Closing the Scale Gap with Digital Infrastructure

By TPN Editorial TeamNovember 11, 2026

As upstream process intensification drives cell culture productivity to record highs, downstream bottlenecks and data silos are creating a critical scale gap across the biomanufacturing lifecycle.

During the past decade, the bio/pharma industry has made significant strides forward in bioprocessing, with advancements in cell line engineering, media optimization, and upstream process intensification affording companies impressive early-stage yield improvements. However, these early-stage productivity gains are not being easily scaled into later operational stages, exposing a scale gap that needs to be addressed to ensure commercial-scale manufacturability is realized.

The Upstream/Downstream Scale Disconnect

“The most significant operational friction in today’s bioprocessing workflows occurs in downstream processing, quality control, and the handoffs between upstream, downstream, and data systems,” notes Jean-Marc Cappia, Head of Market Development for DSP Intensification & Digitalization at Sartorius. “While upstream process intensification has successfully increased cell culture productivity and product titers, downstream purification and QC activities often struggle to keep pace.”

Cappia emphasizes that the scale gap is exacerbated when upstream and downstream development are decoupled early in a program’s lifecycle. “A key driver of this friction is the increasing complexity of biologics manufacturing combined with fragmented workflows and data silos,” he explains. “Many organizations still optimize upstream and downstream operations separately, resulting in bottlenecks when higher upstream productivity overwhelms downstream capacity or creates new scalability challenges.”

From a fluid dynamics and hardware perspective, scaling high-density perfusion to commercial bioreactors reveals physical limitations that do not exist at smaller volumes, observes Gerben Zijlstra, Bioprocess Solutions Expert at Sartorius. “When perfusion is applied to intensify the main bioreactor at 2,000 L scale and beyond, at high cell density, the scalability of the current cell retention systems is problematic,” he says. “The reason is that a certain filter area per bioreactor volume must be scaled up linearly. Where large volume bioreactors are still relatively compact, the number of hollow fiber filters and therefore circulation loops required for alternating tangential flow (ATF) and especially tangential flow filtration (TFF) becomes very hard to manage in terms of bioreactor connections, filter/pump footprint, and tubing complexity.”

Furthermore, bridging the automation scale gap between bench-scale research and production-scale automation requires early planning around data structures, Zijlstra continues. “Problem at bench, up to pilot scale, is that proper end-to-end automation is cumbersome due to physical connectivity, data connectivity, and metadata harmonization challenges,” he adds. “At commercial scale, dedicated distributed control systems (DCS) can solve integration challenges, but at the cost of bespoke, extensive, and expensive programming and configuration.”

Avoidable Scale-Up Errors in Early Development

As a candidate molecule transitions across clinical phases toward commercial release, the core objective shifts from rapid bench-scale proof-of-concept to long-term manufacturability and cost efficiency, Cappia reveals. Mistakes made during early-stage development often widen the scale gap when tech transfer to large facilities begins, he notes.

“One of the most common mistakes organizations make early in development is optimizing for short-term performance without considering commercial manufacturability,” Cappia warns. “Examples include selecting processes that are difficult to scale, failing to characterize critical process parameters, overlooking downstream purification requirements, or developing upstream processes that generate titers beyond what downstream operations can efficiently handle.”

Delaying digital architecture decisions early in development also creates severe hurdles during commercial scale-up, Cappia asserts. “Companies may delay digitalization, data standardization, and end-to-end process integration, making technology transfer and scale-up more challenging later,” he says. “As a result, issues such as reduced yields, process variability, increased costs, and delayed commercialization often emerge during scale-up.”

Zijlstra highlights that automation discrepancies across manufacturing tiers create substantial rework during scale transfer. “A major challenge can be if the automation, including recipes, differs from scale to scale,” he observes. “When performing tech transfer, recipes may need to be fully re-developed, requiring substantial requalification.”

Technological Approaches to Bridging the Scale Gap

However, the bottlenecks being experienced during downstream processing are not simply related to the size of a facility, Cappia stresses. “The downstream bottleneck is increasingly both a capacity and a technology challenge, but today it is primarily a capacity issue that is being exposed by technological limitations,” he states. “Over the past decade, advances in cell culture, media optimization, and process purification have dramatically increased upstream titers. However, downstream operations — particularly chromatography — have not improved at the same pace, creating throughput constraints.”

Relying solely on adding larger chromatography columns to bridge the gap between upstream and downstream processes is physically and economically unsustainable, Cappia remarks. “The underlying issue is not simply a matter of adding more equipment. Traditional chromatography, multiple purification steps, and extensive QC [quality control] requirements can limit productivity, increase costs, and reduce facility flexibility,” he says. “The long-term solution will require technological innovation, including higher-capacity resins, membrane chromatography, continuous downstream processing, automation, and better integration of upstream and downstream development.”

To overcome the physical scale barriers in upstream cell retention, Zijlstra highlights novel fluidics strategies that reduce hardware complexity at larger scales. “Developing high cell density cell cultures with better filterability (high viability, low cellular debris, lower or different anti-foam use) is the first target to enhance ATF and TFF performance,” Zijlstra explains. “Secondly, both reversed TFF and counter-current TFF seem to offer substantially reduced hollow-fiber filter area requirements, allowing high cell density perfusion scale-up to 2,000 L and 4,000 L scale.”

Leveraging Digital Infrastructure

Advanced digital infrastructure is emerging as a key mechanism for ensuring process parity across development scales, Cappia reveals. “The industry is moving steadily toward predictive biomanufacturing, but it is not yet at the point where digital twins, AI-driven control, and PAT are routinely preventing bottlenecks across commercial operations,” he notes. “Today, the most advanced biopharma companies are successfully using these tools in targeted applications. However, widespread adoption remains limited by data quality, system integration challenges, regulatory requirements, and the complexity of biological processes.”

Deploying digital models helps technical teams foresee how bench-scale dynamics will manifest in commercial bioreactors, Cappia asserts. “Over the next five to ten years, companies that combine PAT, advanced analytics, real-time monitoring, and integrated process data are expected to significantly reduce deviations, accelerate troubleshooting, and optimize end-to-end process performance,” he projects.

Zijlstra outlines how digital twins are already helping bridge operational gaps by accurately modeling physical phenomena like filter fouling across scales. “Advanced process modeling, including the use of AI, can be used to predict filter fouling based on process conditions. The other way around, these models can also predict how to run the process to minimize ATF filter fouling while maximizing product yield and quality,” he specifies.

“At manufacturing scale, many larger biopharma companies are already using multivariate data analytics (MVDA) tools like SIMCA Online to compare ongoing batch data against ‘golden batch’ results,” Zijlstra continues. “Using digital twins and AI-driven control is within reach, as long as the models are explainable to regulatory agencies.”

The Scalable Facility of the Future

Solving the bioprocessing scale gap will result in a new generation of commercial facilities, Zijlstra confirms. “The bioprocessing facility of the future will be a compact, intensified, and continuously operated facility that is fully automated using PAT technology and largely operating autonomously,” he predicts. “It will operate with a limited number of operators providing raw materials, exchanging consumables, and harvesting drug product.”

Concurring with Zijlstra, Cappia also expects facilities to be smaller, more modular, and flexible in the future. “Additionally, intensified upstream and downstream operations will be developed as an integrated system, ensuring that productivity gains in one area do not create bottlenecks elsewhere,” he says.

However, true scalability relies on establishing an integrated digital backbone early in development, Cappia asserts. “Rather than relying on large, fixed manufacturing assets and manual interventions, future facilities will operate as digitally integrated environments where equipment, sensors, analytical systems, and manufacturing execution platforms continuously exchange data in real time,” he summarizes. “This shift will move manufacturing from a reactive model toward one that is increasingly predictive and self-optimizing.”