirzepatide against semaglutide, the best one-lock drug, same patients, same conditions, 72 weeks. The two-lock group lost about 20% of their body weight. The one-lock group lost under 14%. Adding the second signal didn’t nudge the number. It produced roughly half again as much weight loss.

Two levers, pulled together, beat one. That’s the engine of everything that follows.

Why not just give more of the one-lock drug? Because it slams into a familiar wall. The same signal that kills appetite also, pushed hard, produces the nausea these drugs are infamous for — and the queasiness climbs right along with the dose. There’s only so far you can crank a single lever before patients can’t stomach it, literally.

Pulling two different levers gets around this. You can set each one at a moderate level — below the point where it makes someone sick — and their effects still add up to something bigger than either could reach alone. Better yet, where the two locks sit side by side on the same cells, as they do in parts of the brain and the pancreas, the combined nudge can exceed the sum of its parts.

There’s even an early hint that the second signal cleans up after the first. In animals, switching on the GIP lock blunts the very nausea the GLP-1 lock provokes — without touching the weight loss. The clinical picture is murkier. People on tirzepatide do report somewhat less nausea than those on the one-lock drug, but the effect is modest and still argued over. Either way the logic holds: several signals at gentle settings beat one signal floored.

A lock that helps whether you open it or jam it shut

Here’s where it gets genuinely weird — a puzzle the field still hasn’t cracked. Tirzepatide works by switching the GIP lock on. But a rival drug from the biotech company Amgen does the exact opposite: it jams the GIP lock shut. Blocks it instead of turning it.

By every intuition, one of these should flop. Instead, in early trials, Amgen’s drug — MariTide — lands in the same 20%-weight-loss neighborhood as tirzepatide. Switching GIP on and shutting GIP off help about equally. That should be impossible if the signal behaved like a simple lever.

No one knows why for sure. The leading guess is that flooding the body with a constant GIP signal eventually wears the lock out — the cells, overstimulated around the clock, just stop answering — so that relentless switching-on and outright blocking end up at nearly the same place by opposite roads. Other researchers suspect GIP does different jobs in different tissues, and the two drugs exploit different ones.

The debate is unsettled, which is why the “block it” approach gets its own section later. For now, take it as a flag planted over how much about these systems is still genuinely unknown — even as the drugs built on them rake in billions.

If two locks beat one, the obvious next question is whether three beat two. The drug in trials now adds a third.

The third lock belongs to glucagon — insulin’s mirror image, the hormone that orders the liver to dump its stored sugar and pushes blood sugar up rather than down. Bolting a sugar-raising signal onto a diabetes drug sounds like sabotage. It isn’t, and the reason comes from glucagon’s other, less famous job. Beyond telling the liver to release sugar, glucagon tells the body to spend energy — it revs up the liver’s metabolic rate and drives stored fat to be broken down and burned.

So it attacks weight from a completely new angle. The first two locks mostly get you to take in less. The glucagon lock gets the body to burn more.

And here’s the trick that keeps it safe: deliberate cancellation. Glucagon’s tendency to push blood sugar up is offset by the GLP-1 signal riding in the very same molecule, which pushes it back down. The two effects on sugar roughly neutralize each other, and the fat-burning benefit is left standing on its own.

The drug furthest down this path is retatrutide — a single shot, once a week, that fits all three locks at once. Its numbers are the biggest anyone has seen. In TRIUMPH-1, the phase 3 trial of more than 2,300 people with obesity that Lilly reported in May 2026, those on the highest dose lost an average of 28.3% of their body weight at 80 weeks. By the two-year mark, participants with severe obesity on the top doses were averaging over 30% — territory once reached only by surgery.

The trajectory is impossible to miss. One lever delivered around 15%. Two delivered above 20%. Three are now landing past 28%, and beyond 30% at the longest follow-up. Each signal you add buys another increment of weight loss.

The idea of folding several hormones into one molecule wasn’t obvious, and it has named authors. It was pioneered in the 2000s by a metabolism researcher, Matthias Tschöp, working with a peptide chemist, Richard DiMarchi. What they saw is a fact about the hormones themselves — and it’s also why this whole strategy belongs to peptides and not to pills.

GLP-1, GIP, and glucagon aren’t strangers to each other. They’re cousins — members of one family of signals built from similar runs of amino acids, so alike that GLP-1 and glucagon are literally cut from the same parent molecule. Their locks are related the same way. That family resemblance is what lets a chemist do something otherwise impossible: build a single hybrid chain that borrows features from two or three of these hormones at once, so one molecule fits several locks. And because the chain is assembled link by link, you can tune it — pressing one lock hard and another gently, in whatever ratio works best. Tirzepatide, for one, is deliberately lopsided, leaning harder on one of its two locks than the other.

A small-molecule pill — the cheap, rugged kind of drug described at the outset — almost never pulls this off. To hit three targets, a pill maker would usually need three separate compounds, each with its own dose, its own schedule, its own rate of clearing from the body, all juggled in one patient at the same time. And finding a tiny chemical that fits even one of these big protein locks is hard enough; finding one that fits three, in a controlled ratio, borders on impossible. A peptide is long enough to carry several addresses written into one chain.

That’s the philosophy behind the next wave of drugs. These molecules aren’t so much discovered as composed — deliberately written to speak several hormonal languages at once, in a chosen accent, from a single shot taken once a week. The lizard hormone and the fatty tail were about making a peptide survive. This is about making a peptide do several things on purpose at the same time.

That’s the shift that turns peptide medicine from a salvage operation into a design discipline.

Far Beyond the Scale

Every drug named so far got approved for one narrow job. First to steady blood sugar, then to trim weight.

Nobody scripted what came next: aim the same weekly injection at an organ that has nothing to do with appetite, and surprisingly often, something improves. The heart. The kidneys. The liver. The throat during sleep.

Each finding arrived as its own headline. But together they quietly reframed the central question of the field.

The weight loss that made these drugs famous may turn out to be the least interesting thing about them. The real prize may be in organs far from the bathroom scale.

The turning point was the heart. It has a name: the SELECT trial.

SELECT enrolled 17,604 people who were overweight and had established heart disease but — importantly — did not have diabetes, the condition these drugs were built for. Over several years, those on weekly semaglutide suffered about 20 percent fewer heart attacks, strokes, and deaths from heart disease than those on placebo. The rate fell from 8 percent to 6.5.

Two details made cardiologists sit up.

First, the gap between the groups opened early, well before most participants had lost much weight. Second, when statisticians traced how much of the benefit came from slimming, only about a third of it tracked with shrinking waistlines. The rest came from somewhere else.

In other words, the drug was guarding the heart by some route that didn’t run through the scale at all.

That distinction turned out to be worth billions. And it’s because of an accident in how health coverage is written.

In the US, Medicare — the government health plan for older Americans — is barred by a decades-old law from paying for drugs used purely for weight loss. The rule was written in the era of dangerous diet pills. But a drug that prevents heart attacks is another matter entirely.

So when regulators expanded semaglutide’s approval in March 2024 to include cutting cardiovascular risk, they handed Medicare a door it was now allowed to walk through. By one analysis, roughly 3.6 million Medicare beneficiaries — people with both excess weight and heart disease — became eligible for coverage almost overnight.

The SELECT result didn’t just add a medical use. It changed the drug’s identity in the eyes of the people who pay the bills: from a lifestyle indulgence into a treatment for the leading cause of death.

The kidneys came next. They’re a hard place to win.

They scrub the blood through millions of microscopic sieves. Once those sieves scar over, the damage is mostly one-way, and for decades barely any drug could even slow the slide.

A trial called FLOW set out to test whether semaglutide could. It enrolled 3,533 people with both type 2 diabetes and failing kidneys — a combination that often ends in dialysis or a transplant. Those on the drug were about 24 percent less likely to suffer a major kidney crisis or to die of kidney or heart causes, and their kidney function declined measurably more slowly.

It was the first time a drug of this class had been put to a dedicated kidney test. And it passed convincingly enough to be hailed as a new option for an organ that had long been running out of them.

The liver told a similar story. In millions of people with obesity, fat builds up inside the liver until it inflames the tissue and starts to scar it. This is a progressive disease, abbreviated MASH, that can harden into permanent cirrhosis, and until very recently it had no approved medicine at all.

A trial named ESSENCE biopsied the livers of nearly 1,200 patients before and after treatment. After about sixteen months on semaglutide, the inflammation had cleared — without the scarring getting worse — in roughly 63 percent of them, against 34 percent on placebo. In many, the scarring itself eased.

On the strength of those biopsies, semaglutide in 2025 became the first drug of its kind cleared to treat the disease.

Then came a condition that sounds mechanical, not metabolic. In obstructive sleep apnea, the soft tissue of the throat — padded with extra fat in people carrying excess weight — collapses during sleep and blocks the airway. The sleeper gets jolted half-awake dozens of times an hour, sometimes hundreds. The standard remedy is a mask strapped to an air pump, worn all night.

In a trial of tirzepatide, the two-hormone drug, up to half of patients improved so much they no longer crossed the threshold for the disorder, averaging some 25 fewer breathing interruptions per hour of sleep. In December 2024 it became the first medicine ever approved for sleep apnea — a problem doctors had always met with machines and surgery, never a pill or a shot.

Why should one gut hormone reach so far?

The answer lies in the lock itself. It isn’t confined to the pancreas. It studs cells across the body — the heart, the kidney, the blood vessels, the brain. A key can only turn a lock that exists, and this key’s locks are nearly everywhere.

But a second answer is only now coming into focus, and it turns on inflammation. The body’s inflammation system is the same alarm it uses to fight infection — swelling, heat, immune cells rushing to a wound. In obesity and in aging, that alarm gets stuck softly on, smoldering at a low level throughout the body and slowly damaging arteries, kidneys, and other tissue. It turns out the drug’s lock sits on immune cells too, and switching it on quiets that smolder.

And here’s the tell. In both mice and people, the calming of inflammation shows up within hours — long before any weight comes off. Which hints that the drug isn’t helping these organs only by shrinking fat. It’s directly turning down a destructive process that excess fat had turned up.

Where “helps everywhere” hits a wall

All of which made the brain the obvious next bet. It’s also where the story has fallen short.

There were real reasons to hope. The same chronic inflammation that scars arteries also shows up in the aging brain. Some researchers had even nicknamed Alzheimer’s “type 3 diabetes,” for the way brain cells seem to lose their grip on fuel.

Two large trials, EVOKE and EVOKE+, put the idea to a hard test. They enrolled about 3,800 people in the early stages of Alzheimer’s and gave them the swallowed form of semaglutide for two years. The results, reported in December 2025, were a plain disappointment. The drug did not slow the decline in memory and daily function any better than placebo. A few markers of the disease shifted in a hopeful direction, but the thing that mattered — the erosion of the mind itself — did not budge.

The expanding-indication story, in other words, is a hypothesis, not a law. “Helps everywhere” is a hope. And the brain just showed it has limits.

That failure is clarifying, because it sharpens the deepest question hanging over the whole class.

One reading of the heart, kidney, and liver results is that semaglutide is a kind of Swiss-army knife — separately useful against a long list of unrelated ailments. The other reading is stranger, and reaches a lot further: that these aren’t unrelated ailments at all, but many faces of a single underlying problem — the slow, systemic wear of carrying too much fuel.

On this view, excess weight isn’t the disease. It’s the most visible symptom. The same metabolic strain that thickens the waistline also inflames the arteries, scars the kidney, and fattens the liver, quietly, for years. If that’s right, the drug isn’t treating a dozen conditions. It’s easing one upstream cause that wears a dozen downstream names.

And the Alzheimer’s miss marks the boundary of the idea: not everything that goes wrong downstream can be undone by relieving the source.

Which picture is closer to true — versatile tool, or single root cause — is genuinely unsettled. But it’s the question on which the next decade of this medicine, and the sweeping claims now being made for it, will turn.

The Hundred-Billion-Dollar Race

These drugs have created one of the richest commercial prizes in the history of medicine. The chase for it now pulls in nearly every large drugmaker on earth.

Goldman Sachs and McKinsey expect annual sales of obesity and diabetes medicines in this class to hit roughly $100 billion by 2030. J.P. Morgan puts the wider category closer to $200 billion. Take a moment and let that sink in. That approaches the size of the entire global market for cancer drugs — the field that has dominated pharmaceutical spending for a generation.

A kind of medicine that barely existed a decade ago is on track to become the single biggest in the world. And the contest to own it is being fought on three fronts at once: science, factories, and very large checks.

For now, two companies hold almost all of it. Eli Lilly sells tirzepatide, the two-hormone injection. Novo Nordisk sells semaglutide, the original blockbuster.

But the balance between them has tipped hard, and fast.

As noted at the outset, Lilly crossed a trillion dollars in market value in late 2025 — the first drug company ever to do so. Novo went the other way. By the spring of 2026 its value had more than halved from its peak, to around $200 billion, and it had fallen out of the world’s ten most valuable companies entirely. The company warned that its sales would actually shrink in 2026 — its first annual decline since 2017 — squeezed by competition and by pressure on U.S. drug prices. Lilly now takes an estimated 60 percent of the obesity market to Novo’s 40, helped by a head-to-head trial in which its two-hormone drug clearly out-slimmed Novo’s single-hormone one.

The lesson the whole industry drew from that reversal is blunt. In this market, leads are not safe. Everything rides on what each company launches next.

Lilly’s next bet is the triple-hormone injection retatrutide — the one that adds a third signal to burn more energy rather than just eat less. The phase 3 results, beginning with TRIUMPH-4 in obesity and knee osteoarthritis (December 2025, ~28.7% loss at 68 weeks) and the much larger TRIUMPH-1 in obesity (May 2026, 28.3% at 80 weeks across more than 2,300 patients, climbing past 30% in severe obesity at the two-year mark), are the strongest anyone has seen from these drugs. That’s territory once reached only by surgery. The remaining phase 3 readouts — TRIUMPH-2 in diabetes and TRIUMPH-3 in established heart disease — are due later in 2026, with Lilly planning an FDA submission as soon as the end of the year.

Lilly’s second bet attacks the other great weakness of these drugs: the needle. Its candidate orforglipron is a once-daily pill — and importantly, not a peptide at all. It’s one of the small, rugged chemical compounds — the aspirin-like kind that survives the stomach and can be brewed cheaply by the ton. That single fact lets it sidestep both problems that make peptide pills so awkward. It doesn’t need the elaborate protective packaging an oral peptide requires. And it doesn’t carry the punishing manufacturing cost of a hand-built chain.

The trade-off is potency. In late-stage testing it drove about 27 pounds of loss, near 11 to 12 percent — less than the injections deliver. But it’s a pill, easy to make and easy to ship, and U.S. regulators approved it in April 2026 after an expedited review. For reaching the hundreds of millions of people who will never inject themselves, that matters far more than the last few percentage points.

Behind both bets sits a third contest most patients never see: who can physically make enough of the drug. Lilly has committed more than $50 billion to U.S. manufacturing in recent years, including a 2025 plan to build four new plants for $27 billion. Capacity is not a side issue here. It’s a competitive weapon. For years demand has outrun supply, and the company that can fill the most pens wins the most patients.

Novo Nordisk is fighting back on every one of these fronts, and has stumbled on several. Its headline next-generation drug, CagriSema, pairs semaglutide with a second ingredient that mimics amylin — another of the body’s own fullness hormones, released by the pancreas alongside insulin. The hope was a 25 percent weight loss that would leapfrog Lilly.

The reality disappointed twice. When the first big trial reported in December 2024, the number came in at 22.7 percent — real, but below the company’s own guidance, and the stock fell sharply on the news. Worse, a later trial pitting CagriSema directly against tirzepatide failed to prove it was even as good as Lilly’s drug. Novo filed for approval anyway in December 2025, but the molecule meant to restore its lead had instead underlined how far ahead Lilly had pulled.

Novo’s other irons are still in the fire. It already has a pill on the market: a higher-dose tablet form of semaglutide, sold as the Wegovy pill, which U.S. regulators cleared as the first oral drug of its kind for weight loss. It launched in early 2026 at about $149 a month for those paying cash — a fraction of the four-figure list prices these drugs debuted at. Novo is also developing a newer molecule, amycretin, that combines GLP-1’s action with that same amylin signal in a single compound, available as both a weekly shot and a daily pill; mid-stage results in late 2025 were solid enough to push it into final-stage trials in 2026.

And to make all of this, Novo’s parent foundation bought the contract manufacturer Catalent for $16.5 billion, with the drug company itself paying $11 billion to take over three of its fill-and-finish plants outright — a scramble for capacity to match Lilly’s.

The Challengers

Below the two giants, a pack of challengers is each attacking from a different angle, each betting there’s room for a drug that does one specific thing better.

The most distinctive comes from Amgen. Its candidate, MariTide — which blocks the GIP signal rather than switching it on — is built for a once-monthly injection rather than a weekly one. It pulls that off by fusing the active piece to an antibody, one of the large, long-lived immune proteins built for stability. Where the fatty-tail trick buys a peptide about a week in the body, riding an antibody stretches it to a month. Mid-stage results showed weight loss around 20 percent, and a large final-stage program is now running. A single shot a month, if it holds up, is a real selling point for the people who dread needles.

A very different kind of challenger is Viking Therapeutics, a small California biotech with no marketed products at all — the kind of firm that, if its drug works, tends to get bought. Its molecule, VK2735, hits the same two hormone targets as Lilly’s tirzepatide, and comes as both an injectable and a pill. The injectable entered large final-stage trials in obesity and diabetes through late 2025 and early 2026, with thousands of patients enrolled. Structure Therapeutics, another biotech, is chasing the same goal as Lilly’s pill from a different lab bench: a once-daily, non-peptide tablet. In mid-stage testing it produced up to about 15 percent weight loss, and it’s set to enter final trials in the second half of 2026.

Not every challenger is aiming at the bathroom scale. Zealand Pharma, a Danish firm, is working with the German company Boehringer Ingelheim on a drug called survodutide. It hits the GLP-1 and glucagon targets. The target disease isn’t obesity but fatty liver disease, a progressive scarring condition that drives a large share of liver transplants. In mid-stage trials, survodutide cleared the disease in a striking share of patients. It’s now in final-stage testing for the liver rather than for weight.

Which is a reminder that the prize here is not one market but many. A latecomer can win by picking the right organ.

Buying In

For the drugmakers that missed the start, the fastest way in has been to buy.

Roche, the Swiss giant, jumped in late 2023 by buying a small startup, Carmot Therapeutics, for $2.7 billion, picking up a portfolio of injected and oral candidates in one stroke. In 2025 it spent up to $5.3 billion more in a partnership with Zealand Pharma for an amylin-based drug.

Pfizer’s path has been more painful. It bet on a daily pill of its own, danuglipron, and stumbled with it over and over: the twice-daily version was abandoned in 2023 after more than half the patients quit over side effects, and the once-daily version was scrapped in April 2025 when a single patient showed signs of liver injury. Having failed to build, Pfizer bought — winning a tense bidding war in November 2025 to acquire the obesity startup Metsera for as much as $10 billion. It beat out Novo Nordisk, which had bid against it until U.S. antitrust regulators objected to a deal that would have put yet another obesity asset in Novo’s hands.

Roche and Pfizer are just the largest names in a broader rush. The wave of deals reflects a simple calculation: buying a credible program, even an expensive one, beats being shut out of a $100 billion market.

What Will Decide It

Several distinct battles will pick the winners, and they’re being fought in parallel.

One is the pill versus the shot. With both a peptide tablet and a non-peptide tablet now approved, the daily pill that travels and stores easily could vastly widen who takes these drugs — even if it works a little less well than the weekly injection. Another is dosing convenience itself: the once-monthly shot against the once-weekly. A third, easy to overlook, is raw manufacturing muscle, where the tens of billions being poured into new plants will determine who can actually meet demand.

The last battle is about price, and that clock is already ticking. Patents don’t last forever, and when they lapse, cheaper copies arrive and sales can fall off a cliff. Semaglutide’s protection has already begun expiring outside the United States — in Canada, India, Brazil, and China during 2026 — opening those markets to copies now. The lucrative U.S. market is shielded longer, until roughly 2032, behind a dense thicket of patents. And because these drugs are peptides rather than simple chemicals, copycats can’t be rubber-stamped the way ordinary generics are. They have to clear their own trials as “biosimilars” — near-copies that take longer to reach pharmacies and tend to lower prices more gently. Even so, the era of $1,000-a-month pricing is ending, pushed down by competition and by cash-pay tablets at a tenth of that.

The milestones to watch are now close together. The first oral drugs have just been approved and are launching into 2026. The triple-hormone injection’s remaining trial results land through 2026, with a possible launch to follow. Newer molecules enter final testing the same year. And the first real patent pressure on the original blockbusters has already begun abroad, building toward the U.S. cliff at the decade’s end.

The hundred-billion-dollar race has no settled finish line. Only a widening field — and two leaders who have learned that being first guarantees nothing.

The next generation already in trials

Underneath the commercial fight, the next generation of drugs is taking shape — each one a deliberate fix to a specific limit of the first wave.

The biggest single move is to layer in a different lever. GLP-1 is not the only signal the body uses to say enough. At the very moment it releases insulin, the pancreas releases a second fullness hormone called amylin — and it works through an entirely separate set of locks, clustered mostly in the brainstem rather than spread across the many organs GLP-1 reaches. Different signal, different road, which is exactly why you can layer it onto GLP-1: the two effects stack instead of colliding. The deeper interest in amylin is that it may restore a conversation obesity broke — a fullness signal called leptin that fat tissue uses to report up how much fuel is stored, and that the brain goes partly deaf to in obesity. Amylin appears to re-open that channel.

Two long-acting amylin drugs are now well along. Novo Nordisk’s cagrilintide, a once-weekly shot, drove about 12% weight loss on its own in late-stage trials. Zealand Pharma’s rival petrelintide delivered a comparable result in mid-stage testing. The headline isn’t the magnitude — it’s the tolerability. The nausea and vomiting that shadow the GLP-1 drugs were barely there. At the most effective dose, one trial recorded no vomiting at all, no one quit over stomach trouble, and side effects landed close to placebo. Paired with GLP-1, amylin pushes total weight loss toward the top of the range while keeping patients comfortable enough to actually stay on it. That pairing is what Novo’s CagriSema was meant to be, and what amycretin — the single-molecule GLP-1-plus-amylin Novo now has in final trials — is the better-engineered version of.

A second fix attacks a problem the headline numbers hide. When the body sheds weight this fast, a quarter to forty percent of the loss is lean tissue — muscle and the scaffolding around it. Much of that is just the expected shrinking of an over-large frame. But for older patients, losing muscle can mean trading obesity for frailty. The fix being chased is a second drug that guards muscle while the appetite drug strips the fat — by jamming the body’s own brake on muscle growth, a signal called myostatin. The lead candidate is bimagrumab, an antibody built to plug the receptor myostatin docks into. In a mid-stage trial, adding it to semaglutide produced about 22% weight loss — and roughly 92% of what came off was fat, a far cleaner result than the appetite drug alone. Eli Lilly bought the company behind it for nearly $2 billion in 2023. The vision is a prescription that’s really a small stack: one drug to quiet appetite, another to protect the muscle, so patients lose the right kind of weight.

The third front isn’t about biology at all. It’s about reach. In real-world tracking, roughly half of patients have stopped taking these drugs within a year, and among those taking them for weight rather than diabetes, closer to two-thirds quit. A medicine that only works while you take it, aimed at people who keep stopping, has a reach problem before anything else. The fixes are mostly the convenience plays already running: cheap daily pills for the many who’ll never inject themselves, and once-a-month shots to thin the schedule for those who will. The science isn’t glamorous. But the arithmetic is decisive. A drug a hundred million people might plausibly start — and actually stay on — is worth far more than a slightly stronger one most of them abandon.

Taken together, the next-generation pipeline is concrete, dated, and already in human trials. Triple-hormone drugs (retatrutide) past 28% in phase 3. Amylin combinations dropping the side-effect bill. Muscle-sparing partners in late-stage. Once-monthly shots in phase 3. Oral peptides in patients for the first time. None of these requires a scientific miracle to land. Most read out by 2027 or 2028. This is the floor of what the next five years brings to medicine.

Beyond Metabolism

Beneath all of that — beneath the commercial race, beneath the next-generation pipeline — sits a much larger story. The drug class is starting to look like something none of its original architects bet on, and the GLP-1 era is starting to look like the prologue to a peptide-design era that may dwarf it.

Two shifts in particular are turning a metabolic drug class into something the field is only just beginning to grasp. The first is what’s already happening with the existing drugs in tens of millions of people: GLP-1s are turning up in places they were never designed to reach. The second is that the drug class itself is being reinvented from scratch.

The asymmetric bets

The signal pulling the field’s largest checks isn’t another weight-loss number. It’s what’s surfacing in the periphery — places GLP-1 was never designed to reach but seems to help anyway. Each of these is asymmetric: a moonshot if it lands, nearly free to attempt because the drugs already exist and are already in tens of millions of people. Any one of them, if it pans out, expands the market for these molecules by a category.

The longevity bet. It took a few years for the framing to assemble, but it’s now in the literature: GLP-1 agonists are the first commercially available drugs touching multiple hallmarks of biological aging at once. Chronic low-grade inflammation, insulin resistance, mitochondrial dysfunction, visceral fat — the underlying processes that grade how fast a body wears out. The mortality data is starting to make the picture concrete in unexpected places. A 2025 study of cancer patients on immunotherapy found GLP-1 users had a 31% lower five-year mortality — 32% died over five years versus 45% of non-users. In the SELECT cardiovascular trial, the heart benefit started showing up before the weight came off, hinting the drug works on tissues directly, not just through pounds lost. Nature Biotechnology, in late 2025, framed the question that’s now driving the conversation: are GLP-1s the first longevity drugs? It’s no longer a fringe claim. It’s a paper title in a top journal — and the trials needed to settle it are starting.

The addiction bet. Patients on GLP-1s started reporting, unprompted, that they drank less, smoked less, used drugs less. In March 2026, Washington University researchers published the largest dataset on the question to date. GLP-1 users had a 15–20% lower risk of developing substance use disorders across opioids, cocaine, nicotine, alcohol, cannabis, and other major substances. For people already addicted, the effect was sharper: lower rates of drug-related death, overdose, hospitalization, and suicide attempts. The mechanism is the same one that quiets food noise — GLP-1 reaches into the brain’s mesolimbic reward system and damps the dopamine that hijacked behavior runs on. In vervet monkey studies, semaglutide didn’t make alcohol aversive. It made it boring. The drug noise quieted the same way the food noise had. Dedicated trials in alcohol use disorder are now running, and several companies are positioning addiction as a primary indication rather than a serendipitous side effect.

The cancer-prevention bet. Even more unexpected. A 2026 analysis of the TriNetX health network found that diabetic patients started on a GLP-1 had a 70–80% lower relative risk of liver cancer compared to patients started on insulin. At ASCO 2026, a study reported colorectal cancer rates 36% lower in GLP-1 users than in aspirin users — and 42% lower in the higher-risk subgroup. Real-world data is now showing reduced metastatic progression in obesity-linked lung, breast, colorectal, and liver cancers. Meta-analyses of the long-standing thyroid-cancer concern, the one the FDA’s sternest warning is built on, continue to find no matching human signal at population scale despite the rodent data. Whether the protective effect is anti-inflammatory, metabolic restraint, or something the field hasn’t named yet, the trajectory is clear: one of the largest cost lines in medicine, cancer, may be quietly held back by a drug class no one designed for it.

The neurodegeneration bet. Novo Nordisk’s EVOKE and EVOKE+ phase 3 Alzheimer’s trials read out the first major test of whether semaglutide can slow cognitive decline in 2026. The phase 2 results were modest. But the underlying bet — that anti-inflammatory and metabolic benefits travel into the brain — is the same one that put GLP-1 and triple-agonist drugs into Parkinson’s, ALS, and stroke-recovery trials over the past two years. Whether the first generation cracks Alzheimer’s or not, the door it has opened toward treating brain disease through metabolism is wider than it has ever been, and the next-generation drugs entering these trials carry triple-hormone designs that the EVOKE molecule didn’t.

Put these together and the asymmetric thesis comes into focus. A drug class built to lower blood sugar is turning into something that may, depending on which readouts land, also turn down inflammation, addiction, cancer risk, brain disease, and the speed of biological aging itself. That is not what was promised when the first GLP-1 drug was approved in 2005. It’s a much larger story than the one the field has been telling. And the cost of being wrong on any of these bets is small, while the upside on any one of them being right is enormous.

Peptides from scratch — the next class

The bets above are still about this drug class, hammered into new uses. The deeper change is that the drug class itself is being reinvented.

Every drug to this point traces back to a molecule the body already builds. The deepest break now underway abandons that constraint entirely. Scientists have begun designing brand-new peptides from nothing — sketching, on a computer, a molecule that has never existed in any living thing, purpose-built to grab onto a chosen target. The approach is called de novo design, Latin for “from scratch.” Its leading figure is David Baker, a biochemist at the University of Washington who shared the 2024 Nobel Prize in Chemistry for it. That prize honored two faces of one revolution. One half went to learning to read nature’s proteins — predicting by computer the shapes they fold into, the feat behind AlphaFold. The other half, Baker’s, went to learning to write entirely new ones.

The tool that made designed peptides practical works in a way anyone who has watched an AI conjure a picture from a text prompt will recognize. Baker’s program, RFdiffusion, starts with a formless cloud of amino acids — the molecular equivalent of visual static — and step by step sculpts that noise into a precise new shape, contoured to clasp whatever target the designer names. His lab built molecules that lock onto a series of hormone signals — among them glucagon — gripping them more tightly than any computer-designed binder ever had, and often succeeding with as few as a single design tested, rather than the tens of thousands a traditional drug hunt would screen. A process that once took years of trial and error can now take one try.

That capability is pulling in extraordinary sums on a clear bet: that the GLP-1 windfall isn’t a fluke but a template. Proof that a single well-made peptide can be worth tens of billions, and that AI can now turn out such peptides to order. The boldest wager is Xaira Therapeutics. It launched in April 2024 with more than $1 billion in hand — among the largest sums ever raised to start a drug company, and the biggest first commitment in the history of its lead backer, the venture firm ARCH Venture Partners. Xaira is the largest of a cluster of new AI-driven design firms, and it was built directly on Baker’s work — he’s a co-founder. It’s run by Marc Tessier-Lavigne, a former head of Genentech’s research and a former president of Stanford. The thesis is that drug discovery, long a matter of screening millions of compounds and hoping, is turning into a design discipline, with peptides as its first great product.

The reason the revolution is landing on peptides specifically returns to the two extremes the field has long been stuck between. The small chemical pills can be made by the ton, but they’re too tiny to get a grip on many disease targets — the large, smooth protein surfaces biology uses to do its work offer no pocket for a little molecule to wedge into. Antibodies can blanket those surfaces, but they’re enormous, costly, and impossible to swallow. A peptide sits right in the gap. Big enough to drape across a broad target and hold on where a pill can’t, yet small enough to build by chemistry — and, the part that matters now, small enough for a computer to model atom by atom and design with real confidence. That’s the sweet spot AI has reached first.

What this opens isn’t a new drug. It’s a new way of finding drugs. The targets a computer-designed peptide can hit include the ones the field has spent decades calling “undruggable” — protein–protein interactions that drive cancer, transcription factors that drive autoimmune disease, intracellular targets that current biologics can’t reach. If the GLP-1 era taught the industry that one well-built peptide can be a trillion-dollar molecule, the next era is the field building those peptides on demand, against targets nobody had a way to hit before.

The next ten years, in plain terms

The peptide era didn’t arrive. It started.

The first hit is real and proven. The second wave — gentler drugs, smaller pills, once-monthly shots, muscle-sparing combinations — is already in trials. The asymmetric bets, the ones that could change what whole categories of human suffering look like, will read out over the next two or three years from drugs already in tens of millions of people. And underneath all of it, the field is learning to build peptides from scratch against targets it had spent decades calling unreachable.

Step back, and what’s happening is simpler than the noise around it suggests. One of medicine’s oldest standing problems — the slow metabolic damage that quietly drives heart attacks, strokes, dementia, addiction, and a long stretch of cancers — has, for the first time, an instrument that touches it directly. Not perfectly. Not for everyone. But directly, in a single shot a week.

A child born today will likely live in a world where treating obesity is as ordinary and durable as treating high blood pressure. Where the cravings behind addiction have a brake their doctor can prescribe. Where lower rates of heart attacks, strokes, and several major cancers are listed among the routine effects of the most prescribed drug class of their generation. And where aging itself is partly treated, rather than only endured.

That’s the prize on the table. It isn’t immortality. It’s something more humble and more sweeping — a quiet rewrite of what we accept as the background conditions of being alive.

The lizard hormone was the prologue. The book is being written now.