☰

Cenexi

Sterile injectable output across four sites, one data spine

Industry
Pharmaceutical Contract Manufacturing
Headquarters
Fontenay-sous-Bois, Île-de-France, France
Public information as of
January 2026

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

Strategic priorities

Cenexi runs four sterile injectable sites — three in France and one in Belgium — with the Fontenay-sous-Bois headquarters serving as the ampoule centre of excellence and Braine-l'Alleud dedicated to oncology and lyophilized vials. The group was acquired by Gland Pharma, a subsidiary of China's Fosun, in April 2023, and leadership has framed a 4IN strategy covering Innovation, Internationalization, Intelligentization and Integration to lift the European business to double-digit EBITDA by FY27.

The most immediate operational question sits at Fontenay. In June 2025 the French regulator ANSM closed a GMP inspection at the site with 11 observations and a 3 to 12 month corrective-action window; the public reporting on the inspection places documentation and data integrity at the centre of the findings. A separate lyophilizer breakdown at Braine-l'Alleud in late 2024 underlines how thin the margin is between routine production and a revenue-impacting stop.

Cenexi Services, the roughly 110-strong formulation and transfer unit that interacts with every customer, runs against 124 country registrations and a stream of new biologic programmes. The stated FY27 shift toward mAbs, ADCs and peptides makes the speed of a tech transfer — and the data trail it leaves — a competitive variable rather than a back-office one, and shifts the priority onto paperless development and an MTP-based (Module Type Package, the VDI/VDE/NAMUR 2658 standard for plug-and-produce automation) modular production model at the Hérouville site.

A €60 million CAPEX plan is the visible part of the answer. The Braine-l'Alleud backup lyophilizer entered service in the first half of 2025, a new high-speed ampoule line at Fontenay was commissioned in January 2025, and additional line replacements are scheduled across the four sites through FY27. The digital work that lets those assets run as a connected network rather than four independent ones is the less visible companion programme.

Challenges we see

  • Operations Reliability

    Keeping lyophilizers and high-speed lines available

    A lyophilizer breakdown at Braine-l'Alleud in late 2024 caused a temporary setback, and a separate machinery breakdown at the Fontenay plant was reported as a drag on FY25 EBITDA. The €60 million CAPEX programme is replacing old lines; a new high-speed ampoule line at Fontenay entered commercial production in January 2025 and a backup lyophilizer unit was installed at Braine-l'Alleud in the first half of 2025.

    Where unplanned downtime on a sterile line is measured after the fact, the population under review is the whole batch. Reading the equipment's electrical and mechanical signatures continuously narrows that population to the deviation window, and the same data feeds the engineering team that plans the next replacement.

  • Quality Regulatory

    Documenting quality work so an inspection can read it

    ANSM issued 11 observations following a June 2025 GMP inspection at the Fontenay facility, with a 3 to 12 month CAPA window. Cenexi has run the same site under sustained FDA monitoring for the infant formula corrective actions and operates in the EU regulatory environment for sterile injectable release.

    Where evidence of a quality decision is compiled from paper records by hand, the time between the work and the record that proves it is the inspector's reading window, and each transfer is a step to verify. Generating the record from the work as it is done makes that window short and the evidence chain readable end to end.

  • Data Integration

    Putting lab and plant data on the same path

    Cenexi Services manages roughly 124 country registrations and handles the tech transfer from development to manufacturing for every customer. The four sites run on different equipment generations and the Cenexi Services analytical and formulation team operates against instruments that pre-date the parent's data infrastructure.

    Where each site and each development unit keeps its own data, the same assay reads as a different artefact in each system and a transfer between them needs reconciliation. A shared model described once lets both ends of a tech transfer point at the same thing, and pre-empts the integration tax that arrives with every new acquisition.

  • Operations Process

    Building the data needed for a biologics portfolio

    The Hérouville site is positioned as the partner for innovative biotechs and vaccines, with the Sterile Product Development Centre sitting alongside commercial capacity. The biologics portfolio adds mAbs, ADCs and peptides to a base that has historically been high-volume, low-margin sterile liquids.

    Where a new biologic process is characterised batch by batch, the first commercial runs are also the first time the process is exercised at scale. A model that runs the process virtually before the physical run makes the early commercial scale-up a matter of execution rather than discovery, and the data collected along the way belongs to the next programme as well.

  • Compliance Cybersecurity

    Securing plant networks as Gland integration widens the perimeter

    The integration with Gland Pharma's global IT infrastructure extends the connected footprint, and the Fontenay site handles narcotics and high-potency substances that are subject to additional controlled-substance reporting. The EU NIS2 directive applies to manufacturing organisations of Cenexi's size from the 2024 transposition date.

    Where plant networks were originally designed as isolated, the parent's global estate brings a route in that the original design did not account for. Designing the segmentation and remote-access model for the new perimeter up front is cheaper than rebuilding it after a finding, and the IEC 62443 framework gives that design a standard to point at.

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. Reading asset health from sterile lines while they run

    Recent lyophilizer and ampoule-line breakdowns at Braine-l'Alleud and Fontenay disrupted production and were a stated drag on FY25 EBITDA. The €60M CAPEX programme replaces the lines but does not, on its own, instrument the lines that are still in service.

    Streaming vibration, temperature and pressure signatures from critical equipment into a monitored model lets a deviation be flagged against the machine that is running, and gives the maintenance team the same view of the asset the operator has at the panel.

    • Gland Pharma Q2 FY25 press release, 4 November 2024
    • Gland Pharma Q4 FY25 result update, brokerage note

  2. Generating GMP evidence from the work rather than after it

    The 11 ANSM observations at Fontenay in June 2025 and the 3 to 12 month CAPA window put documentation and data integrity at the centre of the next inspection. The current model of compiling evidence from paper records leaves each transfer as a separate step to verify.

    Capturing the quality event at the source and writing it into the batch record as data with its own audit trail shortens the release path and gives the next inspection a continuous record rather than a periodic reconstruction.

    • Gland Pharma ANSM inspection update, 4 June 2025
    • Gland Pharma 47th AGM, 28 August 2025

  3. Speeding tech transfer through a shared development-to-manufacturing model

    Cenexi Services supports roughly 110 experts across formulation, analytical development and industrial transfer, and 100 percent of Cenexi customers interact with the unit. Manual data entry and disconnected systems between the development and site teams lengthen time-to-market for biotech partners.

    A shared ontology for formulation, process and analytical data lets development and manufacturing reference the same artefact, which lets a tech transfer be reviewed and approved against the record rather than against a re-typed copy.

    • Cenexi Services overview, cenexi.com
    • Gland Pharma investor presentation on the Cenexi acquisition, 29 November 2022

  4. Simulating biologics processes before the first commercial batch

    The shift toward mAbs, ADCs, peptides and vaccines adds processes that are more variable than the high-volume sterile liquids that have historically been Cenexi's base. The first commercial runs at scale are the first time the process is exercised at full tonnage.

    A process model fitted to historical development runs lets the team run virtual trials before the physical run, hold yield sensitivity to upstream parameters, and bring the first commercial batch closer to the development profile.

    • Cenexi in a nutshell, cenexi.com
    • Gland Pharma Q2 FY25 press release, 4 November 2024

  5. Producing regulatory and tech-transfer documents from the underlying records

    CAPA documents, country registration updates, tech-transfer reports and periodic reviews all consume specialist time across the four sites, and much of that time goes on assembling and checking documents rather than on the technical judgement inside them.

    Narrow AI agents can draft from source records, check completeness against the site template before a document enters the review queue, and identify every controlled document a standards change touches, with a named reviewer approving each output.

    • Gland Pharma 47th AGM, 28 August 2025
    • Cenexi Services overview, cenexi.com

What we'd propose

  • Digital CDMO

    Real-time asset health monitoring for sterile lines

    We instrument critical equipment across the four sites, bring its process and condition data into one time-series model, and run anomaly detection against the machine that is currently running, so a maintenance signal arrives while the line is still turning.

    • Equipment data acquisition

      Getting data off the asset

      Connect controllers, sensors and condition-monitoring instruments through OPC UA (Open Platform Communications Unified Architecture) or MQTT so vibration, temperature, pressure and current signatures leave the equipment in a documented, vendor-neutral form rather than staying inside a closed controller.

    • Anomaly detection against a running asset

      Alerts while the line runs

      Build the normal operating envelope for each critical asset from historical runs, then flag drift against the current shift so the maintenance team sees the signal in minutes rather than after the next breakdown.

    • Maintenance and replacement planning layer

      From alert to plan

      Surface the deviation in the maintenance management system with the asset's history and the engineering team's replacement schedule attached, so the next conversation about a line is informed by what the line has actually been doing.

    • Condition signals arrive against the asset that is running, not the one that already stopped.
    • Maintenance, engineering and operations look at the same asset data rather than three separate extracts.
    • The €60M CAPEX replacement plan is informed by what the lines that are still in service are actually doing.

    We've done this before

    See how we do Digital CDMO
  • Digital Lab

    Paperless quality documentation and digital batch release

    We integrate the quality control lab's instruments and its LIMS or ELN (Laboratory Information Management System or Electronic Lab Notebook) so quality events reach the batch record as data carrying their own audit trail, taking paper-handling and re-keying out of the release path.

    • Instrument integration

      Results captured at source

      Connect balances, particle counters, HPLC (High-Performance Liquid Chromatography) and environmental monitors so results are captured with instrument identity, method version and timestamp attached, instead of being read off a screen and typed into another system.

    • Lab to batch record data flow

      Lab result to batch record

      Map the lab's sample and result records onto the manufacturing batch record so a release decision can be traced back to the specific analytical run that produced each number, across the Fontenay, Braine-l'Alleud, Hérouville and Osny sites.

    • 21 CFR Part 11 audit trail

      Electronic records that hold up

      Implement electronic signature, versioning and audit-trail handling to 21 CFR Part 11, the US rule on electronic records and signatures, and the equivalent EU Annex 11 expectations, so the evidence chain reads as a single record during the next ANSM inspection.

    • Fewer manual transfers between the lab and the batch record, and fewer of them to verify.
    • Each ANSM observation about documentation has a named, timestamped event in the system to point at.
    • Release waits on the analytical result, not on the paperwork that follows it.

    We've done this before

    See how we do Digital Lab
  • Digital CDMO

    MTP-based modular production for biologics partners

    An MTP (Module Type Package, the VDI/VDE/NAMUR 2658 standard for plug-and-produce automation) framework for the Hérouville site, with vendor-neutral OPC UA information models written into procurement requirements, so adding a new biotech partner's process is a configuration exercise rather than an integration project.

    • Reference architecture for the modular site

      One documented module pattern

      Specify how a process module, its supervision layer and the site orchestration layer connect, including the protocol choices and segmentation model, so every new module is built against the same target.

    • Module Type Package definitions

      Plug-and-produce modules

      Write the MTP information models for the unit operations the Hérouville site hosts, so a new piece of equipment can be described by a published MTP rather than re-engineered each time.

    • Process orchestration layer

      Recipe and alarm handling

      Provide centralised supervision, recipe execution and alarm handling across modules, with a single HMI (Human-Machine Interface) pattern, so operators moving between modules see a consistent surface.

    • A new biotech partner's process is configured against existing modules rather than engineered from scratch.
    • The Hérouville site can hold a portfolio of small-batch biologics programs without each one consuming its own automation overhead.
    • The same MTP pattern is reusable when other sites add new production capacity.

    We've done this before

    See how we do Digital CDMO
  • Digital Lab

    Digital twin for biologics process characterisation

    A process model fitted to historical development and clinical-batch data for the biologics portfolio, run ahead of the physical batch to test upstream parameters and yield sensitivity, with the model's outputs reviewed against the next physical run.

    • Process simulation engine

      Predicting batch outcomes

      Build a mechanistic model for the mAb, ADC and peptide processes that Cenexi Services is transferring, fitted to historical runs, so the team can run sensitivity analyses before committing to a set of operating parameters.

    • Real-time data integration

      Live data into the model

      Connect the laboratory and pilot bioreactors' data streams into the model so that each new run refines the parameter estimates and the next physical run starts from a tighter prior.

    • Golden batch comparison

      Benchmarking against the best run

      Compare each new batch's trajectory against the historical best-performing run, so the team can see early in the batch whether the current run is tracking the expected profile.

    • The first commercial runs at scale start from a parameter set the model has already exercised.
    • Development data continues to earn its keep across the portfolio rather than sitting in a project archive.
    • Yield sensitivity to upstream parameters is known before the next physical experiment.

    We've done this before

    See how we do Digital Lab
  • Agents

    AI agents for regulatory and tech-transfer documentation

    Narrow, reviewable agents that take the document-assembly part of regulatory and tech-transfer work: drafting CAPA updates from the underlying records, checking a registration file against the country template, and finding every controlled document a change in the standard touches. A named person approves every output.

    • Drafting from source records

      First drafts from system data

      Generate the first draft of a CAPA update, a tech-transfer report or a periodic review directly from the underlying system records, so the author edits and judges rather than assembles.

    • Template and completeness checking

      Gaps found before review

      Check a submitted document against the country template and the site's own checklist, returning missing or inconsistent sections before the document enters the human review queue.

    • Change impact search across the document set

      Which documents a change touches

      When a regulatory standard, method or specification changes, retrieve every controlled document that references it and rank them by how directly they are affected, so the update scope is known on day one.

    • Review queues move faster because documents arrive complete.
    • The scope of a standards change is established by search rather than by recollection.
    • Every output is traceable to the source records it came from and signed off by a named reviewer.

    We've done this before

    See how we do Agents

Digital maturity: today and target

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

Source: A4BEE analysis of public sources
Asset Connectivity 25 → 75
The four sites run on a mix of equipment generations, including the older lines at Fontenay that the €60M CAPEX programme is replacing. The 2025 commissioning of the new high-speed ampoule line and the Braine-l'Alleud backup lyophilizer is an opportunity to set the connectivity standard for the next replacement wave.
Data Integrity & Compliance 35 → 90
The 11 ANSM observations from the June 2025 Fontenay inspection and the 3 to 12 month CAPA window place documentation and data integrity at the centre of the regulator's current view of the company. The work is ahead of the company rather than behind it.
Process Analytics 20 → 70
The current visibility into the biologics processes being transferred into the Hérouville site is largely batch-level rather than continuous, and the digital twin programme is described in the strategic intent rather than in operating practice.
Cybersecurity Posture 30 → 80
The integration with Gland Pharma's global IT infrastructure extends the connected footprint, and Fontenay's handling of narcotics and high-potency substances raises the bar for evidence. The current design reflects the pre-integration estate.
Technology Transfer Automation 40 → 85
Cenexi Services runs 124 country registrations and a continuous flow of new biotech programmes against an instrument and data estate that pre-dates the parent's infrastructure. The bottleneck is the manual points between development and manufacturing.
Manufacturing Flexibility 35 → 80
The transition to high-value biologics requires a manufacturing model that can take on a new partner's process without re-engineering the automation each time. The Hérouville MTP programme is the start of that model, not its completion.

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 Cenexi, 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].