On the same road, a Chinese company has already published its data in NEJM

RiboX was founded in 2021 and operates across Shanghai, the US and Israel, signing its announcements "Shanghai and Cambridge, Massachusetts". Its three co-founders are Chen Lingling in the circular RNA field, Dan Peer in targeted lipid nanoparticle delivery, and current CEO Zhang Weiyi — each of the two core technology stacks corresponds to a founding figure in its field. Cumulative financing per third-party databases is about US$110 million, with investors including Qiming Venture Partners and Boehringer Ingelheim Venture Fund; this figure comes from external databases rather than company disclosure.

The company also has an asset that entered the clinic earlier, RXRG001, the world's first circular RNA therapy to receive FDA IND clearance (October 2024), for radiation-induced xerostomia — using circular RNA to make salivary glands express more aquaporin, with first patient dosed in March 2025. One replenishes a protein locally, the other builds cells in vivo; the two directions are far apart but share one technology base. The platform is not a single point; this is the most solid part of the company's valuation narrative.

Note two things. First, China is where this product was first tested in humans, but not its registration home — data are generated in China, registration is pursued in the US. Second, the dosing method actually in use now is intravenous infusion; subcutaneous and outpatient dosing are development goals, not accomplished facts.

The company's wording is "the world's first targeted lipid nanoparticle-delivered circular RNA in vivo CAR therapy to receive FDA clearance to enter the clinic". The sentence is built from four stacked qualifiers; remove any one of them and it becomes wrong. Below are four facts of "it is not the first".

Four earlier entities

The last one matters most. The differentiation of circular RNA relative to linear mRNA is not a claim unique to RiboX — on February 9, 2026 Lilly acquired Orna for up to US$2.4 billion, buying precisely the combination of circular RNA plus lipid nanoparticles, and its lead program addresses the same target and the same class of indications.

Status of lipid nanoparticle-route in vivo CAR-T

There is also a group of players on the lentiviral route: Lilly acquired Kelonia for up to US$7 billion on April 21, 2026, AbbVie's collaboration with Umoja totals up to US$1.4 billion, and Interius is also advancing, mostly in oncology. In China, CSPC's SYS6055 Phase 1/2 has completed registration with 374 planned patients, currently publicly listed as not yet recruiting.

What deserves the most attention in this table is a divergence in technological philosophy. CPTX2309 treats "transient expression" as a safety selling point, while RiboX treats "durable expression" as circular RNA's core advantage — two opposite directions. In building CAR-T in vivo, the longer expression lasts, the deeper efficacy may theoretically be, but the lower the controllability. HN2301 has already shown that "transient expression can also wipe out circulating B cells", so what additional benefit "more durable" buys becomes a question RXIM002 must answer for itself. Until it has its own data, nobody can answer that for it.

As of June 2026, China had approved 9 ex vivo CAR-Ts, with public tender prices ranging from RMB 999,000 to RMB 1.29 million. All are manufactured domestically, one batch per patient, all rely on qualified infusion centers, and patients have to go through the whole chain of blood collection, shipment to the plant, manufacturing, waiting, lymphodepletion and reinfusion. The first Commercial Health Insurance Innovative Drug List included 5 of them; in the 2026 list adjustment, expert review was completed at the end of July, but which cell therapy products enter negotiation has not been made public.

The choice of lead indication is pragmatic. In China's 2025 guideline for adult primary immune thrombocytopenia, second line already includes thrombopoietin receptor agonists and rituximab, and also incorporates new thrombopoietic drugs and small-molecule kinase inhibitors. But the guideline also separately defines a "refractory" population: those who have failed first-line treatment, thrombopoietic drugs and rituximab, or who failed splenectomy or relapsed after it. This definition in itself proves that a gap exists.

The cost is a small population. The two studies in China and the US together plan 54 patients, and the lead indication cannot support a commercial story. The real commercial value lies in expansion — public information shows the company is also looking at autoimmune hemolytic anemia, membranous nephropathy, idiopathic inflammatory myopathies, systemic sclerosis and other directions. So this asset's commercial model should be built on "whether outpatient dosing can be achieved". But a reminder: for now this is only a development goal, and the China investigator-initiated trial still uses intravenous infusion.

The most noteworthy passage in the company's announcement is that, before filing with FDA, this product had treated patients in a Chinese investigator-initiated trial, and the company submitted complete data on all treated patients to FDA as part of the application. According to RiboX, based on the completeness and adequacy of this data, FDA allowed it to use an accelerated dose-titration scheme in clinical development and to include a subcutaneous formulation, and mentioned the possibility of outpatient dosing.

With the evidentiary boundaries drawn, this path is still worth studying, because the conditions under which it works are reusable.

The first condition is that the treatment paradigm for the indication is consistent in China and abroad. For immune thrombocytopenia and B-cell-mediated autoimmune diseases, second line in both China and the US is thrombopoietin receptor agonists plus rituximab, with no standard-of-care discontinuity like in oncology. With a consistent paradigm, the obstacles to extrapolating Chinese data are small. For an indication where Chinese and foreign standards of care differ greatly, the same data might not be treated the same way.

The second condition is that the data were submitted completely — data on all treated patients were submitted, including safety and early efficacy, not cherry-picked highlights. The third condition is the center's credibility — the trial was conducted by Shanghai Ruijin Hospital, appearing as sponsor, a genuine plus for the credibility of the data. From first dose in China in December 2025 to US IND clearance in August 2026 took only about eight months.

The cost is hidden in the structure. The two studies in China and the US are two independent single-country studies, not one multi-regional trial, and their populations differ. More importantly there is the anomalous structure: China generates the data, the US does the registration. If it is to launch in China too, either the US data must flow back, or a separate Chinese registration study must be started — both approaches cost additional money and time, and there is currently no public timetable.

When discussing in vivo CAR-T, people habitually talk about efficacy and safety and rarely about manufacturing. But for this technology route, manufacturing is precisely its reason for existing.

Ex vivo CAR-T is one batch per patient: one batch, one release and one traceability chain per patient; capacity equals an upper limit on the number of patients, and costs hardly fall with scale. The in vivo route is a batch-manufactured finished formulation: it can be stocked and scaled up, and unit cost has room to fall with scale. This is the most fundamental economic difference of this route.

The difficulty is real too. Circular RNA has to solve circularization efficiency and purity, lipid nanoparticles have to solve encapsulation efficiency and particle size consistency, and with two processes stacked, controlling batch-to-batch consistency is not easy. Public information can currently only confirm that it has supported drug supply for the Chinese investigator-initiated trial; the US trial has no sites yet and plans to start in February 2027, so one cannot say the company is already supplying clinical material to both China and the US — IND clearance only shows that the process data were sufficient to support advancing the study, not that US clinical batches have been released. Process data, manufacturing site and whether it is in-house or outsourced have not been disclosed either.

One suggestion here: rather than continuing to make currently unverifiable claims such as "circular RNA is more stable than linear mRNA", build a cost model as soon as possible — putting the unit treatment cost of the in vivo route side by side with ex vivo CAR-T. That is the most powerful and least refutable value argument for this asset, and also exactly what payers most want to see.

The regulatory milestone has been delivered; the clinical milestone still has to wait, and the gap in between is precisely a high-risk period for financing and business negotiations — because the strongest card it can play is a regulatory clearance, not a set of clinical numbers. And a Chinese peer in the same race has already produced the latter.

Fortunately the next value inflection point is very clear: the first public disclosure of data from the Chinese investigator-initiated trial. Once it reports B-cell depletion depth, levels of in vivo-generated CAR-positive cells and safety numbers, the pricing logic for this asset will switch from "platform narrative plus regulatory milestone" to "clinical evidence". Until then, it is mainly propped up by M&A prices in the race — and those prices are indeed not low: AbbVie paid US$2.1 billion for Capstan, and Lilly paid up to US$2.4 billion for Orna and up to US$7 billion for Kelonia.

One more thing should be said plainly: the company has not published any go or no-go thresholds. With low funding visibility, the absence of explicit termination discipline is not a small matter.

RXIM002 is easily misread in two ways.

One is to pull the words "world's first" out of their qualifiers and use them on their own. The facts are: the first human data for lipid nanoparticle in vivo CAR-T came from another Chinese company and were published in the New England Journal of Medicine; someone else entered the clinic earlier in autoimmunity; China's first in vivo CAR-T clinical approval belongs to someone else; and even for "circular RNA CD19 in vivo CAR-T for autoimmunity", Lilly has one in hand too.

The other is to underestimate it because it is not first. The company has genuinely accomplished three things: assembling its own capabilities in both technology stacks and advancing one asset from each into the clinic; making data generated by Chinese physicians a core component of a US IND application; and completing the whole process from first dose to overseas clearance within eight months.

What it lacks now is also very specific — just three things: a set of clinical numbers it can make public, a freedom-to-operate conclusion based on a comparison, and a visible funding window. None of the three is a technical problem — and a peer has already shown the market, with one paper in the New England Journal of Medicine, what the first of them should look like.