The cell and gene therapy CDMO market is valued at roughly $6 to $8 billion today. By 2034, analysts project it will hit somewhere between $52 billion and $88 billion. That's a CAGR of 24 to 28 percent, depending on whose numbers you trust. Over 2,000 clinical trials are running globally. Ninety percent of biotech companies depend on contract manufacturers to make their products. Lonza, Catalent, Thermo Fisher, Samsung Biologics, WuXi AppTec, AGC Biologics, Charles River Labs, Resilience, OmniaBio... the list of players grows every quarter.
Everyone talks about capacity. Everyone talks about capital. Everyone talks about viral vector supply constraints and cooling investor sentiment.
Almost nobody talks about the thing that will actually determine who wins and who chokes: training.
I've spent years building and fixing training programs inside CDMOs. I've watched companies raise hundreds of millions of dollars to build gorgeous facilities, buy cutting-edge equipment, and then struggle to get products out the door because they couldn't train operators fast enough. I've seen the same mistakes repeated at site after site, company after company. And I'm watching the industry barrel toward a 7 to 10x scale-up without solving the most fundamental problem it has.
You cannot manufacture cell and gene therapies without qualified people. Full stop. Every new facility, every new client, every new product, every scale-up requires trained operators who do not yet exist. The CDMOs that figure out training will dominate this market. The ones that don't will drown in deviations, audit findings, and client attrition.
Let me walk you through why this problem is so much harder than most people realize.
Why CDMOs Are Fundamentally Different from Sponsor Companies
If you've only built training programs at sponsor companies, you're going to get wrecked at a CDMO. I don't say that to be dramatic. I say it because I've watched it happen.
At a sponsor company, you make your product. Your SOPs. Your batch records. Your quality system. One set of expectations, one set of standards, one product family (usually). Training is hard, but the scope is bounded. You know what you're training people to do because you know what you're making.
A CDMO makes everyone else's products. That changes everything. Your operators need to execute different processes for different clients under different quality agreements with different batch records and different expectations, sometimes in the same week. Your training system has to handle all of that complexity without collapsing under its own weight.
The mental model shift is enormous. At a sponsor company, training is about depth. At a CDMO, training is about depth AND breadth AND the ability to switch contexts without making errors. That is a fundamentally different challenge, and most training systems aren't built for it.
Multi-Client Chaos
Here's what the daily reality looks like. Client A requires aseptic gowning qualification every six months. Client B says annually is fine. Client C wants a specific gowning procedure that differs from your house standard. Client D has batch records that use terminology your operators have never seen because they came straight from a European sponsor's internal system.
Every quality agreement is different. Every client audit checklist is different. Every set of expectations around what "trained" means is different. Some clients want documented OJT with three successful witnessed performances. Others want classroom training plus a written exam. Others want all of that plus periodic requalification on a cadence that doesn't match anyone else's.
Your training system has to track all of this without errors. Your operators have to know which set of rules applies to which product on which day. Your trainers have to be fluent in multiple clients' expectations simultaneously.
This is not a minor administrative challenge. This is the core operational complexity of CDMO training, and most companies are managing it with spreadsheets and prayer.
The Turnover Crisis
Let's talk about the number that keeps CDMO training managers up at night: 30 to 50 percent annual turnover in manufacturing operations.
Think about what that means. If you have 100 operators and you're turning over 40 of them per year, you need to fully train 40 new people every year just to stay even. You're not building capacity. You're not advancing your capabilities. You're running on a treadmill.
And it gets worse. The people leaving aren't random. They're often your most experienced operators, the ones other companies poach because they're already trained. So you lose your best people and replace them with people who know nothing. The knowledge walks out the door and the training burden walks in.
The real cost isn't just recruiting and onboarding. It's the deviations from undertrained staff. It's the batch failures. It's the client dissatisfaction. It's the audit findings. It's the experienced operators picking up slack and burning out and then leaving too.
This is a vicious cycle, and most CDMOs are stuck in it. They underfund training because they're focused on capacity and revenue. The underfunded training leads to poor retention. The poor retention leads to more training burden. The increased training burden on an underfunded system leads to worse outcomes. Repeat.
North America dominates the CGT CDMO market at about 41 percent market share in 2024, concentrated in hubs like Massachusetts, California, and North Carolina. Those hubs are also the most competitive labor markets in the industry. Operators in the Research Triangle or Cambridge can walk across the street and get a raise. If your training program is bad and your culture is worse, they will.
Core vs. Client-Specific Training Architecture
The solution to multi-client complexity starts with architecture. You need a two-tier training system: core competencies that apply to everything, and client-specific modules that layer on top.
Core training covers your house standards. Cleanroom behavior. Aseptic technique. Your quality system. Your deviation process. Your CAPA system. Equipment operation for shared platforms. GMP fundamentals. Data integrity. These don't change based on which client's product is on the line.
Client-specific training covers everything unique to a particular program. Their batch record format. Their in-process testing requirements. Their specific acceptance criteria. Their terminology. Their escalation expectations.
The architecture has to be modular. When a new client comes on board, you shouldn't be rebuilding training from scratch. You should be plugging in a new client-specific module on top of your existing core. When an operator moves from Client A's product to Client B's product, the delta should be clear and bounded: here's what's different, here's what's the same, here's what you need to learn.
Most CDMOs I've seen don't have this architecture. They have a pile of SOPs, a pile of batch records, and a vague expectation that trainers will figure it out. That works at 50 people. It collapses at 200.
The Multi-Product Training Matrix
Cell therapies. Gene therapies. Viral vectors (lenti, AAV, adeno). Plasmid DNA. mRNA. Each product type has different manufacturing platforms, different critical process parameters, different failure modes, and different regulatory expectations.
Now multiply that by the number of clients. Now add the fact that operators may need to work across product types depending on scheduling. Now consider that some of these products are autologous (patient-specific, no room for error, chain of identity is everything) and some are allogeneic (scaled, but still complex).
Your training matrix becomes enormous. An operator qualified on AAV viral vector production for Client A is not automatically qualified for lentiviral vector production for Client B. The core skills overlap, but the specifics diverge significantly. Managing this matrix, knowing who is qualified on what, who needs requalification, who can flex to a different suite, is a massive data management challenge.
And it's a challenge that directly impacts your scheduling flexibility and your capacity utilization. If only 6 of your 30 operators are qualified on a particular client's process, you have a staffing bottleneck that no amount of equipment can solve.
Rapid Onboarding at Scale
The CGT CDMO market is growing at 24 to 28 percent annually. New facilities are coming online. Existing facilities are expanding. That means hiring. A lot of hiring.
I've worked with CDMOs that were onboarding 15 to 20 new operators per month. Think about the logistics of that. Classroom training. Cleanroom gowning qualification. Aseptic technique assessment. Equipment training. SOP read-and-understand. OJT on live batches (where you need qualified trainers supervising, which pulls them off production). Documentation training. Safety training.
A typical onboarding timeline to full qualification in CGT manufacturing is 3 to 6 months. During that time, the new hire is consuming trainer bandwidth, producing at reduced capacity, and representing elevated risk on every batch they touch.
When you're hiring 20 people a month, you need an onboarding machine. Standardized curricula. Dedicated training suites (if you can afford to take capacity offline). Simulation and dry-run capabilities. Clear milestone gates. And trainers. Lots of trainers.
Which brings us to the next problem.
Building and Retaining Trainers When Trainers Also Leave
Your trainers are your most experienced operators. They're the ones who know the processes cold, who can spot errors before they become deviations, who can teach the nuances that don't live in any SOP.
They're also the people most likely to get poached. A senior operator with training experience at a CGT CDMO is one of the most valuable people in this industry. Everyone wants them. And the typical CDMO response, giving them a "trainer" title with no meaningful pay increase and double the workload, is a recipe for losing them.
You need a trainer development and retention strategy that's as deliberate as your operator development strategy. That means real career paths. Real compensation adjustments. Reduced production burden (you can't train effectively if you're also running batches full-time). Recognition. And backup. Because if your training program depends on three people and two of them leave, you're in crisis.
I've seen CDMOs lose a single key trainer and have it cascade into qualification gaps across multiple client programs. That's a fragility you cannot afford in a market growing this fast.
Surviving Client Audits of Your Training Program
Every client audits your training program. It's one of the first things they look at, and one of the easiest places to find findings.
The questions are predictable. How do you ensure operators are qualified before they work on my product? How do you handle requalification? What happens when an operator makes an error? How do you track training currency? Can you show me the training records for the operators who worked on my last three batches?
The CDMOs that survive these audits have documentation systems that can answer those questions instantly. They can pull up any operator's training record, show the qualification history, show the requalification dates, show the trainer who signed off, and tie it all back to the specific batch records.
The CDMOs that struggle are the ones scrambling to compile records from six different systems, discovering gaps they didn't know existed, and making promises about corrective actions they'll struggle to implement.
Here's the thing about CGT client audits: the stakes are higher than traditional pharma. Many of these products are for patients with no other options. The sponsors are often small biotechs with everything riding on their CDMO's performance. They audit hard because they have to. Your training program needs to hold up under that scrutiny.
Campaign Manufacturing and Knowledge Decay
Most CGT CDMOs run campaign manufacturing. You make Client A's product for a few weeks, then clean and changeover, then make Client B's product for a few weeks. This is efficient from a facility utilization standpoint. It's terrible from a training standpoint.
An operator who ran Client A's process in January might not see it again until April. Three months of not performing a process means skill decay. The batch record might look familiar, but the muscle memory is gone. The little tricks that make the process run smoothly, things like knowing that a particular bioreactor runs slightly differently, or that a certain cell line is finicky about temperature ramps, those details fade.
You need a system for managing knowledge decay between campaigns. Refresher training before campaign starts. Pre-campaign briefings. Quick-reference guides. Buddy systems pairing experienced and less-experienced operators. And honest assessment of whether someone who hasn't run a process in six months is truly still qualified.
Tech Transfer: The Training Component Nobody Plans For
Every new client program starts with a tech transfer. The sponsor (or another CDMO) transfers the manufacturing process to your facility. There are technology transfer protocols, process characterization studies, engineering runs, PPQ batches.
What's almost always underplanned is the training component. The tech transfer team focuses on getting the process to work in your facility. They optimize parameters. They troubleshoot equipment differences. They qualify analytical methods. But the systematic training of operators on the transferred process? That's often an afterthought.
I've seen tech transfers where the "training" consisted of operators watching the sponsor's scientists run two batches, then being expected to execute independently. No structured curriculum. No competency assessment. No documentation of what was taught. Just "you saw it, now do it."
With over 2,000 clinical trials running globally and 90 percent of biotechs relying on CDMOs, the volume of tech transfers is only increasing. Every single one of them has a training component that needs to be planned, resourced, and executed with the same rigor as the process transfer itself.
Deviation Management Across Multiple Clients
Deviations are where training failures become visible. And at a CDMO, they become visible to multiple clients simultaneously.
When an operator makes an error on Client A's batch, that deviation gets investigated, reported to Client A, and potentially triggers a CAPA. But here's the question the smart clients ask: "If this operator made this error on my product, have they made similar errors on other products? Is this a systemic training gap?"
You need deviation trending that can identify training-related root causes across your entire operation, not just within a single client's program. If three different operators are making similar errors on buffer preparation across three different client programs, that's not three isolated deviations. That's a training gap in your core competency program.
CDMOs that can connect those dots proactively, before clients connect them during audits, build trust and retain business. CDMOs that treat every deviation as an island lose clients.
The Viral Vector Manufacturing Bottleneck
Limited viral vector manufacturing capacity is consistently cited as one of the key restraints on the CGT market. AAV vectors, lentiviral vectors... demand far outstrips supply. New facilities are being built, but they take years to come online and qualify.
What doesn't get enough attention is that viral vector manufacturing is also where training gaps are the most acute. The processes are complex. The analytical testing is specialized. The regulatory expectations are evolving. And the talent pool is tiny.
There simply aren't enough people in the world who know how to manufacture viral vectors at GMP scale. Every new facility that comes online needs to train operators largely from scratch because you can't hire experienced vector manufacturing operators at scale. They don't exist in sufficient numbers.
This is where the training crisis meets the capacity crisis head-on. You can build the facility. You can buy the bioreactors and the chromatography systems and the fill-finish equipment. But if you can't train operators to run those systems reliably and compliantly, the facility is an expensive warehouse.
How AI and Automation Are Changing Training Requirements
The CGT CDMO market is starting to see real AI and automation integration. Bioprocessing optimization through machine learning. Predictive models for media formulation and cell culture monitoring. Batch yield prediction algorithms that can flag problems before they become failures.
Cellares' Cell Shuttle is one of the most interesting developments: a robotic, AI-enabled platform for autologous cell therapy production that automates much of the manual processing. It's a glimpse of where the industry is heading.
But here's what people miss about automation: it doesn't eliminate training needs. It transforms them. Instead of training operators to perform manual cell processing steps, you're training them to operate, monitor, troubleshoot, and maintain automated systems. Instead of teaching aseptic technique for open manipulations, you're teaching system validation concepts and alarm response protocols.
The AI and automation layer also introduces new training requirements around data integrity, electronic batch records, algorithm validation, and exception handling. When the automated system flags an anomaly, the operator needs to understand what they're looking at and make the right decision. That requires deeper process understanding, not less.
CDMOs investing in automation need to invest equally in the training infrastructure to support it. The technology is only as good as the people operating it.
Training Metrics CDMOs Actually Need
Most CDMOs track training completion rates. That's necessary but nowhere near sufficient. Here's what you actually need to measure:
Time to full qualification. How long does it take a new hire to become independently qualified on their first client process? On their second? Track this over time. If it's getting longer, your onboarding system isn't scaling.
Training-related deviation rate. What percentage of your deviations have training gaps as a root cause or contributing factor? This is your most important quality metric for training effectiveness. If it's above 15 percent, you have a problem.
Qualification coverage. For each client process, how many qualified operators do you have? What's your minimum staffing threshold? Are you at risk of not being able to run a campaign because you don't have enough qualified people?
Trainer capacity utilization. How much of your trainers' time is spent on training vs. production? If it's above 70 percent production, your training is getting squeezed.
Requalification currency. What percentage of your operator qualifications are current vs. expired or expiring within 30 days? If you're constantly playing catch-up on requalification, your system is reactive, not proactive.
Turnover cost per operator. Calculate the fully loaded cost of losing and replacing an operator, including recruiting, onboarding, reduced productivity during training, and increased deviation risk. When leadership sees that number, training investment conversations get easier.
Building Culture When Operators Work on Someone Else's Product
This is the soft problem that underlies everything else. At a sponsor company, operators feel ownership of the product. They know what it does. They know who it helps. They feel connected to the mission.
At a CDMO, operators work on someone else's product. Often multiple someone elses. The product changes every few weeks. The sense of ownership and connection is harder to build.
But it's not impossible. The best CDMOs I've seen build culture around craftsmanship and service. The pride isn't in "our product." It's in "we can make anything, and we make it right." It's a manufacturing excellence culture where the quality of the work is the point, regardless of whose name is on the batch record.
That culture starts with training. When you invest in people's development, when you give them career paths, when you treat training as an investment rather than a compliance checkbox, people feel valued. They stay longer. They perform better. They care more.
In a market projected to grow 7 to 10x over the next decade, with Asia Pacific growing fastest and new entrants appearing regularly, the CDMOs that build this culture will have a durable competitive advantage that no amount of capital can replicate.
The Path Forward
The CGT CDMO market is at an inflection point. The demand is real. The clinical pipeline is massive. The technology is advancing. But the growth projections everyone is excited about assume something that isn't guaranteed: that the industry can train enough people to do the work.
Reduced investor interest is already restricting expansion in some areas. When the investment does flow, it tends to go toward facilities and equipment, things you can photograph and put in a press release. Training infrastructure is invisible until it fails.
I'm not saying the sky is falling. I'm saying that the CDMOs that treat training as a strategic capability, not an overhead cost, will be the ones that capture the growth. They'll onboard faster. They'll retain better. They'll have fewer deviations. They'll pass client audits without scrambling. They'll be able to flex operators across programs and maximize their capacity utilization.
And the CDMOs that keep treating training as a necessary evil? They'll choke on their own growth. They'll build beautiful facilities and fill them with undertrained operators who generate deviations that erode client trust. They'll watch their best people leave for competitors who invest in development.
The $52 billion question isn't whether the market will grow. It's whether the industry can build the workforce to support that growth. Training is the answer. It's time to treat it that way.
Ready to Fix Your CDMO Training Program?
If any of this hit close to home, you're not alone. I've built the CDMO Training Playbook, a comprehensive ebook covering everything in this article and more, with frameworks, templates, and implementation guides you can put to work immediately. Download it free at gxpframe.com.