State of the Industry 2026 Virtual Panel: Advanced Therapies (pt. 2)

Eytan Abraham (EA): We’re all aware of the macroeconomic headwinds: the funding environment is still difficult and won’t return to the pandemic levels of capital flow, which in hindsight were unsustainable. The challenge now is to bring down costs and shorten turnaround times, especially for autologous therapies and viral vector manufacturing. Current costs per patient just aren’t sustainable. The opportunity — and our responsibility — is to transition from what is still a cottage industry into something more like monoclonal antibodies: scalable, robust, and accessible. The good news is that clinical outcomes remain strong. Approvals are increasing, therapies are moving into earlier lines of treatment, and promising data continues to emerge. That positive clinical trajectory is what will keep driving us forward.

Susan D’Costa (SDC): Viral vectors have taken a reputational hit recently, particularly AAV, following high-profile safety concerns, and investors have grown skeptical. But I think it’s critical not to give up on these platforms. They are naturally suited for payload delivery and still hold enormous promise, especially in areas like ophthalmology and CNS. The path forward is in refining serotypes, payloads, and manufacturing quality. At the same time, we need to think in terms of balance: viral vectors alone won’t solve every challenge, but they remain an essential part of the toolkit alongside non-viral modalities, mRNA, and other therapeutic approaches.

Matthew Hewitt (MH): Cost of goods (COGS) is a constant concern, but we need to be realistic: reducing COGS doesn’t necessarily reduce therapy price. There are examples already on the market where pricing hasn’t shifted even when COGS did. What really matters is patient value. Some therapies face demand constraints because patients don’t see enough benefit to switch from standard of care. In contrast, certain cell therapies are growing rapidly because the value proposition is simple and compelling: take this therapy or you don’t survive. That’s not the model we want for every product, but it highlights the importance of clearly defining value from the patient’s perspective.

Christian Cobaugh (CC): I think delivery is key to both cost and value. Many cell therapies today are as much a procedure as a product. If innovations in delivery can make them more product-like — for instance, more targeted injectables that eliminate the need for mobilization or conditioning — we’ll not only reduce COGS but also improve adoption by patients and payers. We’re also starting to see hybrid approaches, combining viral and non-viral elements. That diversity of delivery technologies could tip the balance in favor of broader patient uptake.

Emmanuel Abate (EA2): We shouldn’t expect one workflow to dominate; there will be a diversity of approaches. What excites me is that, despite challenges, the number of patients treated continues to grow. At Cytiva, our consumables business reflects this directly: more patients in trials and on marketed therapies means greater demand. CAR-T, oligos, and even viral vectors are seeing progress, and non-viral approaches like mRNA are scientifically strong even if adoption is still catching up. The trajectory is positive.

Vikas Gupta (VG): I’d just add a note of caution on RNA. Recent political statements in the U.S. questioning RNA platforms — even if walked back — have cast some doubt in the minds of investors and start-ups already struggling with limited runways. That creates additional headwinds for an emerging modality that still needs time to mature. RNA processes are still evolving, with many submodalities and delivery carriers being tested, but the field needs patience to reach full potential. Pressure from private equity ownership or short funding cycles risks forcing progress faster than is realistically possible.

Daniella Kranjac (DK): As the only investor here, I try to look at the field through two lenses: where the science is breaking new ground and where it is becoming investable and scalable. Many of today’s challenges — from regulatory recalibration after recent safety concerns to CMC comparability and raw material bottlenecks — aren’t really scientific; they’re economic and operational. We’re also still short of real commercial successes, which puts a chilling effect on investment.

From an investor’s standpoint, the real opportunities lie with platforms and infrastructure that can both deliver clinical impact and withstand the capital and regulatory intensity of scaling. Advanced therapies are no longer “science projects.” They’re starting to look like industries, which means they require supply chains, standards, and sound economics. The companies that can industrialize while keeping patient benefit front and center will be the winners.

EA: bluebird bio is a great example. Clinically, their therapy worked, patients were waiting, and risks were manageable. Yet the company struggled commercially until private equity stepped in. That shows two things: strong clinical data creates opportunity, but legacy players don’t always succeed at scaling. The challenge isn’t efficacy; it’s supply chain, cost of goods, scalability, and profitability.

MH: Sarepta illustrates this tension too. Their revenues doubled after expanding indications, which showed demand was there. But the FDA relationship and patient population complexities created turbulence. It highlights that gene therapies aren’t biologics. Regulators and the industry still need to adapt to therapies that permanently change biology, not just transiently affect it. To their credit, regulators have shown pragmatism in certain cases — bluebird’s Skysona, for example — but market confidence remains fragile.

SDC: Commercialization doesn’t equal penetration. Roktavian is approved for hemophilia, but uptake has been slow because patients are already reasonably managed by existing therapies, and the gene therapy comes at a very high price. Contrast that with Zolgensma: despite known risks, families take it because the alternative is a child’s death. It’s about risk–benefit and unmet need. Sarepta’s safety events caused outsized concern partly because of heightened scrutiny on viral vectors, but patient advocacy groups continue to push strongly because the need is real.

DK: Exactly — commercial success is not just regulatory approval. Investors see this on the tools side as well: incremental improvements won’t cut it. A new bioreactor, for instance, needs to be an order of magnitude better. The same goes for scaling technologies — we need real step changes, not just tweaks.

EA: The bigger issue is adoption. Many innovative tools already exist, but too often we still see manual, open processes carried into commercialization. Until the industry adopts state-of-the-art systems, scalability will lag.

SDC: That’s partly a matter of timing. Companies under pressure to commercialize often lock in with half-developed processes. Some hemophilia gene therapies are still made in roller bottles or wave bags because of the rush to market.

MH: Technology isn’t the only barrier; value proposition is important. Severe hemophilia patients already consume hundreds of thousands in healthcare annually, but they’re stable. Why would they take a one-shot gene therapy if they’re told 10 more are in the pipeline? That’s why products like Yescarta succeeded despite clunky early manufacturing: the patient proposition was clear. If the choice is “take this or die,” adoption is rapid. Where current standard of care suffices, the hurdle is much higher.

DK: From an investor’s perspective, we don’t invest directly in therapeutics; our focus is on enabling tools and technologies. We tend to react to what’s happening across biotech and ask: What platforms will be needed to make those new modalities commercially viable?

That said, there are certainly niche therapeutic applications where advanced therapies make sense, but when we’re evaluating opportunities, we’re also looking at scalability and breadth. A platform that works across multiple modalities — say, for both cell and gene therapies, or even with biologics — is a far stronger investment case than one with very narrow utility.

EA2: It’s interesting, because we often think of gene therapy in the ultra-rare, life-or-death context, but the irony is that the most widely delivered advanced therapy in history is the mRNA COVID vaccine — a modality that barely existed five years ago. It showed what’s possible: rapid, low-cost manufacturing at global scale. That’s why some of us get frustrated by the hesitancy around mRNA outside of vaccines. The science has proven itself — the adoption just hasn’t followed yet.

VG: I agree but would add some nuance. COVID vaccines were an outlier. While the processes may be faster, the costs are still significant. Critical raw materials for RNA production remain scarce and expensive, and only a handful of suppliers dominate the market. Startups working in RNA are heavily dependent on CDMOs, and the pricing pressure is intense. Until that stabilizes on both the developer and CDMO side, it will continue to slow the pace of development.

EA: Most new modalities start in life-or-death settings because that’s where payers will reimburse almost anything. But we’re already seeing movement beyond that. Autoimmune diseases are an emerging target, and in South Korea, Medipost has treated tens of thousands of patients with a cell-based therapy for knee conditions — hardly life-or-death. The issue isn’t clinical potential; it’s whether we can manufacture at a cost that payers will support and patients will accept. My belief is that we’ll see more non-lethal indications over time, but only if we solve the COGS and scalability challenges first.

EA: Absolutely. It’s a tragedy that many ultra-rare conditions still don’t have viable treatments. Look at the case of SCID, where Orchard ultimately returned the license to UCLA, and patients were left stranded. We need to find a way to deliver therapies for those ultra-rare populations, and I do believe there’s a business case if we can approach it correctly. Initiatives like FDA’s platform designation are part of the solution. No one player will solve it alone, but as an industry we all share some responsibility to make sure these patients aren’t left behind.

MH: We’ve already seen some of this thinking on the cell therapy side. Early CAR-T approvals went after small, narrowly defined populations and then expanded stepwise through supplementary BLAs. On the other end, multiple myeloma programs have shown what happens when you go after a much larger, more homogeneous patient group: demand skyrockets, and the manufacturing side struggles to keep up.

At Charles River, we work with both ends of the spectrum — ultra-rare “n-of-one” therapies and larger patient populations. In conversations with the FDA, EMA, and MHRA, one of the biggest questions raised was whether we need to formally stratify indications into non-rare, rare, ultra-rare, and n-of-ones. That matters because not everything makes sense for a CDMO model. Running a program that only requires 20 or 30 doses per year is not viable for a commercial facility. We need alternative pathways for those patients, whether through academic manufacturing, distributed models, or new regulatory frameworks.

CC: I’d push back a little, especially on how platform designations are currently applied. Today, they’re still very restricted — usually reserved for innovators or CDMOs already running commercial programs. If we extended that designation into earlier stages of development and manufacturing, it could make ultra-rare programs far more feasible. The biology is often validated, the safety is understood, and the real barriers are economic and operational.

For that to work, we’ll need regulators to embrace more uniform, flexible approaches — for example, accepting closed, modular systems that can operate outside of a full cleanroom and potentially at the point of care. Democratizing platform designations and modernizing facility expectations could open the door for ultra-rare and n-of-one therapies.

SDC: Sarepta once held a platform designation for its AAV program, which allowed them to move quickly into a suite of limb-girdle muscular dystrophy programs without repeating every process validation step. But when Elevidys ran into safety issues, the FDA revoked the platform designation. Overnight, those limb-girdle programs were essentially orphaned, leaving patient advocacy groups devastated.

That’s the risk when platform designations are too narrow and tied to a single program. At CDMOs, we try to apply our own “platform-like” approaches for early programs, taking an agnostic, pre-validated process straight into GMP. The result may not be perfectly optimized for every construct, but it allows a therapy to reach patients quickly with limited doses. It’s a “you get what you get” situation, but it can be life-changing for those first-in-human cases.

Stewart McNaull (SM): From what I’ve seen in cell therapy, I don’t think one universal platform is realistic. CDMOs need flexibility. You might be able to define a narrow market segment and build a dedicated process around that, but you won’t get enough customers to fill it broadly. What’s encouraging, though, is the progress in equipment technologies. Groups like Ori, Cytiva, Avectus, and others are bringing forward innovations — but for these to take hold, it really does take a three-way partnership among the innovator, the technology company, and the CDMO. You need upfront proof of concept, diligence, and investment before innovators will adopt. It may mean more money and a few extra months of development, but once it’s up and running, the payoff is a more automated, integrated, and scalable process.

From my perspective, “platforming” is most practical at the level of unit operations and analytics. If you can standardize and modularize individual steps, then you can piece them together flexibly as different programs advance. That’s the kind of pragmatism that will allow us to move programs quickly without forcing a one-size-fits-all approach.

MH: I actually think we focus too much on manufacturing. In biologics, for instance, no one really cares which specific transfection device you used — what matters is the product that comes out the back end, and whether it meets analytical release criteria. We’ve demonstrated this ourselves: testing a CRISPR nuclease knockout across six different transfection systems, we saw nearly identical efficiencies. The device didn’t matter; the analytics did. That’s why I believe analytics should be where we invest in standardization and platforms. Right now, the field suffers from a lack of transparency. Anyone who’s looked at a BLA filing knows how heavily redacted the CQA tables are. We can’t build true platforms if everyone keeps methods and acceptance criteria secret. Biologics succeeded because of much greater data sharing — advanced therapies need to move in that direction.

VG: Analytics is exactly where we’ve been focusing. Through an FDA grant administered via MIT, we’ve been developing a continuous RNA manufacturing platform. But instead of defining continuous strictly in terms of upstream and downstream processing, we put the emphasis on inline analytics. That’s what innovators consistently told us was holding back faster production and faster release.

We’ve been sharing our data with the FDA regularly for two years, and now the Gates Foundation is supporting us as well. Early results are promising. I genuinely believe inline analytics is the most powerful lever we have to accelerate innovation and make RNA therapies scalable.

SDC: We’ve been investing heavily in synthetic DNA as an alternative to plasmid DNA. The expectation is that it should reduce cost of goods and accelerate timelines — no master cell banks required, and potentially an easier regulatory path. The challenge is that the data so far show it’s not yet “like-for-like.” If productivity is even cut in half, you need double the reactor volume, which cancels out much of the cost advantage once you factor in consumables and media. Some approaches, like doggybone DNA, still rely on plasmid DNA as a starting point, so they don’t fully solve the problem. Others, like rolling-circle methods, may hold more promise. But we’re not there yet. Synthetic DNA needs to close the performance gap before it can deliver on its potential.

EA: Another example is the persistence of adherent-based processes. Today, there are commercial products — MSCs, lentiviral vector-based therapies — still being manufactured in adherent systems like cell stacks and factories. It’s understandable, because that’s how those programs began, and it’s hard to switch midstream. But from a cost, safety, and scalability standpoint, it makes no sense. These products need to transition to suspension systems in bioreactors. That shift alone would yield major gains in efficiency and consistency.

EA2: Automation is critical, not just conceptually but in terms of how much and how quickly we invest. At Cytiva, we’re pushing hard on producer cell lines for viral vectors like HEK293, which we see as a real step forward even if adoption is difficult. Advances in CAP/CC, lipid nanoparticle design, and AI-enabled high-throughput screening also have strong potential. But analytics remains a bottleneck. Release testing for therapies like CAR-T still takes too long. Faster, more robust analytics could unlock tremendous efficiencies.

MH: For many commercial programs, viral vector supply is a limiting factor. Companies have been public about not having enough to support their cell therapy pipelines. Closed processing is another inevitability. Regulators are already asking companies without closed systems, “When are you going to close this process?” That’s becoming a compliance expectation.

Beyond that, we’ll eventually need full manufacturing automation. Process automation at the bench will help, but when you start talking about 40,000 doses a year, you’re dealing with millions of logistics “touches.” That’s not sustainable with human labor alone. Robotics will be required, whether linear-track or autonomous, and distribution of manufacturing capacity may become just as important as scaling it centrally.

AI also has a role to play. I’m skeptical about its use in GxP decision-making — regulators hold people accountable, not algorithms — but it can accelerate SOP and risk-assessment drafting. One area where I think AI could be transformative is scheduling. Airlines routinely overbook by 20–30% and still keep flights running smoothly. We’ll need similar intelligent scheduling models to manage capacity for advanced therapy manufacturing.

CC: I’ve spent much of the past decade in the non-viral world, particularly with mRNA medicines. The reality is that IV-injected LNPs still only reliably reach a couple of tissues. There are breakthroughs — Capstan, for example, has shown promise in retargeting, though it remains to be proven clinically. Interestingly, Capstan’s competitors are pursuing engineered viral vectors. If those vectors can be made safer and their integration more precisely controlled, we could turn currently complex cell therapy “procedures” into medicines that can be administered like routine infusions. Personally, I hope ex vivo approaches eventually fade. They remain brutally invasive, not just in terms of toxicities but real mortality risk. The future is productized therapies that don’t require turning the patient into the incubator.

SDC: We’re still missing the right stakeholders at the table. Early in viral vector development, immunologists were nowhere in the conversation, and we paid the price when immune complications surfaced. Viruses evolved to cause disease, not deliver payloads, so of course they default to the liver. That means we need to be engineering capsids for more specific tropisms — detargeting from the liver and homing to the right cells or tissues. In my last company, we exploited cerebrospinal fluid dynamics for delivery into deeper brain regions. The science is moving in that direction, but it requires broader collaboration across immunology, fluid dynamics, bioengineering, and more.

MH: As an immunologist, I’ll echo that. AAV’s core limitation isn’t just liver tropism; it’s that it’s a weak integrator. That’s why multiplicities of infection are so high and why lentivirus remains the preferred vehicle for integration. AAV is well-tolerated at low doses but becomes immunogenic at high MOI, and once you generate neutralizing antibodies, you may lose redosing for life — that itself is a kind of adverse event.

Ex vivo avoidance is attractive, but in vivo transduction creates its own risks. When those edited cells expand inside the patient, they can release unpredictable waves of cytokines. LNPs aren’t a panacea either; they can be acutely toxic at higher doses, and sometimes there isn’t any pathology to explain sudden mortality. The trade-offs between viral and non-viral platforms remain complex, and regulators are rightly demanding deeper insertional analyses as we push forward.

EA: There’s no silver bullet. Solutions will remain indication-specific, varying by target tissue, patient age, and context. We’ve seen major efforts to reengineer AAV tropism; they haven’t panned out yet, but the biology is simply hard. Ultimately, different vectors and chemistries will coexist, and the right answer will always depend on the clinical problem being solved.

SM: It’s hard not to see in vivo as the logical endgame. If you compare it with the burden of ex vivo — chemotherapy, patient prep, cleanroom processing — the idea of doing more of this directly in the patient makes a lot of sense. However, ex vivo therapies remain essential today.

On the autologous versus allogeneic question, allo clearly has advantages: simpler scheduling, more efficient manufacturing, and potentially lower costs. Whether those savings will translate into lower drug prices is less certain, but from an operational standpoint it’s very attractive. The question is when —and if — we’ll see the clinical data and commercial adoption catch up.

MH: The economics are complicated. Take Mesoblast: their full course of treatment runs about $1.4 million, more than some autologous CAR-Ts. That shows allo isn’t automatically cheaper. Roughly 80% of these therapies are still donor-derived, which creates ongoing donor campaign costs. And the upfront spend for an allo program is much higher: engineering, process validation, and large-scale runs all before you’ve treated a patient.

There are also practical questions: will hospitals accept the liability of holding frozen stocks onsite? How efficiently can product be allocated across treatment centers? None of these are insurmountable, but they add friction.

That said, I’m not discouraged. Remember, all the first-generation autologous CAR-Ts failed too, mostly for lack of durability. It took second-generation constructs to succeed. I think allo is following a similar trajectory — the early programs are rocky, but I fully expect the next wave to succeed.

EA: It’s important to remember that “allo” is a huge umbrella. At one end, you have MSCs, which have been used for decades with clean safety records. At the other end are highly gene-edited allo CAR-Ts that raise real safety and durability concerns. We shouldn’t lump them all together.

Using Mesoblast as an example, part of their cost problem is manufacturing. They’re still relying on adherent systems like cell factories, which makes COGS unsustainably high. Moving to scalable suspension bioreactors would significantly reduce costs, even if prices don’t fall in parallel.

The broader vision is compelling: hospitals pulling a vial from a freezer and treating patients as with traditional biologics. We’ll get there — but it will take both technological advances and proof of clinical durability before allo can really claim its place alongside or even ahead of autologous therapies.

SM: I haven’t thought about sustainability in the cell therapy context as much as in large-scale biologics. The scale is so different. For example, an autologous cell therapy batch might generate enough single-use plastic to cover a conference room table an inch or two deep. In contrast, a 2,000-liter antibody batch would fill that same room almost to the ceiling. So, while plastic waste is real, the volumes in cell therapy are modest, and I’m not sure it’s the most pressing sustainability concern here.

EA2: I agree. Sustainability is a huge theme across pharma — most major players have made strong commitments, and at Danaher we certainly have as well. But in advanced therapies, the small production scales mean this isn’t where the main sustainability battle will be fought. Plastic is a critical enabler of these therapies — it makes small-scale, end-of-one, and highly flexible manufacturing possible. Over time, I think we’ll find smarter ways to manage or even reduce plastic use, but for advanced therapies today, I don’t see it as the central issue.

Mechanical recycling — repurposing plastics into things like pallets or benches — is one path, though not a perfect solution. A more forward-looking approach is developing next-generation plastics that are biodegradable or made from eco-friendly sources. Beyond plastics, we shouldn’t forget the role of chemistry — everything from membranes to resins could be redesigned with sustainability in mind. Finally, logistics are a big issue. Our industry relies heavily on air freight, which contributes significantly to the carbon footprint, and that’s an area where we can and should do better.

MH: In Europe, regulators are already putting pressure on suppliers through taxes on single-use plastics, so the financial incentive is real. We tried regrinding cartridges used in endotoxin testing, but impurities remain, and sterilizing reground plastic consumes more energy than it saves. Instead, we launched a recycling program: spent cartridges are now turned into things like park benches. Other forms of sustainability are equally important, like moving away from animal-derived raw materials. Recombinant replacements for serum or horseshoe crab blood not only make processes more reliable; they also reduce ecological impact. But with single-use systems in manufacturing, we must be realistic — patients aren’t going to accept therapies produced in recycled plastic.

DK: Years ago, when I was at a major biopharma, an internal analysis showed the company was using more plastic in its cafeterias than in its single-use manufacturing systems. That doesn’t mean we shouldn’t do better, but it reframes the discussion. You also need to consider the full life cycle: stainless-steel systems consume vast amounts of water, chemicals, and energy for sterilization, which can be more damaging overall. It’s not as simple as “plastics bad, steel good.” You need a true mass balance.

CC: Energy use is probably the bigger lever, especially with personalized medicines. Maintaining cleanrooms with constant air exchanges is far more resource-intensive than the small amount of plastic used per batch. Facility design and the shift to isolator-based systems — smaller, more contained, less energy-intensive — could deliver far greater environmental benefits.

SM: Automation can help us reduce operator involvement and allow for smaller, lower-grade cleanrooms. But retrofitting existing facilities takes time and capital, so it’s something the industry needs to plan for deliberately.

SDC: We shouldn’t forget buffers. Even though they’re contaminated, there may be opportunities to recycle or reclaim them rather than dumping them into wastewater streams. There’s ongoing research, especially in Europe, into how recovery and reuse strategies could make bioprocessing more sustainable.

EA: For me, two things stand out. First, the lack of data sharing — everyone keeps information siloed, which slows collective progress. Second, the absence of standards. We don’t have common frameworks across the field, and that’s a big gap. The more we can drive standardization and share even basic data, the faster we’ll move forward together.

SDC: I agree, and I’ll add one that bothers me: vector genome measurements. We all talk about genome copies per milliliter as though they’re comparable across studies, but they’re not. Different groups use different amplicons, so the numbers aren’t apples-to-apples. Yet we still cite them side by side as if they are. That lack of a common denominator makes it very difficult to assess doses or draw meaningful comparisons across programs.

MH: It’s not just genome copies — we can’t even agree on the method. Multi-probe dPCR shows that only about two-thirds of “full” particles are actually infectious. Neutralizing antibody assays vary wildly; at Penn, we used NAB50, Spark used NAB99, which meant their “clean” patients had no detectable antibodies at all, while ours had higher thresholds. Without standards, the same data can be interpreted very differently. On top of that, we need to better educate investors. In oncology, they’ve come to expect high response rates, but in gene therapy, the bar should be different. Sometimes stable disease is a success, but if investors label such outcomes as failures, it risks stigmatizing entire modalities, as we’ve seen with lentiviral vectors.

EA2: Part of the problem is the volatility in the U.S.: policy swings, leadership changes, shifting priorities. It makes it very hard for investors to know what will stick and what won’t, and that uncertainty filters into our space.

CC: Public financing has become a total wild card. The pullback came at perhaps the worst time — venture capital is already risk averse, and early partnering has slowed. For those of us building cutting-edge platforms, we’re going to have to balance that exposure with steadier plays, just to weather this funding climate. Patience and capital efficiency will be critical until financing stabilizes again.

To see how these advanced-therapy challenges compare with trends in biologics, small molecules, and drug product manufacturing, explore the full Nice Insight 2026 State of the Industry report.