Key Takeaways
Modeled aggregate viral vector capacity may exceed demand through 2031, but specialized capabilities are not interchangeable.
The 2025 contraction in the gene therapy pipeline was concentrated in preclinical programs, while clinical-stage counts increased.
Program failures, strategic exits, and portfolio reprioritization can each remove expected manufacturing demand.
CDMOs increasingly compete on process fit, scale-up, purification, analytics, flexibility, and economics rather than capacity alone.
The Capacity Boom Meets a Different Market
The number of current good manufacturing practice (cGMP) viral vector manufacturers worldwide rose from about 30 at the end of 2018 to more than 60 by 2022. By the beginning of 2024, five manufacturers accounted for about one-third of estimated production potential, but the wider market included many more facilities built or expanded in response to anticipated growth in cell and gene therapy.1
The resulting footprint is now being tested against a demand curve that has not developed as smoothly as earlier expansion plans implied. A 2025 capacity model described by BioProcess International projected that viral vector demand could rise from roughly 2,000 batches to 8,000 by 2031 under a straight-line forecast, although the presenter expected a more modest leveling at about 3,000 batches.1 The same analysis found that available capacity could meet projected need and then exceed it, subject to broad assumptions about facility utilization, vector flexibility, and willingness to serve different programs.
Those assumptions matter. The model treated each suite as producing one batch per month for 10 months each year, assumed that all contract development and manufacturing organizations (CDMOs) could produce all vector types, and reserved 25% of production capability for non-GMP work. In practice, most providers operate within narrower areas of expertise, which means that a headline surplus does not guarantee that every program can secure the right process, scale, analytical package, or slot.
The correction therefore has two dimensions. The sector is carrying more physical capacity than before, while buyers remain exposed to shortages in specific technical capabilities. A 2025 review of the adeno-associated virus (AAV) CDMO market characterized capacity as no longer the primary constraint and instead highlighted limited specialized expertise in upstream scale-up, downstream purity, and regulatory execution.2 The market has moved from a general race to add suites toward a more selective contest over which suites can reliably support which products.
Pipeline Attrition Reshapes Demand
Manufacturing demand begins with development programs, but pipeline totals can obscure where contraction is occurring. The American Society of Gene and Cell Therapy (ASGCT) and Citeline counted 2,040 gene therapies in development in the fourth quarter of 2025.3 The ASGCT definition includes genetically modified cell therapies and programs using several delivery and manufacturing approaches.
The distribution by stage was more revealing than the total. From the first to the fourth quarter of 2025, preclinical gene therapy programs fell from 1,432 to 1,227. Over the same period, phase I programs increased from 350 to 397, phase II programs from 319 to 355, phase III programs from 41 to 49, and pre-registration programs moved from 13 to 12. The year’s contraction was thus concentrated in the preclinical segment, while the number of clinical-stage programs continued to rise.
That pattern continued to evolve in early 2026. ASGCT and Citeline reported 2,132 gene therapies in development in the first quarter, including 1,269 preclinical, 423 phase I, 368 phase II, 57 phase III, and 15 pre-registration programs.4 The report stated that program counts increased at every stage compared with the prior quarter. These figures do not establish a simple rebound in viral vector demand, because the data set includes genetically modified cell therapies and other gene therapy approaches. However, they do show why pipeline attrition should be examined by stage, technology, and manufacturing requirement rather than treated as a single market-wide decline.
Individual program decisions can still remove meaningful expected demand. Pfizer reported in June 2024 that its phase III CIFFREO study of fordadistrogene movaparvovec for Duchenne muscular dystrophy did not meet its primary endpoint or key secondary endpoints. The candidate appeared in Pfizer’s May 2024 pipeline and was absent from the company’s April 2025 pipeline update.5–7
Other changes reflected strategy rather than a disclosed trial failure. In December 2024, Sangamo reported that Pfizer had terminated its collaboration for giroctocogene fitelparvovec and did not plan to submit marketing applications or pursue commercialization.8 In July 2025, Sarepta said it would pause several programs, including most gene therapies in development for limb-girdle muscular dystrophy, while refocusing its pipeline primarily on small interfering RNA programs.9
For manufacturers, these cases represent distinct forms of attrition: clinical failure, termination after development work, and portfolio reprioritization. Each can eliminate forecast batches, validation work, analytical demand, or a prospective commercial campaign. Their effects are not captured adequately by a single count of active programs. Capacity planning must account for the probability that demand will disappear, move, or change form before commercialization.
Capacity Must Fit the Program
Viral vector demand is segmented by vector type, process architecture, scale, dose, and development stage. AAV, lentiviral vector, and oncolytic-virus processes do not draw on identical production assets or expertise. Even within AAV manufacturing, differences in producer cells, transfection systems, serotypes, purification strategies, and analytical requirements can limit practical interchangeability.2,10
The 2025 capacity model projected that AAV would account for more than twice the demand represented by lentiviral vectors and oncolytic viruses combined.1 It also projected a 31% compound annual growth rate for commercial demand over seven years, compared with 4% for clinical demand. Yet clinical programs still represented most demand in the model, and the underlying estimates depended on trial timing, subject counts, dosing, and production assumptions.
Clinical and commercial work also place different demands on a network. Early programs may require smaller, flexible campaigns and close process-development support. Commercial supply requires reproducibility, dependable scheduling, validated methods, and economics that remain workable across repeated campaigns. The transition from one stage to the next is therefore not simply a request for more liters. It is a change in operating model, quality expectations, and financial exposure.
This is why aggregate utilization can coexist with bottlenecks. A facility may have available cleanroom time but lack a suitable platform, scale, downstream train, analytical capability, or team for a particular program. Conversely, a technically capable facility can remain underused when the programs it was built to serve are delayed or discontinued. The more specialized the asset, the more consequential the mismatch between nominal capacity and executable demand becomes.
The Network Is Correcting Selectively
Recent facility decisions illustrate a market adjusting through consolidation, exit, and continued targeted investment. Thermo Fisher Scientific said it would close a viral vector facility in Lexington, Massachusetts, transfer its programs to Plainville, and eliminate positions across its Massachusetts network. The company described the changes as resource optimization following expansion at Plainville.11
Rentschler Biopharma took a different path. In January 2025, it announced its withdrawal from advanced therapy production at its Stevenage, United Kingdom, facility and cited slower-than-expected growth and demand in the cell and gene therapy market that did not meet expectations. The company said it would focus on areas of greater demand, with biologics remaining central to its operations.12
Expansion has not stopped altogether. ProBio opened a 128,000-ft2 facility in Hopewell, New Jersey, in June 2025 for plasmid DNA and viral vector manufacturing. At its opening, the company said GMP AAV production was planned for the third quarter of 2025 and lentiviral vector services for the first quarter of 2026.13
These decisions do not support a simple story of uniform contraction. They show a network being reshaped around company strategy, location, platform, and the perceived quality of future demand. Some capacity is being consolidated into larger sites, some providers are leaving the segment, and others are still placing selective bets on integrated capabilities. The central question is no longer whether the industry has enough rooms and bioreactors. It is whether each asset has a defensible role in the programs that remain.
Technical Execution Becomes the Differentiator
When available capacity was scarce, access itself had strategic value. In a looser market, buyers can place greater weight on performance. The AAV manufacturing review by Kaushik identified upstream scalability, downstream purity, and regulatory scrutiny as critical constraints and argued that success depends on strategic alignment, disease-area specialization, speed, quality, and shared commitment.2
Upstream platforms must generate enough material without creating downstream burdens that erase the benefit. BioProcess International reported that movement from adherent to suspension culture, combined with automation and integrated processing, can support commercial-scale production, while purification remains a substantial technical challenge because of vector complexity.1 This makes process fit more important than a facility’s stated maximum scale.
Downstream performance can influence yield, impurity clearance, consistency, cycle time, and the number of batches required. Analytical readiness determines how quickly process changes can be understood and how confidently comparability and release questions can be addressed. These capabilities affect both development speed and the amount of effective capacity a facility can deliver from its installed footprint.
Platform approaches can reduce repeated development work, but they must remain adaptable to product-specific needs. There has been movement toward more standardized, plug-and-play solutions for AAV manufacturing along with a need for increased yields and lower cost of goods as viral vector therapies expand beyond ultrarare indications.10 The competitive advantage lies in using standardization to shorten development while preserving the control needed for a particular vector and product.
Sourcing Strategies Must Absorb Uncertainty
The capacity correction also changes the make-or-buy calculation. An industry survey found that 44% of respondents used CDMOs to augment internal capabilities, about 25% lacked internal CGT manufacturing capabilities, and about 20% outsourced all manufacturing.10 The results describe a market already using several sourcing models rather than one dominant approach.
Internal manufacturing can provide direct control over process knowledge, scheduling, and long-term supply, but it also concentrates fixed-cost and utilization risk. Outsourcing can preserve flexibility and reduce capital commitments, but it requires disciplined technology transfer, governance, and oversight. Hybrid models distribute work across internal and external networks, which can improve resilience while increasing coordination demands.
The most durable strategy is likely to match commitments to evidence. Early programs can preserve optionality through flexible external capacity and clear transfer rights. As clinical evidence, process maturity, and demand visibility improve, companies can make more specific investments in internal assets, dedicated suites, or long-term external relationships. This stage-gated approach treats manufacturing capacity as a portfolio decision rather than a one-time forecast.
Economics Move to the Center
Facility utilization is only one part of the economic problem. A manufacturer can fill suites and still struggle if campaigns require excessive development effort, low-yield runs, extensive deviations, long release timelines, or repeated process changes. The relevant measure is productive, reliable output rather than scheduled occupancy.
AAV economics are also frequently conflated with therapy pricing. Gangurde and Winitsky presented an illustrative estimate of approximately $2 million for a 200-liter current good manufacturing practice AAV drug-product batch, including manufacturing and release testing.14 Under their stated assumption of 200 doses, the estimated manufacturing cost was about $10,000 per dose. They used the example to argue that the prices of gene therapies largely reflect total development costs rather than manufacturing cost alone.
Their proposed responses included streamlining process and analytical development, applying chromatography parameters across serotypes where feasible, and using improved platform technologies to avoid unnecessary reinvention. These measures matter to both developers and CDMOs because better yields, shorter cycle times, and more transferable methods can lower cost while freeing effective capacity.
The correction therefore places commercial discipline alongside technical capability. Providers need to understand which services create repeatable value, which platforms can support multiple programs, and where bespoke work is justified. Developers need manufacturing strategies that can survive delays, discontinuations, and changes in dose or market size. Both sides benefit when contracts, process-development plans, and capacity commitments recognize those uncertainties early.
A More Disciplined Manufacturing Market
The viral vector sector is emerging from its expansion phase with more infrastructure, more experience, and a clearer view of where manufacturing remains difficult. The adjustment is visible in selective closures and exits, but also in continued investment by companies that see an opportunity to build differentiated platforms.
Pipeline data likewise resist a single verdict. Preclinical gene therapy programs contracted during 2025, while clinical-stage counts increased, and the first quarter of 2026 brought higher counts across all stages in the ASGCT and Citeline data set.3,4 At the program level, however, failed trials, terminated collaborations, and portfolio reprioritization continue to remove demand that once appeared in manufacturing forecasts.
The next phase will reward capacity that is technically specific, economically productive, and flexible enough to follow credible programs. Facilities still matter, but their value depends on the process knowledge, analytical strength, quality systems, and operating judgment behind them. The winners after the boom will be the organizations that can convert installed assets into reliable supply without assuming that every pipeline entry will become a commercial product.
References
Abbott, Josh. “Rentschler Pulls Out of CGT Business.” BioProcess International. 31 Jan. 2025.
Gene, Cell, & RNA Therapy Landscape Report Q4 2025. American Society of Gene and Cell Therapy and Citeline. Jan. 2026.
Gene, Cell, & RNA Therapy Landscape Report Q1 2026. American Society of Gene and Cell Therapy and Citeline. Apr. 2026.
Chen, Maggie. “Overcoming AAV Manufacturing Challenges: Movement Toward Plug-and-Play Solutions.” BioProcess International. 9 Feb. 2023.
Gangurde, Rajiv, and Steve Winitsky. “Gene Therapy: Are High Costs and Manufacturing Complexities Impeding Progress?” Parexel. 31 Jul. 2024.
Kaushik, Rahul. “The AAV CDMO Market in August 2025: Navigating the Crossroads of Capacity, Complexity, and Cost.” Cell & Gene Therapy Insights. 11: 939–943 (2025).
“Pfizer Pipeline.” Pfizer. 30 Jul. 2024.
“Pfizer Provides Update on Phase 3 Study of Investigational Gene Therapy for Ambulatory Boys with Duchenne Muscular Dystrophy.” Pfizer. 12 Jun. 2024.
“Pfizer Pipeline.” Pfizer. 29 Apr. 2025.
“ProBio Opens Flagship U.S. Plasmid & Viral Vector Manufacturing Facility in Hopewell, New Jersey to Advance Cell and Gene Therapy.” ProBio. “27 Jun. 2025.
Sangamo Therapeutics, Inc. “Form 8-K.” U.S. Securities and Exchange Commission, 22 Dec. 2024.
“Sarepta Therapeutics Announces Strategic Restructuring and Pipeline Prioritization Plan to Maintain Long-term, Sustainable Growth and Provides Update on ELEVIDYS Label.” Sarepta Therapeutics.16 Jul. 2025.
Scott, Cheryl. “Modeling Viral Vector Capacity: Healthy Growth by 2031.” BioProcess International. 25 Jun. 2025.
Taylor, Nick Paul. “Thermo Fisher Eliminating 160 Jobs, Closing Massachusetts Facility.” MedTech Dive. 15 Nov. 2024.
