☰

Aerbio A/S

Treating gas fermentation as a queryable process

Industry
Industrial Biotechnology (Single-Cell Protein)
Headquarters
Copenhagen, Denmark
Public information as of
February 2026

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

Strategic priorities

Aerbio A/S produces Proton™, a single-cell protein (SCP) made by aerobic gas fermentation using hydrogen-oxidising bacteria, principally Cupriavidus necator, that convert captured CO2, hydrogen and oxygen into a microbial biomass. The company emerged in 2024 from a management buy-out of Deep Branch Biotechnology and runs a pilot facility at the Brightlands Chemelot Campus in Geleen, Netherlands, that produces around 200 kilograms of Proton™ per month. A €50 million Series A round was resumed in 2025 to fund the company's next step: a 'Market Launch Facility' designed for 250 to 1,000 tonnes per year.

The economic case rests on reaching price parity with fishmeal without a sustainability premium. The company's CTO identifies gas-liquid mass transfer — the rate at which hydrogen and CO2 dissolve into the fermentation broth — as mission-critical, and the company's analysis states that energy and hydrogen together account for 60 to 70 percent of operating expenditure. Process yield, energy scheduling and the stability of the bacterial culture are therefore the operational levers that determine whether the unit economics work.

The regulatory case rests on European Food Safety Authority (EFSA) and UK Food Standards Agency (FSA) Novel Food and Novel Feed approvals, which require a continuous, auditable record of batch consistency and traceability from the gas source to the protein powder. The REACT-FIRST consortium — Drax as CO2 supplier, BioMar as aquafeed customer and Sainsbury's as downstream retailer — adds a second reporting load: roughly £3 million of UKRI grant funding with milestone, life-cycle assessment and carbon-intensity reporting attached. The same data estate has to satisfy both regulators and partners.

Across the pilot plant, the design of the new facility and the consortium, three capabilities recur: a bankable process model that lets the company reason about vessel design before concrete is poured; instrumented bioreactors whose off-gas and broth signals are read as data rather than observed by eye; and a shared data model that ties a kilogram of Proton™ back to its gas source, its fermentation run and its downstream batch record.

Challenges we see

  • Operations Manufacturing

    Sizing the next vessel on physics rather than intuition

    Aerbio's CTO identifies gas-liquid mass transfer — how quickly hydrogen and CO2 dissolve into the fermentation broth — as mission-critical. At larger scales gas bubbles rise too quickly or coalesce, the surface area for transfer falls and the bacteria are starved. The company is designing a Market Launch Facility intended to produce 250 to 1,000 tonnes a year, and the vessel chosen now determines operating cost for a decade.

    Where reactor design is sized from intuition and pilot data alone, the new vessel is either over-specified — wasting capital — or under-specified — creating a future bottleneck — and the answer is rarely visible until the tank is built and running.

  • Operations Manufacturing

    Holding a bacterial culture stable through volatile hydrogen supply

    Green hydrogen from electrolysis is scarce and expensive today, and blue hydrogen from natural gas compromises the carbon-neutral positioning. The company states that energy and hydrogen together account for 60 to 70 percent of operating expenditure, and a fluctuation in either can crash a biological culture or destroy unit economics.

    Where fermentation runs on a fixed schedule regardless of energy price, the highest-cost hours become the production hours, and demand response moves from being an option to being the operating model.

  • Digital Integration

    Joining pilot, academic and consortium data into one estate

    Operations span Copenhagen (HQ and strategic functions), Geleen (pilot plant at Brightlands Chemelot) and academic partners including the Synthetic Biology Research Centre at Nottingham. Genomic and metabolic models developed in Nottingham run separately from the SCADA (Supervisory Control and Data Acquisition) systems that hold the pilot-plant data, and consortium partners exchange quality certificates, carbon-intensity figures and feed-trial results by email and spreadsheet.

    Where the gas source, the fermentation run and the downstream batch record sit in separate places, the line of evidence a regulator or a partner needs is reconstructed by hand for each request, which makes a continuous record harder to maintain than a periodic one.

  • Compliance Regulatory

    Keeping batch consistency measurable for Novel Food approval

    EFSA and FSA Novel Food dossiers require proof of batch-to-batch consistency with statistical evidence that the process is in control, plus full traceability from the CO2 source through to the protein powder. Using CO2 captured from industrial flue gas raises questions about impurities entering the food chain that the dossier has to address for every batch.

    Where batch protein content and amino-acid composition are read after each run rather than tracked as a continuous statistical signal, the evidence of consistency has to be compiled for each submission, and a single out-of-range batch can hold up a filing.

  • Operations Manufacturing

    Catching contamination before a tank has to be dumped

    Aerobic gas fermentation is highly prone to contamination by faster-growing wild microbes. A single contamination event requires dumping the entire tank, costing days of production and the materials that went into it.

    Where contamination is noticed by eye after it has taken hold, the batch is already lost; the difference between an early signal and a late observation is the difference between a saved batch and a written-off one.

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. A digital twin for the next-generation vessel

    The Market Launch Facility is in design, and the vessel geometry, impeller design and gas-sparging arrangement chosen now will set hydrogen mass-transfer efficiency — and therefore energy use per kilogram of protein — for the lifetime of the plant.

    Computational fluid dynamics (CFD) and metabolic models brought into one simulation environment let the engineering team test reactor geometries, impeller designs and gas-sparging rates in software, so the vessel that gets built is the one the model says will deliver the required mass-transfer rate, not the one that looked right in pilot.

    • Aerbio's CTO, on mass transfer as mission-critical
    • Aerbio Series A overview deck, 2024

  2. Scheduling fermentation against the energy market

    Energy and hydrogen together account for 60 to 70 percent of operating expenditure, and the company is explicit that it must reach price parity with fishmeal without relying on a sustainability premium.

    Connecting the bioreactor control system to real-time electricity pricing and hydrogen-supplier availability lets fermentation intensity, aeration and feeding track the cheapest energy windows while keeping the bacterial culture stable, so the same biology is run at a lower unit cost.

    • Aerbio Series A overview deck, 2024 — OpEx composition
    • Aerbio's CTO, on hydrogen availability

  3. A shared data backbone for pilot, consortium and regulator

    Aerbio's fermentation runs sit in pilot-plant SCADA, genomic and metabolic models sit in academic systems, and partner-facing data — carbon intensity scores, batch certificates, feed-trial results — is exchanged by spreadsheet and email with Drax, BioMar and Sainsbury's. EFSA, FSA and UKRI each want a view over the same underlying record.

    A shared data model for gas, run, batch, certificate and life-cycle metric, with role-based access for partners and read-only access for regulators, lets the same data answer questions from the bioreactor team, the consortium coordinator and the regulator without being recompiled for each.

    • REACT-FIRST consortium partners (Drax, BioMar, Sainsbury's)
    • UKRI / Innovate UK grant reporting requirements

  4. Predicting contamination from off-gas and broth signals

    Aerobic gas fermentation is highly prone to contamination by faster-growing wild microbes, and a contamination event requires dumping the entire tank — a direct cost in lost production.

    Pattern-recognition models trained on off-gas composition (CO2 production, O2 consumption, respiratory quotient), broth pH, dissolved oxygen and optical density can flag a contamination event hours before it becomes visible, giving operators the choice of corrective action or early termination rather than a guaranteed write-off.

    • Cupriavidus necator H16 physiology and fermentation literature
    • Aerbio technical description, gas-fermentation scale-up

  5. AI agents for Novel Food dossier and grant reporting

    EFSA and FSA Novel Food submissions, UKRI milestone reports and consortium life-cycle assessments each draw on the same fermentation and batch data but are currently compiled for each request.

    Narrow AI agents can draft the repetitive sections of the Novel Food dossier from the source records, check a submission against its template before review, and assemble the UKRI milestone report from operational data, with a named person approving every output before it leaves the company.

    • EFSA Novel Food application requirements
    • UKRI / Innovate UK grant reporting obligations

What we'd propose

  • Digital Lab

    A digital twin for gas-fermentation scale-up

    A simulation environment that combines computational fluid dynamics for gas-liquid mass transfer with metabolic models for Cupriavidus necator, so the reactor geometry, impeller design and gas-sparging arrangement for the Market Launch Facility are sized against a model rather than against pilot intuition.

    • CFD reactor model

      Modelling gas-liquid mass transfer

      Build a computational fluid dynamics model of the proposed vessel that predicts mass-transfer coefficient (kLa) for hydrogen and CO2 across candidate geometries, so the configuration that goes to fabrication is the one the physics supports.

    • Metabolic state soft sensor

      Estimating what cannot be measured

      Combine the metabolic model with the live broth readings to estimate the intracellular carbon-to-nitrogen ratio in real time, so the control system can keep the organism in its protein-producing state rather than drifting into polyhydroxybutyrate (PHB) accumulation.

    • Scale-up scenario library

      Configurations tested before fabrication

      Run parametric studies comparing reactor geometries and operating conditions, so the engineering decision is a comparison of modelled outcomes rather than a single chosen configuration.

    • The reactor that gets built is the one the model says will deliver the required mass-transfer rate.
    • Capital risk moves from a single physical configuration to a set of modelled ones.
    • The same model becomes a defensible reference for Series A and project-finance diligence.

    We've done this before

    See how we do Digital Lab
  • Digital CDMO

    Energy-aware fermentation scheduling

    A control layer that brings real-time electricity pricing and hydrogen-supplier availability into the bioreactor control system, so fermentation intensity, aeration and feeding track the cheapest energy windows while keeping the bacterial culture stable.

    • Energy market data feed

      Live price and availability

      Connect the control system to electricity-market pricing and hydrogen-supplier availability feeds, so the bioreactor control loop knows the cost of the next hour's energy as well as it knows the next hour's setpoint.

    • Adaptive setpoint control

      Production shaped to energy cost

      Modulate aeration, agitation and feed rate against the energy-price signal within the biological envelope, so production shifts toward low-cost windows without compromising the culture.

    • Hydrogen supply scenarios

      Resilience to price spikes

      Run scenarios against hydrogen supply interruption and price spikes so the schedule can absorb shocks, and the transition from blue to green hydrogen becomes a planned change rather than an emergency.

    • Energy cost per kilogram of protein follows the cheapest available window, not the average.
    • The same control logic absorbs the eventual switch to green hydrogen without a redesign.
    • Energy scheduling becomes an investor-visible operating capability, not a hidden cost.

    We've done this before

    See how we do Digital CDMO
  • Enterprise AI

    A consortium data backbone for pilot, partners and regulator

    A shared data platform that defines gas source, fermentation run, batch, certificate and life-cycle metric once, then loads Aerbio's pilot-plant data, partner-shared CO2 and feed-trial information, and regulator-facing records against the same model.

    • Shared process ontology

      One agreed set of terms

      Define gas, run, batch, certificate and life-cycle metric as explicit entities with agreed relationships, so a query written once returns comparable answers across Aerbio, Drax, BioMar and the regulator instead of four dialects of the same data.

    • Partner and regulator portals

      Role-based access by design

      Expose the model through role-based portals so BioMar reads batch certificates, Drax verifies CO2 offtake and EFSA or FSA reads the regulatory record, each with an access scope that matches their agreement.

    • Automated carbon intensity and LCA

      Sustainability as a live number

      Calculate carbon intensity and life-cycle assessment from the same underlying data, so the figure quoted to Sainsbury's, to UKRI and to investors is the current one rather than a quarterly estimate.

    • Partner and regulator questions are answered from one record rather than recompiled per request.
    • Carbon-intensity reporting becomes a live output of operations rather than an annual exercise.
    • New partners, markets and grant milestones attach to the same model rather than triggering a new data exchange.

    We've done this before

    See how we do Enterprise AI
  • Digital Lab

    Predictive contamination and process-deviation detection

    A monitoring layer that reads off-gas composition, broth signals and trend patterns against a 'golden batch' envelope, so a contamination event or metabolic drift is flagged against the live run hours before it becomes visible to operators.

    • Off-gas signature analysis

      Listening to the culture

      Track CO2 production rate, O2 consumption and respiratory quotient (the ratio of CO2 produced to O2 consumed, which shifts when the organism is stressed) against the expected trajectory, so a contamination event or metabolic switch shows up in the off-gas before it shows up in the broth.

    • Multi-sensor trend correlation

      The pattern, not the single reading

      Correlate pH, dissolved oxygen, temperature and optical density trends against historical runs so a drift that no single sensor would catch is surfaced as a single alert.

    • Tiered escalation workflow

      Right signal to the right person

      Distinguish minor drift (operator notification) from critical anomaly (automatic protective action), with every alert carrying the data behind it for the post-event review.

    • The window between an emerging contamination event and a write-off decision widens from minutes to hours.
    • The cost of a lost batch is reduced or avoided, and the data behind the alert is available for the regulatory record.
    • Operators work from alerts generated from the data rather than from visual checks of a tank.

    We've done this before

    See how we do Digital Lab
  • Agents

    AI agents for Novel Food dossier and grant reporting

    Narrow, reviewable agents that take the repetitive part of regulatory and grant reporting: drafting sections of the EFSA and FSA Novel Food dossier from source records, checking a submission against its template, and assembling UKRI milestone reports from operational data. A named person approves every output.

    • Drafting from source records

      First drafts from system data

      Generate the first draft of a Novel Food dossier section or a UKRI milestone update from the underlying fermentation, batch and consortium records, so the author edits and judges rather than assembles.

    • Template and completeness checking

      Gaps found before review

      Check a submission against the relevant EFSA, FSA or UKRI template and the site's own checklist, returning missing or inconsistent sections before the document enters the human review queue.

    • Change-impact search across records

      Which records a change touches

      When a process parameter, gas source or specification changes, retrieve every controlled record that references it so the scope of a dossier update is established on day one rather than during the next filing window.

    • Submissions reach review complete, so the time between data and filing shortens.
    • The scope of a process or specification change is established by search rather than by recollection.
    • Every output traces back to the source records it came from and is signed off by a named reviewer.

    We've done this before

    See how we do Agents

Where QB Systems fits

Alongside our services we build QB Systems, hardware and software for bioprocess control. QB Systems is a product brand of A4BEE Sp. z o.o.

Applications
Bioreactors

Software-defined control for a bioreactor — a new QB vessel, an upgrade to one you have, or a retrofit of the existing PLC.

Automated sampling

Automated sampling from 4–18 sources, aseptic-capable and up to 72 hours unattended. Works with any vendor's bioreactor.

Buffer & media preparation

Automated preparation of growth media and process buffers, so a recipe runs the same way every time without fixed infrastructure.

Deployment
Retrofit

Existing equipment keeps running; QB takes over the PLC, or reads from it without touching control.

Scale
Benchtop (1–8 L)

Glass vessels with the complete hardware and software stack. This is the core range for development work.

Digital maturity: today and target

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

Source: A4BEE analysis of public sources
Bioprocess digital twin 25 → 80
Metabolic models exist for the chassis organism but are not yet integrated with industrial process control, and there is no computational fluid dynamics capability for the reactor design in flight.
Cross-site data model 30 → 85
Operations span Copenhagen, Geleen and Nottingham, with consortium partners exchanging data by spreadsheet; a shared model across the pilot plant, the consortium and the regulator is ahead of the company rather than behind it.
Adaptive process control 35 → 80
Basic bioreactor control is in place at the pilot plant, but there is no demand response against energy pricing and no real-time connection to hydrogen-supplier availability.
Predictive analytics 22 → 78
There is no contamination prediction today, and off-gas and broth signals are largely read by operators looking at the tank rather than by a model looking at the data.
Quality and compliance systems 28 → 85
The EFSA and FSA Novel Food record has to be assembled continuously rather than at filing time, and the consortium reporting load sits on the same underlying data; the systems to do that are not yet in place.
IT and OT convergence 30 → 80
The pilot-plant SCADA (Supervisory Control and Data Acquisition) sits largely isolated from the academic models and the consortium exchange; the Market Launch Facility is a chance to set the integration standard for the next stage of the business.

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 Aerbio 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].