Newsletter · · Ashutosh Agarwal

China Becomes Big Pharma's Pipeline as Keytruda Faces a DeepSeek Moment - Biotech M&A and Drug Pricing - Week of September 18, 2026

Biotech M&A and Drug Pricing for the week of September 18, 2026: nearly half of global drug-licensing deals now involve a Chinese company, a China-tested rival gave Merck's Keytruda its DeepSeek moment, the XBI is up about 70% as money rotates out of AI, a Morgan Stanley banker says a large deal is more likely than in years, and Blackstone is funding biotechs to stay independent.

Biotech M&A and Drug Pricing

Week of September 18, 2026: China Becomes Big Pharma's Pipeline as Keytruda Faces a DeepSeek Moment

Here is the uncomfortable idea that ran through this week's podcasts: the drug industry's answer to its patent cliff is increasingly being invented somewhere else. Nearly half of the world's drug-licensing deals last year involved a Chinese company, up from one in six just four years ago, and the sums are getting enormous. Bristol-Myers put as much as $15.2 billion behind one Chinese partner in May; Pfizer followed weeks later with a deal worth up to $10.5 billion.

At the same time, the best-selling medicine on Earth is being quietly menaced by exactly this trend. And in a twist nobody was talking about a year ago, biotech stocks are having their best run in years, up 70%, partly because investors are fleeing expensive artificial-intelligence names and parking money in a sector that was left for dead.

A quick note on who's talking, because it matters. This week's most authoritative voices were an actual dealmaker (Blackstone's life-sciences chief), a Wall Street banker who runs healthcare M&A, and a sell-side biotech analyst, all clearly labeled below. The color commentary came from journalists at Bloomberg and The Wall Street Journal.

TL;DR

  • China is now the pipeline. Roughly half of global drug-licensing deals last year involved a Chinese company (up from 16% in 2022), and 2026 is on pace to break last year's record of 45 U.S.–China deals. The prices are real: $15.2 billion (Bristol-Myers) and $10.5 billion (Pfizer), both this year.
  • Keytruda has a "DeepSeek moment." Merck's ~$31-billion-a-year cancer drug, which loses U.S. patent protection in 2028, is being challenged by a China-tested rival built on Keytruda's own recipe with a new target, and Merck is already fighting back with a defensive drug combination.
  • The beaten-down sector is ripping. The big biotech index (XBI) is up ~70% over the past year versus ~20% for the S&P 500, helped by money rotating out of pricey AI stocks. The catch: biotech is speculative, not a safe haven.

What's New

1) China has quietly become Big Pharma's research-and-development department.

The single most important podcast of the week was Bloomberg's Big Take Asia, "Big Pharma Needs New Drugs. China Has Them," with reporters Amber Tong and Robert Langreth (journalists, so read this as reporting, not a stock call). The numbers they laid out reframe the whole patent-cliff story.

"Last year, nearly half of global pharmaceutical licensing deals involved Chinese companies. That's up from just 16% in 2022."

A licensing deal is simpler than a takeover: a big drugmaker pays for the rights to develop and sell a medicine that a smaller company discovered, and the seller hands over the formula and the data. The appeal, the reporters explained, is that "maybe only a small percentage of that money is paid up front, so it's much cheaper": you pay a little now and a lot later only if the drug works. The pace has gone vertical: "U.S. and Chinese drug makers have struck at least 32 licensing deals so far this year through July... on track to surpass last year's record of 45," per Bloomberg Intelligence. And the checks are huge: Bristol-Myers's collaboration with Hengrui Pharmaceuticals worth up to $15.2 billion in May, and Pfizer's $10.5 billion cancer-drug deal with a different Chinese company weeks later, in which the Chinese partner even runs the early trials.

Why China? Speed. "China can move drugs through early development faster because it has huge patient pools and large research hospitals," plus regulatory reforms that now mirror the U.S. FDA's rules. One company they cited, Sichuan Kelun Biotech (whose experimental cancer drug Merck is now testing), sits inside a group whose parent "just a decade ago... was best known for manufacturing IV fluids."

The worry, voiced by both the reporters and unnamed lawmakers: the U.S. is training its own future competitor. As Langreth put it, the fear is that "there'll be Chinese drug giants... making most of the money from these things in the future and not doing the deals, because we've trained them." Over the summer, members of both the Senate and House introduced bills to subject China biotech investments to the same scrutiny as semiconductors and AI.

2) The best-selling drug in the world just got its "DeepSeek moment," and Merck is already counter-punching.

This is the patent cliff made concrete. On WSJ's Take On the Week, "A Breakthrough Cancer Vaccine Fueled a Biotech Rally. Can It Last?", Wall Street Journal Heard on the Street columnist David Wehner (a pundit) explained that Merck's Keytruda "sells over $30 billion a year and just works on a lot of different cancers." Then came the scare:

"We saw data that came out from a company that partnered with a Chinese biotech that basically showed better improvement to cancer patients using the formulation that Keytruda had, but locking it in with a new target. And suddenly the market was like, oh boy, we have a problem."

Wehner literally called it a "DeepSeek moment," a nod to the Chinese AI model that rattled the tech market. Merck's stock fell, though "not as much, because they're a big pharma company." The debate isn't settled: the key survival data (whether patients actually live longer) is still pending, and the trials were run in China, so there's argument about whether the result repeats in U.S. and international patients.

The sell-side view added the piece Wehner didn't: Merck's defense plan. On BiotechTV - News, H.C. Wainwright senior biotech analyst Mitchell Kapoor (sell-side) was discussing Summit Therapeutics' rival drug, partnered with China's Akeso, which posted strong survival data. His take: "I think it's better than Keytruda." But he immediately pointed to Merck's response: "something we're watching is Merck's defense strategy that has emerged recently: sac-TMT... just added on to Keytruda." In plain terms, Merck is bolting a second cancer drug onto Keytruda to protect the franchise. Kapoor said the China data was "so closely comparable... it's kind of scary."

Both loss-of-exclusivity threads point at 2028, when Keytruda's U.S. patent protection ends. For a company that booked over $31 billion from that one drug last year, that is the number that keeps Merck shopping.

3) The left-for-dead sector is up 70%, and AI is part of the reason.

Here's the surprise. Wehner noted that XBI, the big biotech exchange-traded fund, "is up about 70% over the past year versus about 20% for the S&P 500." Part of that is simple recovery: the index has only just "caught up with its high from '22" after a brutal multi-year slump. But part is a rotation nobody predicted:

"The AI trade is becoming more and more linked to the healthcare trade. As the market gets more concerned about AI valuations... they're looking for alternatives."

In other words, some of the money getting nervous about sky-high AI stocks is landing in healthcare because it "just hadn't run up" and looked cheap. Wehner was careful to draw a line, though: that defensive logic applies to big, steady pharma names like Merck and Pfizer. Small and mid-sized biotech is still "a speculative investment," with "huge jumps, huge drops." His advice for anyone who isn't a specialist was refreshingly humble: spread your bets or just buy the index, because even seasoned biotech hedge funds "spread their bets."

4) A Wall Street banker says a genuinely big deal is finally back on the table.

For two years the M&A story has been all bolt-ons and no blockbusters. That may be shifting. On Squawk on the Street (the September 15 "10AM Hour" from the Morgan Stanley Health Care Conference), Morgan Stanley's co-head of M&A (a banker, so talking his own book) said healthcare deal volumes were pacing well past last year's, and then went further than bankers usually do:

"I wouldn't say for sure that a large deal will happen. But I think there's much more likelihood than there would have been three or four years ago."

His reasoning was the patent cliff by another name: "certain of the pharmas have some very real growth challenges that... a lot of times they have sought to address by putting two companies together, running more efficiently, and kind of resetting the base." He described dinner "with about 15 or 20 of the leading CEOs in health care" where M&A was "very much top of mind," and summed up the mood with a gardening metaphor: the industry has gone "from green shoots to actually seeing the plants now."

On which giant is doing the buying, the sell-side was blunt: Eli Lilly, flush with GLP-1 (weight-loss drug) cash. Kapoor at H.C. Wainwright ran through Lilly's recent shopping, the "multibillion-dollar" Colonia acquisition, plus ORNA and a company called Sail, layered on top, all to build out "in vivo CAR-T," a next-generation cancer and immune therapy. "Lilly with GLP-1 cash to spend on everything... they've already had 2 deals, so maybe they're going back for more." He also flagged AbbVie's purchase of Capstan in the same race.

5) The counter-move: some biotechs are being funded to not sell.

The most interesting operator voice this week argued the opposite of the M&A boom. On The BioCentury Show, "Ep. 119 - Blackstone's Galakatos on funding biotech's hardest mile," Nick Galakatos, head of Blackstone Life Sciences (an actual dealmaker, so treat this as insider mechanics), described a business built around helping promising biotechs stay independent instead of being swallowed.

The model is "scale capital": very large, mostly non-dilutive financing (royalties, loans, milestones rather than new shares) that lets a company fund expensive late-stage trials without selling itself. His marquee example was Alnylam: a $2 billion package that was only $100 million in dilutive equity, the rest a ~$1 billion royalty on its cholesterol drug Leqvio (partnered with Novartis), up to $150 million for two other products, and a loan of up to $750 million. The point, Galakatos said, is to give a company "a stronger negotiating hand with a pharmaceutical company that is interested in acquiring them," or to skip the sale entirely. He named Apogee as a current example, which needed "$1 and $1.5 billion" to launch its eczema drug on its own.

Why does this matter for the patent cliff? Because every biotech that funds its own path to profitability is one fewer easy target for a cash-rich acquirer. Galakatos noted how rare that path is: very few independent biotechs ever get "north of the $20 billion market cap range" (Amgen, Gilead and Vertex being the exceptions that made it).

The Debate

The long-running argument is M&A supercycle (the buying has to continue) versus the risks: overpaying, integration failures, and regulators blocking deals. This week the podcasts leaned hard toward the bull case, but the most thoughtful pushback came from an unusual angle: whether M&A creates value at all.

The bull case was everywhere. Half a trillion dollars of sales rolling off patent; Keytruda's 2028 cliff made vivid; a banker saying a big deal is more likely than it's been in years; Lilly on a buying spree. As one Biotech Hangout analyst put it, we're having "one of the best M&A years ever in biotech," driven by "borderline unprecedented LOEs [losses of exclusivity] across large pharma over the next five, ten years." The pressure to replace revenue is not going away.

The sharpest bear argument wasn't about antitrust or price; it was about the math of dealmaking itself. On Biotech Hangout (Episode 195, a panel of sell-side analysts and investors), one panelist, a former biotech finance chief, laid out the uncomfortable arithmetic. Buy a drug early and cheap, before it's proven, and "maybe pay $2 billion, $3 billion, $4 billion" while accepting roughly "35, 40%" odds of success. Buy it late and de-risked and "you need to be... a 90%, 92%... free throw shooter," at a much higher price. His verdict was stark:

"M&A usually is value defeating... We do believe you should go early. Do six deals and not one. Hope for two of them to work and then they're going to work in an outsized way."

And he punctured a comforting myth about targets: buyers "sometimes ultimately buy who you're able to buy," because "we always think that these companies have 100 levers they can pull. That's not the way it works at all." The best assets often aren't for sale.

With so much of the action now in licensing deals and mid-sized bolt-ons rather than mega-mergers, the antitrust and FTC leg of the bear case was not where the smart voices focused.

The Names in Play

  • Merck (MRK): the week's central name. Keytruda (~$31B in sales, 2028 U.S. patent cliff) faces a China-tested challenger built on its own formula; Merck is defending with a sac-TMT-plus-Keytruda combination. Still the most motivated buyer at the table.
  • Eli Lilly (LLY): the serial acquirer, using GLP-1 cash to stack up in-vivo CAR-T deals (Colonia, ORNA, Sail). Sell-side expects it to keep going.
  • Bristol-Myers (BMY) and Pfizer (PFE): the two headline China licensers this year, at up to $15.2B and $10.5B respectively.
  • Novartis: the activism story (see below); still the cautionary tale of buying unproven science.
  • Summit Therapeutics: its Akeso-partnered drug posted China survival data an H.C. Wainwright analyst called "better than Keytruda," which is precisely what has Merck playing defense.
  • Alnylam and Apogee: the poster children for the "stay independent" alternative to being acquired, via Blackstone's scale-capital financing.
  • AbbVie (ABBV): bought Capstan in the same in-vivo CAR-T race.

Read-Throughs

  • China is the dominant read-through, but as a licensing partner, not a takeover target. The deal flow (32 deals through July, on pace to beat 45) is a tailwind for U.S. pharma pipelines and a headwind for the long-term U.S. innovation lead. It's also a live political risk, with scrutiny bills now in Congress.
  • A shareholder revolt at Novartis shows the market policing bad deals. On Biotech Hangout, the panel detailed an Artisan Partners letter to Novartis "raising concerns about board oversight of acquisitions," pushing for an acquisition committee and a compensation overhaul, and "criticizing the chair," declaring the dealmaking "party... being over." The trigger was Novartis's roughly $12 billion Avidity Biosciences purchase (far above its usual "$5 billion and below" bolt-ons), after which the Street "removed about $5 billion in sales from the model" and shaved its long-term earnings-growth estimate "by about 200 basis points to 4%." A reminder that investors will punish a cliff-driven deal that misfires.
  • The "who stays independent" question is a new lens on the cliff. Blackstone's alt-capital model (above) means the pool of easy targets may be smaller than the bulls assume: some of the best assets are being financed to remain sellers-of-last-resort, not first choice.
  • Bankers plainly benefit. Morgan Stanley's own M&A chief described a humming pipeline, though that is a description of a busy desk rather than a stock idea.

What Changed

Last week was about who is doing the buying: the discovery that big pharma had shrunk to just 45% of deal spending (from 69%) as mid-caps crowded in, set against Novartis's $40 billion wipeout on three trial failures and a hard September 29 tariff deadline.

Three things moved this week.

The Novartis story escalated from an accident to a governance fight. Last week it was a bad run of data; this week shareholders drew blood, with Artisan Partners publicly challenging the board's dealmaking and pay. The cost of getting a cliff-driven acquisition wrong is no longer just the stock price; it's the chair's job.

The China thread got hard numbers and a specific victim. Recent weeks flagged China as a rising source of licensed molecules in the abstract. This week put real money on it ($15.2B Bristol-Myers, $10.5B Pfizer; 32 deals and counting) and a concrete casualty: a direct, China-tested threat to Keytruda, the single biggest franchise on the cliff.

And the mood on the sector flipped from defensive to euphoric. The pricing-policy dread that dominated the late-summer podcasts (tariffs, "most favored nation," Medicare pricing programs) gave way to a 70% XBI rally fueled partly by investors fleeing expensive AI stocks. That's a healthier backdrop for deals, but as more than one voice warned, a hot tape is exactly when overpaying becomes easy.