Ask anyone who negotiates drug licences for a living what changed in the last two years, and you will usually get the same answer, delivered with a half-embarrassed laugh: we used to fly to Asia to book manufacturing capacity. Now we fly there to buy the pipeline. For two decades, every claim that Asia was becoming an originator of medicines rather than a manufacturer of them ran into one stubborn objection: show me the money. Not the manufacturing contracts, not the clinical trial enrolment numbers, not the patent filings. The money. The upfront cheques that global pharmaceutical companies write when they genuinely believe a molecule invented somewhere else is worth owning.
Over the trailing 24 months, that ledger has filled up faster than anyone in this industry predicted, including the people signing the cheques. Western pharma is no longer partnering with Asia in the polite, capacity-buying sense that the word "partnership" used to carry in this region. It is buying its next portfolio from Asia. The evidence is not a mood or a conference theme. It is a set of numbered entries with dollar figures attached, and this story is an audit of those entries: what was licensed, what was actually paid on signature versus promised on contingency, what the buyers were purchasing at the level of modality, and, just as revealing, what the sellers refused to sell.
One rule governs everything that follows. Deal terms in this business are often only partially disclosed. This publication reports what the parties have disclosed, marks the rest as undisclosed, and never estimates an upfront. The ledger is powerful precisely because it does not need embellishment.
Every audit starts with an opening position, and the opening position here is worth stating plainly because it explains why the last 24 months feel like a rupture rather than a trend line.
In 2021, cross-border out-licensing from Greater China biopharma companies totalled roughly 13.9 billion US dollars, according to data provider PharmCube. That figure was respectable and largely ignored. Asia-origin licensing was a curiosity: a Duality Biologics here, an early antibody-drug conjugate there, deals that Western business development teams treated as inexpensive optionality rather than core pipeline strategy. The exception that proved the rule sat in Japan, where Daiichi Sankyo's DXd antibody-drug conjugate platform had already extracted a 1.35 billion dollar upfront from AstraZeneca in 2019 for what became Enhertu, and a 4 billion dollar upfront from Merck in late 2023 for three more ADC programmes. Those deals were treated as a Daiichi story, not an Asia story. That reading turned out to be wrong. They were the precedent.
The ignition event for the current cycle, if one must be chosen, arrived in September 2024, when Akeso's ivonescimab, a PD-1 and VEGF bispecific antibody discovered and developed in China, beat pembrolizumab head to head on progression-free survival in a Chinese lung cancer trial. Pembrolizumab is the best-selling drug in the history of the pharmaceutical industry. A molecule from Zhongshan had just outperformed it in a randomised study. Summit Therapeutics had licensed ivonescimab's Western rights in December 2022 for 500 million dollars upfront, a figure that looked aggressive at signing and looked like the bargain of the decade 21 months later. Every global business development head watched that readout and reached the same conclusion at the same time: the discount on Asian science was a mispricing, and mispricings do not survive being noticed.
What followed, across the 24 months from October 2024 to September 2026, is the densest run of Asia-origin out-licensing the industry has ever recorded.
The first landmark entry came in November 2024, when Merck licensed LM-299, a PD-1 and VEGF bispecific from LaNova Medicines, for 588 million dollars upfront and up to 2.7 billion dollars in milestones. The upfront mattered more than the total. Merck was not buying an option. It was paying more than half a billion dollars, on signature, for a preclinical-adjacent Chinese asset in the exact class that had just embarrassed its flagship product. A month later, Merck returned for Hansoh Pharma's oral GLP-1 candidate at 112 million dollars upfront and up to 1.9 billion dollars in milestones, an entry that signalled the buying would not be confined to oncology.
Then the entries stopped being landmarks and became a pattern. In May 2025, Pfizer paid 1.25 billion dollars upfront, at the time the largest upfront ever paid for a Chinese-origin asset, plus up to 4.8 billion dollars in milestones and a 100 million dollar equity investment, for 3SBio's PD-1 and VEGF bispecific SSGJ-707. In June 2025, Bristol Myers Squibb paid BioNTech 1.5 billion dollars upfront for shared rights to BNT327, another bispecific in the same class. The footnote on that entry belongs in the Asia ledger even though the counterparty was German: BNT327 originated at Biotheus in Suzhou, which BioNTech had acquired months earlier. The molecule's economics changed hands twice inside a year, and both times the underlying value was Chinese chemistry.
July 2025 introduced a new deal architecture. GSK and Hengrui Pharma signed a portfolio agreement covering a lead COPD asset and options across up to eleven further programmes. The structure carried 500 million dollars upfront against roughly 12 billion dollars in potential milestones, and it represented the largest China-to-West multi-asset licensing deal executed to that point: pharma licensing an entire pipeline segment rather than a single molecule. The same month, the ledger recorded its most significant Indian entry, examined later in this piece, when AbbVie paid Glenmark's innovation arm 700 million dollars upfront for a trispecific antibody.
By year-end, the annual totals had lost contact with every prior benchmark. Cross-border licensing deals between Greater China-based companies and global drugmakers reached 137.7 billion dollars in 2025 by PharmCube's count, a near tenfold increase on 2021, spread across 186 out-licensing transactions. A second data house, VBData, counted the year differently, at roughly 135.7 billion dollars across 157 deals. The two figures disagree at the margin because the counters apply different inclusion rules, a point this audit returns to in its footnotes, but they agree on the only thing that matters: the order of magnitude changed.
Then 2026 opened, and the ledger accelerated again. On 30 January 2026, AstraZeneca and CSPC Pharmaceutical Group announced a strategic collaboration built on CSPC's long-acting peptide platform, valued at up to 18.5 billion dollars, with a 1.2 billion dollar upfront that set a new record for the largest upfront in a Chinese out-licensing deal. In May, Hengrui and Bristol Myers Squibb reached a 15.2 billion dollar multi-field collaboration. Between those bookends, the mid-size entries kept landing at a cadence that would each have been front-page news three years ago. Innovent licensed a preclinical oncology and immunology asset to Eli Lilly in February for 350 million dollars upfront against a total value of 8.85 billion dollars. In May, Innovent signed again, this time with Pfizer, taking 650 million dollars upfront against a 10.5 billion dollar total for twelve early-stage oncology projects. Madrigal Pharmaceuticals licensed experimental liver-disease programmes from Suzhou Ribo for 60 million dollars upfront and up to 4.4 billion dollars in potential payments. Sino Biopharmaceutical signed with Sanofi in a deal worth up to 1.53 billion dollars, and Antengene with UCB at approximately 1.18 billion dollars, contributing to a China-origin out-licensing total that reached 52 billion dollars from 41 transactions before February had even closed.
By mid-year, VBData put the total value of China's innovative drug business development deals at 106.3 billion dollars for the first half of 2026 alone, with Chinese companies occupying eight of the top ten global out-licensing deals by value. Read that sentence again. Eight of ten. The global league table of pharmaceutical asset sales is now, by value, an Asian league table with two guests.
A ledger rewards the reader who knows which column to trust, and in pharmaceutical licensing the honest column is the upfront. Milestone totals are marketing. They are real contractual obligations, but they are contingent on futures that mostly do not arrive, and the industry's own structuring habits admit as much. Across the recent deal record, upfront payments have averaged roughly seven percent of headline deal value, with milestone-heavy structures dominating. An 18.5 billion dollar headline is a press release. A 1.2 billion dollar wire transfer is a belief.
So what has happened to the belief column? It has moved faster than the headline column. The average upfront fee on China out-licensing deals reached 77.7 million dollars in early 2026, double the 38.8 million dollar average of 2025 and roughly three times the 2021 level, according to Pharmcube data. Average total deal size hit 1.3 billion dollars in early 2026, up 76 percent from 2025 and about six times the 2021 average. The direction of travel is unambiguous: buyers are shifting money from the contingent column into the committed column, which is precisely what buyers do when they stop treating a market as a lottery and start treating it as a source.
The mechanism behind the shift is a closing discount. Clinical-stage Chinese assets have historically traded at upfronts 60 to 70 percent below Western comparables, a gap rooted in domestic capital scarcity, geopolitical risk pricing and the simple novelty of the sellers. That gap was the entire arbitrage, and arbitrages die in public. As one Macquarie Capital analysis put it, Chinese firms are asking for higher valuations as demand for their assets, and recognition of their quality, has improved. Tom Barsha, who heads Asia-Pacific M&A at BofA Securities and has advised on several of these transactions, expects the total value of out-licensing deals to double again over the next 18 to 24 months, with global pharma focused on identifying its next generation of pipeline in China through a widening variety of transaction structures.
There is a second, quieter reading of the upfront column. It measures not only what buyers believe but what sellers no longer need. In the first wave of these deals, Chinese biotechs licensed early because domestic capital markets were shut and an upfront was survival money. The 2026 entries increasingly read differently. Companies like Hengrui, CSPC and Innovent are profitable or well-capitalised institutions negotiating from strength, retaining co-development rights, taking equity components, and in some cases declining to sell at all until an asset carries later-stage data that multiplies its price. The ledger's growth is therefore compounding on both sides: buyers paying more per asset, and sellers timing their sales better.
Aggregate numbers say the region became a source. Only the modality breakdown says a source of what, and here the ledger is remarkably concentrated. The buyers are not purchasing Asia's whole pipeline. They are purchasing four things.
The first is the bispecific antibody, above all the PD-1 and VEGF class that ivonescimab legitimised. Merck's LaNova deal, Pfizer's 3SBio deal and BMS's BNT327 arrangement are all, at bottom, the same purchase: a hedge against the possibility that the most valuable franchise in oncology is about to be succeeded by a design that Chinese labs industrialised first. No other single target class has moved this much upfront capital this quickly in the industry's history.
The second is the antibody-drug conjugate, where the concentration is even more extreme. Chinese biotechs now account for nearly 90 percent of all global ADC licensing activity, an inheritance of the engineering-intensive, iteration-heavy discovery culture that Chinese oncology built through the 2010s, and of the Daiichi precedent that proved ADC platforms command platform prices. When Western pharma wants a conjugate now, the default first call is to Shanghai or Suzhou, not Boston.
The third is metabolic disease, which is where the single largest cheques have landed. The CSPC and AstraZeneca peptide collaboration, Merck's Hansoh oral GLP-1 licence and Regeneron's separate Hansoh arrangement all chase the same prize: the obesity market's next generation of oral, long-acting and combination therapies. The GLP-1 and obesity space alone produced at least six multi-billion dollar deals in a single year, and Asia's peptide chemistry and manufacturing depth, the very capabilities built during its contract-manufacturing era, turn out to be exactly what the class requires. The factory skills did not disappear when the region became an originator. They became the originator's moat.
The fourth is nucleic acid medicine, the newest column and the fastest-growing. Madrigal's Ribo licence for RNA interference programmes, and the Korean RNA entries discussed below, mark the point where Asia's platform credibility extended beyond antibodies into the modality Western investors long assumed was their protected ground. Ribo's own disclosure ambitions, articulated around its Hong Kong listing, capture the new confidence: these are no longer companies grateful to be bought from.
What the buyers are conspicuously not buying, at least not yet at scale, is early cell therapy, gene editing and the frontier platforms where clinical validation remains thin. The ledger is rational. It pays for classes where Asian data has already cleared Western bars.
A regional audit that stopped at the Chinese border would repeat the laziest error in Western coverage of this story, because the second-fastest-growing account in the ledger is Korean.
South Korean drug licensing deal value reached 7.68 billion dollars in 2025 by GlobalData's count, a 113 percent increase over 2024, with the out-licensing of Korean assets to international firms surging 180 percent, an increase of roughly 5.1 billion dollars, powered by billion-dollar agreements with buyers such as Eli Lilly and GSK. Lilly committed 630 million dollars for OliX Pharmaceuticals' MASH candidate in February 2025 and a separate 1.3 billion dollar arrangement for Rznomics' RNA-based gene therapies in May. Add ABL Bio's April 2025 licence of its Grabody-B blood-brain-barrier shuttle platform to GSK, at 38.5 million pounds upfront against a total of up to 2.075 billion pounds, and a distinct Korean signature emerges: platforms and delivery technologies rather than single assets, sold earlier and priced on engineering rather than on clinical endpoints. Korea is monetising the picks and shovels of the next modality cycle.
Japan's account reads differently again, because Japan wrote the ledger's preface. Daiichi Sankyo's ADC platform deals remain, per upfront dollar, among the most successful out-licensing transactions any company anywhere has ever signed, and they established the crucial precedent that a single Asian platform could anchor a top-five pharma's oncology strategy. Japan's current-window entries are quieter, partly because its majors increasingly appear on the buy side of the very same ledger, competing with Western pharma for Chinese and Korean assets. That role reversal, Asian pharma outbidding Western pharma for Asian science, may be the most underreported line item of the whole period.
And then there is India, which contributed the window's most instructive single entry. In July 2025, AbbVie licensed ISB 2001, a trispecific antibody for multiple myeloma developed by Ichnos Glenmark Innovation, for 700 million dollars upfront and up to 1.225 billion dollars in milestones, plus tiered royalties. The upfront was the largest ever paid for an Indian-origin novel biologic, and it landed a decade after most of the industry had quietly concluded that Indian innovation would remain a generics-adjacent story. One deal does not make India a source. But 700 million dollars, paid on signature by one of the most diligence-heavy buyers in the industry, ends the argument that it cannot be one.
Every ledger has a shadow ledger: the rights that stayed home. And the sophistication of this cycle shows most clearly in what Asian sellers withheld.
The Glenmark entry is again the cleanest illustration. AbbVie took ISB 2001 rights for North America, Europe, Japan and Greater China. Glenmark retained the asset for emerging markets, including India itself. The logic is worth spelling out because it recurs across the window's smartest deals. The retained territories are the ones where the seller's own commercial infrastructure, pricing knowledge and regulatory relationships are genuinely superior to the buyer's, and where keeping the asset converts a licensing deal into the foundation of a future commercial-stage company rather than a one-time monetisation.
Chinese sellers have pursued the same instinct through different structures. Hengrui's GSK portfolio deal and Innovent's Pfizer arrangement both preserve substantial Chinese-market economics for the originator, and across the window's larger transactions the standard template has become ex-Greater-China rights out, home rights retained. The NewCo model refines the instinct further: sellers place ex-China rights into a purpose-built, Western-financed vehicle, take equity in it, and thereby keep exposure to the asset's global upside instead of capping their participation at milestones. Option-to-license mechanics have proliferated alongside these structures, letting sellers charge buyers for the right to decide later, which is what negotiating leverage looks like when it is written into contract architecture.
The pattern answers the question sceptics reasonably ask of the whole thesis: if Asian companies were merely arbitraging a valuation gap, they would sell everything, everywhere, as fast as possible. They are doing the opposite. They are selling the territories they cannot yet serve and industrialising the ones they can. That is not the behaviour of a region cashing out. It is the behaviour of a region building holding companies.
An honest audit discloses its own limitations, and this one has three.
The first is that the ledger's totals depend on who is counting. PharmCube's 137.7 billion dollars across 186 deals for 2025 and VBData's 135.7 billion across 157 describe the same year through different inclusion rules on options, NewCo formations and platform collaborations. This publication reports both rather than adjudicating between them, because the divergence is itself information: the deal universe has become large and structurally varied enough that reasonable counters disagree.
The second is disclosure itself. A substantial share of the window's transactions, including several involving major multinationals, disclosed no upfront at all, and some disclosed nothing beyond existence. Those entries appear in this ledger as what they are, undisclosed, and readers should treat every industry-wide average built on disclosed figures, including the averages quoted above, as describing the visible portion of a partially visible market. This publication does not estimate upfronts, and readers should be suspicious of any outlet that does.
The third footnote is political, and it is the one that could reprice the entire book. The same 24 months that produced this ledger produced escalating American scrutiny of pharmaceutical dependence on Asia, from tariff investigations to legislative pressure on research partnerships. So far the money has voted the other way: every quarter of policy noise has coincided with a new record in deal flow, and buyers have responded to political risk not by retreating but by restructuring, through NewCos, equity vehicles and jurisdictional engineering. But a ledger records the past. It does not indemnify the future, and a genuine regulatory rupture between Washington and Chinese biotech remains the one scenario under which the trailing 24 months would read, in retrospect, as a peak rather than a beginning.
Strip away the deal-by-deal detail and the audit closes on a single finding. In October 2024, the proposition that Asia-Pacific had become the pharmaceutical industry's originating region was a thesis. By September 2026, it is an accounting identity. Asia-origin transactions now represent roughly a third of global out-licensing value, the largest upfront in the industry's recent history was paid for a Chinese platform, the largest ADC franchise in the world is Japanese, the fastest-growing platform-licensing account is Korean, and the largest upfront ever paid for an Indian biologic cleared 700 million dollars. Manufacturing partner to originator, inside a single cycle, exactly as the region's own executives kept insisting would happen and exactly as fast as everyone else insisted it could not.
The next 24 months will test a different question, one this ledger can only hint at. Selling molecules is the second-highest rung on the value ladder. The highest is keeping them: taking an Asia-origin asset through global registration under the originator's own name, and booking the revenue rather than the royalty. The territorial retentions, the NewCo equity stakes and the rising confidence in the upfront column all suggest the region's leading companies understand the difference and are financing the climb. When the first of them completes it, the story will no longer be a ledger of what Western pharma paid for Asian science. It will be a ledger of what Asian pharma kept.
That edition is already being written. This one merely records where the money moved while the industry was still deciding whether to believe it.
**All deal values reflect company disclosures. Where parties have not disclosed terms, they are reported as undisclosed. No upfront payments have been estimated**
