☰

Biond

Bridging chip prototype and industrial production

Industry
Biotechnology
Headquarters
Delft, Netherlands
Public information as of
January 2026

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

Strategic priorities

Biond — formally BIOND Solutions BV, commercially known as Bi/ond — built silicon-based organ-on-chip platforms for drug discovery, with the MUSbit muscle-on-chip and inCHIPit platforms the most visible products. The spin-off from Delft University of Technology combined two decades of semiconductor microfabrication know-how with 3D tissue biology, and participated in the MAGIC project consortium for rare-disease modelling. As of April 2025 the founders announced the company was ceasing operations, citing the difficulty of moving from prototype to a globally scaled business in the current funding environment.

The relevant data sits at the boundary of biology and electronics. A chip run produces stimulation traces, contractility readouts, microscopy images and patient-derived induced pluripotent stem cell metadata, and the value of the platform to a pharmaceutical partner is the integrated record. As of 2024 the company reported that its product was being adopted in cardiac and neuromuscular disease programmes, and that the technology had raised over $4 million in funding across its life.

The wider market gives the work its commercial shape. The global organ-on-chip market was valued at roughly $126 million in 2024 and is forecast to reach about $950 million by 2030, a compound annual growth rate of around 40 percent, with the drug-discovery segment growing fastest as pharmaceutical companies look for human-relevant pre-clinical data that does not depend on animal models.

Inside that market, the practical digital question is the same as in any instrument-heavy biotech: how to bring stimulation traces, microscopy, lot and donor metadata and quality records into one model that an analyst, a quality team and a partner can query together. The rest of this page is a read of where Biond stood on that question and where the technology the company worked on is likely to go next.

Challenges we see

  • Engineering Manufacturing

    Combining microelectronics with living tissue in one device

    Organ-on-chip platforms pair high-performance semiconductor components with aqueous microenvironments in which human tissue has to survive for days or weeks. The company described balancing microelectronics with the requirements of tissue and muscle development as the primary technical hurdle.

    Where chips and biology share the same substrate, the population to qualify is the whole device, not the integrated circuit alone. Qualifying each assembled device rather than the chip lot narrows that population to the units that actually carry the biology.

  • Biology Operations

    Keeping 3D tissue viable past the first week

    The team reported that 3D tissues cultured in vitro remain viable for a limited window, which constrains chronic dosing studies and long-term toxicity work that pharmaceutical partners expect to run. As of 2024 the state of the art in long-term ex vivo culture still imposes mechanical stress on tissue slices and uneven nutrient and oxygen gradients within air-liquid interface systems.

    Where a long dosing study needs the same tissue to remain physiologically representative for weeks, the device's perfusion, mechanical stimulation and environmental control become the variables the protocol has to control, and the same variables have to be readable as data while the run is in progress.

  • Operations Scale-up

    Carrying an instrument company from prototype to a global product

    Organ-on-chip companies move from a handful of working prototypes to a standardised, manufactured product that has to behave the same in a partner's lab as it does in the developers'. The capital required to build that supply chain and the regulatory package to go with it is several multiples of what a proof-of-concept round supports. As of April 2025 Biond's founders announced the company was ceasing operations, citing the difficulty of completing that transition in the prevailing market.

    Where the same instrument ships to multiple partners, reproducibility becomes a property of the released product and of the data trail behind it, not a property of the bench prototype. Building that reproducibility into the platform before the first commercial release is what keeps partner data comparable across sites.

  • Compliance Regulatory

    Producing the validation evidence pharmaceutical partners expect

    Pharmaceutical partners are governed by GxP (Good Practice) regulations. New pre-clinical platforms such as organ-on-chip systems must demonstrate that their readouts are as reliable as those produced by the animal models they are intended to replace, and industry-wide standards for microfluidic data interoperability and digital twin validation are still being defined.

    Where the evidence a regulator wants is generated device by device, the validation package reads from the device's own data and the standards it ships with, and that package has to exist for every shipped unit rather than being assembled once at the platform level.

  • Organisation Process

    Bridging engineering and biology in the same team

    Biond reported a roughly even split between engineers and biologists on staff, which the team treated as a feature of the technology rather than a coordination cost. Where two disciplines meet on the same device, the language used to describe a parameter, a quality event or a release decision has to be shared across the two.

    Where engineering and biology are each naming the same run with their own vocabulary, the run record has to carry both vocabularies as data so that an engineering alert and a biology observation refer to the same event. A shared vocabulary in the platform's own records is what makes a 50/50 team a feature rather than a coordination problem.

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. Putting stimulation, microscopy and donor data into one record

    A typical organ-on-chip run produces electrical stimulation traces, contractility readouts, microscopy images and donor-derived induced pluripotent stem cell metadata, and they are recorded in separate places. Combining them into a single analytical view is a manual exercise that takes weeks per study.

    An agreed ontology for chip run, stimulation protocol, microscopy observation, donor line and quality event lets all of those data sources load against one model, so a partner query written once returns comparable answers across studies and sites instead of a hand-built extract for each one.

    • Biond website, https://www.gobiond.com
    • MarkNtel Advisors, Global Organ-on-Chip Market Research Report, 2025

  2. Reading tissue condition while the chip is still running

    Tissue viability in 3D culture drifts over days, and the indicators that matter for a chronic dosing study are typically only reviewed at the end of the run. Long-term viability limits the kinds of study a platform can support and the length of any dosing protocol a partner will design against it.

    Streaming microscopy and sensor data from the chip into a time-series store, and watching each running chip against its own historical envelope, lets viability drift be flagged during the run and gives the next study a usable upper bound on duration rather than a guess after the run is over.

    • NIH PMC, A Microfluidic Cancer-on-Chip Platform Predicts Drug Response Using Organotypic Tumor Slice Culture, 2022
    • Biond blog, https://www.gobiond.com/author/nikolas/

  3. Bringing the chip onto the same validation backbone as a regulated assay

    Pharmaceutical partners are used to reading validation evidence for an assay in a defined format. An organ-on-chip platform ships with its own evidence, generated on its own instruments, and that evidence is read against standards that are still being defined across the industry.

    Capturing every chip run against a defined validation schema — including stimulation parameters, environmental conditions and the quality events of every run — gives a partner a familiar evidence package to read and gives a regulator a consistent format to assess across studies.

    • Insightace Analytic, Organ-on-Chip Market Review Report 2025-2034
    • Biond website, https://www.gobiond.com

  4. Reproducing a single run across manufacturing batches

    Where every chip is hand-assembled and the biology varies by donor line, two runs intended to be replicates can drift in stimulation response and viability. A partner reading across runs has to assume the noise is in the biology, when some of it sits in the device.

    Recording each assembled chip's electronic and microfluidic test results against its serial number, and folding those test results into the run record, lets a partner separate device noise from donor noise and lets the manufacturer see when a device batch is producing a wider response distribution than its peers.

    • Biond funding announcement, March 2022, https://www.gobiond.com/2022/03/08/biond-secures-over-4m-in-funding/
    • Biond blog, https://www.gobiond.com/tag/ooc/

  5. Bridging prototype data and partner-facing reporting

    The data a researcher wants during development is not the data a quality team or a partner wants to read. Researchers want raw traces; partners want summary metrics against their own assays. The conversion between the two is currently done by hand for each study.

    Defining the summary metrics the partner reads once, at the data model, and computing them from the underlying traces as part of the run record, lets the same data set serve development and partner reporting without a parallel document trail.

    • Silicon Canals, Delft-based biotech company Bi/ond is set to shut shop, 2025
    • Biond blog, https://www.gobiond.com

What we'd propose

  • Enterprise AI

    Ontology-based data backbone for organ-on-chip studies

    We design the data model and the ingestion layer for an organ-on-chip research operation, so that stimulation traces, microscopy observations, donor line metadata and quality events from MUSbit and inCHIPit platforms all load against one shared ontology that partners and analysts can query together.

    • Shared organ-on-chip ontology

      One agreed set of terms

      Define chip run, stimulation protocol, microscopy observation, donor line and quality event as explicit entities with agreed relationships, so a query written once returns comparable answers across studies and partner sites instead of two dialects of the same table.

    • Ingestion from chip, microscopy and LIMS

      Loading from every source

      Build ingestion for stimulation controllers, microscopy stations and laboratory information systems, with schema validation at the boundary so a malformed record is rejected at the door rather than corrupting a downstream study.

    • Analytics and partner-facing views

      Questions answered without IT tickets

      Expose the data set through dashboards and a retrieval layer so partners, quality and research staff can ask questions of the combined data set without commissioning a new extract for each one.

    • Partner queries are written once against the shared model and run as needed, rather than rebuilt for each study.
    • Stimulation, microscopy and donor data are correlated as data rather than reconstructed by hand.
    • New assays and additional data sources attach to the model rather than triggering another integration.

    We've done this before

    See how we do Enterprise AI
  • Digital CDMO

    Time-series monitoring and run-time viability tracking

    We stream chip sensor and microscopy data into a time-series store and run run-time anomaly detection against each chip's own envelope, so that tissue viability drift is visible while the run is in progress rather than at the end of the study.

    • Streaming from chip controllers

      Data off the device

      Connect stimulation controllers and onboard sensors through OPC UA (Open Platform Communications Unified Architecture) or MQTT so stimulation parameters and sensor readings leave the chip as documented, vendor-neutral time series.

    • Per-chip operating envelope

      Normal for this chip

      Build the normal operating envelope for each chip type from historical runs and watch each running chip against its own envelope, so a drift in stimulation response or viability surfaces while the run is in progress.

    • Run-time alert and study summary

      What changed in this run

      Generate an alert when a chip's signal drifts outside its envelope during the run, and assemble a study summary at the end that compares the run against its own envelope rather than against a generic baseline.

    • Tissue viability drift is visible during the run rather than only at the end of the study.
    • Each study produces a comparison against its own envelope, so week-to-week variation is readable.
    • Future chip generations can be compared against the same envelope framework with the same metrics.

    We've done this before

    See how we do Digital CDMO
  • Digital Lab

    Validation package for organ-on-chip readouts

    We define the validation schema an organ-on-chip platform presents to pharmaceutical partners, and instrument every chip run so that the schema is populated by the run itself rather than assembled afterwards.

    • Validation schema definition

      What a partner reads

      Define the validation schema for stimulation parameters, environmental conditions and quality events at the platform level, in a form a pharmaceutical partner's quality organisation can read against existing assay validation expectations.

    • Run-time evidence capture

      Evidence assembled as the run runs

      Capture the evidence the schema calls for directly from the chip and its supporting systems during the run, with audit trail back to the device and method version that produced each value.

    • Per-study validation dossier

      One document per study

      Assemble a per-study validation dossier from the run records and the schema, so each study ships with its own evidence package rather than depending on a separately maintained document.

    • Partners receive evidence in a format they can read against existing assay validation expectations.
    • Each study's evidence package is generated by the run, not reconstructed afterwards.
    • Validation work scales with the number of studies rather than the number of document authors.

    We've done this before

    See how we do Digital Lab
  • Digital CDMO

    Per-chip device record for assembled platforms

    We record every assembled chip's electronic and microfluidic test results against its serial number, and feed those results into the run record, so device-to-device variation is visible alongside the biology it is being asked to explain.

    • Per-chip test record

      One record per assembled device

      Capture each chip's electronic and microfluidic test results against its serial number during production, so the device's test history follows it into every study it participates in.

    • Run record joined to device record

      Run and device in one view

      Join each study's run record to the device record of the chip it ran on, so partner analysis can read stimulation response against the device batch that produced it rather than against a generic baseline.

    • Cross-batch comparison view

      See when a batch drifts

      Compute a per-batch response distribution across studies and surface batches whose distribution is wider than their peers, so the manufacturer sees device drift as soon as a study reports it.

    • Device noise is visible in the same view as biology noise, so a wide response distribution can be attributed to the right place.
    • A device batch producing unusual results is flagged from the studies that used it, not after the fact.
    • The same device record feeds manufacturing, quality and the partner-facing view.

    We've done this before

    See how we do Digital CDMO
  • Enterprise AI

    Partner-facing metrics computed from raw run data

    We define the summary metrics partners read once, at the data model, and compute them from the raw traces during the run, so the same data set serves development and partner reporting without a parallel document trail.

    • Partner-facing metric catalogue

      What a partner reads

      Define the summary metrics a partner reads — stimulation response, viability, contractility, donor response profile — once at the data model, with the calculation and the source data version pinned to the metric definition.

    • Metric computation during the run

      Metrics from raw traces

      Compute the catalogue from the raw traces as part of the run record, so the metrics a partner reads are the metrics the run produced, not a separate calculation done by hand afterwards.

    • Run summary assembled from metrics

      One summary per run

      Assemble a per-run summary from the catalogue, ready to share with a partner, with the underlying traces still reachable from the summary for the partner's own analysis.

    • Partner-facing metrics are computed once at the model and reused across studies.
    • The run summary is generated from the raw traces, not assembled by hand.
    • A partner can drill from the summary into the traces without leaving the platform.

    We've done this before

    See how we do Enterprise AI

Digital maturity: today and target

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

Source: A4BEE analysis of public sources
Data interoperability 30 → 80
Stimulation, microscopy, donor and quality data lived in separate systems for a chip-prototype organisation of Biond's scale. A shared ontology would have made cross-study analysis possible without a separate extract for each question.
Cloud readiness 45 → 85
Chip and microscopy data can grow quickly into the terabyte range per study. A cloud-agnostic ingestion layer would have let the same workload run on the cloud a partner preferred and on local infrastructure for sensitive donor data.
Process automation 40 → 80
Chip loading and stimulation protocol setup were performed largely by hand. Closing that loop around a documented standard operating procedure would have reduced variation between runs and shortened onboarding for new laboratory staff.
Regulatory compliance 50 → 90
Pharmaceutical partners read pre-clinical platforms against GxP (Good Practice) expectations. A validation schema populated by the run itself would have made partner review faster and would have set a format a regulator could read across studies.
Asset lifecycle 25 → 75
Each assembled chip carried its own calibration, configuration and test history, but the history was not in a system the manufacturing, quality and partner sides could all read. A platform-level lifecycle record would have shortened the path from assembled device to partner-ready evidence.
Workforce digital literacy 60 → 85
A 50/50 split between engineers and biologists gives a small organisation two strong disciplines and two vocabularies for the same chip. A shared data model would have made the shared vocabulary a property of the platform rather than a coordination cost.

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 Biond, 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].