Why Insulin Costs So Much in America — And Almost Nothing Everywhere Else

Insulin was first sold under a patent assigned for $1, and a 2018 study estimated that a year's supply could be produced and sold at a profit for somewhere between $48 and $133 per patient. Yet in 2018 the average US manufacturer price was $98.70 per standard unit, against $8.81 across 32 comparison countries. This article covers the biology that makes insulin non-negotiable, how the American price gap formed, what has changed since 2023, and what remains unresolved.

Educational content, not medical advice. Talk to your doctor before changing any medication, treatment, or routine. Information here reflects current research but every person's situation differs.

Key takeaways

• Banting, Best, and Collip assigned their insulin patent to the University of Toronto for $1 in early 1923, but insulin was still manufactured and sold commercially from the start.
• People with type 1 diabetes make little or no insulin and cannot go without it; rationing for cost is documented, and it can end in diabetic ketoacidosis.
• A 2018 study in BMJ Global Health estimated that biosimilar human insulin could be sold profitably at $48 to $71 per patient per year and most analogues at $78 to $133.
• The US price is set in two layers: a high list price that patients with deductibles often pay, and a lower net price after confidential rebates.
• The FTC alleged in 2024 that the three largest pharmacy benefit managers (PBMs) built a rebate system that rewarded high list prices; two of the three had settled by mid-July 2026.
• Since 2023, Medicare caps insulin at $35 per month, and the three main manufacturers cut list prices on key products by 70 to 78 percent.
• Large gaps remain for uninsured people, for patients on some private plans, and in how quickly biosimilar competition arrives.

A $1 patent, and what actually happened next

The story people tell is simple: two Canadians found a miracle, gave it away, and a greedy industry stole it back. The record is more interesting, and it explains a lot about why the price problem is structural rather than a morality play.

Frederick Banting was a surgeon with no research track record when he pitched John Macleod, a physiology professor at the University of Toronto, on an idea for isolating the pancreas's internal secretion. In the summer of 1921 Macleod gave him lab space and a student assistant, Charles Best. By January 1922 the team was treating Leonard Thompson, a 14-year-old boy dying of diabetes, and the results spread through the medical world within months. James Collip, a biochemist, joined to purify the extract.

The patent came next. On January 1, 1923, Banting, Best, and Collip assigned the patent on the extraction process to the University of Toronto's Board of Governors for one dollar each, according to the university's Thomas Fisher Rare Book Library records. The stated goal was to stop anyone from securing a profitable monopoly on the process.

The patent did not mean "free"

Here is the part that gets lost. The University of Toronto's own Connaught Laboratories could not make enough good extract, and a shortage hit Toronto in spring 1922. On May 20, 1922, Eli Lilly and Company received exclusive rights to manufacture insulin for the United States and Central and South America, with the two sides agreeing to share knowledge on production problems. A $1 patent kept the process out of any single owner's hands. It did not turn insulin into a charity product.

The 1923 Nobel Prize in Physiology or Medicine went to Banting and Macleod, who each shared their prize money with Best and Collip respectively. The point of the patent assignment was access to the method, not a promise about what future companies would charge for future products.

From animal insulin to engineered insulin

For six decades, insulin came from the pancreases of cattle and pigs. That changed in 1982, when the FDA licensed Humulin, developed with Genentech and brought to market by Lilly, as the first medical product made with recombinant DNA technology, according to the National Science Foundation. Researchers inserted the human insulin gene sequence into bacteria, which then produced large quantities of insulin matching the molecule humans make. If you want the cellular machinery behind that, our explainer on how proteins are made covers the ribosome step that the bacteria are doing on the company's behalf.

The next wave was "analogues," insulins with small amino acid changes that make them act faster or last longer. Humalog and Lantus, launched in the 1990s and around 2000, are the best-known examples. Analogues proved convenient and became the products around which the modern market, and its price disputes, formed.

What insulin does, and why it cannot be skipped

Insulin is a hormone made by beta cells in the pancreas. Think of it as a key that unlocks cells so they can take up glucose from the blood. Without the key, glucose piles up in the bloodstream while the cells themselves starve. For a broader tour of hormones as chemical messengers, see our piece on how hormones work, and for the storage side of the story, how the body stores and releases glucose.

In type 1 diabetes, the immune system destroys the beta cells. Without injected or pumped insulin, the body switches to burning fat for fuel and produces acidic compounds called ketones. When they accumulate, the blood turns acidic, a condition called diabetic ketoacidosis (DKA), which can cause coma and death. There is no dietary or exercise workaround for a person who makes no insulin.

Type 2 diabetes is different: cells respond poorly to insulin, and the pancreas often keeps producing some. Many people with type 2 are treated with other medicines first, and some eventually need insulin. Our articles on insulin and diabetes and the biology of type 2 diabetes cover the distinction in depth.

How many people are affected

The Centers for Disease Control and Prevention (CDC) estimates that 40.1 million Americans, about 12.0 percent of the population, had diabetes in 2023. Of the 29.1 million with a diagnosis, 2.1 million had diagnosed type 1 diabetes. Not everyone with diabetes uses insulin, and no single authoritative national count of insulin users is cited here because the commonly repeated figures vary by source and year.

Rationing: what the evidence shows

In June 2017, Alec Smith, 26, was found dead in his apartment. His birthday had come on May 20, and he had aged off his mother's insurance. Without insurance, his diabetes supplies were estimated at about $1,300 a month, mostly insulin. The marketplace plan he found carried a $450 monthly premium and a $7,600 annual deductible. He died of diabetic ketoacidosis, and his family believes he was rationing, using less than he needed to stretch his supply, according to KFF Health News. The pen he was using was empty.

One case is not a statistic, but the statistic exists. A survey of patients at the Yale Diabetes Center, published in JAMA Internal Medicine in December 2018, found that one in four respondents reported using less insulin than prescribed because of cost. Those who underused insulin were more likely to have poor blood-sugar control, 43 percent versus 28 percent. The study came from one center and cannot prove cause and effect, but the pattern fits what clinicians describe.

What insulin actually costs to make

The widely cited estimate comes from a 2018 paper in BMJ Global Health by Gotham and colleagues, with support from Health Action International's ACCISS project. The authors estimated the price at which a producer could sell a year's supply and still turn a profit, building up from active ingredient costs, excipients (inactive ingredients), formulation, transport, development, regulatory costs, and a margin. According to the study summary, the estimates were:

Product type	Estimated profitable price, per patient per year
Biosimilar human insulin	about $48 to $71
Most analogue insulins	about $78 to $133
Glargine (long-acting analogue)	about $78 to $108

Two caveats matter. First, these are estimates of a sustainable price under competition, not an audit of any company's books. The authors assumed generous development costs and noted that prices could fall further with more competitors. Second, insulin is more than a vial of protein. Pens, needles, cold storage, and quality control are real costs, which is why the study includes them and why a figure like "pennies" is too glib.

The takeaway is not that insulin is free to make. It is that the distance between a competitive production-based price and what the US market charged cannot be explained by manufacturing alone. The same study described the global market as dominated by three multinational companies, and its abstract put their combined share at 96 percent. That concentration matters for everything that follows.

How a US insulin price is actually set

Most goods have one price. US prescription drugs have at least two, and insulin's problems sit in the gap between them.

The list price (technically the wholesale acquisition cost) is the sticker price set by the manufacturer. The net price is what the manufacturer actually keeps after rebates paid to middlemen. Think of a hotel's rack rate versus what a corporate travel desk pays: almost nobody pays the posted rate, except the traveler who has no contract and walks in off the street.

For insulin, the person walking in off the street is a patient with no insurance, or with a high deductible or coinsurance based on list price. The FTC's 2024 complaint alleges that the largest PBMs shifted the burden of high list prices to exactly these patients.

Who sits between the factory and the patient

Player	Role	How it gets paid
Manufacturer (Lilly, Novo Nordisk, Sanofi)	Makes the insulin, sets list price	Net price after rebates
Pharmacy benefit manager (Caremark, Express Scripts, OptumRx)	Negotiates with manufacturers for insurers, decides which drugs are on formularies	Rebates, fees, and other revenue linked to list price, per the FTC
Insurer or employer plan	Pays part of the bill	Premiums, minus any rebate pass-through
Pharmacy	Dispenses	Dispensing fee and margin
Patient	Pays deductible, coinsurance, or the full price	Out-of-pocket

The three PBMs named by the FTC handle about 80 percent of US prescriptions, according to the agency. That makes them a bottleneck for any manufacturer that wants broad coverage, and it gives them leverage in rebate negotiations.

The rebate trap, in numbers

A toll road offers a useful comparison. Imagine the toll collector is paid a percentage of the posted toll rather than a flat salary. The collector has no reason to push the posted toll down, and every reason to want it higher, even if each driver's actual trip cost is unchanged through some side discount.

The FTC's complaint describes something like this for insulin. PBMs earn revenue tied to a drug's list price. The agency alleges that starting around 2012, PBMs threatened to exclude drugs from their formularies unless manufacturers paid larger rebates, and manufacturers raised list prices to fund those rebates. A lower-list-price insulin could lose its place to an identical product with a higher list price and a bigger rebate.

Here is a simplified illustration. The numbers are hypothetical and do not describe any company's actual contract.

Scenario	List price per vial	Rebate	Manufacturer's net	Price a patient in a deductible pays
Before	$300	$180 (60%)	$120	$300
After list increase	$400	$280 (70%)	$120	$400

The manufacturer's net revenue is identical in both rows. The rebate pool shared with the PBM grew by $100 per vial. The only party who is worse off is the one paying list price. This is the mechanism critics call a perverse incentive, and it explains why list prices could climb for years while net prices barely moved.

Both sides dispute the story

PBMs reject the framing. PBMs say they lower drug spending for consumers and have disputed the FTC's characterization of their role. The FTC's administrative complaint, filed on September 20, 2024, names the three PBMs and their group purchasing organizations, and the commission voted 3-0, with two commissioners recused. As of mid-July 2026, Express Scripts and Caremark had settled, with OptumRx the only remaining respondent. The Caremark order requires patient out-of-pocket costs to be based on the contracted amount less rebates rather than list price. Settlements are not findings of liability, so the allegations should be read as allegations.

A price history that is hard to explain by cost

The FTC's complaint cites two examples. Humalog's list price rose from about $21 in 1999 to more than $274 by 2017, an increase of over 1,200 percent. Novolog U-100 rose from $122.59 in 2012 to $289.36 in 2018, more than doubling.

Product	Earlier list price	Later list price	Source
Humalog	about $21 (1999)	more than $274 (2017)	FTC complaint, 2024
Novolog U-100	$122.59 (2012)	$289.36 (2018)	FTC complaint, 2024

An independent measure points the same direction. The Health Care Cost Institute (HCCI) found that per-person insulin spending for people with type 1 diabetes roughly doubled between 2012 and 2016, from about $2,900 to $5,700, and concluded the increase was primarily driven by higher prices, according to CBS News's report on the analysis. Those figures represent gross spending and do not account for rebates or coupons, which can lower costs for some patients.

A caution about older figures: widely repeated percentages such as "168% to 325%" for 2012 to 2016 do not appear in the sources checked for this article, and they are omitted here. The documented story is already stark without them.

Patents, "evergreening," and the slow arrival of competition

A vial of insulin from 2000 would not be patent-protected in 2020, in theory. In practice, manufacturers filed new patents on devices, formulations, and manufacturing processes, extending the period in which rivals could not easily enter.

The advocacy group I-MAK analyzed Sanofi's Lantus in a 2018 report. It found that 95 percent of the patent applications on Lantus were filed after the drug was already on the market. I-MAK argues that this "thicket" delayed competition. Sanofi and other manufacturers argue that follow-on patents reward real improvements and that filing before approval is not always possible. Readers should treat I-MAK's analysis as an argument from an advocacy organization, supported by public patent records.

The regulatory quirk that delayed biosimilars

Insulin is a protein, and proteins are what regulators call biologics. For decades, though, US insulin was approved as a conventional drug, which meant the biosimilar pathway Congress created in 2010 did not apply to it. In March 2020, insulin products were moved into the biologic category. A biosimilar is a biologic that is highly similar to an approved reference product with no clinically meaningful differences in safety or effectiveness. It cannot be an exact copy, since large proteins made in living cells are never perfectly identical batch to batch, but it can compete on price.

The first US interchangeable biosimilar insulin was Semglee (insulin glargine-yfgn), approved by the FDA on July 28, 2021, as a substitute for Lantus. "Interchangeable" means a pharmacist may substitute it without calling the prescriber. Rapid-acting biosimilars, which reference products such as Novolog, arrived later.

Competition has been slow to register. Biosimilar uptake depends on formulary decisions by the same PBMs whose rebate structures reward the high-list-price products, which is one reason a technically available cheaper product does not automatically reach patients.

What other countries pay, and what that does and does not prove

A RAND Corporation analysis commissioned by the Department of Health and Human Services' Office of the Assistant Secretary for Planning and Evaluation (ASPE) compared insulin prices using 2018 sales data. Per STAT's summary, the average US price per standard unit, RAND's common measure that lets vials, pens, and cartridges be compared, was $98.70.

Comparison (2018, price per standard unit)	Price	US price as multiple (calculated)
United States, all insulin	$98.70	1.0
Canada	$12.00	about 8.2
United Kingdom	$7.52	about 13.1
Australia	$6.94	about 14.2
All other OECD countries combined	$8.81	about 11.2
US rapid-acting vs other OECD	$119.36 vs $8.19	about 14.6
US intermediate-acting vs 32 other countries	$73.56 vs $5.98	about 12.3

The multiples are simple divisions of the published prices. STAT's own phrasing was that US prices were "often five to 10 times higher" than abroad two years earlier, so the ratio depends on which comparison and which year you pick. What is not in dispute is the direction and the size: the US was an outlier across product types.

Why the gap exists

RAND's comparison is a list-price comparison. Net prices in the US are lower than list, and the gap between list and net is exactly the rebate system described above. Cross-country price comparisons therefore overstate what a typical insured American pays and understate what an uninsured one pays.

Other wealthy countries negotiate or regulate as a bloc, though the methods differ. In Germany, under a 2011 law known as AMNOG, an independent nonprofit and a federal regulator assess each new drug's added benefit, and the price is then negotiated with the statutory health insurance funds that cover more than 90 percent of the population, with an arbitration board for deadlocks. In Canada, the pan-Canadian Pharmaceutical Alliance negotiates on behalf of public drug plans, and the Patented Medicine Prices Review Board can order reductions for patented drugs it judges excessively priced, as summarized by The Incidental Economist. Because many insulin products are long past their original patents, patent-focused price review is only part of the picture in any country; competition among suppliers and bulk purchasing matter too.

A fair reading is that the United States combined a concentrated supply side, a fragmented purchasing side, and a rebate system tied to list price. Few other wealthy countries had all three at once.

What changed from 2023 to 2026

After years of public pressure, a burst of change arrived. Most of it targets what the patient pays at the counter or what list prices say, rather than rebuilding the incentive structure.

Date	Change	What it does	Limits
2023	Medicare caps insulin at $35 per month under the Inflation Reduction Act	Medicare.gov says a one-month supply of each Part B or Part D covered insulin product costs no more than $35, with no deductible	Medicare only; a private-insurance version was dropped in the legislative process
Mar 2023	Lilly announces 70% list-price cuts and a $35 cap for commercially insured and uninsured patients	Humalog list price falls to $66 per vial	Participating pharmacies; manufacturer program, not law
Jan 1, 2024	Medicaid rebate cap removed	Rebates can exceed the drug's price, so high list prices lose their payoff	Interacts with Medicaid only
Jan 2024	Novo Nordisk cuts list prices up to 75%; Sanofi cuts Lantus 78%	NovoLog $72 per vial; Lantus $64 per vial	List prices; patients on net-price plans see less change
2026	California's CalRx glargine pens	$55 for a five-pack of 3 mL pens, interchangeable with Lantus; vials still pending	State-branded product; one state's market
2026	FTC settlements with Express Scripts and Caremark	Orders on rebate practices and patient cost-sharing basis	OptumRx pending as of mid-July 2026

The list-price cuts in the table come from ABC News's reporting on the 2023 and 2024 announcements. That same reporting explains a motive that is easy to miss: a 2021 law removed the cap on the rebates drugmakers pay Medicaid, so a high list price with fast-rising increases had become costly. Analysts estimated the cuts would let Lilly avoid about $430 million in rebates in 2024, Novo Nordisk about $350 million, and Sanofi about $560 million. The cuts were partly generosity under pressure and partly arithmetic.

California's initiative shows another route. The state partnered with Civica Rx, a nonprofit, and Biocon Biologics to sell CalRx insulin glargine pens at no more than $55 per five-pack. Civica's suggested retail price for a 10 mL vial is capped at $30, but vials had not yet launched at the time of checking.

State laws add another patchwork. Many states have passed insulin copay caps, with limits that vary by state. These laws apply to state-regulated plans and do not reach every employer plan.

What has not been fixed

Several problems remain even after the cuts and caps.

• Uninsured people depend on manufacturer programs, retail coupons, or newer low-cost products rather than a legal guarantee.
• The list-versus-net structure is still there. Cuts to list prices reduce the sticker price, but the rebate machinery that rewarded high list prices has not been repealed by statute.
• Newer, more complex products, such as insulin pumps and continuous glucose monitors, carry costs that a $35 insulin cap does not touch.
• The 2018 comparison is dated. Prices have changed since, and no figure in this article should be read as today's international ratio.
• Biosimilar uptake hinges on formulary behavior, so a cheaper approved product does not guarantee a cheaper pharmacy counter.

For context on how total health spending compares, our ranking of countries spending the most on healthcare shows that insulin is one slice of a larger national pattern. The mechanics of deductibles and coinsurance are explained in how insurance actually works.

Common misconceptions

"Insulin is expensive because it is hard to make." The production-cost estimates above say otherwise, though producing insulin with devices and quality control is not free. The price gap is mostly about market structure.

"The price of insulin is the same for everyone in America." It is not. A person in Medicare, a person on a plan with a $35 cap, and an uninsured person buying at list price can face very different amounts for the same product.

"The inventors gave away insulin." They assigned a process patent to a university for $1 each to block a private monopoly on the method. Within a year, commercial manufacturing under license began.

"Cutting list prices solves it." It helps the patients who pay list price. It does not by itself change how PBMs are paid, which is why regulators focused on rebate practices.

"The US just pays for the world's drug innovation." Insulin's core discovery is a century old, and the 2018 cost estimate above already assumed generous development costs. Whether US prices fund innovation in newer drug classes is a separate debate that this article does not settle.

FAQ

Q: How much does it cost to make a vial of insulin? A: No official audit exists. A 2018 BMJ Global Health study estimated that a year's supply of biosimilar human insulin could be sold profitably for about $48 to $71 per patient, and most analogues for about $78 to $133, including a profit margin. Those are estimates of a competitive price rather than measured factory costs.

Q: Why is insulin so much cheaper in Canada or Australia? A: Those countries negotiate or regulate prices at a national or provincial level and rely less on rebate-based pricing. RAND's 2018 comparison found average prices of $12.00 per standard unit in Canada and $6.94 in Australia, against $98.70 in the United States, using list-level prices.

Q: What is a pharmacy benefit manager? A: A PBM is a company that administers prescription drug benefits for insurers and employers, negotiating with manufacturers and deciding which drugs a plan covers. The three largest handle about 80 percent of US prescriptions, according to the FTC.

Q: Is insulin still expensive in the United States in 2026? A: It depends on who is paying. Medicare caps insulin at $35 a month, major manufacturers cut list prices by 70 to 78 percent on key products in 2023 and 2024, and some state and manufacturer programs cap out-of-pocket costs. Uninsured patients and those on some plans can still face high costs.

Q: What is a biosimilar insulin? A: It is a biologic that is highly similar to an approved insulin with no clinically meaningful differences in safety or effectiveness. The FDA approved Semglee as the first interchangeable biosimilar insulin in July 2021, and rapid-acting biosimilars have followed.

Q: Did the Inflation Reduction Act cap insulin prices for everyone? A: No. The $35 monthly cap applies to Medicare. A version covering private insurance was removed during the legislative process.

Q: What should someone struggling to afford insulin do? A: This article cannot give personal advice. People who cannot afford insulin should speak with a clinician or pharmacist promptly, since rationing carries real medical risks, and ask about manufacturer programs, state caps, and low-cost options that may apply to their situation.

The takeaway

Insulin shows what happens when a cheap-to-make, life-or-death product moves through a pricing system built around percentages of list price. The molecule did not change, the science did not get harder, and the patent was assigned for $1 a century ago. What changed was who stood between the factory and the patient, and how they were paid.

The recent cuts, caps, and settlements show that the system responds to pressure, and some of that response is arithmetic as much as conscience: removing the Medicaid rebate cap made high list prices costly to maintain. The test of the next few years is whether the rules that reward high list prices change in law and in contract, or whether the price falls only until the next incentive pushes it back up.

---

Related reading

• The Vitamin D Paradox — Why It's So Important and So Misunderstood
• Why Vanilla Is So Expensive — The Biology and Economics of the World's Second-Priciest Spice
• The Science of Exercise — Why It Works and How Much You Actually Need
• Why Wagyu Beef Is So Expensive — The Science of the World's Most Marbled Meat
• Why Aging Is a Disease — And Could We Actually Cure It?