☰

AGC Biologics A/S

Harmonising data across a global CDMO

A global CDMO with five manufacturing sites and a new Yokohama Smart Factory, moving toward paperless operations

Industry
Biopharmaceutical CDMO
Headquarters
Copenhagen, Denmark
Public information as of
January 2026

A4BEE prepared this analysis from publicly available sources. It reflects our own reading of AGC Biologics A/S's published strategy and is not endorsed by, or produced in cooperation with, AGC Biologics A/S. Company website

Strategic priorities

AGC Biologics is a global biopharmaceutical CDMO formed through the convergence of Asahi Glass (Japan), CMC Biologics (US), Biomeva (Germany), and MolMed (Italy). The company employs approximately 2,000 people across Copenhagen (HQ), Boulder and Longmont (Colorado, divested 2024), Seattle (Washington), Milan (Italy), and Chiba (Japan). The Yokohama greenfield Smart Factory is under construction for 2025-2027 completion. Services span mammalian cell culture, viral vectors for cell and gene therapy, and mRNA. Notable clients include Roche for commercial manufacturing.

The CDMO manages five production sites with fundamentally different levels of digital maturity and different legacy vendor ecosystems. Paper-based batch records and disconnected LIMS/ELN systems are prevalent at the Milan CGT site and Bothell mammalian operations, while Yokohama is being designed from scratch with full Industry 4.0 architecture. The FDA issued Form 483 observations to the Bothell site relating to electronic data controls — a reflection of the paper-to-digital transition that is underway but incomplete.

The AGC Group CTO has set a target of 5,000 DX-trained employees, acknowledging that the gap between shop-floor operators and digital strategists is the primary constraint on digital transformation execution. The Yokohama Smart Factory is positioned as the flagship that demonstrates what full MTP-based modularity, digital twin capability, and AI-powered QC looks like from day one — and that becomes the reference site for the group's wider transformation.

Challenges we see

  • Digital Integration

    Heritage IT/OT systems across five sites preventing cross-site process visibility

    AGC Biologics was formed through multiple acquisitions — Asahi Glass, CMC Biologics, Biomeva, MolMed — each bringing different vendor ecosystems, SCADA systems, LIMS implementations, and data formats. The Milan CGT site, Copenhagen mammalian operations, Seattle, and Chiba sites do not natively communicate.

    When the Milan site runs a viral vector campaign and Copenhagen runs a monoclonal antibody campaign, the process knowledge generated at each site stays at that site. A sponsor expecting cross-site technology transfer receives two disconnected data environments — the 'unified global ecosystem' described in the strategy is not yet reflected in the data architecture.

  • Compliance Regulatory

    Paper-based laboratory operations creating FDA inspection risk

    Critical data regarding cell titers and media composition is still recorded in paper logbooks or isolated Excel spreadsheets at multiple sites. The Bothell site received FDA Form 483 observations relating to quality control and procedural controls for electronic data — indicating that paper-based processes are a regulatory risk, not just an operational inefficiency.

    Where critical manufacturing data is recorded on paper, the data integrity required by 21 CFR Part 211 cannot be automatically verified. Each regulatory inspection carries the risk that paper records will be found to have gaps, illegible entries, or transcription errors — the same pattern related to the FDA observations at Bothell.

  • Operations Manufacturing

    Large-scale Colorado assets misaligned with market demand after divestment

    The acquisition of large-scale Boulder and Longmont facilities brought bioreactor assets designed for high-volume blockbuster drugs. The market has shifted toward personalised medicine and orphan drugs requiring mid-scale flexibility. The Colorado divestment (2024) signals financial friction from this mismatch, while the remaining sites must handle increased volumes of complex molecules — bispecifics and trispecifics — with mid-scale equipment.

    When mid-scale sites absorb volume from divested large-scale facilities, the existing mid-scale equipment may not have the digital connectivity to provide the process telemetry data that operators need to run complex molecules at the limits of their performance envelope.

  • Operations Operations

    DX-savvy talent gap between operators and digital strategists

    The AGC Group CTO has identified a critical gap between shop-floor operators who run the equipment and the digital strategists who design the transformation programme. The 5,000 DX-trained employee target has not been achieved, and the gap means that digital solutions are designed without sufficient input from the people who will use them.

    When digital solutions are deployed without adequate operator input during design, the result is interfaces that do not match the way the work is actually done — producing 'Digital Hesitancy' where operators prefer the familiar manual method, and undermining the ROI of the digital investment.

  • Digital Integration

    Customer portal strategy expanding the OT attack surface

    The Digital-First strategy includes friendly customer portals for real-time data exchange with sponsors. As more process data moves from internal OT networks to externally accessible platforms, the attack surface for intellectual property theft and production sabotage increases — while the legacy PLC and SCADA infrastructure at older sites lacks modern encryption.

    When process data from a client's manufacturing campaign is accessible through a customer portal, the security of that data depends on the security of the entire chain — including the OT network at the manufacturing site, which may run legacy PLCs that were not designed for an interconnected environment.

Opportunities, by urgency and business impact

Each bubble is one opportunity, numbered to match the list below. Further right means it bites sooner; higher means a bigger effect on the business. A bigger bubble means a bigger implementation effort.

Source: A4BEE analysis of public sources

  1. Paperless QC laboratory and electronic batch records

    Critical biological KPIs are tracked in Excel separate from live process trends, paper-based batch records require manual transcription prone to error, and the Bothell site received FDA Form 483 observations relating to electronic data controls.

    Implement an integrated Electronic Batch Record system linked to DCS that enforces ALCOA+ data integrity principles, automates data capture from instruments, and enables real-time compliance verification during production — eliminating the regulatory risk that paper records create at inspection.

    • Bothell FDA Form 483 observations, 2023-2024
    • 21 CFR Part 211 electronic records requirements

  2. Cross-site data harmonisation and the Global Data Lake

    Systems in Milan, Copenhagen, and Seattle do not natively communicate due to heritage silos, preventing the creation of a Global Data Lake and making cross-site process trending nearly impossible for technology transfer optimisation.

    Build an ontology-based Industrial Data Platform that harmonises data from heterogeneous sources across all sites — SCADA, LIMS, ELN — into a unified semantic model enabling cross-site comparison, Golden Batch trending, and technology transfer that does not start from scratch at each receiving site.

    • Milan MolMed heritage system documentation
    • CMC Biologics cross-site technology transfer protocols

  3. Yokohama Smart Factory with full Industry 4.0 architecture from day one

    The Yokohama greenfield facility opening 2025-2027 represents a once-in-a-generation opportunity to implement full Industry 4.0 from the ground up, but without proper architecture it risks inheriting the data silo problems of the existing Western sites.

    Position Yokohama as the Digital Twin Flagship with MTP-based modularity ensuring equipment from Cytiva, Sartorius, and Thermo Scientific integrates into a single Plug and Produce ecosystem with AI-powered QC from day one — and as the reference site that demonstrates what the group's transformation looks like when it works.

    • AGC Biologics Yokohama Smart Factory project scope
    • Industry 4.0 greenfield implementation best practices

  4. CGT manufacturing automation for viral vector cost reduction

    CGT manufacturing in Milan is still largely manual despite supporting 12 commercial products, while the target of USD 1,000 per patient viral vectors requires massive automation to achieve the throughput needed for commercial success.

    Implement robotisation and cobots for cleanroom operations to eliminate human variability in viral vector filling and cell isolation, with vision systems for real-time monitoring of foam and cell morphology in complex bioreactor runs.

    • Milan CGT site 12 commercial products operational scope
    • EVP Luca Alberici cost reduction targets

  5. Legacy equipment digital retrofitting at Chiba and remaining mid-scale sites

    Many legacy microbial skids in Chiba and Seattle lack integrated digital control, while the remaining mid-scale sites must handle increased volumes of complex molecules after the Colorado divestment without capital for full equipment replacement.

    Retrofit existing legacy bioreactors and chillers with smart controllers enabling closed-loop regulation, OPC UA connectivity for unified data acquisition, and integration into the modern data platform — at a fraction of the cost of replacement equipment.

    • Chiba site legacy equipment inventory
    • AGC Biologics mid-scale site capacity requirements post-divestment

What we'd propose

  • Digital Lab

    Paperless QC Laboratory and Electronic Batch Records

    We transform paper-based QC laboratories at Bothell, Milan, and Copenhagen into integrated digital environments with automated data capture from instruments, real-time compliance verification during production runs, and electronic batch records that enforce ALCOA+ data integrity principles from the first entry to the final release decision.

    • Electronic Batch Record deployment

      Paper batch records replaced with structured digital records

      Deploy an eBR system linked to DCS that presents operators with structured digital records at each production step — replacing handwritten logbook entries with electronic data capture that enforces data完整性 and creates a complete, tamper-evident record automatically as each step is executed.

    • Instrument data auto-capture

      HPLC, cell counters, and nutrient analysers connected

      Integrate laboratory instruments — HPLC systems, automated cell counters, nutrient analysers — directly into the eBR through validated instrument drivers, so that analytical results flow into the batch record without manual transcription and without the transcription errors that manual entry introduces.

    • Real-time compliance verification

      OOS conditions detected at the time they occur

      Configure real-time specification checking within the eBR that flags out-of-specification conditions at the moment they are recorded rather than at the end-of-batch QA review — enabling immediate investigation and reducing the scope of a batch hold when a deviation is confirmed.

    • FDA Form 483 observations are addressed at the root cause — the paper-based record system — rather than patched with corrective actions that leave the underlying vulnerability in place.
    • Batch release time is reduced because the eBR data is available for QA review immediately after batch completion rather than requiring days of data transcription and reconciliation.
    • The ALCOA+ compliant eBR satisfies 21 CFR Part 211 electronic records requirements without requiring retrospective validation of paper-based records.

    We've done this before

    See how we do Digital Lab
  • Enterprise AI

    Ontology-Based Industrial Data Platform for Cross-Site Harmonisation

    We build a semantic data lakehouse that harmonises heterogeneous data sources — SCADA, LIMS, ELN, eBR — from Milan, Copenhagen, Seattle, and Chiba into a unified ontology-driven model enabling cross-site comparison, Golden Batch trending, and technology transfer that is powered by historical data rather than starting from scratch.

    • Heritage system connectors for each site

      Every site connected regardless of vendor

      Build validated data connectors for each heritage platform — Siemens PCS 7 at Copenhagen, Emerson DeltaV at Seattle, Rockwell at Milan, and Yokogawa at Chiba — that ingest process data into the unified semantic model without requiring replacement of the existing control systems.

    • Golden Batch comparison engine

      Best-performing batches identified and benchmarked

      Develop a Golden Batch analysis capability that correlates process parameters across all sites to identify the conditions that produce the highest titer or best quality outcome for each molecule type — giving process development teams evidence-based targets rather than historical precedent.

    • Cross-site technology transfer workspace

      Transfer packages built from unified data, not from scratch

      Create a structured technology transfer workspace where the sending site can share process data, batch records, and analytical methods with the receiving site through the platform — reducing the time to reach comparable performance at the new site by providing historical reference data that was previously inaccessible.

    • Technology transfer between AGC sites is no longer a blank-sheet exercise — each transfer benefits from the accumulated process knowledge of all previous campaigns across the network.
    • The Golden Batch analysis identifies the conditions that produce the best outcomes for a given molecule, giving process development teams a quantitative target rather than a best-effort estimate.
    • The unified data model is the prerequisite for the Yokohama Smart Factory to serve as the group's digital reference site — sharing data standards and analytical approaches with the older sites rather than operating in isolation.

    We've done this before

    See how we do Enterprise AI
  • Digital CDMO

    Yokohama Smart Factory with MTP-Based Modular Architecture

    We design and implement the digital backbone for the Yokohama Smart Factory using Module Type Package standards to ensure that equipment from Cytiva, Sartorius, and Thermo Scientific integrates into a single Plug and Produce ecosystem from day one — creating the reference site for the group's wider transformation.

    • MTP equipment library for Yokohama equipment roster

      All Smart Factory equipment described in standard format

      Create MTP-compliant equipment packages for the complete Yokohama equipment roster — bioreactors, chromatography skids, media prep systems, and fill-finish equipment — so that new equipment additions are configured through the MTP library rather than requiring custom integration engineering for each supplier.

    • Digital twin of Yokohama reference processes

      Simulation platform operational from day one

      Build a digital twin of the key mammalian cell culture and viral vector processes that will run at Yokohama, calibrated against the process data from Copenhagen and Milan — enabling operators to practice and optimise in simulation before the physical facility is commissioned.

    • AI-powered in-process QC analytics

      Real-time batch quality prediction from process telemetry

      Deploy ML models that analyse in-process parameters — glucose consumption rate, lactate profiles, amino acid concentrations, off-gas composition — to predict final batch quality attributes in real time, enabling operators to make steering decisions before the batch trajectory is committed.

    • Yokohama operates from day one with the MTP-based architecture that existing Western sites should have had — avoiding the 18 months of custom integration work that each legacy site now requires.
    • The digital twin enables Yokohama to demonstrate process performance to clients and regulators before a single physical batch is run, accelerating regulatory acceptance of new processes.
    • The Smart Factory becomes the training ground where operators from existing sites learn the MTP-based architecture that the group will roll out across the network.

    We've done this before

    See how we do Digital CDMO
  • Digital Lab

    CGT Cleanroom Automation and Robotics for Viral Vector Manufacturing

    We implement robotic automation for the Milan CGT manufacturing suite — covering viral vector filling, cell isolation, and quality control sample preparation — to eliminate human variability, reduce contamination risk, and achieve the throughput required for commercial-scale operations at the USD 1,000 per patient cost target.

    • Cobot integration for cell isolation

      Automated cell isolation replacing manual pipetting

      Deploy cobots for the cell isolation steps that currently require skilled operators to perform repetitive manual pipetting — reducing process time, eliminating operator variability, and freeing skilled staff for the process development work that requires human judgement.

    • Vision system for bioreactor monitoring

      Foam and morphology monitored automatically

      Implement computer vision monitoring of cell culture in the bioreactor — detecting foam spikes, cell aggregation patterns, and morphology shifts in real time — so that operators receive automated alerts when conditions deviate from the expected trajectory rather than relying on periodic manual observation.

    • Automated viral vector fill-finish

      Robot filling line for vials and syringes

      Integrate a robotic filling station for viral vector drug product into the Milan CGT suite, with automated vial handling, aseptic filling, and lot traceability — replacing the manual fill operations that limit throughput and introduce contamination risk.

    • The per-patient viral vector cost target becomes achievable because robotic automation reduces the labour content of each dose — the key enabler of the USD 1,000 per patient ambition.
    • Contamination events that currently require investigation, batch rejection, and re-run are reduced because the robotic filling line eliminates the human error vectors that manual operations introduce.
    • The computer vision system provides continuous monitoring data that improves the digital twin model — each run generates data that makes the next run more predictable.

    We've done this before

    See how we do Digital Lab
  • Digital CDMO

    Legacy Equipment Digital Retrofit for Mid-Scale Site Capacity

    We retrofit existing legacy bioreactors, fermenters, and auxiliary equipment at Chiba and other mid-scale sites with smart controllers enabling closed-loop process regulation, OPC UA connectivity for unified data acquisition, and integration into the modern data platform — at a fraction of the cost of replacement equipment.

    • Smart controller retrofit for legacy bioreactors

      Legacy bioreactors connected without full replacement

      Install smart controller units on legacy bioreactor and fermenter assets at Chiba that provide modern control algorithms, local data logging, and OPC UA connectivity — enabling the legacy assets to participate in the site-wide data platform without the capital cost of full equipment replacement.

    • OPC UA data acquisition from legacy PLC equipment

      All equipment visible in the unified data platform

      Deploy OPC UA collectors at Chiba and Seattle that aggregate data from legacy PLCs — temperature controllers, utility monitoring systems, media prep skids — and forward them to the ontology-driven data platform, completing the site-wide data picture that was previously limited to modern DCS-connected assets.

    • Remote monitoring dashboard for legacy equipment

      Operators see legacy equipment status from the same screen

      Create a unified monitoring dashboard that integrates legacy equipment status alongside modern DCS data — giving operators a single view of the entire manufacturing campaign regardless of the vintage of the underlying equipment.

    • Mid-scale sites gain the digital visibility needed to run complex molecules at the limits of their performance envelope — without the capital cost of equipment replacement that the post-divestment budget cannot absorb.
    • The OPC UA retrofit establishes the data foundation for predictive maintenance on legacy assets, extending their productive life and deferring the replacement decision.
    • Chiba's digital capability becomes comparable to the group's best sites, enabling it to bid for more complex work that was previously reserved for Copenhagen or Milan.

    We've done this before

    See how we do Digital CDMO

Digital maturity: today and target

Scored out of 100 across six dimensions. The target is what AGC Biologics A/S's own published ambition implies — not a perfect score.

Source: A4BEE analysis of public sources
Data Integration 35 → 85
Milan, Copenhagen, Seattle, and Chiba each run different SCADA systems, different LIMS platforms, and different ELN implementations. The cross-site data integration that the Global Data Lake strategy requires has not been built — each site is a data silo, and the technology transfer process between sites starts from scratch rather than from accumulated historical data.
Paperless Operations 40 → 95
Paper-based batch records and disconnected laboratory instruments are prevalent across Milan CGT and Bothell mammalian operations. The Bothell FDA Form 483 observations relate to paper-based processes and regulatory risk — the transition from paper to digital is underway but has not reached the level of maturity required for regulatory confidence.
Process Automation 45 → 90
Milan CGT manufacturing is still largely manual despite supporting 12 commercial products. The cobot and robotic automation described in the CGT strategy has not been implemented — the human variability that manual operations introduce is the primary constraint on achieving the USD 1,000 per patient cost target.
IT/OT Convergence 30 → 85
Legacy PLCs and SCADA systems at older sites lack modern encryption and network segmentation. The customer portal initiative is expanding the accessible data surface before the underlying OT security architecture has been hardened — creating a gap between digital strategy and security reality.
Real-Time Analytics 40 → 90
Process data is collected by DCS and SCADA systems but not analysed in real time for predictive purposes. The Golden Batch trending capability does not exist across sites — best practice identification depends on periodic manual review rather than continuous algorithmic comparison.
Equipment Connectivity 50 → 90
Modern DCS-connected assets at Copenhagen and Seattle are well-integrated, but legacy assets at Chiba and the periphery of each site are not connected to the data platform. The MTP-based equipment library for Yokohama has not yet been extended to existing sites, meaning each new equipment addition requires custom integration work.

Check this yourself

Our Service Portal has free self-assessments and market comparisons. These are the ones that line up with what we've read above — no sales call required.

Think we've read this right?

Talk to us

Related reading

This is an independent analysis prepared by A4BEE from publicly available information as of January 2026. It reflects A4BEE's own interpretation and opinion, is not affiliated with, endorsed by, or verified with AGC Biologics A/S, and may be incomplete or inaccurate. All company names and trademarks are the property of their respective owners. To request a correction or removal, contact [email protected].