America's China Drug Habit: A $132 Billion Pipeline and a $6 Billion Peptide Boom

On Sunday the FDA rejected ivonescimab, a lung cancer drug invented by China's Akeso and licensed to Summit Therapeutics. Summit said the agency would not approve it without a statistically significant survival benefit. It was one of the most closely watched tests of whether a drug built and first proven in China can win a US approval.

The rejection lands in the middle of two booms. Big pharma is buying Chinese molecules faster than ever to refill pipelines ahead of a patent cliff. And Americans are injecting Chinese-made peptides, some sold legally through compounding pharmacies, many sold online as "research chemicals." We built a model to put numbers on both. Our base case: China-invented drugs could reach about $132 billion a year in peak sales, and the US peptide market is worth about $6.3 billion a year.

What the market had priced in for Summit

The FDA's problem was survival in Western patients. In the global HARMONi trial, the overall survival hazard ratio was 0.98 in the Western subgroup, meaning no benefit at that point, Fierce Pharma reported. Later data cuts improved it to 0.76, but the agency stuck to its requirement.

Investors were braced for it. Leerink had given approval a 40% chance. Summit shares fell 13% in the week before the decision, closing Friday at $14.15, and Jefferies had expected a further 5% to 10% drop on a rejection, according to Investing.com. On Summit's $11.3 billion market cap, that implies investors valued US approval in this narrow use at only about $2.1 billion. Most of Summit's value rests on what comes next: first-line lung cancer data from the HARMONi-3 trial, starting around year-end.

The delay also helps rivals. Merck, Pfizer (with China's 3SBio), Bristol Myers and BioNTech, and AbbVie are all pushing their own PD-1xVEGF drugs into global phase 3 trials. Pfizer's candidate also came from China.

Big pharma's China shopping list

Licensing deals for Chinese drugs reached $139.4 billion in potential value by Oct. 8, already more than all of 2025, according to PharmCube data compiled by APICK. In 2024 the total was $51.9 billion, per Life Science Daily. At this year's pace, 2026 comes to about $181 billion, up a third.

The shift is sharpest at the top. Among drugmakers worth $50 billion or more, Chinese assets made up about half of licensing deals with upfronts of $50 million or more in the first nine months of 2026, and about two-thirds of the upfront dollars. In 2025 those shares were 39% and 36%. The pull is speed and cost: a phase 1 trial takes about seven months in China against 17 or more in the US, and costs 32% to 52% less.

Washington has noticed but left the door open. Treasury's October proposal to restrict US investment in Chinese biotech explicitly exempts licensing deals.

$1 billion-plus China licensing deals signed in 2026, by buyer. Revenue is the latest 12 months; market caps are as of Nov. 13. Source: Life Science Daily deal tracker, Compound analysis.

Two things stand out. AstraZeneca and Bristol Myers have signed China deals with potential value equal to roughly a third of their annual revenue. AstraZeneca's biggest, worth up to $18.5 billion, is for CSPC's long-acting obesity peptides. But the cash paid so far is tiny: every company's upfronts are below 0.7% of its market cap. The exposure is in the pipeline, not on the balance sheet.

What China-invented drugs could be worth

Goldman Sachs says $440 billion of drug sales lose patent protection through 2032. If drugmakers replace 60% of that with drugs they buy from outside, and China supplies half of those, in line with its current share of big-pharma deals, China-invented drugs reach about $132 billion a year in peak sales. At an assumed 12% royalty, about $16 billion a year would flow back to Chinese companies, and Western licensees would keep about $116 billion.

Sensitivity: peak annual sales of China-invented drugs, by share of patent-cliff sales replaced (columns) and China's share of the replacements (rows). Base case in bold.

The headline deal numbers overstate the money changing hands now. Only about 5.8% of potential value is paid upfront, so this year's record works out to roughly $10.5 billion of actual cash for Chinese licensors. The rest depends on milestones that many programs will never reach.

The other China drug trade: peptides

On Oct. 28, HHS Secretary Robert F. Kennedy Jr. said the government would not take action against licensed pharmacies and doctors that compound six unapproved peptides, including BPC-157 and TB-500, if they meet 10 conditions. The policy went against the FDA drug center's scientific recommendation, law firm ArentFox Schiff notes. One stated aim is to pull buyers away from foreign "research use only" peptides.

Hims & Hers shares rose 6% that day, TIKR reported, and the company is now offering BPC-157 and TB-500 to customers. The stock has since fallen below where it started, including a drop after third-quarter results.

How big is the market? Novo Nordisk's CEO said in January that as many as 1.5 million Americans may be using compounded GLP-1 drugs, at around $199 a month. That alone is about $3.6 billion a year. Analysts put the market for other peptides at $2.2 billion to $3.3 billion. Together that gives our base case of about $6.3 billion a year. One physician quoted by the Los Angeles Times put the gray market at $7 billion to $10 billion. Using that figure instead gives an upper bound of about $12 billion.

Most of the raw material comes from China. US imports of hormone and peptide compounds from China doubled to $328 million in the first nine months of 2025, according to customs data cited by GZERO. Against our $6.3 billion market, every $1 of Chinese imports ends up selling for about $14 in the US.

The supply chain has a dark side. Chainalysis found peptide vendors received $32 million in crypto in the first quarter of 2026, up 700%, and projects more than $100 million this year, Fortune reported. In October, another analytics firm found that of more than 150 Chinese chemical makers it tracked accepting crypto, about 75% listed both fentanyl precursors and peptides on their websites.

Who wins and who is exposed

Hims & Hers: Even if it captured 15% of a newly legal compounded-peptide market, which we put at about $0.8 billion a year, that would add about $120 million in revenue, roughly 4% of its current sales. The upside is real but small.

Eli Lilly and Novo Nordisk: If 30% of compounded GLP-1 users would otherwise pay about $500 a month for branded drugs, compounding costs the two companies about $2.7 billion a year. That's about 3% of Lilly's revenue over the latest 12 months.

Big pharma: AstraZeneca, Bristol Myers, Lilly and Pfizer have the most riding on Chinese science. A US rule that limited Chinese clinical data, or a rejection pattern like Summit's, would hit their pipelines years before it showed up in their earnings.

Download the China drugs and peptides TAM model (Excel)Our market-based model of China-invented drug and US peptide market sizes and US drug company exposure. Every input is editable and every sourced figure is cited, with sensitivity tables.Download · Free

What to watch

• Summit on Monday: the first trade since the rejection, then HARMONi-3 squamous data around year-end.
• Treasury's biotech investment rule: comments close Dec. 7, and a final rule is expected by March 13, 2027. Licensing deals are exempt in the draft.
• Peptide oversight: Democrats take control of both chambers' committees in January, and the HHS policy is a likely oversight target. The FDA is funded only through Dec. 11.
• Trump-Xi: the leaders are expected to meet Thursday, with the trade truce expiring in January.

How we estimated this: the deal, import, user and price figures come from the sources linked above. Market data is as of the Nov. 13 close, before the FDA decision. The share of the patent cliff replaced, China's share of replacements, the royalty rate, milestone payout odds, gray-market conversion, Hims' market share and the branded GLP-1 price are our assumptions. Deal values are potential totals, mostly contingent milestones. Gray-market estimates are weakly sourced, so we use the $7 billion to $10 billion figure only as an upper bound.