☰

Anivia Medical

Connectivity and scale-up for life-support devices

Industry
Medical Devices
Headquarters
Taipei, Taiwan
Public information as of
January 2026

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

Strategic priorities

Anivia Medical, the 2024 rebrand of Asia Pacific Medical Technology Development, runs a Dual-Track strategy around its SG1000 pump console: an Open System for US and EU markets and a vertically integrated line under KYHT (Zhuhai) Biotech for China. The Chinese subsidiary received NMPA registration for its centrifugal pump in August 2025, placing commercial scale-up on the immediate roadmap. Investors led by Panacea Venture are watching revenue traction now that the company has moved from development to commercial stage.

The console is engineered around CAN bus and RS-485, suited to a moving pump but invisible to hospital networks. Competitor systems are moving the other way: LivaNova's Essenz integrates with HIS and PACS, and Getinge's Cardiohelp supports Talis RemoteView for cloud-based fleet management. Electromagnetic compatibility aside, leaving telemetry behind the bus puts Anivia outside the tenders that require EMR integration. The PMA clinical pathway for the neuroprotection programme raises the same point in the data language clinical sites are already required to speak.

Pilot capacity is 200 catheters and 10 extracorporeal systems a day out of Zhuhai, sized for trials rather than commercial demand. Manufacturing relies on kitted assembly with end-of-line inspection, with quality confirmed after the unit is built rather than measured as it is built. Manufacturing in China under NMPA, in Taiwan under TFDA and in the United States under FDA creates a third concurrent set of documentation work, which the company is currently keeping on paper.

Three priorities sit on the critical path together: getting device telemetry outside the bus without triggering a Significant Change filing, scaling the Zhuhai line to commercial volumes with evidence the regulators can read, and shortening the data lock at the end of the PMA trial. The same underlying capability, getting instrument and process data out of equipment and into shared systems, holds all three together.

Challenges we see

  • Digital Integration

    Getting telemetry out of a life-support console

    The SG1000 console communicates internally over CAN bus and RS-485 with no Wi-Fi, Bluetooth or cellular telemetry specified. The pump exposes ultrasonic flow, bubble, pressure and temperature data and stores whole-case timeline graphs that today are retrieved manually.

    Where device data stays behind an industrial bus, sharing it with the rest of the hospital depends on a physical step that the clinical team can skip. Translating the bus to a network protocol at the edge opens the same data to fleet management, EMR integration and remote service without redesigning the cleared device.

  • Operations Manufacturing

    Scaling from pilot to commercial manufacturing

    Public disclosures place current pilot output at 200 catheters and 10 extracorporeal systems daily, with assembly based on kitted sub-assemblies from external vendors and inspection at the end of the line. August 2025 NMPA registration of the centrifugal pump in China moves the line from trial supply to commercial demand.

    Where end-of-line inspection is the main quality check, the population under review is every unit. Recording torque, test bench and operator data at each station narrows that population to the units where a critical parameter actually deviated and produces the device history record the regulators expect in the same step.

  • Compliance Regulatory

    Keeping a single quality record across FDA, NMPA and TFDA

    Anivia's regulatory footprint spans NMPA for the Zhuhai line, TFDA for Taiwanese R&D and FDA for US sales. Specifications, procedures and validation protocols must hold across all three to keep the version being built in Zhuhai aligned with what was approved elsewhere.

    Where the same change has to be reflected in three separate quality systems, the risk of drift scales with the number of hand-offs between them. A shared version of controlled documents and an audit trail that records who changed what and when keeps the linkage explicit instead of relying on it being maintained in parallel.

  • Operations Service

    Replacing scheduled visits with continuous device telemetry

    Marketing material describes data-driven predictive maintenance grounded in Weibull reliability models and long-term testing. There is no real-time field telemetry pipeline behind the claim, so service today rests on scheduled visits and reactive dispatch.

    Where service is scheduled rather than triggered, the trigger available is whatever a technician sees on arrival. Streaming motor current, bearing temperature and battery state to a monitored model lets the service team respond on a signal rather than after a failure.

  • Digital Evidence

    Accelerating PMA evidence collection from clinical sites

    The PMA pathway for the neuroprotection programme requires valid scientific evidence from controlled trials. With the console relying on CAN and RS-485, US site data is captured by Case Report Forms that are transcribed from the device screen, and R&D teams in Taiwan review site data weeks after it is recorded.

    Where trial data is typed by hand from a screen, the time between an event and a usable record is bounded by the transcription cycle, and the chain of custody is a paper artefact. Reading the same parameters directly from the instrument and writing them to an immutable record shortens that cycle and keeps the data readable end to end.

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. Bridging the console bus to a network protocol at the edge

    Critical parameters on the SG1000 are exposed on CAN and RS-485 and have to be retrieved by manually plugging in a USB drive. Hospital networks, EMR systems and service teams all sit on the other side of those protocols.

    An external medical-grade gateway that reads the existing bus, wraps the values in a documented network message and forwards them over Wi-Fi or cellular gives Anivia's existing devices fleet management and EMR access without redesigning the cleared console.

    • Anivia, Products & Solutions (SG1000 communications: CAN bus, RS-485)
    • Getinge, Talis RemoteView with +ACG (competitor fleet connectivity)
    • LivaNova Essenz Perfusion System (competitor HIS/PACS integration)

  2. Reading the Zhuhai assembly line as data while it runs

    Pilot output is 200 catheters and 10 systems a day built from kitted sub-assemblies with inspection at the end of the line. NMPA registration of the centrifugal pump in August 2025 moves that line into commercial-volume demand.

    Instrumenting the torque stations and test benches and streaming the values into a manufacturing execution system produces the device history record for every serial number in the same step as the build and gives operations a yield view that today is compiled after the shift.

    • Anivia, Research & Development (pilot capacity figures)
    • Anivia, About (kitted assembly and end-of-line inspection model)
    • Anivia, News (August 2025 NMPA registration, KYHT Zhuhai)

  3. Treating the cross-border quality system as one digital thread

    Anivia is operating in three regulatory jurisdictions with separate, largely paper-based quality records. A design change in the US has to reach the Zhuhai floor and the Taiwan R&D operation with the same version, date and rationale.

    A single shared QMS with one document store and one audit trail, configured to route approvals per jurisdiction, lets the three operations reference the same current version of each artefact and removes the manual propagation between them.

    • Anivia, About (US, Taiwan and China operations; regulatory footprint)
    • Anivia, Research & Development (vendor-built sub-assemblies, manual QC)

  4. Reading deployed consoles on the schedule their bearings actually follow

    Service today rests on scheduled visits and reactive dispatch; the predictive maintenance framing in marketing material is grounded in static lab data because real-time field telemetry is not yet collected.

    With motor current, bearing temperature and battery telemetry leaving the console on a regular cadence, a monitored Weibull model becomes a live signal rather than a forward-looking claim, and the service team can plan a visit against a bearing curve rather than a calendar.

    • Anivia, Research & Development (Weibull reliability and predictive maintenance framing)

  5. Collecting PMA trial data directly from the instrument

    The PMA pathway needs valid scientific evidence from controlled trials, and Case Report Forms are transcribed from the SG1000 screen at each US site, so the R&D team in Taiwan reads the data weeks later and the audit trail is the form itself.

    Capturing the same parameters directly from the device, transmitting them to an immutable clinical data set, and running automated integrity checks at the boundary turns Case Report Forms into derived products of a record the regulator can read in full.

    • Anivia, About (PMA pathway and clinical evidence objective)
    • Anivia, Research & Development (clinical workflow and trial signal)

What we'd propose

  • Digital CDMO

    Edge gateway for the SG1000 console

    A medical-grade edge gateway that connects to the SG1000 on its existing CAN and RS-485 ports, translates the bus payloads into a documented network protocol and forwards them over Wi-Fi or cellular, so fleet management, EMR integration and remote service become properties of the installed base rather than of a future device revision.

    • Bus to network translation

      Reading the existing console without redesigning it

      Parse the existing CAN and RS-485 payloads, expose them as a documented set of parameters, and wrap them in TLS 1.3-encrypted MQTT or HL7 messages, so the gateway reads the cleared device as data rather than altering its control surface.

    • Connectivity options for hospital and service use

      Wi-Fi, cellular and isolated network profiles

      Provide Wi-Fi and cellular connectivity with hospital network profiles and a separate service profile, so the same gateway can join a hospital LAN for clinical use and reach a service operations dashboard without crossing between them.

    • Security and access controls aligned with medical device guidance

      FDA 2023 cybersecurity considerations from day one

      Implement authentication, signed firmware and least-privilege access following the FDA 2023 cybersecurity guidance for medical devices, so introducing connectivity does not introduce a new attack surface for a cleared life-support console.

    • Existing cleared consoles gain fleet management and EMR access without a Significant Change filing.
    • Service and clinical teams read the same parameters from one source.
    • Predictive maintenance becomes a live signal rather than a forward-looking statement.

    We've done this before

    See how we do Digital CDMO
  • Digital CDMO

    Smart manufacturing for the Zhuhai assembly line

    An Industry 4.0 programme for the Hengqin facility: instrumenting the assembly stations, streaming the values into a manufacturing execution system, and producing the device history record as the unit is built, so quality evidence for FDA, NMPA and TFDA is generated by the line rather than compiled after it.

    • Station-level data acquisition

      Torque, test results and operator at the station

      Capture torque values, test bench results and operator IDs at each assembly station against the serial number being built, using industrial IO-Link or OPC UA (Open Platform Communications Unified Architecture) so the data leaves each station in a vendor-neutral form.

    • Manufacturing execution with device history record

      One record per serial number

      Aggregate the station-level data in a manufacturing execution system and write a device history record per unit, with the components, parameters, operator and time assembled in one place, satisfying the record expectations of FDA 21 CFR Part 820 and the corresponding NMPA requirements.

    • Yield and operations visibility

      Production metrics in real time

      Expose yield, scrap and station-level throughput as dashboards so operations leadership can plan scaling decisions on the same data the regulators will see, rather than on a post-shift manual report.

    • Quality evidence is generated in line with the build, not rebuilt for inspections.
    • Yield and scrap are visible by the hour, not by the shift.
    • The same production architecture scales as pilot volumes move to commercial.

    We've done this before

    See how we do Digital CDMO
  • Digital Lab

    One quality system for FDA, NMPA and TFDA

    A shared quality management platform with a single document and audit store, configured approval routes per jurisdiction, and a complete change history for every controlled artefact, so the three regulatory records reference the same current version and a design change propagates to all three in one motion.

    • Shared document and record store

      One current version of each artefact

      Hold specifications, procedures, validation protocols and training records in one store with version control and an immutable change history, so the Zhuhai floor, the Taiwan R&D team and the US operation all read the same current version of each artefact.

    • Jurisdiction-aware approval workflows

      One workflow, multiple paths

      Route design changes and document revisions through the right combination of FDA, NMPA and TFDA approvers for the artefact, while keeping one shared record, so the approval lineage is consistent across jurisdictions without copy-paste between separate systems.

    • Audit-ready change history

      Traceable end to end

      Produce the audit trail for a given device, document or change in one query, with approver identity, timestamp and reason for every state transition, so inspections can be answered from the record itself.

    • A design change in the US appears in Zhuhai and Taiwan at the same revision, in one workflow.
    • Three regulatory teams read the same current version of every controlled artefact.
    • Inspections are answered from the record itself, not reconstructed for them.

    We've done this before

    See how we do Digital Lab
  • Enterprise AI

    Direct clinical data capture for the PMA programme

    A clinical data pipeline that takes device parameters directly from the SG1000 at US trial sites, transmits them under HIPAA to an immutable data set, and runs integrity checks at the boundary so the Case Report Form becomes a derived product of a record the regulator can read in full.

    • Site-side data capture

      Reading from the console, not from the screen

      Deploy the SG1000 edge gateway or a paired secure USB-to-cloud uploader at US trial sites so device-side parameters and trial-side identifiers travel together, removing the manual Case Report Form transcription step entirely.

    • Immutable clinical data set with integrity checks

      One record, ALCOA+ from the device

      Aggregate the site data in a HIPAA-compliant data lake with an append-only model, validate every record against the protocol, and flag missing or anomalous records to the trial monitor before they become audit findings at database lock.

    • Trial dashboards and site-to-R&D feedback

      Trial status visible across continents

      Surface trial progress, parameter distributions and integrity exceptions to R&D leadership in Taiwan and trial operations in the US in one view, so the iteration cycle between sites and the algorithm team is measured in days instead of weeks.

    • Data lock at the end of the trial shortens because the record is already there.
    • Regulators read the evidence end to end; the chain of custody is the data set itself.
    • R&D iteration across continents runs on days, not on transcribed forms.

    We've done this before

    See how we do Enterprise AI
  • Agents

    Agents for cross-jurisdiction regulatory document work

    Narrow, reviewable agents that take the repetitive part of cross-border regulatory documentation: drafting change-control and validation documents from source records, checking completeness of submissions before review, and finding every controlled document a specification change touches. A named reviewer approves every output.

    • Drafting from source records

      First drafts from the QMS, not from scratch

      Generate the first draft of a change-control summary, a validation summary or a protocol update directly from the underlying QMS and laboratory records, with the same document references a human reviewer would manually assemble.

    • Submission completeness checking

      Gaps found before review

      Check the submission against FDA, NMPA and TFDA checklists before a document enters the human review queue, surfacing missing sections, inconsistent dates or cross-references that would otherwise be caught in review.

    • Impact search across the controlled document set

      Which documents a specification change touches

      When a specification or standard changes, retrieve every controlled document that references it across the three jurisdictions and rank them by how directly the change affects each, so the update scope is known on day one.

    • Review queues move faster because submissions arrive complete and consistent.
    • The blast radius of a specification change is established by search, not by recollection.
    • Every output is traceable to the source records it came from and approved 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 Anivia Medical's own published ambition implies — not a perfect score.

Source: A4BEE analysis of public sources
Device connectivity 20 → 75
Communications on the SG1000 are CAN bus and RS-485 with no documented wireless telemetry; competitor consoles are moving toward hospital network and cloud integration.
Manufacturing digitalization 30 → 85
Zhuhai runs kitted assembly with end-of-line inspection on pilot volumes, and the August 2025 NMPA registration moves the line toward commercial demand.
Quality system integration 35 → 80
Anivia operates in FDA, NMPA and TFDA jurisdictions with separate, largely paper-based quality records that have to agree manually today.
Data analytics capability 25 → 80
Predicted maintenance and clinical evidence generation both need data to leave the equipment that produces it; today that data is collected by hand.
Supply chain visibility 30 → 70
Vendor-built sub-assemblies are inspected at receipt rather than observed in production, leaving tier-2 and tier-3 changes invisible until they affect a built unit.
Field service maturity 25 → 75
Service is scheduled rather than triggered because real-time field telemetry is not yet collected from the installed consoles.

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 Anivia Medical, 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].