The last essay in this series said China beat India at India's own game: the antibody-drug conjugate was the second-mover's dream modality, and China ran the second-mover playbook — industrialise the proven, engineer don't invent — better and faster than India, whose antibody base should have made it the natural player. This essay is what comes after that, and it is worse for the doctrine. Because in bispecific antibodies, China did not run India's game better. It stopped playing India's game and started playing the West's. A Chinese company built a genuinely novel molecule and ran it head-to-head against Keytruda — the best-selling drug in the history of the industry — and won. (On progression-free survival, in a Chinese trial, with the survival verdict still open; hold that bound, it matters.) And then the tell arrived: the Western majors began licensing Chinese molecules, for billions, to build their own next-generation pipelines. China is no longer the second-mover manufacturing the proven modality. It is the first-mover designing the next one, and Merck is the customer. This is the essay where the second-mover window India has been waiting to slip through is quietly closing — because the seat India wanted is being taken by an ecosystem that invents and industrialises at the same time.
Two targets, one molecule
A bispecific antibody does what a normal antibody cannot: it grabs two different things at once. A T-cell engager, the most established form, holds a cancer cell with one arm and an immune T-cell with the other, dragging the immune system bodily onto the tumour and forcing the kill. A checkpoint bispecific like ivonescimab blocks two escape routes at once — the PD-1 brake the tumour uses to hide from immunity, and the VEGF signal it uses to grow its blood supply. One molecule, two jobs, engineered to do together what two separate drugs do apart.
The class is real and approved. It began with blinatumomab — Amgen's Blincyto, the first-in-class T-cell engager, approved in 2014 — and has become a wave: teclistamab for myeloma, tarlatamab for the notoriously untreatable small-cell lung cancer, glofitamab and mosunetuzumab for lymphoma. These are not incremental. Tarlatamab reaches a cancer that had barely moved in thirty years. And the modality is brutally hard to make: bolt two different antibodies into one molecule and the protein chains mispair, yielding up to ten different molecular species where you wanted one, so the entire field runs on clever engineering fixes — "knobs-into-holes," common light chains — just to assemble the thing correctly. Like the ADC, the bispecific is an engineering-heavy modality. Unlike the ADC, the frontier of it is no longer only Western.
The ivonescimab moment, stated exactly
In September 2024, at the world lung-cancer congress, Akeso reported the result that reorganised the field. In a Chinese Phase 3 called HARMONi-2, its bispecific ivonescimab, given alone, was compared head-to-head against pembrolizumab — Keytruda — given alone, in first-line PD-L1-positive lung cancer. Ivonescimab roughly doubled progression-free survival: a median of 11.1 months against 5.8, a hazard ratio of 0.51, published in The Lancet. It was the first time in the checkpoint era that any drug had beaten Keytruda head-to-head in a Phase 3. China's regulator approved it first-line in April 2025.
Now the bound, because the bound is not a footnote — it is half the truth. That win is progression-free survival, not overall survival; it is a Chinese trial in a Chinese population; and it was ivonescimab as monotherapy. The survival data — does it make people live longer, the endpoint that actually matters — is still immature in HARMONi-2 and, so far, only a positive trend. And when Akeso's partner Summit ran the global trial, HARMONi, in a different setting, the result split: progression-free survival was met decisively, but overall survival came in at a hazard ratio of 0.79 with a p-value of 0.057 — a miss on statistical significance — and Summit's stock fell about thirty per cent in a day. The FDA has signalled it wants a clean survival win before approval. There is even a geography split inside the same drug: the Chinese version of that second-line setting hit overall survival; the global one did not.
So state ivonescimab precisely, because getting it wrong sinks the essay: it is a genuinely novel Chinese molecule that beat the best-selling drug in the world on progression-free survival in a Chinese Phase 3, and whose global, live-longer verdict is still unproven. Not a slam dunk. But — and this is the load-bearing word — not a copy. China did not fast-follow Keytruda. It built a different molecule with a novel mechanism and put it in the ring. Whether ivonescimab ultimately wins on survival is contested. That China produced a credible best-in-class challenger to Keytruda at all is not.
The tell: the West is now shopping in China
If one contested trial were the whole case, this would be a thin essay. It is not, because the strongest evidence is not the trial — it is where the smart money went afterward, and that evidence is airtight.
Ivonescimab's result set off a scramble among the Western majors for the same class of molecule. And every asset they reached for was Chinese. In November 2024 BioNTech acquired Biotheus outright for its PD-L1/VEGF bispecific, eight hundred million dollars upfront; Bristol Myers Squibb later partnered on the same molecule. That same month Merck licensed LaNova's PD-1/VEGF bispecific for five hundred and eighty-eight million upfront against up to roughly three billion. In May 2025 Pfizer licensed 3SBio's PD-1/VEGF bispecific in a deal worth up to six billion dollars. Three of the largest pharmaceutical companies on earth, chasing the mechanism ivonescimab validated, and not one of them built it themselves or bought it from a Western biotech. They all went to China.
That is the graduation, and it is proven not by an efficacy argument anyone can dispute but by revealed preference. When Merck and Pfizer and BioNTech decided they needed a next-generation checkpoint bispecific, the place they went to buy one was China. China is no longer the ecosystem that manufactures the proven modality cheaply. It is the ecosystem that designs the modality the West then licenses. More broadly, China's biotech out-licensing hit record levels in 2025 — bispecifics and T-cell engagers a large slice of it — and China now files roughly twice as many first-in-human trials of next-generation antibodies as the United States and Europe combined. Those specific figures are trade-tracker estimates and should be read as such, but the direction they point is the same direction the three named deals point, and the three named deals are not estimates.
The window this closes
Here is why bispecifics matters more to the India thesis than its own market size, and why it belongs at the end of the series rather than the middle.
The entire therapeutics doctrine this publication has built rests on a single structural assumption: that there is a second-mover seat — that the frontier ecosystem (the West) invents the modality, retires the discovery risk, and then a disciplined, cost-competitive follower (India) industrialises the proven version and wins on cost. Every India-wins essay in this series — CAR-T, GLP-1 — depends on that seat existing and staying open. The seat exists only if two things stay true: the frontier stays somewhere India can watch and wait, and the frontier ecosystem does not also hold the cost advantage.
Bispecifics breaks the second condition, and that is fatal to the seat. China now holds the frontier and the cost base at once. It invents the novel bispecific and it manufactures it cheaply. When a single ecosystem occupies both the discovery frontier and the low-cost production floor, there is no gap in between for a second-mover to slip into — the thing India was going to do after China invented it, China already does itself, cheaper than India could. The second-mover window is not a permanent feature of the world. It exists in the space between a high-cost inventor and a low-cost follower, and in bispecifics China has closed that space by being both. The doctrine did not fail here the way it failed in ADCs, by India running the play too slowly. It expired — the play itself stopped being available.
India's position: the back of the room
Which leaves India, in the hottest antibody modality in oncology, further back than anywhere else in this series.
There is no indigenous, India-developed, clinical-stage novel bispecific. Not a slow one, not a preclinical one worth naming — effectively none. The single marquee "Indian" bispecific tells the whole story by where it lives: Bicara Therapeutics, developing an EGFR×TGF-β bifunctional antibody, was founded out of Biocon — and is domiciled in Cambridge, Massachusetts, listed on Nasdaq, with American science and an American cap table. India's best bispecific bet had to become American to happen. The talent was Indian; the company is not. Beyond that, Dr Reddy's is making a Keytruda biosimilar — a copy of the old drug, not a bispecific — and Aurigene offers bispecific contract manufacturing, which is the shovel, not the mine.
So India in bispecifics has no equivalent even of the thin doors it holds in ADCs. In ADCs there was at least Zydus's world-first ADC biosimilar — a real, if narrow, first. In bispecifics there is no biosimilar-first-mover, no novel programme, only CDMO capacity and a diaspora company that left. It is the thinnest position in the entire therapeutics thesis.
The access inversion, at its sharpest
And the human end of it is the sharpest in the series. Blinatumomab — the founding bispecific, a decade old, a standard of care for a lethal leukaemia — is not even registered in India. An Indian patient who needs it obtains it, if at all, through named-patient import: paperwork, special permission, a drug the rest of the world has had since 2014 treated as a bespoke exception. The novel bispecifics now redrawing oncology worldwide are, for Indian patients, import documents. India in this modality is neither the designer (China) nor the affordable-maker (its own biosimilar machine has not reached here) but the importer — the patient at the end of the line, not the producer anywhere in it. That is as far back as a country can stand in a modality, and it is where India stands in the one the whole industry is racing toward.
Where it breaks
The honest bounds, because the flagship claim rests partly on a contested trial.
The ivonescimab survival story is genuinely unresolved, and if the global overall-survival data reads out clearly negative, the single most dramatic line in this essay — a Chinese molecule beat Keytruda — loses its flagship. It must always be written as the bounded thing it is: a progression-free-survival win, in China, with the live-longer verdict open. The good news for the essay's spine is that it does not depend on ivonescimab winning — it depends on China having produced a credible enough challenger that the West went shopping in China, and the three multibillion-dollar deals prove that regardless of how HARMONi's survival curve finally reads.
The China deal-share statistics are trade-press estimates, not audited figures, and are flagged as such; the named deals carry the argument, not the aggregates.
And the "window closes" claim is a directional argument, not a theorem. India could still find niches — bispecific biosimilars as the first generation ages off patent, the access-and-affordability door it has not yet built, a diaspora founder who brings a molecule home the way Bicara's talent could have. The window closing is a strong tendency, not a sealed door. But the burden has flipped: it is now on the optimistic case to show where India's bispecific seat is, and today it cannot.
Close
Read the two China essays together and the arc completes. In antibody-drug conjugates, China beat India at India's own second-mover game — ran the industrialise-the-proven playbook faster than the country whose doctrine it is. In bispecifics, China stopped playing that game and started playing the West's — inventing a frontier molecule that beat the best-selling drug in the world on progression-free survival, and then selling the next generation of such molecules to Merck and Pfizer and BioNTech. The seat India has been waiting to sit in — arrive after the invention, industrialise the proven modality, win on cost — closes precisely when one ecosystem learns to hold the frontier and the cost floor at once, and in bispecifics China has done exactly that. India's honest place in the hottest antibody modality on earth is the back of the room: not the designer, not the affordable-maker, but the importer filing paperwork for a decade-old drug. The lesson the whole series has been circling lands hardest here, and it is not comfortable: the second-mover window is not a birthright and not permanent, and the country that assumed it would always be open to catch is, in this modality, watching it shut — while the molecules go, one licensing deal at a time, from Chinese labs to Western balance sheets, with India nowhere on the invoice.
The sixth modality essay in the Atoms and Cells therapeutics thesis and the completion of the China-competitive cluster begun in "The Payload War." Receipts verified and dated, with the ivonescimab efficacy story deliberately bounded. Ivonescimab (Akeso; PD-1/VEGF; a genuinely novel Tetrabody molecule, not a Keytruda copy): HARMONi-2 (China, monotherapy vs pembrolizumab, PD-L1+ NSCLC) reported median PFS 11.14 vs 5.82 months, HR 0.51 (0.38–0.69), p<0.0001, at WCLC on 8 September 2024 and in The Lancet — the first head-to-head Phase 3 win over Keytruda, but a progression-free-survival win, in a Chinese population, with overall survival immature (HR ~0.777 at ~39% maturity, a non-significant positive trend); NMPA first-line approval 25 April 2025. The global HARMONi trial (2L EGFR-mutant NSCLC + chemo) topline 30 May 2025 met PFS (HR 0.52) but missed OS significance (HR 0.79, p=0.057), and Summit's stock fell ~30%; the FDA has signalled a significant OS result is needed. A geography split exists: the China 2L EGFRm trial (HARMONi-A) hit OS (~HR 0.74) while the global HARMONi did not. Akeso→Summit license 6 December 2022, $500M upfront, up to $5.0B. Class validation: blinatumomab (Blincyto, CD19×CD3) FDA 3 December 2014; teclistamab (Tecvayli, BCMA×CD3) 25 October 2022; tarlatamab (Imdelltra, DLL3×CD3) accelerated approval 16 May 2024; glofitamab (Columvi) June 2023; mosunetuzumab (Lunsumio) 2022. The Western scramble for Chinese PD-(L)1/VEGF bispecifics — the airtight evidence — BioNTech–Biotheus (BNT327/PM8002) 13 November 2024, $800M upfront / up to $950M, with BMS later partnering; Merck–LaNova (LM-299) November 2024, $588M upfront / up to ~$3.3B; Pfizer–3SBio (SSGJ-707) May 2025, up to ~$6B — every asset Chinese. China out-licensing ~$136B in 2025 with bispecifics/TCEs ~30% of the mix and roughly twice the US+EU next-gen-antibody first-in-human filings (trade-tracker estimates, attributed; consistent with "The Harvest Illusion"). Bispecific manufacturing is a genuine engineering moat (chain mispairing yielding up to ~10 species; knobs-into-holes and common-light-chain fixes across many formats). India: no indigenous clinical-stage novel bispecific; Bicara Therapeutics (BCA101, EGFR×TGF-β), founded from Biocon, is US-domiciled and Nasdaq-listed; Dr Reddy's–Alvotech is a pembrolizumab biosimilar; Aurigene provides bispecific CDMO, not discovery. Access: blinatumomab is not registered in India (named-patient import only); Indian pricing kept qualitative for lack of a citable figure. Deal values are ceilings; China share figures are estimates; the ivonescimab survival verdict is genuinely open. The thesis in one line: in bispecifics China moved past second-mover to frontier — inventing the molecule the West now licenses — which closes the second-mover window India's doctrine depends on, and leaves India, in the hottest antibody modality on earth, the importer rather than the maker.