Pfizer will pay $330 million upfront for exclusive rights outside Greater China to three preclinical CAR-T programs from Starna Therapeutics, a Chinese RNA-medicine developer. The deal includes up to $2.3 billion in development and commercial milestones, for a potential total of about $2.6 billion, plus tiered royalties on sales. Pfizer announced the agreement on November 18.
The deal was reported with the three programs and the milestone structure. The headline number is large. The strategic bet is larger. The programs are a form of CAR-T that is meant to work without removing a patient's cells. Instead of extracting T cells, engineering them in a lab, and reinfusing them after chemotherapy conditioning, an in vivo approach delivers genetic instructions inside the body and reprograms the cells where they are.
What in vivo CAR-T changes
Conventional CAR-T therapy is a bespoke manufacturing process. A patient's T cells are collected, genetically modified to recognize a target, expanded outside the body, and returned. The process takes weeks, requires specialized facilities, and usually includes conditioning chemotherapy that clears space for the engineered cells.
An in vivo approach tries to collapse that sequence. Starna's platform uses RNA-based medicines delivered by lipid nanoparticles, the same broad family of delivery technology that made mRNA vaccines possible. The instructions are transient: the reprogrammed cells express the engineered receptor for a period, then the RNA degrades. That is different from the permanent edit a viral vector makes in ex vivo CAR-T.
The mechanism matters for how the therapy is used. A chimeric antigen receptor is a synthetic receptor that redirects a T cell toward a chosen target. In cancer, the target is often a protein on a tumor cell. In autoimmune disease, the target is often a protein on B cells, and the goal is to deplete the B cells that drive the attack on the body's own tissue. A transient receptor may be enough to clear a wave of B cells, and the immune system can then rebuild.
The tradeoff is delivery. A lipid nanoparticle has to reach the right cells in the right tissue and deliver enough RNA to matter. The liver, the spleen, and the immune system all present different barriers. Pfizer's manufacturing experience with RNA is the reason the company can argue it has an edge in the delivery step.
Why autoimmune disease
The programs target autoimmune disease, not cancer. That is the fastest-moving frontier for CAR-T. The same cell-engineering logic that lets a T cell attack a tumor can be used to eliminate the B cells that produce autoantibodies in conditions such as lupus. Early academic studies in lupus have shown deep remissions in some patients.
Autoimmune disease is also where the safety questions are sharpest. Novartis and Bristol Myers Squibb recently paused trials of CAR-T therapies in autoimmune conditions over safety concerns. The pauses do not invalidate the approach. They do show that the risk profile in autoimmune patients, who may have less advanced disease than cancer patients, is still being defined.
An in vivo approach could change that calculus if it works. A transient, repeatable dose may be easier to stop than a permanent engineered cell population. It may also be easier to give earlier, before a patient has accumulated organ damage. Those are hypotheses, not results.
The structure of the deal
Pfizer gets the three programs and rights to Starna's technology in inflammation, immunology, and oncology. It also gets options to license additional projects and to extend its territory into Greater China. Pfizer will lead clinical development outside China and plans to manufacture initial trial supplies in the United States.
Starna is an RNA-medicine developer working on in vivo cell reprogramming. The company describes its platform as a way to reprogram cells inside the body, and it has been expanding from oncology into autoimmune disease. The programs Pfizer licensed are preclinical, which means they have not yet entered human trials.
That stage matters for how the deal should be read. A preclinical licensing deal is a claim about a platform, not a product. The upfront payment buys access to the platform and the option to develop it. The milestones are paid only if the programs advance. The royalties arrive only if a product reaches the market, as Reuters reported.
The economics of ex vivo CAR-T explain the interest in a simpler model. A patient-specific batch requires a dedicated manufacturing slot, a chain of custody from the clinic to the lab and back, and a team that manages the apheresis, the engineering, and the infusion. The cost of goods is high, the logistics are complex, and the treatment is concentrated in centers that have built the infrastructure. An off-the-shelf RNA dose would not eliminate those costs, but it could move the therapy toward a more standard infusion model.
The regulatory path
A preclinical program becomes a clinical program when a company files an investigational new drug application and gets clearance to begin human trials. The first trial for an in vivo CAR-T in autoimmune disease would likely start small, with careful dose escalation and long safety follow-up. Regulators will want to see that the RNA does not produce an uncontrolled immune activation, that the reprogramming is transient in practice, and that the effect on B cells is reversible enough to manage infections.
The timeline is measured in years. A three-program portfolio increases the chance that at least one reaches the clinic, and it also multiplies the cost. Pfizer's development organization will have to run the trials outside China while Starna continues its own work in the territory it retained.
The China licensing wave
The Starna agreement is at least Pfizer's fifth major alliance with a Chinese company since the beginning of 2025. The company's earlier deals include a large collaboration with Innovent and a licensing arrangement with 3SBio. Pfizer is not alone, industry coverage noted. AstraZeneca, AbbVie, Gilead Sciences, and Bristol Myers Squibb have made in vivo cell-therapy acquisitions in the past two years, and Eli Lilly announced two such buyouts in 2026.
The pattern is a response to two pressures. The first is scientific: Chinese biotech companies have built credible platforms in cell therapy and RNA delivery, and the preclinical data can be compelling. The second is commercial: licensing a preclinical program is cheaper and faster than building the platform from scratch, and it spreads the risk across more shots on goal.
The in vivo field is still early, which is why the deals are structured as options on a platform. Eli Lilly's two 2026 buyouts and the earlier acquisitions by AstraZeneca, AbbVie, Gilead, and Bristol Myers give the field a set of comparable bets. Pfizer is arriving later than some of its peers, and it is arriving through a license rather than an acquisition. That lowers the upfront cost and leaves the integration risk with the partner.
The pattern also has a political dimension. BIO, the biotechnology industry association, has proposed a five-year plan to support U.S. biotech competitiveness against China. A deal that sends $330 million to a Chinese company and keeps development and manufacturing in the United States sits awkwardly between the two goals. Pfizer's answer is that the manufacturing and the clinical development stay domestic, and the license is a way to access science that would take years to reproduce.
What preclinical means for a patient
Nothing in this deal changes care today. The three programs are years from an approval, and most preclinical programs do not make it. A patient with lupus or multiple sclerosis should not read the announcement as a treatment that is about to arrive.
The deal does signal where the industry thinks the next generation of cell therapy will be fought. The ex vivo model works, but it is expensive, slow, and concentrated in a limited number of centers. An in vivo model that could be dosed in an outpatient setting would change the economics and the reach. That is the prize, and it is why the upfront payment is large relative to the stage of the programs.
The safety bar is also higher in autoimmune disease than the enthusiasm suggests. The Novartis and Bristol Myers pauses are a reminder that reprogramming the immune system in a patient who is not facing a terminal cancer is a different risk calculation. Regulators will ask for long follow-up, and the trials will be watched closely.
The delivery problem
The deal's success will be decided in the laboratory, and the decisive question is delivery. RNA is fragile. A lipid nanoparticle has to protect it, reach the target tissue, and get it into the right cells without triggering an immune reaction that undermines the treatment. The published preclinical work in the field is promising, and it is early.
Pfizer's bet is that its RNA manufacturing and formulation experience shortens the path. Starna's bet is that its delivery platform is the differentiator. The milestones are structured so that both companies share the risk if the platform does not translate.
If it does translate, the change could be structural. A CAR-T that is manufactured as a dose rather than as a patient-specific batch would be cheaper to scale, easier to distribute, and available in places that cannot support an ex vivo cell-therapy center. That is the version of the future the deal is paying for.
The bet behind the number
The $2.6 billion headline is a sum of possible payments, most of which may never be made. The $330 million upfront is the real commitment, and it buys Pfizer a place in a race that several larger competitors have already entered.
The science will be judged by whether the delivery works and whether the safety profile holds in autoimmune patients. The business will be judged by whether Pfizer can turn a platform into a pipeline before the competition does. The patients who might benefit are several trial phases away from knowing.
The structure of the payments reflects that uncertainty. A large upfront payment is usually a signal that the licensee expects the platform to move quickly. A milestone-heavy deal is a signal that the licensee wants to share the risk. This deal has both: a substantial upfront and a milestone pool that is nearly seven times larger. Pfizer is paying for access and optionality, not for a finished product.
Primary sources
- BioPharma Dive, for the deal terms, the three preclinical programs, the in vivo technology, and the competitive context.
- FierceBiotech, for the transaction coverage and the strategic framing of Pfizer's first in vivo CAR-T deal.
- Miami Herald, for the Reuters report on the licensing agreement and the milestone structure.