☰

BEOnchip

Making organ-on-chip data ready for the next workflow

Industry
Organ-on-chip devices and microphysiological systems
Headquarters
Zaragoza, Spain
Public information as of
January 2026

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

Strategic priorities

BEOnchip was founded in 2016 as a spin-off from the University of Zaragoza and now sells the Be-series of microfluidic organ-on-chip devices into academic and pharmaceutical labs. The company sits inside a fast-growing market: the global organ-on-chip field is projected to expand from about USD 157 million in 2024 to roughly USD 950 million by 2030, a compound annual growth rate of 35.1 percent.

The near-term regulatory environment is shifting around it. In April 2025 the US Food and Drug Administration set out a plan to phase out mandatory animal testing for certain drug classes in favour of New Approach Methodologies, which is exactly the regulatory category that organ-on-chip platforms are built to serve. The European Commission's parallel €79 billion annual cost figure for Adverse Drug Reactions gives the policy its economic weight.

The operational direction follows from there. BEOnchip's recent technical notes — on anaerobic intestinal environments, volatile organic compound detection and biolaminin vascular adhesion, all published between June and October 2025 — point to a portfolio that is becoming more instrumented and more dependent on data that moves between devices, software tools and customers' laboratory information systems.

Consortium work shapes the funding and the standards. BEOnchip is a partner in the European UNLOOC, PRIME and Moore4Medical projects, with UNLOOC alone bringing together 68 organisations across ten countries. The same projects also set the data, interoperability and validation requirements that future customers will expect organ-on-chip platforms to meet.

Challenges we see

  • Digital Integration

    Bringing Be-series process data into the rest of the lab

    Each Be-series device generates its own readings — flow, shear, oxygen, image time-courses — and currently these streams stay close to the device rather than loading into a shared laboratory information system or analytics layer. As more customers ask for evidence trails, the absence of a single data model for an experiment becomes visible.

    Where experiment data lives in device-local stores, building the evidence trail a regulator or a customer asks for is a separate exercise for every question, so the same data is reconstructed more than once for each reviewer.

  • Operations Manufacturing

    Predicting shear and perfusion behaviour in larger organoid models

    When organoid size grows, parameters such as pressure and liquid flow resistance scale in ways that differ from single-chip prototypes, which makes different-sized organoids hard to treat as experiment triplicates. The October 2025 technical note on Biolaminin 521 in vascular chips addresses one part of the same problem.

    As the portfolio moves toward larger and more complex organoid models, the ability to predict flow and shear in advance shifts from a research convenience to a design input, and the experiments that confirm it become fewer and more targeted.

  • Operations Manufacturing

    Carrying oxygen, gradient and flow control away from external benchtop rigs

    Current organ-on-chip workflows depend on external pumps, sensors and readout equipment. The June 2025 technical note on anaerobic intestinal environments and the PRIME project's tubeless, contactless microfluidic approach point toward removing that dependence.

    When control and readout move inside the device, the operating envelope is set by the chip design rather than by what the surrounding rig can deliver, which changes how a deployment looks in a customer's lab.

  • Operations Manufacturing

    Keeping cell adhesion stable under physiological shear

    Human umbilical vein endothelial cells detach under dynamic flow in vascular chip models. The October 2025 Biolaminin 521 note describes how Biolaminin coatings and a controlled ramp-up in shear help cells adhere through the first culture stages.

    As vascular and multi-organ models become routine, the experimental conditions under which cells stay attached become a defined operating envelope rather than a per-experiment judgement, and the protocols that hold them in place are easier to repeat.

  • Quality Manufacturing

    Detecting microchannel defects during microfluidic chip production

    Microfluidic biochips can fail from dielectric breakdown, hydrophobic layer damage, electrode shorts or parasitic leakage. Published microfluidic literature describes these as silent failure modes that human inspection misses during production.

    When defect detection sits at the end of the production line, the units that deviated from specification have already been built, so moving the check earlier narrows the population of suspect chips rather than the whole batch.

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. Unifying Be-series and lab data into one model

    Each Be-series device, plus the bioreactors, plate readers and balances around it, produces data in its own structure. There is no shared ontology that lets a customer query an experiment across instruments or hand the same record to a regulator and an analyst.

    An ontology-driven data platform that defines assay, device, run, sample and result once lets data from every Be-series device and the surrounding lab load against the same model, so the combined record can be queried and audited in one place.

    • BEOnchip about-us page, accessed January 2026
    • BEOnchip technical notes, June–October 2025

  2. Building a digital twin of organ-on-chip experiments

    Multi-organ and rare-disease models depend on combinations of flow, oxygen, gradient and cell type that have to be tuned experimentally today. Failed experiment cycles are common because the parameter space cannot be searched by hand.

    A simulation layer that holds the physics of flow, diffusion and cell behaviour lets new experimental conditions be predicted before chips are run, so failed cycles drop and the experiments that do run are more targeted.

    • BEOnchip, Organ-on-chip development: ORCHID Project
    • A4BEE case study, Creating Digital Twin to execute simulations of biopharma processes

  3. Designing IT and OT connectivity for new instrument generations

    Be-series devices and the benchtop equipment around them use a mix of vendor protocols, and customer laboratories often run them as standalone units. As new generations ship with onboard sensing, the absence of a shared connectivity story becomes visible to the customer.

    An IT/OT reference architecture — protocol standards, network segmentation and equipment data contracts agreed before new generations ship — lets interoperability arrive with the device rather than being negotiated per customer deployment.

    • BEOnchip, Microfluidic Chip Design Customization Guide, December 2024
    • BEOnchip about-us page, accessed January 2026

  4. Retrofitting legacy lab equipment around the Be-series

    Customer laboratories frequently run Be-series devices alongside air-gapped bioreactors, plate readers and balances from other vendors. The path from instrument to scientist is broken at the boundary where data leaves these devices.

    Connecting legacy instruments through vendor-neutral IoT gateways with OPC UA (Open Platform Communications Unified Architecture) closes the gap from source to scientist, so a Be-series run can be joined to the surrounding lab data without manual transcription.

    • BEOnchip, Vision 2030 roadmap and Digital Hesitancy analysis
    • A4BEE case study, Restoration: Establishing or Re-establishing a Complex Lab Ecosystem

  5. AI vision for microfluidic chip production inspection

    Microfluidic biochips carry defects — dielectric breakdown, hydrophobic layer damage, electrode shorts — that human inspectors cannot see at production speed. As the Be-series moves toward higher-volume production, the inspection step is the bottleneck.

    An AI vision system that watches the chip during production can flag the defects human inspection misses in real time, so the population of suspect chips narrows to the ones the model flags rather than the whole batch.

    • Microfluidics-Based Biochips: Technology Issues, Implementation Platforms, and Design-Automation Challenges, ResearchGate
    • A4BEE case study, Bioreactor Control: Computer Vision for Non-Invasive Foam Management

What we'd propose

  • Enterprise AI

    Ontology-driven data platform for the Be-series ecosystem

    We build a shared data platform that defines the entities BEOnchip and its customers share — assay, device, run, sample, result — once, and loads data from the Be-series devices and the surrounding lab instruments against that single model.

    • Shared organ-on-chip ontology

      One agreed set of terms

      Define assay, device, run, sample and result as explicit entities with agreed relationships, so a query written once returns comparable answers across Be-series devices and the lab equipment around them instead of two dialects of the same table.

    • Pipelines from Be-series and lab instruments

      Loading both sides

      Build ingestion for Be-series device output and for the bioreactors, plate readers and balances around them, with schema validation at the boundary so a malformed record fails loudly instead of silently.

    • Analytics and retrieval on top of the model

      Questions answered without IT tickets

      Expose the model through dashboards and a retrieval layer so BEOnchip's application scientists and a customer's quality team can query the combined data set without commissioning a new extract for each question.

    • Integration work happens once against a shared model instead of once per point-to-point interface.
    • Customers can hand the same record to a regulator and to their own analyst without rebuilding it.
    • New Be-series generations attach to the model rather than triggering another migration.

    We've done this before

    See how we do Enterprise AI
  • Digital Lab

    Digital twin for organ-on-chip experiment design

    We build a simulation layer that models the physics of flow, diffusion and cell behaviour in the Be-series, so new experimental conditions can be predicted before chips are run and failed cycles drop.

    • Predictive physics model

      Flow and shear in advance

      Implement a model of flow, shear stress and diffusion calibrated against existing Be-series data, so a proposed organoid size, channel geometry or media composition returns a predicted outcome before any chip is loaded.

    • Simulation workflow engine

      Runs planned, not improvised

      Centralise the inputs, outputs and acceptance criteria of a simulation run so experimentalists can move from a parameter set to a planned chip run without rebuilding the workflow each time.

    • Lab data integration

      Simulation tied to real runs

      Connect the simulation layer to Be-series data so a completed run can be compared with its prediction, and the prediction model is updated as new evidence arrives.

    • Fewer experiment cycles end in failure because the parameter space is searched in simulation first.
    • Multi-organ and rare-disease models reach a working condition in fewer physical attempts.
    • Predictions and measured outcomes live side by side, so the model improves as the portfolio does.

    We've done this before

    See how we do Digital Lab
  • Digital CDMO

    IT and OT connectivity design for Be-series generations

    An architecture and standards package for the next Be-series generations: protocol choices, network segmentation and equipment data contracts agreed before the devices ship, so interoperability is designed rather than negotiated per customer.

    • Reference architecture for the device family

      One documented data path

      Specify how Be-series devices, benchtop equipment, laboratory information systems and enterprise tools connect, including the segmentation model, so each new generation builds toward the same target.

    • Equipment data contracts

      What each device publishes

      Write the OPC UA information models and topic structures into product requirements, making interoperability a design condition rather than an integration project after handover.

    • Segmentation and security baseline

      Safe from day one

      Define network segmentation and remote-access rules before new generations ship, so customer deployments do not have to be re-architected against a live production line.

    • Interoperability is designed with the device rather than built after customer handover.
    • The same architecture and procurement language is reusable as new Be-series generations ship.
    • Customer laboratories receive a platform that joins their existing lab information system rather than sitting beside it.

    We've done this before

    See how we do Digital CDMO
  • Digital Lab

    OPC UA retrofit for legacy lab equipment in customer deployments

    We connect the air-gapped bioreactors, plate readers and balances that sit around the Be-series in customer laboratories through vendor-neutral IoT gateways, so the data path from instrument to scientist is closed without replacing equipment that works.

    • OPC UA gateway deployment

      Standardised industrial connectivity

      Deploy IoT gateways speaking OPC UA (Open Platform Communications Unified Architecture) for real-time data streaming from legacy bioreactors, plate readers and balances into the shared data platform.

    • Vendor-agnostic integration

      Multi-protocol device support

      Build custom drivers and connectors for instruments from multiple vendors so the data platform receives a consistent record regardless of the underlying protocol.

    • Network orchestration with secure segmentation

      OT layer kept separate

      Reconfigure the connectivity layer with strict network segmentation so the operational technology layer remains isolated while the data path from edge to cloud remains open.

    • Customer laboratories reach 100 percent connectivity for the instruments that matter to a Be-series run.
    • Vendor lock-in drops because the data platform accepts multiple instrument protocols.
    • Manual transcription between instrument and record falls out of the workflow.

    We've done this before

    See how we do Digital Lab
  • Digital CDMO

    AI vision for microfluidic chip production inspection

    A computer vision system that watches Be-series microfluidic chips during production and flags the defects — dielectric breakdown, hydrophobic layer damage, electrode shorts — that human inspection cannot see at line speed.

    • Real-time defect detection

      AI-powered visual inspection

      Use image processing models trained on chip imagery to detect dielectric breakdown, hydrophobic layer damage and electrode shorts as the chip moves through production, with a confidence score attached to each flag.

    • Adaptive control logic

      Response tuned to the defect

      Implement control logic that adapts to the defect class — separating, rerouting or reworking a flagged chip rather than treating every signal the same way.

    • 24/7 line monitoring

      Continuous production oversight

      Deploy camera stations and intelligent algorithms for round-the-clock inspection, so the production line is watched continuously rather than sampled.

    • Defects invisible to the human eye are caught at the line rather than downstream.
    • The population of suspect chips narrows to the units the model flags, not the whole batch.
    • Inspection throughput rises because the line is watched continuously instead of by sampled checks.

    We've done this before

    See how we do Digital CDMO

Digital maturity: today and target

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

Source: A4BEE analysis of public sources
Lab data integration 35 → 80
Be-series devices and the lab equipment around them use vendor-specific data structures, and there is no shared ontology across them today. The target is a single model that experiment records, lab instruments and customer information systems can all load against.
Simulation and digital twin 25 → 70
Organ-on-chip parameter space is searched experimentally today, and failed experiment cycles are common in multi-organ and rare-disease work. The target is a simulation layer that predicts outcomes before chips are run.
IT and OT connectivity 30 → 75
Be-series devices and customer-side legacy equipment are connected through bespoke integrations. The target is a reference architecture with protocol and segmentation standards agreed before new generations ship.
Production inspection 38 → 78
Microfluidic biochips carry defects that human inspection misses at production speed. The target is continuous AI vision inspection across the line.
Workforce digital readiness 42 → 72
Advanced laboratory settings report a 57 percent digital skill gap, and BEOnchip's leadership has pointed to a wider trust gap in digital systems. The target is a workforce that can adopt new digital tooling as part of routine practice.
Data-driven experiment design 32 → 74
Experimental conditions for organ-on-chip runs are decided from prior experience today. The target is a workflow where simulation, prior data and incoming Be-series readings together guide the next run.

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 BEOnchip, 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].