Keytruda's 2028 Patent Cliff Is Reshaping Where New Drugs Get Made

Merck's Keytruda, the best-selling cancer drug in the world, pulled in $31.7 billion in a single year. The same drug loses US patent protection in December 2028, the day cheap copies can enter the market and erase one of the largest single revenue streams in the pharmaceutical industry. With more than $300 billion in annual prescription drug sales at risk of losing patent protection before 2030, the industry is running out of time to replace its own blockbusters with its own laboratories. A blockbuster is industry shorthand for a drug that clears about $1 billion a year in sales.

That math is the engine of a geographic shift in pharmaceutical invention. Bringing one new drug to market costs about $2.2 billion and roughly eight years from first human trial to approval, a timeline Merck, Bristol Myers Squibb, and AbbVie cannot shorten by spending more. The replacement molecule has to be in a human trial today to plausibly backfill a 2028 cliff, and most of Big Pharma's late-stage pipeline is not that far along. The industry is responding by licensing the next generation of blockbusters from labs it does not own, and most of those labs are now in China.

In late 2022, the Chinese biotech Akeso out-licensed ivonescimab, a next-generation cancer drug that combines a PD-1 checkpoint blocker (the same class as Keytruda) with a VEGF inhibitor that starves tumors of blood supply, to Summit Therapeutics for the United States, Canada, Europe, and Japan. The deal carried a $500 million upfront payment, up to $5 billion in milestones, and low-double-digit royalties on net sales. Summit's largest shareholder installed Akeso's chief executive, Michelle Xia, on its board.

GSK struck a similar arrangement with Hengrui Pharma, one of China's largest drugmakers, agreeing to pay $500 million upfront for a slate of up to 12 experimental medicines in respiratory disease, immunology, and oncology. The agreement includes a licensed program for chronic obstructive pulmonary disease (COPD), a progressive lung condition, and gives GSK options on 11 additional Hengrui programs once each reaches Phase I trials in humans.

Goldman Sachs Research reported that 46 percent of new drug molecules entering human trials in the first half of the year it measured originated at Chinese biopharma companies, and that Chinese-originated assets accounted for roughly half of global licensing deal value in the same window. The supply did not appear overnight. Over the past decade, China built out contract-research labs, biologic manufacturing capacity, and a domestic insurance and pricing system that pays for innovative drugs, all of which let Chinese biotechs run clinical trials faster and cheaper than their US counterparts. The result is a deeper bench of late-preclinical and early-clinical compounds at exactly the stage where Western pipelines are thinnest.

The deals give Big Pharma optionality on the molecules it cannot build fast enough internally. They do not solve the underlying revenue problem. When a $30 billion-a-year drug loses patent protection, even a slate of licensed blockbusters has to clear trials, win approval, and ramp sales before the original cash flow disappears. Merck's own annual filing acknowledges the December 2028 expiration of Keytruda's core patent and the scale of the revenue at risk. The structural read: pharma R&D geography is reorganizing under market pressure, not as a strategic contest, and the test arrives in 2028 when the calendar catches up to the math. A successful internal-replacement launch from Merck, Bristol Myers Squibb, or AbbVie before December 2028 would undercut the licensing-necessity claim. Until then, the deals keep coming.