☰

Annexin Pharmaceuticals AB

CMC scale-up through a virtual operating model

Industry
Biopharmaceuticals (Recombinant Protein Therapeutics)
Headquarters
Stockholm, Sweden
Public information as of
February 2026

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

Strategic priorities

Annexin Pharmaceuticals is a clinical-stage Swedish biopharma developing ANXV, a recombinant human Annexin A5 protein, in ophthalmology and oncology indications. The company operates with approximately four internal employees, relying on a global network of contract manufacturing organisations, contract research organisations and clinical sites for fermentation, purification, clinical trials and regulatory work. The 2025 fiscal year ended with financing estimated to last the company into April 2026.

The current operational priority is the Chemistry, Manufacturing and Controls (CMC) work needed to scale ANXV production from pilot scale to 5,000-litre commercial-scale fermentation runs capable of producing three consistent batches for market authorisation. Annexin has reported allocating roughly 17 percent of its budget to pharmaceutical product-related costs for this scale-up. Phase 2a studies in diabetic retinopathy and retinal vein occlusion have already shown promising topline results, with a database lock in the retinal vein occlusion study completed and reported by the company.

Because manufacturing, clinical and stability data all flow through external partners, the digital priority sits in the data plane rather than on any single piece of equipment. Annexin's Stockholm headquarters sees clinical results and financial reports but does not have a direct view into operational technology data at the contract manufacturing sites, and regulatory documentation arrives in PDFs, Excel spreadsheets and paper formats that have to be reassembled by hand into submission packages. A platform that ingests data from the contract sites into one auditable model sits on the critical path for Phase 3 readiness and for the licensing deals the company is pursuing with major pharmaceutical partners.

Challenges we see

  • Operations Manufacturing

    Stabilising recombinant fermentation at commercial scale

    ANXV is produced by expressing human Annexin A5 protein in living E. coli cells. At 5,000-litre scale, factors including oxygen transfer, temperature and metabolic load can produce inclusion bodies or total batch failure. Annexin has reported that 17 percent of its budget is allocated to pharmaceutical product-related costs covering this scale-up work.

    Where fermentation is monitored by sampling after the fact, the unit operation under review is a whole batch; reading process signals as the batch runs narrows that population to the units that actually deviated, which is the difference between a recoverable intervention and a write-off.

  • Digital Integration

    Joining contract manufacturing, research and clinical data into one estate

    Clinical data arrives from sites in London and the United States, and CMC data arrives from contract manufacturing partners, in heterogeneous formats including PDFs, Excel spreadsheets and paper records. The company's senior leadership has stated publicly that potential licensing partners want to review both topline results and detailed patient-level outcomes.

    When each partner delivers data in its own format, the consolidated regulatory submission has to be rebuilt by hand, and the cognitive load of doing so grows with the partner count rather than with the science; an ontology-based ingestion layer turns that work into a per-document review rather than a per-report reconstruction.

  • Compliance Regulatory

    Producing audit-ready traceability across external partners

    Phase 3 readiness and licensing due diligence require traceability from source records to the submission under FDA 21 CFR Part 11 and GAMP5. The company has acknowledged that its virtual model means most operational records sit with external contractors.

    Where data lineage has to be reconstructed across organisations, audit questions are answered from secondary evidence rather than from primary records; making the audit trail a property of the platform rather than an end-of-project assembly changes the question from "can we find this" to "can we show where this came from".

  • Digital Manufacturing

    Reading operational technology signals from contract manufacturing sites

    Stockholm headquarters sees clinical and financial information but lacks real-time visibility into fermentation and purification metrics at the contract manufacturing sites. The company has stated publicly that it collaborates with experienced contractors executing manufacturing and clinical trials on its behalf.

    Where fermentation parameters arrive by exception rather than by stream, the time between a deviation emerging and a write-off decision is compressed into a window that cannot be widened by escalation; instrumenting the line so the same data is visible to the sponsor and to the contract manufacturer makes the time window a design choice rather than a property of the report cadence.

  • Operations Regulatory

    Keeping drug stability evidence continuous through Phase 3

    Drug stability testing across temperature, humidity and light exposure is a core CMC requirement for late-stage approval. The company's virtual model means stability data is reported manually by contract laboratories rather than captured directly from instruments.

    Where stability measurements arrive as periodic contractor reports, the granularity of the evidence is whatever the contract laboratory's cadence can sustain; capturing readings at the instrument and writing them into the regulatory record as data keeps the evidence continuous rather than sampled.

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 fermentation signals as a commercial-scale batch runs

    Recombinant ANXV production at 5,000-litre scale introduces variability in oxygen transfer, temperature and metabolic load that can produce inclusion bodies or total batch failure, with each failure representing a multi-million-euro write-off against a cash runway that runs into April 2026.

    Instrumenting the bioreactor line so process data leaves the equipment in a vendor-neutral form, and running anomaly detection against the running batch, lets a deviation be flagged while the batch is still recoverable, which is the difference between an investigation and a write-off.

    • Annexin interim report, March 2025 (rights issue of approximately SEK 50 million)
    • Annexin corporate materials: "The production process is complicated, which is why recombinant proteins are becoming relatively expensive drugs."

  2. Joining contract manufacturing, contract research and clinical data into one model

    Clinical data arrives from sites in London and the United States, and CMC data from contract manufacturing partners, in heterogeneous formats including PDFs, Excel spreadsheets and paper records. The CEO has stated publicly that licensing partners want both topline results and detailed patient-level outcomes.

    An ontology-based data platform defines assay, batch, lot, site and patient as explicit entities with agreed relationships, then ingests the heterogeneous partner outputs into one model, so regulatory submissions, licensing due diligence and executive review all read from the same evidence.

    • The company's CEO, on licensing partner data review
    • Annexin Pharmaceutical company materials on virtual operating model

  3. Streaming operational technology data from contract manufacturing sites

    Stockholm leadership can see clinical and financial data but cannot see fermentation and purification metrics at the contract manufacturing sites in real time, which compresses the response window for a deviation into the time between contractor report and review.

    An OPC UA (Open Platform Communications Unified Architecture) gateway at the contract manufacturer streams process parameters to the sponsor dashboard with role-based access and an audit trail, so the sponsor and the contractor see the same line data.

    • Annexin website: collaboration with experienced contractors
    • Annexin interim report, March 2025

  4. Capturing drug stability data directly at the instrument

    Drug stability testing across temperature, humidity and light exposure is reported by contract laboratories through periodic manual reports rather than captured at the instrument, leaving the evidence granularity at the contractor's report cadence.

    Connecting stability chambers to a monitoring layer so readings arrive as data with timestamp, instrument identity and method version, and writing the consolidated evidence directly into the regulatory record, keeps the evidence continuous rather than sampled.

    • Annexin interim report, March 2025, on stability and CMC scope
    • Annexin corporate materials on scale-up and analytical methods

  5. Drafting regulatory and licensing documents from source records

    IND-enabling CMC dossiers, periodic regulatory submissions, orphan-drug designation evidence and licensing due-diligence packages all consume specialist time at a four-employee organisation, and most of that time goes on assembling and checking documents rather than on the technical judgement inside them.

    Narrow AI agents can draft the first version of a regulatory or licensing document from the source records on the platform, check a document against its template before review, and find every controlled document a standards change touches, with a named reviewer approving every output.

    • Annexin interim report, March 2025
    • Annexin corporate materials on resource concentration

What we'd propose

  • Digital CDMO

    Real-time fermentation visibility at the contract manufacturer

    We instrument the contract manufacturer's bioreactor line so process data leaves the equipment in a documented, vendor-neutral form, stream into a monitored model, and run anomaly detection against the running batch, so a deviation is flagged while the batch is still on the line.

    • Line data acquisition

      Getting data off the equipment

      Connect controllers, sensors and inspection stations through OPC UA (Open Platform Communications Unified Architecture) or MQTT so process values leave the bioreactor in a documented, vendor-neutral form rather than staying inside a closed controller, with the sponsor and the contract manufacturer both able to read the same values.

    • Deviation detection against a running batch

      Alerts while the batch runs

      Build the normal operating envelope for each critical process parameter from historical runs, then flag drift against the current batch so both Annexin's leadership and the contractor's shift team see a signal in minutes instead of a result in a later report.

    • Sponsor and contractor shared dashboard

      One view, two organisations

      Expose the streamed data through a dashboard with role-based access so Stockholm leadership sees the same line data as the contract manufacturer, with the same audit trail behind every value.

    • A deviation is flagged against the batch that is running, not the batch that has already been written off.
    • Sponsor and contractor read the same line data, so an investigation starts from one record rather than two.
    • The instrumented line is reusable across partner sites, so adding a new contract manufacturer is a configuration step rather than an integration project.

    We've done this before

    See how we do Digital CDMO
  • Enterprise AI

    One data model for contract manufacturing, research and clinical partners

    An ontology-based data platform that defines the entities a virtual biopharma shares with its partners — assay, batch, lot, site, patient — once, then ingests the heterogeneous partner outputs (PDFs, Excel, instrument files) into that single model with full lineage.

    • Shared biopharma ontology

      One agreed set of terms

      Define assay, batch, lot, site and patient as explicit entities with agreed relationships, so a query written once returns comparable answers across contract manufacturers, contract research organisations and clinical sites instead of two dialects of the same table.

    • Heterogeneous source ingestion

      Loading every partner format

      Build ingestion for instrument files, Excel spreadsheets, PDF reports and laboratory information management system outputs, with schema validation at the boundary so a record that does not parse fails loudly rather than silently entering the regulatory evidence base.

    • Regulatory and licensing retrieval layer

      Questions answered without a new extract

      Expose the model through a retrieval layer so regulatory affairs, business development and the chief executive can ask questions of the combined data set without commissioning a new extract for each submission or due-diligence request.

    • Integration work is done once against a shared model rather than once per point-to-point interface.
    • Licensing due-diligence questions are answered from the platform rather than from a manual compilation.
    • New partners attach to the model by ingesting their format, rather than triggering a migration.

    We've done this before

    See how we do Enterprise AI
  • Digital CDMO

    Streaming operational technology data from contract manufacturing sites

    An OPC UA gateway and network architecture at the contract manufacturing site that streams bioreactor and purification data to a sponsor dashboard with role-based access, audit trail and alerting, so Stockholm leadership and the contractor see the same line.

    • OPC UA gateway at the contract site

      One documented data path

      Specify the protocol standards, network segmentation and equipment data requirements for the contract manufacturer's line, making interoperability a contract condition rather than an integration project after handover.

    • Role-based access and audit trail

      Sponsor and contractor on the same line

      Configure role-based access so Annexin's leadership and the contractor's operations team see the same line data with the same audit trail behind every value, with a documented record of who read which parameter when.

    • Tiered alerting

      The signal rises above the noise

      Configure tiered alerting logic that notifies Annexin's leadership of critical process deviations while filtering routine operational noise, so the response window widens from a manual report cadence to a continuous monitor.

    • The sponsor sees the line in real time rather than waiting for the next contractor report.
    • The deviation-to-decision window widens from a manual report cadence to a continuous monitor.
    • The same architecture is reusable across multiple contract manufacturers, so adding a partner is a configuration step.

    We've done this before

    See how we do Digital CDMO
  • Digital Lab

    Capturing drug stability data directly at the instrument

    We connect the contract laboratory's stability chambers and analytical instruments so readings arrive as data with timestamp, instrument identity and method version attached, and write the consolidated evidence into the regulatory record as part of the platform's normal flow.

    • Instrument integration at the contract laboratory

      Results captured at source

      Connect stability chambers, balances and analytical instruments so readings are captured with instrument identity, method version and timestamp attached, instead of being read off a screen and typed into a contractor report.

    • Stability evidence into the regulatory record

      Lab result to submission evidence

      Map the contract laboratory's sample and result records onto the regulatory submission record so a stability claim can be traced back to the specific instrument reading that produced each number.

    • 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, so the stability evidence chain stands on its own during an inspection.

    • Stability evidence is continuous rather than sampled at the contractor's report cadence.
    • An inspection question is answered from the record itself rather than from a reconstruction.
    • The same instrument integration pattern is reusable across multiple contract laboratories.

    We've done this before

    See how we do Digital Lab
  • Agents

    AI agents for regulatory and licensing document work

    Narrow, reviewable agents that take the repetitive part of document work: drafting IND-enabling CMC sections from the platform's source records, checking a document against its template before review, and finding every controlled document a standards change touches. A named person approves every output.

    • Drafting from source records

      First drafts from platform data

      Generate the first draft of an IND-enabling CMC section, a stability summary or a licensing due-diligence package directly from the platform's source records, so the author edits and judges rather than assembles from PDFs and spreadsheets.

    • Template and completeness checking

      Gaps found before review

      Check a submitted document against its template and the company's own checklist, returning missing or inconsistent sections before it enters the human review queue, so a small team spends its review time on the technical judgement inside the document.

    • Change impact search across the document set

      Which documents a change touches

      When a 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 rather than rediscovered during the next audit.

    • A four-person team produces submission-ready drafts rather than reconstructing them from partner inputs.
    • Review queues move faster because documents arrive complete against the template.
    • 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 Annexin Pharmaceuticals AB's own published ambition implies — not a perfect score.

Source: A4BEE analysis of public sources
Data integration 25 → 80
The virtual operating model creates systemic data silos with manual PDF and Excel consolidation across contract manufacturing, contract research and clinical partners; a shared ontology is ahead of the company rather than behind it.
Process automation 20 → 75
Manufacturing, clinical and stability processes rely heavily on contractor manual reporting with limited digital orchestration, leaving automation gains available across every external partner interface.
Real-time visibility 15 → 85
There is no operational technology convergence at the contract manufacturing sites today, and the sponsor's view of fermentation and purification parameters is periodic rather than continuous; closing this gap is a precondition for the Phase 3 work.
Regulatory compliance 40 → 90
GAMP5 compliance is challenging without unified audit trails across external partners, and the manual data pipelines that exist today increase the integrity risk on Phase 3 submissions and licensing due-diligence packages.
Predictive analytics 10 → 70
No digital twin or AI-driven batch prediction is in place; the 5,000-litre scale-up work proceeds as a sequence of physical batches with retrospective analysis rather than as a model-assisted sequence with deviation forecasts.
Partner readiness 35 → 85
Data presentation for licensing partners requires significant manual compilation today and lacks the plug-and-produce infrastructure that the company's stated go-to-market strategy assumes; building that layer is a precondition for the partnership pipeline.

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 February 2026. It reflects A4BEE's own interpretation and opinion, is not affiliated with, endorsed by, or verified with Annexin Pharmaceuticals AB, 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].