One Molecule Is Half of Merck. In 2028, the Patent on It Runs Out.

In 2024, one cancer drug brought Merck $29.5 billion. Not a franchise, not a category — one molecule, pembrolizumab, sold under the name Keytruda. That single product grew 18% in a year and accounted for nearly half of the company's $64.2 billion in worldwide sales.¹ For a pharmaceutical company, that is the closest thing to a printing press the law allows. The catch is printed on the press itself: the core patent on the molecule was filed in 2008, and it expires in 2028.⁵ Merck has roughly three years before the law that prints the money starts handing the plates to everyone else.

The comfortable story is that Keytruda is Merck's moat — a deep, defended position in oncology that competitors cannot dislodge. That story is true and it is the problem. A moat that is one molecule wide is not a moat. It is a single bridge, and the company has parked nearly half its revenue on it the year the toll authority is set to expire.

Why the same drug is the asset and the threat

Keytruda's strength is genuinely extraordinary. It is approved across a sprawling set of cancers — lung, melanoma, cervical, head and neck, liver, bladder, the MSI-H tumor types, and more, as both monotherapy and in combination.² Each indication is a separate clinical trial, a separate label, a separate sales channel. That breadth is exactly what a durable franchise should look like: not one bet, but dozens, all riding the same molecule. The trouble is the last three words. Every one of those indications depends on the same compound, and the compound has one composition-of-matter patent, filed in 2008, expiring in 2028.⁵ Diversifying across cancers does not diversify away from the molecule. When the molecule goes generic, all 38-plus indications go generic at once.

This is the anatomy of a concentrated moat. The thing that makes Keytruda a fortress — one mechanism that works across an unprecedented range of tumors — is the same thing that makes it a single point of failure. The breadth is real, but it is breadth of *use*, not breadth of *protection*. And protection is the only kind that survives 2028.

What the 2028 date actually triggers

It is tempting to picture 2028 as a single cliff edge, with revenue intact one day and gone the next. The mechanism is messier and, for Merck, worse in its certainty. The composition-of-matter patent — the one that says nobody else may make this molecule — expires in 2028, and the intravenous version is projected to lose U.S. protection that same year.⁶ Biosimilar developers are not waiting; multiple programs are in development with FDA submissions possible as early as 2026 to 2027, staged to launch the moment the wall comes down.⁶ Merck's own SEC risk disclosures concede the direction of travel: the company expects U.S. Keytruda sales to begin declining in January 2028 as government pricing under the Inflation Reduction Act takes hold, with a further decline upon loss of exclusivity.⁸

Merck has built a wall of patents around the core — by one count, 129 applications, half of them filed after the drug reached market, layering method-of-use, dosing, and formulation claims that extend toward 2036.⁵ That wall slows the erosion; it does not stop it. Those later patents protect specific *uses* and *formulations*, not the molecule. A biosimilar maker who clears the 2028 composition patent can compete on the indications that aren't separately fenced — and on price, immediately. The cliff is real for the core. What follows is a slower bleed, indication by indication, as each protective layer falls in its turn.

The sequel that buys time but not safety

Merck's answer is a lifecycle-management move with a quiet kind of cunning. In September 2025 the FDA approved Keytruda Qlex, a subcutaneous formulation — pembrolizumab paired with berahyaluronidase alfa-pmph — administrable in as little as one minute, across the same solid-tumor indications as the IV drug.³⁴ It is the first and only subcutaneous immune checkpoint inhibitor, and Europe's regulators have signaled approval too.⁴ The strategy is to migrate patients off the one-hour IV infusion and onto a one-minute shot *before* biosimilars arrive — and crucially, Qlex carries its own formulation patents that run past 2028. A would-be competitor now has to match not just the molecule but the molecule-plus-enzyme combination.

The numbers tell you how partial the rescue is. Analyst forecasts have Keytruda peaking around $32.7 billion in 2026, then sliding to just over $7 billion by 2032 once biosimilars are loose.⁷ Qlex is projected to reach $7 billion-plus by 2032 — a real business, and a meaningful offset.⁷ But put the two side by side: a franchise that delivered roughly $30 billion is forecast to land near $14 billion combined. That is not a save. That is cutting the fall in half. Merck's 2025 results already show the seam — combined Keytruda and Qlex sales of $31.7 billion on $65.0 billion in total revenue, with the company restructuring its oncology business as the expiry nears.⁸

Isn't this just what every great drug company faces?

The fair objection is that patent cliffs are routine. Every blockbuster expires; Lipitor did, Humira did, and the companies behind them survived by doing exactly what Merck is doing — extending with new formulations and replenishing with a pipeline. By this reading, Keytruda's 2028 cliff is a known, manageable event, and the alarm is overdone. There's truth in it. Merck is not blindsided; it has years of warning, a subcutaneous bridge in market, and a deep R&D engine that did not, despite some reporting, undergo a strategic retrenchment.²

But the steelman cuts the other way once you weigh the dependence. The danger of a patent cliff scales with how much of the company sits on the one expiring asset, and ~46% of revenue from a single molecule is concentration on a different order than a diversified portfolio losing one of many drugs.¹ The companies that absorbed their cliffs gracefully usually had a successor of comparable size waiting — or a far less concentrated base to begin with. Merck's projected offsets, Qlex included, do not add up to a replacement Keytruda in time.⁷ The risk is not that Merck is unaware. It is that awareness does not manufacture a second $30 billion drug on a regulator's schedule. You cannot pipeline your way out of arithmetic this lopsided in three years.

So which is it — blockbuster or risk? The question contains its own error. Keytruda is not a moat that happens to have a weakness. It is a single magnificent molecule that built a fortress and a fuse out of the same patent, and Merck has staked half a company on the years before the fuse burns down. The subcutaneous sequel is a real and clever way to lengthen the fuse. It is not a way to make the building stop depending on one wall. The most concentrated asset in big pharma is, by that exact concentration, its most exposed — and 2028 is the year the bill for owning only one of anything comes due.