☰

Samabriva S.A.

Digital readiness for hairy root biomanufacturing at 50,000-litre scale

Industry
Biotechnology (Plant-Based Biomanufacturing and CDMO)
Headquarters
Liège, Belgium
Public information as of
January 2026

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

Strategic priorities

Samabriva S.A. is a Belgian biopharmaceutical company commercialising a proprietary hairy root organ culture platform for the production of complex APIs and recombinant proteins. The platform — which avoids the chromosomal instability common in transgenic cell lines — has attracted a partnership with Genethon for cost-effective AAV vector manufacturing and positions Samabriva as a differentiated CDMO in the European market. The company is executing a scale-up roadmap from 25-litre pilot R&D in Amiens to 1,000-litre industrial bioreactors at a new 1,400 square metre GMP-compliant facility in Liège, with a longer-term target of 50,000-litre production to serve Top Pharma CDMO contracts.

The immediate challenge is bridging the data gap between Amiens and Liège. Process data from clone optimisation and Design of Experiment workflows is trapped in Excel spreadsheets and manual lab notebooks. The inability to unify offline analytical results with online sensor data prevents real-time technology transfer and Golden Batch analysis — where a proven production run's parameters are used as a live reference for active batches. The Amiens site also carries legacy digital debt: equipment accumulated since 2011 includes isolated workstations with outdated operating systems that must now integrate with modern automated bioreactors in Liège, while USB-based air-gap transfers create both data integrity and NIS2 compliance risks.

On the manufacturing side, hairy root organ cultures present unique monitoring challenges. The non-uniform oxygen transfer coefficients, shear sensitivity and dense tissue matrices make traditional fermentation sensors inadequate for real-time biomass estimation. Manual sampling — the current fallback — introduces contamination risk that is unacceptable for GMP pharmaceutical production.

Challenges we see

  • Digital Integration

    Connecting cross-site process data as Golden Batch analysis becomes a regulatory expectation

    Process data from clone optimisation and DoE workflows at Amiens is trapped in Excel spreadsheets and manual lab notebooks. The Liège facility's online sensor data is disconnected from the offline analytical results, preventing real-time technology transfer and the Golden Batch overlay that regulatory inspectors increasingly expect.

    Where cross-site data remains siloed, each new production campaign starts from first principles rather than from the accumulated learning of prior runs. An ontology-based data platform means the Golden Batch profile is always live, always current, and always comparable to the active campaign.

  • Operations Manufacturing

    Replacing manual sampling with real-time biomass estimation for GMP bioreactor control

    Hairy root organ cultures in 1,000-litre bioreactors have non-uniform oxygen transfer, shear sensitivity and dense tissue matrices that make traditional fermentation sensors inadequate. Manual sampling is the current fallback but introduces contamination risk unacceptable for GMP pharmaceutical production.

    Where biomass estimation depends on manual sampling intervals, contamination risk and monitoring gaps are introduced by the sampling process itself. Non-invasive computer vision-based monitoring means the biomass estimate is continuous and the contamination risk of manual intervention is eliminated.

  • Digital Integration

    Building NIS2-compliant IT/OT architecture as air-gap security is replaced by connectivity

    The Amiens site relies on air-gap transfers via USB drives between isolated workstations and the Liège bioreactors. Connecting these systems to enable the data integration that the CDMO business requires means replacing the air gap with controlled connectivity, introducing cybersecurity vulnerabilities that NIS2 directive compliance demands are managed.

    Where the air gap is removed without a structured connectivity architecture, the USB-based isolation is replaced by an unmanaged network exposure. A zero-trust IT/OT convergence architecture means the connectivity required for data integration is achieved without the exposure that NIS2 was designed to address.

  • Compliance Regulatory

    Accelerating CDMO client onboarding as each new molecule requires full GMP validation

    Each new CDMO client project requires extensive manual system validation and data partitioning to protect IP and maintain FDA 21 CFR Part 11 compliance. Without containerised IT architecture, the validation overhead for each new client engagement becomes a bottleneck to commercial scaling.

    Where system validation for each new client molecule is从头到尾managed manually, the time to first batch delivery is set by the validation team's backlog, not by the manufacturing readiness of the process. Containerised, MTP-based modular architecture means client onboarding is a configuration exercise, not a construction project.

  • Operations Manufacturing

    Validating 50,000-litre bioreactor design before committing to a capital investment of this scale

    Scaling hairy root cultures from 1,000 litres to 50,000 litres is unprecedented. The fluid dynamics, shear stress, oxygen distribution and tissue density interactions at this scale cannot be reliably predicted from first principles or from smaller-scale data alone. Trial-and-error at this scale is economically prohibitive.

    Where bioreactor scale-up relies on empirical correlation from small-scale data, the first 50,000-litre campaign is effectively a validation experiment with a full commercial batch at stake. Digital twin models built from existing process data mean the scale-up hypothesis is tested in simulation before a single euro of capital is committed.

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. Ontology-based cross-site data platform for real-time technology transfer

    Process data is fragmented between Amiens R&D and Liège manufacturing, trapped in Excel spreadsheets and disconnected instrument workstations. The inability to unify offline analytical results with online sensor data prevents Golden Batch analysis and slows technology transfer for new CDMO client molecules.

    Deploy an ontology-based Industrial Data Platform with automated data pipelines that unify cryopreservation records, offline analytical results and online sensor data into a single namespace accessible across both sites, enabling real-time Golden Batch comparison and technology transfer in hours.

    • Samabriva deep research analysis, 2025

  2. Computer vision bioreactor monitoring for real-time biomass estimation

    Hairy root organ cultures require non-invasive monitoring solutions that traditional fermentation sensors cannot provide. Manual sampling introduces contamination risk that is incompatible with the GMP pharmaceutical quality standard expected by Top Pharma CDMO clients.

    Implement computer vision-based non-invasive monitoring systems that provide continuous biomass density estimation without breaking sterile barriers, eliminating contamination risk from manual sampling and enabling adaptive nutrient feed control.

    • Samabriva deep research analysis, 2025

  3. NIS2-compliant IT/OT convergence for the Liège CDMO facility

    Legacy isolated workstations at Amiens are connected to modern automated bioreactors in Liège via USB-based air-gap transfers. Replacing the air gap with proper connectivity is required for CDMO data integration but introduces cybersecurity vulnerabilities that NIS2 directive compliance demands are addressed.

    Establish a zero-trust IT/OT convergence architecture using OPC UA and VLAN segmentation that enables the data integration the CDMO business requires while achieving NIS2 compliance and protecting Master Working Root Bank data.

    • NIS2 directive compliance requirements, 2024

  4. MTP-based modular architecture for rapid CDMO client onboarding

    Each new CDMO client molecule requires extensive manual system validation and data partitioning. Without containerised IT architecture, the validation overhead becomes the binding constraint on commercial scaling and delays revenue from the Liège facility.

    Implement MTP-based modular production architecture with containerised validated workflows that enable rapid tech transfer for new client molecules — a validated module is configured, not constructed from scratch for each new engagement.

    • Samabriva deep research analysis, 2025

  5. Digital twin for 50,000-litre hairy root bioreactor scale-up validation

    Scaling hairy root cultures to 50,000 litres is unprecedented. The fluid dynamics, shear stress, oxygen transfer and tissue density interactions at this scale cannot be predicted reliably from first principles or small-scale data alone, and trial-and-error at this capital scale is economically prohibitive.

    Build a data-driven digital twin model that merges mechanistic growth kinetics with real-time process data from the 1,000-litre operations to simulate bioreactor performance at 50,000-litre scale, validating the scale-up hypothesis before committing to the capital investment.

    • Samabriva leadership strategic brief, 2024

What we'd propose

  • Enterprise AI

    Ontology-based cross-site data platform for Samabriva operations

    We design and deploy an ontology-based Industrial Data Platform that connects R&D workflows at Amiens with manufacturing operations at Liège through automated data pipelines, enabling real-time Golden Batch analysis, technology transfer in hours and unified biological asset management across both sites.

    • Unified ontology and namespace architecture

      One data language across Amiens and Liège

      Design an ontology-based unified data architecture that maps clone selection, DoE, cryopreservation, bioreactor sensor and offline analytical data into a single semantic namespace accessible from both sites, eliminating the translation layer that currently impedes cross-site data use.

    • Automated data pipeline from all instruments

      Data flows from instrument to platform without manual steps

      Build automated data pipelines that ingest data from all connected instruments at both Amiens and Liège — without requiring engineers to export, format and import — creating a real-time data environment that is always current.

    • Real-time Golden Batch comparison layer

      Every active run compared to the best reference campaign

      Implement a Golden Batch comparison layer that overlays historical reference campaign parameters against live bioreactor data, alerting operators when the active run deviates from the proven envelope with enough lead time to take corrective action.

    • Technology transfer for new CDMO client molecules compressed from weeks to hours through automated data flow.
    • Golden Batch analysis available in real time, enabling corrective action before out-of-spec events occur.
    • Data integrity across Amiens and Liège sites demonstrable for regulatory inspections through complete lineage records.

    We've done this before

    See how we do Enterprise AI
  • Digital Lab

    Computer vision bioreactor monitoring system for hairy root cultures

    We implement computer vision-based non-invasive monitoring systems for the 1,000-litre hairy root bioreactors at Liège, providing continuous biomass density estimation, contamination detection and adaptive nutrient feed control without any breach of the sterile barrier.

    • Non-invasive computer vision deployment

      Seeing biomass through the bioreactor wall

      Deploy machine vision systems at strategic points on the 1,000-litre bioreactors that capture image data through the bioreactor wall or sampling port, using trained ML models to estimate biomass density, morphology and health indicators without introducing any contamination risk.

    • Real-time biomass and contamination monitoring

      Continuous estimates, not interval samples

      Build a monitoring platform that processes computer vision data in real time, providing continuous biomass density estimates, trend visualisation and automated alerts when the morphology or density profile deviates from the proven envelope.

    • Adaptive nutrient feed control integration

      Biomass data feeding back to the process controller

      Integrate the computer vision biomass estimates with the bioreactor control system to enable adaptive nutrient feed regimes that respond to the actual biomass state rather than a scheduled feeding programme.

    • Manual sampling frequency reduced, with corresponding reduction in contamination risk for GMP batches.
    • Biomass monitoring is continuous and real-time, not limited to the sampling interval schedule.
    • Adaptive feeding enabled by real-time biomass data improves substrate utilisation and reduces media cost per gram of product.

    We've done this before

    See how we do Digital Lab
  • Digital CDMO

    Zero-trust IT/OT convergence for NIS2 compliance

    We establish a zero-trust IT/OT convergence architecture for the Amiens-Liège connection using OPC UA for secure machine-to-system communication, VLAN segmentation between IT and OT zones, and continuous monitoring that achieves NIS2 compliance while enabling the data integration the CDMO business requires.

    • OPC UA secure connectivity gateway

      Data flows securely between Amiens and Liège

      Deploy OPC UA connectivity gateways that replace USB air-gap transfers with encrypted, authenticated data exchange between the Amiens legacy systems and the Liège OT network, maintaining data integrity while enabling the automated pipelines the CDMO requires.

    • OT network segmentation and monitoring

      Zero-trust architecture for the OT network

      Implement VLAN segmentation between OT zones, network access controls that authenticate every device, and continuous OT security monitoring that detects anomalous behaviour without disrupting normal manufacturing operations.

    • NIS2 compliance documentation pack

      Audit-ready evidence of security controls

      Produce a structured NIS2 compliance documentation pack — risk assessments, security policies, incident response procedures, supply chain security documentation — ready for regulatory submission without last-minute assembly.

    • USB air-gap removed in favour of managed, monitored, compliant connectivity.
    • NIS2 compliance demonstrable to regulators through structured documentation rather than retrospective evidence assembly.
    • Data integrity between Amiens and Liège protected by encryption and authentication, not by the uncontrolled handling of physical media.

    We've done this before

    See how we do Digital CDMO
  • Digital CDMO

    MTP-based modular production architecture for rapid CDMO client onboarding

    We implement an MTP-based modular production architecture at the Liège facility, enabling containerised validated workflows and Plug and Produce module integration that allows each new CDMO client molecule to be onboarded through configuration rather than construction.

    • Module Type Package library for hairy root platform

      A library of validated process modules

      Develop a library of MTP-compliant module type packages for Samabriva's hairy root platform — seed culture, main bioreactor, harvest, downstream processing — that have been characterised, validated and archived as reusable, auditable units.

    • Containerised client workspace isolation

      Each client molecule in its own validated container

      Deploy containerised environments for each active CDMO client project that provide data isolation, access controls and audit trails satisfying FDA 21 CFR Part 11 requirements without requiring a separate validated system for each client.

    • Rapid tech transfer workflow

      From client molecule to first batch in weeks, not months

      Establish a structured tech transfer workflow that uses the MTP library to accelerate the process characterisation and validation steps for new client molecules, reducing the time to first GMP batch delivery.

    • Client onboarding timeline reduced by replacing validated system construction with validated module configuration.
    • FDA 21 CFR Part 11 compliance demonstrable per client through container isolation, not separate system builds.
    • Commercial revenue from new CDMO clients unlocked faster, reducing the window between facility commissioning and positive cash flow.

    We've done this before

    See how we do Digital CDMO
  • Enterprise AI

    Digital twin for 50,000-litre bioreactor scale-up validation

    We build a data-driven digital twin model for the 50,000-litre hairy root bioreactor scale-up, integrating mechanistic growth kinetics from published literature and Samabriva's own process data with fluid dynamics simulation to predict oxygen transfer, shear stress and tissue density profiles at 50,000-litre scale before the capital commitment is made.

    • Mechanistic growth kinetics model

      The biological engine for the scale-up simulation

      Develop a mechanistic growth kinetics model for Samabriva's hairy root platform that captures the relationships between substrate concentration, oxygen demand, shear stress tolerance and growth rate at the conditions Samabriva's process operates.

    • Computational fluid dynamics integration

      Fluid dynamics at 50,000 litres before the tank is ordered

      Integrate the growth kinetics model with CFD simulation of the 50,000-litre bioreactor geometry to predict oxygen transfer coefficients, shear stress distribution and mixing time at operating conditions, identifying design risks before fabrication begins.

    • Scale-up uncertainty quantification

      Knowing what we don't know about the scale-up

      Apply uncertainty quantification to the scale-up model, identifying which process parameters have the greatest effect on success at 50,000 litres and which require additional small-scale experiments to reduce uncertainty before the campaign begins.

    • 50,000-litre scale-up risk quantified before any capital is committed, enabling an informed go/no-go decision.
    • Critical experiments at 10-litre and 100-litre scale identified from the uncertainty analysis, focusing the experimental programme on what actually matters.
    • Investor confidence in the scale-up roadmap strengthened by a quantitative, data-driven evidence base.

    We've done this before

    See how we do Enterprise AI

Digital maturity: today and target

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

Source: A4BEE analysis of public sources
Cross-site data integration 25 → 85
Process data is fragmented between Amiens R&D and Liège manufacturing in disconnected formats. No unified data platform spans clone selection, DoE, bioreactor sensor and offline analytical data.
Real-time process monitoring 30 → 80
Biomass estimation for hairy root bioreactors relies on manual sampling. No inline computer vision or real-time biomass monitoring is currently deployed at the 1,000-litre scale.
IT/OT convergence and security 25 → 80
USB-based air-gap transfers connect the legacy Amiens isolated workstations to the Liège OT network. NIS2 compliance has been identified as a gap but no structured remediation programme is in place.
Modular CDMO architecture 30 → 80
Each new CDMO client project requires bespoke system validation. MTP-based modular architecture and containerised validated workflows are not yet implemented.
Scale-up digital capability 25 → 75
50,000-litre bioreactor design is at concept stage with no digital twin or CFD-based scale-up model yet developed. Scale-up relies on published correlations and engineering judgement.
Regulatory data infrastructure 40 → 85
GMP data management infrastructure at Liège is in build-out phase. Electronic batch records, audit trail documentation and data integrity controls are being established but are not yet production-ready.

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 Samabriva S.A., 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].