☰

Eleva Biologics

Putting phototrophic bioreactors on a shared data spine

Industry
Biopharmaceuticals (Plant-Based Biologics)
Headquarters
Freiburg im Breisgau, Baden-Württemberg, Germany
Public information as of
January 2026

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

Strategic priorities

Eleva is industrialising Bryotechnology, a moss-based (*Physcomitrella patens*) expression platform, around its lead asset Factor H (CPV-104) for complement-mediated disease. The 2,000L industrial single-use bioreactor programme is funded through a $74M round announced in 2021, with clinical development advancing toward FDA/EMA Phase 1b/2 trials and an alliance with Spanish CDMO 3PBIOVIAN intended to triple GMP capacity.

The scale-up introduces physics that mammalian-cell platforms do not see. Light attenuation in a 2,000L phototrophic culture follows Beer-Lambert dynamics, so viable cell density (VCD) directly changes how much light each cell receives; the host cell protein (HCP) antibody coverage reported in scale-up shifted from 91 percent at 100L to 44 percent at 500L, and the mixotrophic transition (adding sugar to drive yield) further alters the impurity profile. Both are method questions that move onto the critical path the moment a batch is committed to clinical supply.

Operations are now distributed across two regulatory jurisdictions. The Freiburg R&D and pilot footprint remains the centre of gravity, while 3PBIOVIAN in Pamplona is contracted for clinical-grade GMP production under the alliance signed in early 2025. A new Chief Business Officer joined in January 2026, signalling an externalised-manufacturing posture: the platform is being positioned for licensing, so the data exchanged with the CDMO is also the data shown to future partners.

Digital maturity reflects a clinical-stage operation moving from research-grade to GxP-grade infrastructure: bioreactor telemetry is still largely research-style, analytical data from external providers (Alphalyse for LC-MS/MS, BioGenes for ELISA) is reconciled manually in Excel, and 21 CFR Part 11-compliant electronic records are not yet the default. Each of these is the kind of method gap that closing-loop data and paperless evidence chains are designed to address.

Challenges we see

  • Operations Manufacturing

    Holding light intensity even as biomass rises

    Moss cells are phototrophic, so the 2,000L industrial bioreactor has to deliver usable light to the entire vessel volume. The Beer-Lambert law means that as viable cell density rises, light intensity drops exponentially with depth and scattering from the larger multicellular moss structure widens the gradient — leaving regions of the tank underperforming on protein expression.

    Where light delivery is set at the start of a batch and held constant, the operating window the cells see drifts as VCD rises; modulating LED power in step with the live VCD signal narrows that drift and keeps the batch operating inside the envelope it was designed for.

  • Digital Manufacturing

    Reading HCP profile while the moss-specific assay develops

    The transition from phototrophic to mixotrophic cultivation changed the metabolic state of the cells, and HCP antibody coverage has been reported to drop from 91 percent at the 100L scale to 44 percent at the 500L scale. No off-the-shelf HCP kits exist for *Physcomitrella patens*, and a second-generation assay takes one to two years to develop and qualify.

    Where the only readout of a key impurity class is being developed, process parameters and analytical outputs that already exist can be combined into a predictive model so the team has an in-flight signal on impurity risk while the formal assay catches up.

  • Integration Operations

    Sharing process visibility with the Spanish CDMO without losing control

    The 3PBIOVIAN alliance signed in 2025 transfers Eleva's proprietary light-control bioreactor configuration to a third-party GMP facility in Pamplona, while Freiburg retains the responsibility for process design and release. Two sites, two regulatory jurisdictions, one batch record.

    Where process signals stop at the receiving site, deviation handling becomes a teleconference rather than a workflow; streaming the same parameters that the Freiburg team can see to the people running the batch shortens the loop and keeps release evidence coherent.

  • Digital Integration

    Joining external analytical outputs to in-house bioreactor data

    Critical analytical work is performed by external providers — Alphalyse for LC-MS/MS and BioGenes for custom ELISA — and returns in their own formats and timing. Inside Freiburg, bioreactor telemetry and process logs are stored separately again, so a complete batch view is assembled by hand across Excel.

    Where each output arrives in its own format, comparing them and acting on them together is a recurring manual task; a single ontology for assay, specimen, result, run, batch and instrument lets the same query read across providers and in-house sources without per-report reconciliation.

  • Compliance Regulatory

    Producing audit-ready evidence while the program runs

    Factor H (CPV-104) is advancing to Phase 1b/2, the manufacturing footprint includes a Spanish CDMO operating under EMA-aligned inspection, and proprietary *Physcomitrella patens* sequences and process know-how are part of what is being protected. Many of the supporting records — analytical, batch, change, validation — are still paper- or Excel-based.

    Where records are produced for an inspection rather than continuously during the run, evidence has to be reconstructed after the fact and is harder to defend; records produced as the work happens carry their own lineage into an inspection and reduce the time the team spends reconstructing history.

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. Dynamically adjusting light delivery to live viable cell density

    In a 2,000L phototrophic culture, viable cell density increases through the batch and the light reaching each part of the vessel decreases. A static LED setpoint leaves parts of the vessel in effect under-lit, which the moss cells register as reduced expression and which the batch record then has to absorb as heterogeneity.

    Pairing each LED module with the live VCD reading in a closed control loop keeps light intensity inside the envelope the moss needs as the cell density grows, and turns a fixed setpoint into a moving target that is read rather than assumed.

    • Eleva Biologics press release, 'Reaches Industrial Production Scale', 2021
    • Bioprocess International, 'HCP Analysis for a Moss Expression System', 2023

  2. Predicting HCP profile from existing process and analytical data

    A moss-specific HCP assay is in development and the formal readout will not arrive for another one to two years. In the meantime, manufacturing is making decisions about process changes and release against limited impurity evidence.

    Training correlation models against the LC-MS/MS data already being generated for external analyses and against the upstream process variables captured today gives the team an in-flight signal on where HCP risk is rising while the qualified assay is being completed.

    • Bioprocess International, 'HCP Analysis for a Moss Expression System', 2023
    • Eleva Biologics, Technology page

  3. Streaming live process signals between Freiburg and 3PBIOVIAN

    Process parameters from the 3PBIOVIAN bioreactors in Pamplona are not currently visible to the Freiburg team in real time, so deviations are detected after the batch finishes rather than while it is running.

    An industrial data platform that streams the same process signals from both sites into one model lets the Freiburg process team see what is happening at the CDMO as it happens and intervene earlier, rather than relying on post-batch reconciliation.

    • Eleva Biologics, 'Eleva and 3PBIOVIAN sign strategic alliance', 2025
    • Eleva Biologics press release, 14 September 2021

  4. Replacing paper- and Excel-based records with instrument-connected workflows

    Analytical instruments return results that are transcribed or copy-pasted into spreadsheets and batch records; QC review queues are tracked in shared files; and the chain of custody between an analytical run and the resulting release decision depends on the operator at the keyboard.

    Capturing results at the instrument and writing them into the batch record as data, with the audit trail attached, takes transcription out of the release path and produces evidence that stands on its own during a Phase 1b/2 inspection.

    • Eleva Technical Report on HCP Analysis, 2023
    • Eleva Biologics, Strategic Analysis Report, 2026

  5. Designing clinical-stage GxP architecture for a cloud-using CDMO operation

    Manufacturing is now split across Freiburg and a Spanish CDMO and the program is entering Phase 1b/2; many of the records exchanged across that boundary are still paper- or Excel-based, and the cloud infrastructure carrying the data is being asked to be GxP-compliant at the same time as it is being stood up.

    A documented GxP-aligned architecture with role-based access, segregated partner environments and continuous validation evidence under 21 CFR Part 11 turns the cloud into an inspection-suitable substrate for the program rather than a risk to be managed around.

    • Eleva Biologics, 'Eleva and 3PBIOVIAN sign strategic alliance', 2025
    • Eleva Biologics, Strategic Analysis Report, 2026

What we'd propose

  • Digital CDMO

    Closed-loop light and VCD control for phototrophic bioreactors

    An instrumentation and control package for the 2,000L phototrophic bioreactor that joins VCD sensing to LED power modulation and writes the resulting envelope into a per-batch record, so light delivery tracks what the cells actually experience rather than a fixed start-of-batch setpoint.

    • VCD-aware LED control loop

      Light tracks biomass

      Read viable cell density online and adjust LED power in step, using an operating envelope calibrated from runs at 100L and 500L, so light intensity stays inside the design window across the full 2,000L volume as biomass rises through the batch.

    • Per-batch light and VCD record

      Evidence for the release file

      Generate a per-batch trace of LED power, light intensity at multiple depths and VCD over the run, with lineage back to each instrument, so the release record documents what the cells actually received rather than a single setpoint held at start.

    • Digital twin for scale-up calibration

      Run the regime before the vessel is run

      Build a digital twin of the 2,000L vessel that models the Beer-Lambert attenuation at the densities expected for *Physcomitrella patens*, calibrated against 100L and 500L data, so scale-up parameters are tested in simulation before they are run in stainless steel.

    • Light delivery tracks biomass instead of the start-of-batch setpoint, narrowing the within-batch spread that the 2,000L vessel is otherwise most exposed to.
    • The same model and instrumentation that runs the industrial vessel can be reused on smaller vessels once the moss platform licenses out.
    • Scale-up is rehearsed in simulation first, so commissioning time at 2,000L is spent confirming rather than searching.

    We've done this before

    See how we do Digital CDMO
  • Enterprise AI

    An ontology-based platform for moss-process and analytical data

    A shared data model and ingestion layer that joins Eleva's in-house bioreactor telemetry, the analytical outputs from Alphalyse and BioGenes, and the process logs from both Freiburg and Pamplona, so the same query reads across providers and sites.

    • Shared process and analytical ontology

      One agreed set of terms

      Define the entities the moss platform cares about — run, batch, vessel, light regime, VCD trace, assay, specimen, result, instrument, site — with explicit relationships, so a query written once returns comparable answers across in-house and external sources.

    • Pipelines from instruments, providers and sites

      All sources loaded the same way

      Build ingestion for Eleva's bioreactor historians, for the file outputs returned by Alphalyse and BioGenes, and for the operating data the 3PBIOVIAN team shares, with schema validation at each boundary so a malformed record fails loudly.

    • Analytics on top of the combined model

      Questions answered without manual joins

      Expose the combined model through dashboards and a retrieval layer so process, analytical and quality teams can ask questions of the full dataset rather than reconciling per query, and so HCP correlation models can read across the sources that feed them.

    • One query reads across providers and in-house sources instead of joining the same spreadsheets by hand each time.
    • HCP correlation models inherit every new analytical run automatically, so they improve as the program runs.
    • The same model provides the view shown to future Bryotechnology licensing partners during technical due diligence.

    We've done this before

    See how we do Enterprise AI
  • Digital CDMO

    Cross-site process visibility for Eleva and 3PBIOVIAN

    An industrial data platform that streams the same process signals from Eleva's Freiburg site and the 3PBIOVIAN site in Pamplona into one shared view, with role-based access aligned to the responsibilities of each site and to the GxP obligations of a Phase 1b/2 program.

    • Unified multi-site dashboards

      Freiburg sees what Pamplona sees

      Aggregate process data from the Freiburg pilot and the Pamplona CDMO bioreactors into one Grafana-based dashboard, so the process team in Freiburg sees the same parameters as the people running the batch in Pamplona at the same time.

    • Deviation alerts against the active batch

      Alerts against what is running

      Set the deviation logic against a Golden Batch profile per program, so an alert fires while the affected batch is still in progress rather than appearing in a report after the batch has been released.

    • GxP-aligned partner environment

      Inspection-ready by design

      Configure the data path with the segmentation, role-based access and audit-trail handling required for 21 CFR Part 11, so the same platform that supports the alliance is also the one an FDA or EMA inspection can review.

    • Deviations are detected against the batch that is running, not the batch that has been released.
    • Freiburg and Pamplona work from the same dataset rather than from their own extracts.
    • The cross-border data exchange is auditable, which keeps it inside the program's compliance envelope.

    We've done this before

    See how we do Digital CDMO
  • Digital Lab

    Paperless laboratory and QC workflows on instrument-connected data

    A digital lab and quality-control layer that connects Eleva's analytical instruments to the batch record, removes paper logbooks and Excel consolidation from the release path, and produces the records the Phase 1b/2 program needs as the work happens.

    • Instrument-integrated result capture

      Results captured at the source

      Connect chromatography, spectrometry and balance outputs directly into the laboratory record so each analytical result arrives with instrument identity, method version and timestamp attached, rather than being read off the screen and typed into another system.

    • QC review queue and electronic batch records

      Review work tracked, not held in someone's inbox

      Replace shared spreadsheets tracking review status with a configured QC workflow and an electronic batch record that prevents steps from being skipped, captures signatures in place, and produces a complete evidence package for the release decision.

    • 21 CFR Part 11 evidence and ALCOA+ audit trail

      Inspection-suitable records

      Implement electronic signatures, versioning and the ALCOA+ data integrity attributes (Attributable, Legible, Contemporaneous, Original, Accurate, plus Complete, Consistent, Enduring and Available) that the Phase 1b/2 inspection regime expects to see in the laboratory record.

    • The release path runs through instrument records and an electronic batch record rather than through transcription and reconciliation.
    • Phase 1b/2 inspection questions are answered from the system rather than reconstructed from notebooks.
    • QC teams spend less time on consolidation and more time on the judgement calls that require them.

    We've done this before

    See how we do Digital Lab
  • Agents

    AI agents for regulatory, HCP and cross-site documentation work

    Narrow, reviewable agents that take the repetitive part of the documentation load that comes with a clinical-stage, cross-site, novel-host program: drafting the first version of HCP and process sections from underlying system records, checking regulatory templates for completeness before review, and finding every controlled document a standards change touches. A named reviewer approves every output.

    • Drafting HCP and process sections from source records

      First drafts from system data

      Generate the first draft of an HCP, process or analytical method section directly from the bioreactor, instrument and external-provider records, so the CMC author edits and judges rather than assembles the document from raw data.

    • Template and completeness checks ahead of review

      Gaps found before the regulator sees them

      Check a submission section against the FDA, EMA and PEI templates the program must complete, flagging missing or inconsistent fields before the document enters the human review queue.

    • Change-impact search across the controlled document set

      Which documents a change touches

      When a standard, a method or a process parameter changes, retrieve every controlled document that references it and rank them by how directly they are affected, so the scope of an update is known on day one rather than found by recollection.

    • Authoring time on HCP, process and method sections is moved from assembly to judgement.
    • Incompleteness is surfaced before review, not at the inspection.
    • Every output is traceable to the records it was generated 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 Eleva Biologics's own published ambition implies — not a perfect score.

Source: A4BEE analysis of public sources
Process automation 30 → 75
Bioreactor control is largely research-style, with monitoring that depends on operator entry rather than continuous capture from the vessel. The move to 2,000L and the Phase 1b/2 timeline both favour closed-loop control over human-in-the-loop data entry.
Data integration 25 → 80
External analytical outputs from Alphalyse and BioGenes sit alongside Eleva's own bioreactor telemetry in separate systems and are reconciled by hand in Excel. A shared ontology for run, batch, assay, specimen, result and instrument is the missing layer.
Predictive analytics 22 → 72
There is no production Digital Twin of the 2,000L phototrophic vessel and no model-driven HCP prediction in flight; both have been identified as the route to de-risk the 2,000L scale-up and the 1-2 year moss-specific assay development timeline.
Regulatory compliance 38 → 88
Phase 1b/2 inspection expectations and 21 CFR Part 11 obligations have to be met across two regulatory jurisdictions. Laboratory records are still paper- and Excel-based in many places, and the cloud substrate the program runs on is being GxP-aligned.
Multi-site visibility 18 → 78
Freiburg and Pamplona are not yet visible to each other in real time; deviations are detected after a batch finishes. Process transfer documentation is manual and cross-border troubleshooting is constrained.
Cybersecurity 32 → 82
Proprietary *Physcomitrella patens* sequences and process know-how, plus clinical data from a multi-jurisdiction program, require identity-based access controls and a documented architecture before the platform can be licensed out.

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 Eleva Biologics, 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].