Every year, the mRNA-Based Therapeutics Summit gives the field a chance to take its own pulse and reflect on how far RNA therapeutics have come. Last year, the conversation had moved past "can mRNA work?" and into "how far can it go?" This year, the field stopped debating the question and started acting as if the answer is already settled: mRNA is a platform, and now the goal is to expand the indications, manufacture well, and above all, confirm therapeutic efficacy. 

The science is sprinting ahead into oncology, autoimmunity, rare disease, and delivery to tissues we couldn't reach a few years ago. But across nearly every session, the same theme surfaced from a different angle: The bottleneck is no longer ambition. It's clarity, the ability to see precisely what an RNA construct is doing inside a real biological system. 

From promise to platform 

If there was a single message from industry leaders across the sessions, it was that mRNA has crossed from a modality into a true platform. Regulators confirmed it with their own data: The FDA and EMA are both seeing mRNA submissions climb, with clinical activity shifting toward oncology. In vivo CAR-T, in particular, has gone from a curiosity to one of the most-watched applications at the conference, with regulators noting a distinct uptick in submissions. 

Investors echoed the same conviction. One venture panelist called mRNA one of the few genuinely "platformable" modalities, precisely because it's programmable, and pointed to CAR generated in vivo as the most likely bridge from initial proof points with the COVID-19 vaccines to true therapeutics. There's a business reality underneath the science, too: A wave of patent cliffs later this decade is pushing large pharma to acquire RNA and delivery capabilities now, and much of that appetite is aimed squarely at delivery technology. 

Delivery is still the frontier 

For all the optimism, multiple attendees agreed on where the hard problems live. As one panelist framed it, every time you unlock a new tissue, you unlock a whole swath of therapeutics. Getting beyond the liver, intramuscular injection, and immune cells remains the central challenge. 

Much of the delivery conversation is centered on targeted LNPs. We heard detailed work on antibody-conjugated LNPs to reach extrahepatic tissue, including the tradeoffs between one-step conjugation and surface modification, and the observation that smaller particles may be the key to reaching sites like bone marrow through narrow fenestrations. But one of the most useful comments of the summit was a caution: Even "targeted" LNPs still end up in the liver, so specificity can't be an LNP-only problem. The RNA itself must carry some of the targeting logic, such as through miRNA binding sites that silence expression in the wrong cell types. Delivery, in other words, is increasingly a co-design problem between the particle and the sequence. 

New types of RNA 

The field's vocabulary is expanding, too. Although linear mRNA has set the foundation, self-amplifying RNA (saRNA) and circular RNA (circRNA) constructs are continuing to rise in popularity. saRNA is of particular interest because it lets drug developers lower the delivered dose, which matters when working with a narrow therapeutic window. However, speakers were candid that the replicase doesn't last forever, and that ceiling ultimately limits saRNA's durability. circRNA continues to draw interest for its durability and re-dosing potential.  

Multi-RNA therapeutics are emerging as their own category. One team presented a fibrosis program combining two mRNAs and an siRNA that outperformed any single- or double-agent version, a genuinely combinatorial approach to disease biology. From pediatric brain-tumor vaccines to in vivo CAR programs delivered with unmodified RNA, the diversity of constructs on display was a reminder that "mRNA therapeutics" is quickly becoming an umbrella term. 

Regulators are writing the rulebook and asking for better data 

At the conference, drug developers heard directly from the FDA, EMA, and MHRA on how they are approaching guidance and approvals for RNA therapeutics. The regulators were remarkably aligned, both on where they're headed and on what they still can't see. 

All three are fielding a rising volume of non-infectious-disease submissions and pre-submission meetings, and the MHRA signaled that guidance on individualized cancer immunotherapies is expected later this year. But the most important part for drug developers was how explicit the agencies were about the data they want. The FDA noted that HPLC alone isn't enough to distinguish some critical changes and pointed to the need for complementary technologies. The EMA and MHRA want to understand where mRNA-LNPs actually go, how much is released, and how much release is needed for effect. 

Making it faster, and cheaper 

Manufacturing was the other pole of the conversation, and the throughline was speed and cost without sacrificing quality. Several groups showcased synthetic linear DNA as an alternative to plasmid templates, promising templates and the ability to perform in vitro transcription in days to weeks rather than months and, notably, higher IVT yields with lower double-stranded RNA contamination. Others presented systematic "toolbox" approaches to T7 polymerase and UTR selection, and continuous, closed-system manufacturing aimed at cutting the cost of goods and dsRNA at the same time.  

Speed is only half the mission 

The clearest articulation of why all of this matters came from CEPI and its 100 Days Mission, the goal of developing safe, effective, and accessible vaccines within 100 days of identifying a new pandemic threat. It's an ambitious target, and it reframes speed as a public-health imperative rather than a commercial one; the faster a countermeasure is ready, the smaller the outbreak and the more equitable the access. 

But producing a vaccine quickly is only half the problem. Understanding its quality quickly matters just as much, and today, measuring the critical quality attributes of an mRNA vaccine often means running a battery of separate assays, each one adding time and complexity. A 100-day vaccine is only useful if you can also confirm, in something close to 100 days, that it is what you think it is. That is squarely an analytics problem, and it's one Eclipsebio is proud to be working on with CEPI's support: a proof-of-concept exploring how a single sequencing-based workflow can measure multiple critical quality attributes. Using a single sample, this workflow will measure RNA integrity, dsRNA impurities, and capping efficiency in a one integrated process, replacing a battery of separate tests and reducing reliance on complex infrastructure. 

AI is everywhere, but it is only one part of the solution 

Unsurprisingly, AI-driven RNA design — and its limits — was a major topic. Large pharma teams were candid that today's translation-efficiency models are good but not great, that optimizing for IVT yield can produce underperforming transcripts in cellulo, and that the familiar Pareto front of codon adaptation versus folding energy doesn't hold for every protein. There's even an over-optimization paradox: Cleaning a sequence too aggressively can strip out the very signals a vaccine or oncology construct needs. 

The consensus was healthy. As one investor put it, AI is another tool in the toolkit, not the final solution. The frontier isn't a better sequence model in isolation; it is models enriched with real structural and cellular data, and the wet-lab validation platforms to prove the predictions out. 

The common thread: Analytics 

Looking across all of the sessions, a single problem connects many of them. New modalities, targeted delivery, tighter regulatory expectations, and smarter AI all run into the same wall: The analytical methods we use to characterize RNA haven't kept pace with the complexity of what we're now building. 

The evidence was everywhere. One group described spending four months and multiple PhD-level scientists just to qualify a single assay, and they are still wrestling with method variability. A major pharma's analytical lead admitted that for most of the LNP parameters the field routinely measures, we don't actually understand their impact. Additionally, the AI models everyone is excited about are only as good as the experimental data used to train and validate them. 

The challenges facing modern RNA medicine aren't theoretical hurdles. They're analytical gaps. 

Where Eclipsebio fits in 

This is precisely the space Eclipsebio was built for, the gap between what a construct is supposed to do and proof of what it actually does. Our sequencing-based approach is designed to interrogate RNA at a resolution that matches today's constructs: direct detection of dsRNA impurities through eSENSE dsRNA, base-level mapping of RNA integrity and fragmentation, and readouts that connect a sequence to its real translational and biological behavior. We answer exactly the kinds of identity, purity, and potency questions regulators are now asking sponsors. 

We brought that message to the podium twice. In a pre-conference session, we shared how we use AI to design more effective RNA therapies. In the main program, we made the case for pairing that AI-driven design with sequencing-based characterization to de-risk drug candidates and ground and validate the predictions design models make, using deep, direct readouts of a construct, its dsRNA impurities, its integrity and fragmentation, and how faithfully it is translated. That pairing speaks directly to the tension the AI sessions surfaced: A model is only as trustworthy as the experimental data behind it. Those are the same gaps the regulators, the CDMOs, and the AI teams all described from their own vantage points over the course of the summit. 

Why this matters now 

The field has made its decision: RNA is a platform, and it's expanding into new modalities, new tissues, and new indications faster than the surrounding infrastructure can comfortably keep up. The gating factor for the next wave of programs won't be whether the biology is exciting. It'll be whether developers can see their constructs clearly enough to convince themselves, their partners, and regulators that they work. 

That's the clarity we provide at Eclipsebio. If you're navigating the challenges of modern RNA medicine and want to explore how our platform can support your next program, let's talk. Our scientific team would be glad to help you move toward the clinic with confidence.