Merck, Novo Nordisk, and Novartis are not just writing big checks in China. They’re betting on science that hasn’t even reached human trials. The US Treasury is watching closely. The real concern isn’t who owns the asset. It’s who gets the know-how. Washington is worried that co-development deals could give Chinese partners the skills to compete worldwide.
Novartis secured a global license for ABO2203, an mRNA-encoded T-cell engager targeting CD3 and CD19, with a deal potentially worth up to $7.8 billion including $575 million upfront.
New science, new risks
China is no longer just a source for antibody and oncology assets. The latest deals—Novartis with Abogen, Merck with Cybrunch, and Novo Nordisk with an undisclosed partner—cover RNA-encoded T cell engagers, KRAS molecular glues, and GLP-1/GIP dual agonists. Only Novartis’s deal with Abogen had any clinical data, and that came from two ongoing investigator-initiated trials in China. The other two assets are still preclinical. Big pharma is now paying top dollar for science that hasn’t been tested in people.
The Abogen asset is different. It uses mRNA-LNP to tell the patient’s body to make a T cell engager for days or weeks after one treatment. This “vectorized biologic” sits between in vivo CAR-Ts and standard T cell engagers. According to a BioSpace report, ABO2203 is the first mRNA-encoded T-cell engager to reach clinical evaluation for autoimmune diseases. It temporarily prompts the body to make a protein that destroys B cells. Merck’s KRAS G12D molecular glue could be first-in-class, but Martz says details are still unclear compared to Estellas’s work. Novo Nordisk’s deal is less public. It targets a GLP-1/GIP dual receptor agonist, also at the preclinical stage.
Politics and the COINS Act
US officials are not worried about asset shopping. They’re worried about capability transfer. Steve Usdin, BioCentury’s Washington editor, draws a line: licensing deals hand over assets, but co-development deals hand over process, know-how, and learning. That’s how Chinese partners could build companies that compete globally. The Treasury is preparing to add biotech to the COINS Act, the Comprehensive Outbound Investment National Security framework. If co-development deals keep growing, the sector could face tighter rules.
Merck closed an exclusive global license with SciBrunch Therapeutics for SPR2015, a preclinical oral KRAS G12D inhibitor, with a $400 million upfront payment and a total potential value of $2.13 billion. Merck stated it will record the upfront payment as a pre-tax charge in its Q3 2026 financials.
Frank LeDeaux, a former McKinsey leader in China, says Chinese assets are going global. But Chinese companies are not yet global players. Until a Chinese company owns its US profit and loss in the world’s biggest market, instead of just selling assets to Western firms, a true peer to Novo Nordisk is unlikely to come from China soon.
Legal fights and drug pricing rules
The legal battles are heating up. Beam Therapeutics is suing YolTech over trade secrets. The case centers on “negative know-how”—the idea that knowing what fails can help a rival move faster. Beam says former scientist Emma Wang co-founded YolTech while still at Beam and took not just technical data but also key insights into failed approaches. The defense says Beam’s CEO publicly denied theft, that there’s no direct link between Beam’s data and YolTech’s inventions, and that YolTech has its own R&D records. The jury will have to decide if negative know-how counts. Usdin compares it to Cold War tech races.
Drug pricing is also in flux. The GLOBE most-favored-nation model was supposed to save $11.9 billion. Now, after many exemptions, it’s down to $440 million. But the precedent matters. CMS is using its demonstration authority to set international reference pricing—something Congress never passed. Pharma CEOs are now negotiating directly with the White House. The GARD model for Medicare Part D launches the same day as GLOBE. Lawsuits are expected to challenge both.
Optogenetics, which just won the Nobel Prize, is still mostly a research tool. Clinical use is mostly limited to the eye. Karl Deisseroth, Peter Hegemann, and Georg Nagel’s discoveries changed neuroscience. But as Koch points out, turning circuit-level insights into treatments for diseases like Alzheimer’s is still far off. It’s a slow climb, much like the path from CRISPR as a tool to CRISPR as a medicine.
China’s role as the “pipeline’s margin of safety” is running into new barriers. The licensing boom is real. It’s picking up speed. But the lines—political, legal, and commercial—are being drawn now. Treasury rulemakings, court battles in Massachusetts, and year-end drug pricing fights are all shaping the future. The next moves by Washington and the outcome of cases like Beam v. YolTech will decide if this surge in early China deals opens a new era of scientific exchange or builds a new regulatory wall. Companies are racing for the next breakthrough. They’re paying big for unproven science. But no one knows how much capability transfer the US will allow. For more on how companies are changing their operations under this pressure, see what was reported earlier on integrated biologics development.
