Vol.158 Industry Watch 29 | How Should We View China’s “DeepSeek Moment” in Biopharma?
Summary
- China’s innovative drugs now have hard-dollar backing for a “DeepSeek/Sputnik moment,” but the reversal has not yet extended across the entire industry value chain. In 2024, funding flowing into rights to China’s biopharma R&D pipelines accounted for roughly 30% of the global total, versus an average of less than 5% before the pandemic; by the License-out standard of deals with upfront payments above $50M, China’s large transactions accounted for nearly one-third of the global total, spanning assets from preclinical to clinical stages and from small molecules to ADCs and bispecific antibodies. 沈炯 calls the quality “extremely solid,” but 马睿 cautions that this primarily benefits MNCs in urgent need of pipeline replenishment, and may not benefit Chinese Biotechs and their investors proportionately.
- AK112’s head-to-head win over Keytruda shows that Chinese companies may be moving from following global agendas to defining new strategies. The global PD-1 market is worth about $42.4B, while Keytruda contributes roughly $25B a year to Merck; 康方 licensed AK112’s overseas rights to Summit for a total consideration of $5B, including a $500M upfront payment, and Summit’s valuation briefly reached about $20B after data were released in September 2024. The market moved from questioning whether the buyer could afford the deal to debating whether 康方 had “sold it $100B too cheaply,” exposing the gap between value discovery and commercialization capabilities.
- 康方’s key achievement was not discovering an entirely new target, but upgrading the “PD-1 ship” with a PD-1/VEGF bispecific in a crowded field. 杨洁玲 calls it “a modest differentiated exploration at a controllable cost,” more a “strategic success than a scientific one”: mature targets provide the floor, while an upgraded PD-1 could replace existing monotherapies and combination regimens, creating the upside. Only continued validation in U.S. trials and in combination therapies with Pfizer and others can make the claim that the China-U.S. gap has shifted from “five years behind” to “three years ahead” real.
- The industry may be moving from a downward spiral in which payment, supply and capital dragged one another down into a recovery phase driven by strong supply and new payment channels. Domestic “new drugs” were often not truly new, with as many as 60 PD-1/PD-L1 pipelines once said to be competing on identical ground; price cuts under national reimbursement, setbacks in overseas trials and rights deals, and repeated contractions in primary financing and secondary-market valuations compounded the pressure. Companies that survive the winter are shifting from being “half a step ahead” to “ten steps ahead,” but a new virtuous cycle still depends on clinical value attracting commercial insurance, overseas BD and non-U.S. markets.
- China wins on molecule development, trial-and-error efficiency and clinical resources, while the U.S. still controls the source of targets, pricing and commercialization. Drug prices in China are roughly one-fifteenth to one-twentieth of U.S. prices, while the U.S. innovative-drug market is about 30x China’s; along the China Biotech–China Pharma–U.S. Biotech/NewCo–MNC chain, the most valuable downstream economics may still be captured by MNCs. This cycle therefore cannot be judged by License-out totals alone; upfront payments, downstream economics, co-development rights and whether Chinese companies can build their own MNCs matter just as much.
- Antibodies are currently the clearest vehicle for China’s advantage, but that does not mean AI drug discovery has already delivered. Roughly three-quarters of large outbound assets are antibody-related: their development resembles an industrialized process of “mixing and matching,” with rapid iteration and epitopes that are difficult to lock up through patents, making the modality well suited to China’s engineering dividend; small molecules depend more heavily on chemists’ know-how and “artistic” intuition. AI is more likely to accelerate preclinical work for now; getting a drug into humans still requires clinical trials, and many AI-drug pipelines may remain at an early stage, not yet sufficient to replace existing Chinese Biotechs.
- The investment theme is not a broad industry rally, but the “most aggressive competitors” that can produce strong human data with less capital. If commercial insurance uses Class C formularies to absorb high-priced products with high clinical value, the potential incremental market could reach at least RMB1T; overseas MNCs also face a $100B-$350B patent cliff over the next 5 years. The screen is therefore large disease areas, strong execution, deep industry experience and tangible output, while commercial-insurance implementation, overseas deal pricing and AI conversion rates remain variables that must be verified milestone by milestone.
Deep dive
1. Large License-out Deals Turn the “DeepSeek Moment” from Sentiment into Data
The inflection point highlighted at the start of the episode was this: in 2024, funding flowing globally into rights to China’s biopharma R&D pipelines accounted for roughly 30% of the total, versus an average of less than 5% before the pandemic. China has moved from marginal supplier to a market contributing nearly one-third of transactions.
沈炯 tightens the statistical definition to License-out deals with upfront payments above $50M, because these are transactions where “real money has actually been handed to you.” Large China deals surged in 2023 and 2024, and preclinical assets represented a significant share, meaning the story no longer depends only on mature pipelines nearing launch.
The assets are also sufficiently diversified, spanning small molecules, monoclonal antibodies, ADCs and bispecific antibodies. 沈炯 therefore accepts the description of a “Sputnik moment for Chinese Biotech”: judged by deal size, development stage and technology modality, “the quality should be extremely solid.”
2. AK112’s Win over Keytruda Also Shows China Still Sits Upstream in the Value Chain
杨洁玲 first explains why Keytruda is the benchmark: PD-1 is not merely a drug, but more like “a ship” for solid-tumor treatment, capable of carrying chemotherapy, targeted small molecules and other immunotherapies in combination. The global PD-1 market is worth about $42.4B, while Keytruda contributes roughly $25B a year to Merck; the Chinese market is about RMB12B.
At the end of 2022, 康方 licensed AK112’s overseas rights to Summit for total consideration of $5B, including a $500M upfront payment. After head-to-head results against Keytruda were released in September 2024, Summit’s share price surged and its valuation briefly approached $20B, while 康方 had already handed the buyer the core rights in high-price markets.
马睿 retains the dramatic reversal in public opinion: early in the deal, media outlets questioned whether Summit even had $500M on its balance sheet and dismissed the $5B headline value as “a mirage”; after the data emerged, they took Summit’s roughly RMB140B valuation, subtracted the deal consideration and concluded that 康方 had “sold it $100B too cheaply.”
马睿 sees Summit’s Bob Duggan as a sample of value-chain capability. Over his lifetime, he has made roughly 6 cross-industry investments, specializing in entering companies, helping operate them and then selling them; when he bought into Summit at age 74, he ultimately funded the upfront payment himself. “He was simply that confident he could make money from this,” and he did connect a Chinese molecule to the U.S. commercialization system.
3. The PD-1 Ship Is Being Upgraded by Bispecifics, but the Lead Still Needs U.S. Validation
The period from 2010 to 2014 was the era of immune-checkpoint monotherapy; combinations began to work around 2015, and Keytruda plus chemotherapy produced significant results in a major Phase 3 first-line lung-cancer trial in 2018. Its enormous patient population drove a decisive sales advantage and ultimately created an almost winner-take-all market.
With China’s first wave of PD-1 monotherapies launching domestically in 2019 and dozens of similar pipelines crowding the field, 康方 began developing 2 PD-1-based bispecifics. Its monotherapy was approved in 2021, with sales rights handed to 正大天晴, while resources were heavily tilted toward the bispecifics, which began receiving indication approvals in 2023 and 2024.
杨洁玲 defines AK112 as “a modest differentiated exploration at a controllable cost”: mature targets set the floor, while the PD-1/VEGF bispecific opens up the ceiling and could allow an upgraded PD-1 to replace existing monotherapies and combination regimens. It was “a strategic success rather than a scientific one.” If U.S. trials and combination therapies with Pfizer and others reproduce the domestic results, the China-U.S. gap could shift from 5 years behind to 3 years ahead.
The more aggressive proposition is whether, after 2024, the industry will collectively replace the “PD-1 ship” with upgraded PD-1 versions. 康方 already has an approved product, while the comparable molecules purchased from China by Merck and BioNTech remain in Phase 1, creating a temporary but real time advantage.
4. China Spent 20 Years Turning R&D Services into a Strategic Window
Beginning around 2000, China entered modern drug R&D through CROs such as WuXi AppTec. A large population of biopharma undergraduates and graduate students created a cost and talent dividend; China first provided R&D services to global pharmaceutical companies and then developed biosimilar capabilities around 2010.
As PD-1 gained momentum, Chinese companies began developing antibody drugs, with the first products from Innovent, Hengrui and Junshi receiving approval in 2019. From 2021 to 2023, multiple domestic combination regimens were approved. Whether in monotherapy or combinations, China had previously lagged overseas markets by roughly 5 years.
马睿 summarizes today’s advantage as trial-and-error efficiency and clinical resources: smaller Biotechs can “generate molecules without internal friction—very quickly and very cheaply.” Hengrui has roughly 147 pipelines, most of which could rank within the global top 10; its approach is not to place a single bet on one target, but to use its cost advantage to “do them all” and then find ways to transact them externally.
5. License-out Is Shifting from Executing Global Agendas to Influencing Them
Early Chinese outbound assets were plentiful, but the source of innovation was often still overseas. The core ADC paradigm is that “the antibody acts like a missile that finds the tumor, while the toxin on the other end does the killing.” After Daiichi Sankyo’s HER2 ADC DS-8201 showed markedly better clinical results than prior targeted drugs from 2019 to 2022, the world began chasing the modality in earnest.
Two representative 2023 deals were Merck’s acquisition of certain ADC assets from Kelun-Biotech for about $8B and BMS’s acquisition of comparable ADC molecules from Baili Tianheng for roughly $8B-$9B. The industry joked that overseas pharmaceutical companies had come to China to “wholesale ADC drugs in bulk.”
siRNA followed a similar path. Alnylam began exploring drug development in 2000, announced Phase 3 results in a 4,000-patient hypercholesterolemia trial with Novartis around 2020, and released Phase 1 hypertension data at the end of 2023 before entering a $2.8B partnership with Roche; afterward, several comparable molecules from 博望生物 were licensed to Novartis for about $4B.
AK112 is different because the strategic source came from China. Before 康方 released its results, overseas markets had not fully recognized the direction; in August 2024, a small overseas Biotech purchased a comparable molecule from 盈迈昂科, and after the September data, BioNTech and Merck separately bought PD-1/VEGF bispecifics from 普米斯 and 礼新药业. Chinese companies may be starting to influence what buyers choose to pursue.
6. Strong Supply Is Beginning to Break the Downward Spiral of Payment, Supply and Capital
杨洁玲 attributes the past 5 years of winter to the mutual reinforcement of 3 forces: reimbursement negotiations and other measures contracted the payment side; the widely circulated figure of roughly 60 PD-1/PD-L1 pipelines exposed that “China’s new drugs were not actually new”; and the return of some BeiGene rights, Innovent’s loss against Keytruda in a head-to-head trial, and overseas setbacks for Junshi and Hutchmed weakened external recognition.
The result was that “payment did not work, supply did not work, and investment became very difficult”: primary-market financing kept shrinking, while secondary-market assets such as the Hang Seng Innovative Drugs Index continued to fall. Insufficient capital further constrained innovation, forming a complete downward spiral.
Yet the capital accumulated from 2010 to 2020, the engineering dividend and demand from overseas buyers left China with strong supply. Once the contraction in financing eliminated homogenous companies, a small number survived on technology; companies that were “half a step ahead of peers” before could become “10 steps ahead of others” once their peers lost financing.
Outbound activity also evolved from the 2 BeiGene and Legend exemplars in 2022, to bulk ADC transactions in 2023, and then to direct purchases of niche assets, large assets setting the trend and platform-based “designated R&D” in 2024. In some deals, higher upfront payments or improved downstream participation may signal a new balance between asset quality and bargaining power.
7. China Wins on Trial-and-Error Efficiency, while the U.S. Retains Commercialization and Pricing Power
马睿 divides the value chain into 3 parts: biology and target selection are primarily U.S. strengths; molecule development, generation, design, preclinical work and clinical trials are China’s strengths; commercialization is “completely controlled by the U.S.” 康方’s scientific and execution success did not automatically translate into control over end-market value.
U.S. drug prices are roughly 15x-20x those in China, and the innovative-drug market is about 30x China’s; the overall biologics market is about 3x-4x larger. As long as this gap persists, even successful Chinese supply may still see the highest-value economics captured by U.S. channels and MNCs.
马睿 therefore says that the so-called DeepSeek moment “benefits MNCs, and does not necessarily benefit investors or Biotechs”: early investors rarely manage to select the right pipeline in advance, while MNCs can wait until the end of the China Biotech–China Pharma–U.S. Biotech or NewCo–MNC chain and pick up assets that have already been validated.
Hengrui once sold an asset to an intermediary for $25M, after which the intermediary resold it for about $1B. The example illustrates the importance of bargaining power: securing a good price requires the company to recognize the asset’s value, data to persuade the buyer, and negotiations to extract a higher upfront payment and stronger downstream rights.
8. Commercial Insurance and Class C Formularies Could Rebuild Domestic Innovative-Drug Payment
杨洁玲 divides healthcare payment into public insurance and government, out-of-pocket spending and commercial health insurance. Public-insurance funding is expected to remain relatively stable; under a simple assumption that aging and basic consumption gradually approach their ceilings, personal spending will shift toward health, leisure and education over the long term, with a “high degree of certainty” that more will flow toward healthcare.
In 2023, roughly RMB900B flowed into commercial insurance, of which about RMB380B actually went to healthcare, implying an allocation efficiency of only slightly above 40%. Direct out-of-pocket healthcare spending by individuals was about RMB2T, with the 2 categories together representing roughly 6%-8% of total personal spending. If part of the out-of-pocket spending moved into commercial insurance and claims efficiency improved, the potential incremental pool would be at least RMB1T.
The logic of a Class C formulary is not simply to raise drug prices, but to bring products and services that are “expensive, difficult for individuals to pay for, and effective at the same time” into commercial-insurance coverage or related arrangements, using insurance leverage to improve access. 杨洁玲 reduces the central question to 1 line: “Can your product help health insurers attract customers?”
The policy directions mentioned by 马睿 also include opening up reimbursement data, supporting renewal and scope adjustments through real-world evidence, connecting public and commercial insurance systems, and promoting innovative drugs in Belt and Road markets such as ASEAN and the Middle East. But he explicitly retains uncertainty: “It is uncertain whether this draft is actually real.”
9. The Patent Cliff Creates 5 Years of External Buying Demand for Chinese Pipelines
European and U.S. MNCs face estimates of a $100B-$350B patent cliff over the next 5 years. The top dozen or so pharmaceutical companies hold roughly $60B in cash reserves, while 马睿 describes their leveraged purchasing capacity as having “$1T in ammunition.”
The markets that can continue supplying new-drug pipelines remain concentrated in the U.S., Europe, China, South Korea and Japan. China already has a group of 18A- and A-share-listed innovative-drug companies, a large upstream system of instruments, reagents and CROs, and companies that retain gross-margin and pricing advantages; the cost of new-drug development remains below that of most regions.
Buyers’ payment logic is straightforward: they need assets capable of replacing cash flow from expiring patents and meeting European and U.S. clinical and commercial requirements. China has abundant pipelines, rapid development and low costs, making it an important source of late-stage pipeline replenishment for MNCs; but external demand will not provide a floor for assets without clinical value.
10. Antibodies Fit China’s Industrialized Strengths, while Small Molecules Still Depend on Chemical Inspiration
沈炯 estimates that, by value, small molecules account for only about 21% of large outbound deals and, by number, about 25%; the remaining roughly three-quarters are monoclonal antibodies, bispecifics and ADCs. With almost every drug in the 2024 global Top 10 by sales other than apixaban being an antibody, buyers’ willingness to pay premium prices is understandable.
The supply side has not collectively abandoned small molecules: the number of small molecules approved by the FDA in 2023 was roughly double the number of biologics, and even more in 2024; Chinese companies’ FDA submissions were approximately split between small molecules and antibodies, while NMPA data showed even more small molecules. Transaction preferences therefore cannot be attributed simply to R&D volumes.
沈洁? Wait. 沈炯 explains 康方’s win as “strong execution—it moved the fastest,” rather than PD-1/VEGF being an idea nobody had considered. Antibody development has a clear playbook, with structures and targets that can be combined and recombined at scale; put simply, “enough force can produce a miracle,” which fits China’s speed, cost structure and organizational capabilities.
More importantly, the binding epitope between an antigen and antibody “can even basically be considered impossible to protect through patents in the true sense.” U.S. originators therefore have weaker first-mover barriers; small molecules, by contrast, are difficult to design around patents and depend on chemists’ know-how and intuition, with organic synthesis described as “more of an art.”
11. This Is Not Yet an AI-Drug-Discovery Moment, but AI Could Rewrite the R&D Paradigm
Asked whether 康方 was “not really that AI-driven,” 马睿 acknowledged that this breakthrough was not fully an AI drug-discovery moment. If AI had already deeply penetrated biologics, U.S. MNCs might use domestic AI companies directly rather than purchase Chinese Biotech assets; reality shows that China’s clinical development and execution remain faster and cheaper.
For now, AI is more likely to shorten preclinical molecule generation and lower costs, but the clinical process after a drug “enters humans” cannot be skipped, and value growth remains limited when a program reaches Phase 1. Many AI-drug pipelines may still be at an early stage, and there are few teams that genuinely integrate AI with drug development; the sector has not broadly reached the point of being acquired by MNCs.
Over the long term, AI could improve success rates and molecule quality, generating molecules that traditional methods cannot produce. 马睿 moves from AlphaFold 2 and RNA-molecule design to the shift from training scaling to inference scaling; fields with better data, including IVD, imaging and pathology, are “staking out a position first,” while commercial applications remain to be observed.
杨洁玲 pushes the issue toward the R&D paradigm itself: the current process breaks disease down into molecules, cells and animals, then validates each layer, but cannot eliminate the intrinsic differences between cells, animals and humans. Only if AI-friendly data and algorithms emerge that directly observe human disease could the industry escape the old paradigm rather than merely optimize one link within it.
12. Establishing a Bottom Does Not Mean a Broad Rally; Investment Will Reward Only Strong Supply
The team defines the current stage as “a bottom being established and confidence returning across the industry,” based on a floor forming on the payment side, the gradual emergence of new payment channels and stronger supply capabilities. But if confidence becomes overheated, the outcome will still be that “only strong supply survives,” because commercial insurers and MNCs pay only for clinical and commercial value.
马睿’s investment screen is highly specific: “You must invest in innovation, in large disease areas and in the most aggressive competitors.” The ideal target is a team with strong execution and deep industry experience that can generate human antibody data with relatively little capital; a platform must “actually produce things,” rather than exist only in its creators’ imagination.
To capture the maximum value of innovative drugs, China still needs 3 forms of self-liberation: build its own MNCs; increase healthcare-service consumption and the commercial-insurance share while raising rather than continuing to suppress innovative-drug prices; and expand into non-U.S. markets. If AI penetrates drug development further, it could push cost efficiency beyond that of traditional pharmaceutical companies.
The next phase is not predetermined. Investors must watch how commercial insurance cooperates with public insurance, which products enter Class C formularies and how they are priced, as well as what development stages overseas buyers purchase and which transaction and payment structures they use. Only if these checkpoints continue to validate the thesis can “new, strong and correct supply” emerge at scale.