Cell and Gene Therapy’s Manufacturing Bottleneck: The Unsexy Investment Thesis

A Science Story and an Operations Story

More than twenty cell and gene therapies now carry market authorization across the major regulatory jurisdictions, and upwards of two thousand active clinical trials are testing the next wave. Spending on the manufacturing side of this industry reached $5.9 billion in 2023, up 38 percent from the year before, and every credible forecast has that number compounding at somewhere between 17 and 29 percent annually through the mid-2030s, depending on which research house’s model you trust.

On the science, the field is unambiguously working. On the plumbing required to actually make and deliver these products, it is not — and that gap is where the more interesting venture opportunity has been quietly sitting for several years now. It is worth stating plainly why this distinction matters more here than in almost any other pharmaceutical category: a small-molecule drug, once its synthesis route is validated, can be manufactured essentially identically whether the batch size is one kilogram or one thousand kilograms. A living cellular therapy cannot. Scale, in this category, does not simply mean building a bigger version of the same process — it frequently means solving an entirely different and harder engineering problem than the one that got the therapy through its Phase 1 trial in the first place.

Where the Bottleneck Actually Lives

The bottleneck has several distinct anatomical points, and treating it as a single undifferentiated “manufacturing problem” obscures where the actual investable opportunities sit. Viral vector production, the delivery mechanism for most gene therapies, remains inefficient: yields are inconsistent, batch-to-batch variability is high, and scaling a process validated in a research lab up to commercial volume is its own multi-year engineering project, frequently requiring a near-total redesign of the production process rather than a simple scale-up of the original method.

Autologous cell therapies — where a patient’s own cells are extracted, engineered, and reinfused, the model behind most approved CAR-T products — are structurally resistant to the economies of scale that make every other pharmaceutical category cheaper over time, because each batch is, by definition, a batch of one. There is no equivalent of building a bigger reactor to produce more product per unit of labor and overhead; every single patient requires its own complete manufacturing run, with its own quality control, its own chain-of-custody documentation tracking the product from the patient’s own apheresis collection through to reinfusion, and its own opportunity for the kind of process failure that, in a conventional pharmaceutical batch, would simply mean discarding a lot of product rather than failing an individual patient’s only treatment attempt.

And building a new GMP-compliant manufacturing suite from scratch takes eighteen to thirty-six months and tens of millions of dollars before it produces a single usable dose, which means capacity additions will structurally lag clinical demand for the foreseeable future, almost regardless of how much capital gets thrown at the problem. That lag is worth dwelling on: it means that even a company with unlimited capital and a fully validated, FDA-approved therapy cannot simply will more manufacturing capacity into existence on a timeline shorter than roughly two years, a hard physical constraint that has no real analog in most other technology-driven industries, where capacity can often be added far more quickly by simply provisioning more compute or more conventional factory floor space.

The Cost Reality

The cost consequence of all this is stark and worth stating plainly: manufacturing alone can exceed $100,000 per patient for an autologous therapy, a figure that strains even the wealthiest health systems and is simply prohibitive across most of the geographies where OceansGled actually operates. A curative gene therapy that costs more to manufacture than most people earn in several years is not a commercially viable product outside a narrow band of the world’s health systems, no matter how compelling the clinical data. Closing that gap is not a marginal engineering exercise — it is the central commercial problem the entire category has to solve before it can matter to the roughly six billion people living outside the US, Western Europe, and a handful of wealthy Asian markets, and it is worth being honest that no company or technology platform has yet solved it convincingly at scale.

The Investable Infrastructure Layer

This is exactly where the investable infrastructure layer has emerged. A decade ago, a handful of specialist contract development and manufacturing organizations served this category. Today there are more than 125 companies specializing specifically in cell and gene therapy manufacturing, and the market for that manufacturing-services layer is growing considerably faster than the underlying therapy market itself — forecasts put it compounding at roughly 26 to 29 percent annually into the mid-2030s. That is the tell: when the picks-and-shovels layer of an industry is growing faster than the industry it serves, it usually means the bottleneck, not the demand, is the binding constraint, and capital flowing into the bottleneck has an unusually clear thesis behind it, one that does not depend on picking which individual therapeutic asset among dozens of competing clinical programs will ultimately win approval.

That last point deserves emphasis for how it changes risk profile. A venture investor backing a specific gene therapy asset is making a binary bet on that asset’s clinical trial outcome, a genuinely high-variance wager even with excellent diligence. A venture investor backing the manufacturing infrastructure layer is instead betting on the category’s aggregate demand curve, which is diversified across dozens of underlying therapeutic programs and considerably less exposed to any single trial’s binary outcome — a meaningfully different, and in our own assessment more attractive, risk-return profile for a venture fund operating with a twenty-five-company concentration discipline.

Allogeneic Versus Autologous: A Manufacturing Question, Not Just a Clinical One

Within that manufacturing layer, the more durable opportunity sits with allogeneic, “off-the-shelf” approaches over autologous ones, even though autologous CAR-T remains the category with the deepest clinical track record. Allogeneic manufacturing standardizes quality control, eliminates the per-patient batch problem entirely, and is currently estimated to account for roughly 62 percent of manufacturing market share for exactly that reason — a single allogeneic production run, sourced from a healthy donor rather than the patient themselves, can in principle supply dozens or hundreds of treatment doses, converting the economics of the category from the batch-of-one model discussed above into something closer to a conventional biologics manufacturing process.

It is also the more geographically portable model — a regional manufacturing hub producing allogeneic product for a whole population is a far more realistic build for an emerging market than a bespoke, per-patient autologous supply chain requiring apheresis centers within hours of every treating hospital, since apheresis, the process of collecting a patient’s own immune cells for engineering, is itself a specialized procedure requiring equipment and trained staff that most hospitals across our own operating geographies simply do not yet have.

Regulators Are Adapting, Slowly

Regulators are, slowly, adapting to the category’s specific needs rather than forcing it through frameworks built for small-molecule tablets. The FDA announced increased manufacturing flexibility for cell and gene therapy clinical and commercial quality control testing in January 2026, an acknowledgment that treating a living therapeutic product exactly like a chemically stable pill was never going to work — a small-molecule tablet can sit on a shelf for years and be tested against a fixed chemical specification; a cell therapy product has a shelf life measured in hours or days and requires quality-control methods that can return a result fast enough to actually matter before the product itself degrades.

Harmonizing that flexibility across the FDA, EMA, and Japan’s PMDA simultaneously remains the harder, slower problem, and it is a genuine source of cost and delay for any company trying to manufacture and distribute globally rather than market by market — a company that has validated its manufacturing process to FDA standards frequently has to run a separate, only partially overlapping validation process to satisfy EMA or PMDA requirements, effectively multiplying the regulatory burden by the number of major markets a company hopes to serve.

Where We Are Actually Looking

The venture opportunity we find most compelling here is not the tenth CAR-T company chasing the same hematologic malignancy that six better-capitalized competitors are already pursuing. It is the layer underneath: vector yield optimization, closed-system automation that reduces the labor intensity of cell processing, digital-twin process modeling that shortens the eighteen-to-thirty-six-month suite commissioning timeline, and regional manufacturing hubs built to serve populations that the current US- and Europe-centric supply chain has no realistic plan to reach. None of it is glamorous. All of it is the actual constraint standing between a validated clinical result and a patient receiving the therapy, and all of it benefits from the kind of patient, infrastructure-oriented underwriting a concentrated fund is better positioned to provide than a generalist growth investor chasing the next headline clinical readout.

The Closing Argument

Every gene therapy that fails to reach a patient because of a manufacturing queue, rather than a failed trial, represents a clinical success being wasted on an operations problem. For a concentrated fund willing to look past the therapeutic-area headlines, that gap — unglamorous, structurally necessary, and genuinely underinvested relative to the science it supports — is exactly where the discipline of this portfolio should be pointed, and it is a thesis that, unlike a bet on any single therapeutic asset, does not require us to be right about which specific disease target ultimately wins.