Nexans 700 VDC Microgrid: What Changes for Factories and Spares

Nexans has switched part of its own AmpaCity global innovation hub in Lyon, France, onto a 700 V direct-current main distribution bus. The first phase was commissioned and announced on 11 May 2026, developed with Schneider Electric and VINCI Energies Building Solutions, and it is one of the first Current/OS pilots deployed in France. A cable manufacturer re-wiring its own research campus is a small project in megawatt terms. It is a large signal in standards terms, because the equipment that makes a building run on 700 VDC is the same equipment class that data centres and, next, factories will have to buy.
This article separates what the pilot actually demonstrates from what the ecosystem claims, explains the control and protection rules behind it, and then does the part most coverage skips: what a DC bus does to an automation maintenance budget — which parts get more critical, which parts stay exactly as they are, and what to stock while sites retrofit in phases.
Key takeaways
- Nexans leads this project, not Schneider Electric. Schneider Electric supplies the AC/DC interlink converter integrated into the pilot; VINCI Energies Building Solutions contributed the microgrid engineering and design. Nexans is the site owner, the system integrator of its own cable portfolio, and the party that published the result.
- Phase one is a building demonstrator, not a factory. The loads on the bus are USB-C office workstations, lighting, a 14.4 kWh battery and two 30 kW EV chargers. Solar PV comes in phase two. The industrial case is argued by extension, not demonstrated on this site.
- The control signal is voltage, not a network packet. Current/OS droop curves make bus voltage the shared, physical coordination mechanism, so source and load management keep working if communications go down. That is a deliberate design choice and it changes what “edge energy monitoring” means in practice.
- 700 VDC is a nominal value with an operating window. Systems typically run 640–760 VDC, with time-limited excursions to 820 VDC on the source side and 840 VDC on the load side. Data-centre 800 VDC and 1500 VDC values are accommodated inside the 700/1400 VDC hardware limits.
- Protection is where the real change lands. The Current/OS DC zone model grades installations by fault behaviour and required breaker speed: under 50 ms for electromechanical devices and fuses, under 1 ms for superconducting hybrid breakers, under 10 µs for solid-state breakers. AC-rated miniature circuit breakers do not transfer to a direct-current bus.
- Most of the spares list survives. Converters, DC-rated protection, DC bus cable, 24 V control power and the control layer change in importance; drives, PLCs, sensors, condition monitoring and the AC side of the installation remain the bulk of what you order.
What was actually commissioned
The press release is unusually specific for a sustainability announcement, which makes it worth reading as a specification rather than as a headline.
| Element | As commissioned at AmpaCity (phase 1) | Why an industrial reader should care |
|---|---|---|
| Site | AmpaCity, Nexans’ global innovation hub in Lyon; part of the campus converted to a fully DC-operating environment | A live occupied building, not a lab bench — ageing, availability and maintenance access are being measured |
| Main bus | 700 VDC main distribution bus, connected with Nexans DC bus cable | 700 VDC is one of the two nominal values aligned across Current/OS, EMerge Alliance and the Open Data Center Alliance |
| Loads | Office workstations powered over USB-C, lighting, two 30 kW EV chargers | DC-native loads avoid a conversion stage each; USB-C lighting and sockets are the pattern that repeats in offices and control rooms |
| Storage | 14.4 kWh battery on the bus | Small by design: it demonstrates that voltage-band control can ride through without a large battery |
| AC interface | Grid connection through an interlink converter integrated with Schneider Electric | The site is grid-connected, not off-grid; the converter is the single point where AC rules meet DC rules |
| Generation | PV added in phase two | Phase one proves distribution and control, not renewable self-consumption |
| Cable portfolio | Nexans DC Series: enhanced ageing performance, inter-tripping capability, Current/OS interoperability | Cable is the product being validated; interoperability is the claim that third-party certification exists to test |
| Positioning | Described as the first cable manufacturer to integrate a DC microgrid in its own R&D facility; stated to pave the way for future 700–800 VDC standards and directly applicable to data centres | This is the commercial thesis: building-scale evidence now, data-centre specification later |
The quotes in the release map to those roles. Jérôme Fournier, VP Group Innovation at Nexans, frames the pilot as the step that accelerates DC adoption in buildings and credits the partner collaboration for testing under real conditions. Andrea Quarteroni, New Electrical Distribution League Leader at Schneider Electric, describes DC as an opportunity to rethink how energy is distributed and managed inside a building. Eric Plumey, Managing Director of VINCI Energies France Building Solutions, states that the transition to DC buildings needs industry-wide collaboration and that the company contributed microgrid engineering and design expertise to make scalable solutions reach the market. Vincenzo Salmeri, President of Current/OS, positions the site inside the consortium’s pilot programme.
Three corrections worth making before you plan a retrofit
Secondary coverage of this project has drifted in three directions that matter if you are using it to justify a capital decision.
One: the lead is the cable company. The framing “Schneider Electric and partners advance a DC microgrid architecture” inverts the relationship. Schneider Electric is one of several named partners and is credited in the release with the integrated interlink converter. Nexans owns the site, the programme and the cable portfolio under test. That distinction tells you what is actually being proven: a cable and bus infrastructure claim, not a switchgear claim.
Two: it is a building, and the release does not claim a factory. The 14.4 kWh battery and 30 kW chargers are building-scale numbers. A production plant has motor loads, drives, welding, compressors and process heat that a campus office does not. Factory and process adoption is supported elsewhere in the ecosystem — the industrial low-voltage DC standardisation work at IEC and the German DC-INDUSTRIE programme are real — but AmpaCity phase one is not the evidence for it.
Three: “edge energy monitoring integration” is the wrong mental model for the control layer. The published technical rules deliberately make bus voltage the primary control signal. Sources and loads read the voltage band they are connected to and respond along a droop curve. Communications are available and useful — and the consortium’s own materials note that source and load management continue to work when communication is lost. A monitoring gateway is an optimisation on top of a physical control loop, not the control loop itself. If you spec this architecture as an IoT project, you will miss the part that keeps the bus stable when the network is not.
Why a cable maker’s own campus is the interesting part
Demonstrators are cheap. What makes AmpaCity worth reading twice is who is running it and what they are measuring. A cable manufacturer’s exposure in a DC architecture is technical rather than commercial: DC does not have a zero crossing to extinguish an arc, fault currents rise differently, and insulation and conductor behaviour under permanent polarity and power-electronic harmonics is not the same as under 50 Hz. So the “enhanced ageing performance” and “inter-tripping” language in the release is the company addressing the failure modes that kill DC adoption, on its own building, where an outage is embarrassing.
The same reasoning explains the data-centre bridge in the release. A 700–800 VDC bus inside a building and an 800 VDC bus inside an AI factory are the same distribution problem at different scales, and the cable, connector and protection specification transfers more directly than the load profile does.
Voltage is the control signal: how the droop rules work
Current/OS is an open, non-profit partnership; its published material describes more than 160 strategic partners in over 30 countries, and the governance list is the clearest evidence that this is a standards play rather than a vendor protocol. ABB chairs the Technical Committee, Schneider Electric chairs the Certification Committee and leads DC codes and standards strategy, Nexans leads the Cable Group, Siemens chairs a Specialization group, and Eaton holds the protection devices, arc-flash and standardization leads; UL Solutions, Tridonic, Mersen, Jacobs and Uninova appear in the same structure.
The mechanism is simple enough to state in one paragraph and specific enough to design against.
| Nominal DC | Where it sits | Practical note |
|---|---|---|
| 48 VDC | Telecom, legacy control circuits, low-power zones | Familiar to anyone who has maintained a 48 V battery room or a DC control panel |
| 175 VDC | Intermediate tier in the published voltage set | Present in the rules, rarely quoted in project material |
| 350 VDC | Buildings and lighter distribution; aligned with EMerge Alliance and ODCA | The competing building choice; expect mixed 350/700 estates |
| 700 VDC | The AmpaCity choice — buildings, industry, data centres | Typical operating window 640–760 VDC |
| 1400 VDC | Higher-power distribution | Accommodates 1500 VDC data-centre values within hardware limits |
Within a nominal tier, the band is the message. The rules define time-limited operation up to 820 VDC for sources and 840 VDC for loads on a 700 VDC system, and reference IEC TR 63282 for the band concept. Droop curves translate deviation from the nominal band into a power response: how hard a source pushes, how much a load draws, whether storage charges. Those curves are tunable at commissioning and in operation, which means the same hardware can be biased towards self-consumption, towards resilience, or towards grid-service response.
The behaviour that follows is worth stating plainly, because it is the argument for the whole architecture:
- Oversupply: the bus rises, loads run at their maximum capacity and storage absorbs the rest — batteries first, then anything thermal that can be deferred, such as hot water, refrigeration or pre-cooling.
- Undersupply: the bus sags, loads shed in priority order, thermal storage runs idle, and the battery discharges.
- Communication loss: the band and the droop curves still exist, so the system keeps regulating itself.
That is a different reliability story from “an edge gateway reports consumption to a platform”. If you are maintaining the plant, the difference is that a network outage becomes an optimisation problem rather than a power problem.
Protection, earthing and arc-flash: the part that decides your spares list
The technical rules grade a DC installation into zones by how much fault energy is available and how fast the protective device must act. This ladder is the single most useful thing in the document for a maintenance planner, because it converts a vague “DC is dangerous” concern into a device-speed specification.
| DC zone | Source character | Required breaking time and device class | Design consequence |
|---|---|---|---|
| Zone 0 | Unprotected source with high short-circuit current | No overcurrent protection present | High fault current and high arc-flash incident energy; circuits cannot be distributed — they must be segregated in panels or restricted-access rooms |
| Zone 1 | Protected source, high short-circuit power | Under 50 ms — electromechanical breakers and fuses | Multiple sources on a common distribution board; conventional protection practice still applies |
| Zone 2.1 | Protected source, bounded short-circuit power | Under 50 ms, with capacitor stored energy not exceeding 600 J | Sources cannot deliver much more than nominal by design; arc-flash energy lower than Zone 1 but still requires assessment; design by skilled engineers and hard to replicate or upgrade |
| Zone 2.2 | Bounded source, distributed, with fault-current-limiting inductance | Under 1 ms — superconducting hybrid circuit breakers (SCHCB) | Capacitor discharge limited; sources can be distributed; arc-flash energy reduced; selectivity possible but difficult; designs become replicable, suitable for complex industrial installations |
| Zone 3 | Multiple electronic sources | Under 10 µs — solid-state circuit breakers (SSCB) | Fault current ultra-limited; incident energy far below 1.2 cal/cm², so assessment is typically not required; no fault calculation needed, design by nominal current; loop or mesh topologies become easy; residual-current protection adds a layer; mid-point earthing simplifies maintenance |
| Zone 4 | Single electronic source | As Zone 3 | Same consequences at a single-source scale |
Two more mechanisms are worth carrying into a specification. The intertripping wire is deliberately unsophisticated: a 48 V high-impedance signal that, when shorted to 0 V, turns down the connected sources and disconnects the grid feed. It doubles as an emergency shutdown of everything around it for maintenance — safety and control from one simple conductor, which is also a cable product. And the earthing principle is to isolate the DC microgrid from the AC grid, with a TN-S arrangement on the DC side, which is what makes mid-point earthing and simpler maintenance possible in the electronic zones.
The practical conclusion is uncomfortable and useful at the same time: the fast end of the ladder is where solid-state protection and power electronics overlap, and that is the least mature part of the supply chain. ABB states that its Infinitus solid-state circuit breaker is the first to be fully certified to IEC 60947-2 and is in production, which is a real milestone — but it is one vendor’s device against a zone model that assumes fast breaking everywhere, and the wider standardisation of DC protection devices is still catching up. Until then, a real installation is a mix of Zone 1/2 protection practice and Zone 3 ambitions.
Buildings are demonstrated; data centres are where the volume is
AmpaCity is not alone in the Current/OS reference list, and the set of projects is more informative than any single one.
| Project | Type | Published claim | Evidence strength |
|---|---|---|---|
| Wave, Lille | DC office building, VINCI Energies | 20% energy saving | Operating building, stated by the consortium |
| AmpaCity, Lyon | R&D campus, Nexans | 700 VDC bus, USB-C workstations, 14.4 kWh storage, 2×30 kW charging | Commissioned phase one, owner-published |
| MiRaCCLE | DC industrial park, Belgium | Industrial-scale DC district | Consortium reference |
| Shift2DC Stuttgart / Funchal port | Data centre / port | DC distribution for compute and marine electrification | Consortium references |
| N470 highway, Delft / Pulse, TU Delft / The Green Village | Infrastructure and research | Peak feed-in power reduced 2–5×, cited against Circel Pavilion, a.s.r. car park and the N470 | Measured pilots, academic sponsors |
On the data-centre side the pressure is arithmetic rather than environmental. IEA projections cited by ABB in its 21 September 2026 Infinitus Direct Current announcement expect data-centre electricity demand to more than double by 2030, and ABB’s own estimate is that 25–40% of new capacity added in 2030 could use DC distribution. Next-generation AI racks are heading above 1 MW against roughly 200 kW today. ABB, with BCG, puts the efficiency gain from DC distribution at more than 5%, and sizes the commercial consequence: on a 500 MW campus, about 25 MW of extra IT power, or more than US$300 million of potential additional revenue per year. The same release commits ABB to having main components in production within 12 months and the first complete DC-native sites within two to three years, and states that the portfolio extends the NVIDIA collaboration announced in October 2025.
NVIDIA’s 21 October 2026 Europe announcement about the DSX platform supplies the demand-side numbers, measured rather than projected: DSX MaxLPS is stated to use up to 30% less power for the same token output; a Lambda proof of concept on HGX B200 reported 24% more cluster token throughput and 19% more GPU capacity inside a fixed power budget; Nscale, on four GB300 NVL72 racks, reported 49% higher inference throughput and 37% more managed GPU capacity within the same 264 kW. A DSX Flex trial with Emerald AI, EPRI, Nebius and the UK National Grid reported up to 40% cluster power reduction during simulated grid events. Schneider Electric appears there as a partner that built power and cooling reference designs for the Vera Rubin generation and works with AVEVA on digital twins; Hitachi Energy is integrating its energy management system with DSX Flex, with GE Vernova and Siemens Energy on the power-infrastructure side.
Standards work is following the pilots: Current/OS and the Open Data Center Alliance are reported to be engaging with NFPA ahead of the 2029 National Electrical Code cycle; IEC has published standardisation study SRD 63317:2025 on industrial low-voltage DC; and the German VDE-SPEC 90037, driven by the DC-INDUSTRIE and DC-INDUSTRIE2 projects, already specifies industrial LVDC installations. NVIDIA, Google and Microsoft are pushing 800 VDC through the Open Compute Project, with a joint white paper and an LVDC solid-state-transformer specification issued during 2026 and more than 80 equipment vendors reported to be involved.
What actually changes on the spares list
Here is the section that has to be read with an honest eye, because the temptation is to declare a new parts catalogue. The truthful version: a DC retrofit is phased, and during the phase-in most of your consumption is unchanged. What changes is the criticality of a few categories and the appearance of one new category — power conversion.
| Layer | What a DC bus does to it | Stocking consequence |
|---|---|---|
| AC distribution | Stays; the grid link and the interlink converter sit behind it | Unchanged demand for AC circuit protection, e.g. Schneider Electric iC60N 2P C16, iC60N 2P C63, iC60N 4P C63, iC60H 2P C20 |
| Control power (24 VDC) | Gets closer to the main bus; more converter stages upstream of it | Higher criticality: ABL8RPS24100 240 W, TSXPSY1610, TSXPSY1610M, TSXPSY5500M, BMXCPS4002S, BMXCPS3000, AS-P120-000, ABB SD831, ABB SD834 |
| Ride-through and redundancy | The bus can sag or be shed by design; control must not follow | Battery buffering and redundancy stop being optional: ABL8BBU24200 battery control module, 140CPS12400, 140CPS22400 redundant, 140CPS11400C 24 V 8 A |
| DC protection and switching | New duty: no current zero crossing, speed defines arc-flash energy | DC-rated devices, not AC equivalents; see circuit protection and Schneider Electric switching such as TeSys Giga LC1G1504KUEN and LC1G2654KUEN |
| Drives and motor loads | DC-connected operation removes a rectification stage; ABB explicitly lists cooling VSDs and motors fed from DC | Drive spares stay central: ATV930D22N4 22 kW, ATV320U75N4WS 7.5 kW IP65, ATV320U30N4C 3 kW — but DC-input ratings must be confirmed per unit |
| Control and logic | Load shedding by priority needs something to execute priorities locally | Zelio SR3B261BD, SR3NET01BD Ethernet module, 140CPU11302, 140CPU11303C, remote I/O such as Advantys STB |
| Cabling and connectivity | The item under test: DC bus cable, intertrip signalling, USB-C power outlets at the desk | See cables and connectors; intertrip wiring is a new circuit that did not exist in an AC panel |
| Condition monitoring and surge | More power electronics means more harmonics and more transient paths | Bently Nevada 3300/12-01-20-00 AC power supply, MTL IOP HC32, MTL IOP32D |
Two cautions inside that table are worth repeating. First, an AC-rated miniature circuit breaker such as the Acti9 iC60 range is listed here on the AC side of the converter precisely because it must not be assumed onto a 700 VDC bus: breaking an arc without a current zero crossing is a different device problem, and the zone ladder above is a breaker-speed specification, not marketing. Second, a variable frequency drive’s DC bus is not the same thing as being rated for connection to a 700 VDC facility bus; DC-input capability, precharge behaviour and inrush limits have to be read from the unit’s own data sheet before anything is connected.
For the categories where we hold stock, the practical entry points are power supplies, drives and inverters, PLCs and controllers, I/O and function modules, communication modules and the ABB and Schneider Electric brand ranges. If you are working out what a specific part number means before ordering, how to verify an automation part number before you order and what new surplus, used and repair conditions actually mean cover the two questions that come up on every retrofit project.
What does not change
Sensors, encoders, proximity probes, vibration monitoring, safety relays, HMIs and the mechanical side of a plant do not care what the distribution bus is. Neither does the discipline that decides whether a plant stops: knowing which part number you need, having it on the shelf, and being able to confirm a replacement is electrically and mechanically identical. A DC architecture shifts a few categories up the criticality list and adds converters to it; it does not shorten a lead time or remove an obsolescence problem.
That is also the honest answer to the “should we plan for DC now” question. For a building being wired from scratch, or a data-centre expansion, the standards direction is now clear enough to design against. For an existing factory, the near-term change is not the bus voltage — it is the arrival of more power electronics, more storage and more shedding logic inside the same envelope, which is a power-quality, control-power and spares-criticality problem today.
Procurement risk over the next 24 months
- Two building voltages will coexist. 350 VDC and 700 VDC are both aligned nominal values, and the data-centre track pushes 800 VDC. Components that straddle bands through wider hardware limits are the ones that stay serviceable; single-band parts risk becoming a site-specific SKU.
- Fast DC protection is pre-standard. Solid-state and hybrid breakers are the destination of the zone model, with the relevant IEC standard still in progress. Expect vendor-specific modules, short product families and first-generation firmware. Keep AC-side spares normal and treat DC protection as a design decision to revisit at each expansion.
- Converter failure domains are new. The interlink converter is the one device whose outage returns a DC island to whatever its battery can hold. A 14.4 kWh battery in a building is a demonstration; in a plant, ride-through sizing is a calculation you have to own. Redundant control power and battery-buffered 24 V rails are the cheap insurance.
- Phase-in means mixed estates. Almost every real site will run AC and DC distribution side by side for a decade. Spares planning that assumes a clean cutover will over-buy the new world and under-buy the old one. Our guide to managing industrial automation obsolescence with surplus stock is written for exactly this situation, and the Schneider Electric spare parts buying guide covers how series and part numbers behave across long lifecycles.
A practical checklist for engineering and maintenance leads
- Ask any DC proposal which DC zone each circuit falls into, and get the required breaking time in writing. “DC-rated” without a breaking time is not a specification.
- Confirm the nominal voltage and operating band a device supports — 640–760 VDC continuous with defined time-limited excursions, not “up to 800 V”.
- Ask whether source and load management survive loss of communications, and whether droop curves are configurable at commissioning. If the answer is that everything depends on a gateway, you are being sold a monitoring project with a power bus attached.
- Check what intertrip and earthing arrangement is proposed — a 48 V high-impedance intertrip circuit and a TN-S DC side with mid-point earthing are the published patterns.
- Size ride-through for control power and safety-critical instrumentation independently of the site battery.
- Verify interoperability certification rather than brand compatibility claims; the Current/OS Certification Mark exists to cover multi-vendor interoperability, and third-party conformity is the difference between an ecosystem and a proprietary bus.
- Write the mixed-estate spares policy now: which AC parts stay stocked while the DC side is single-vendor, and what triggers re-evaluation.
Numbers to treat with care
Because this is a marketing-adjacent topic, some circulating figures deserve a label rather than repetition. The Nexans release publishes no efficiency percentage for AmpaCity. The “>5% efficiency gain” and the 25 MW / US$300 million campus arithmetic are ABB’s and BCG’s, for data centres, not for this pilot. The 20% saving belongs to the Wave building in Lille. The 2–5× peak feed-in reduction belongs to the Circel Pavilion, a.s.r. car park and N470 projects. The “3–5 percentage point efficiency gain and more than 40% copper reduction” figures attached to 800 VDC in secondary summaries of the NVIDIA/Google/Microsoft work are not stated as headline numbers on NVIDIA’s own 800 VDC page, which describes the mechanism — fewer conversion stages, lower current, less copper and cable bulk versus 54 VDC racks and 480 VAC halls — without publishing a single percentage. Similarly, claims about specific rack generations and 2027 timelines come from analyst summaries of the white-paper series and should be checked against the primary documents before they enter a capital plan.
Where the argument is strong, it is strong without the round numbers: a cable manufacturer has put a 700 VDC bus in its own building under an open interoperability framework whose technical rules already specify the voltage band, the droop mechanism, the protection-time ladder and the earthing scheme. That is what turns a pilot into a procurement question.
FAQ
What is a 700 VDC microgrid and why 700 volts?
700 VDC is one of the nominal direct-current distribution voltages published in the Current/OS technical rules, aligned with the EMerge Alliance and Open Data Center Alliance voltage sets so that building, industrial and data-centre equipment can share a supply chain. It is high enough to keep currents and copper bulk manageable for power distribution and low enough to stay in the low-voltage regulatory world. In practice a 700 VDC system operates around 640–760 VDC, with time-limited excursions to 820 VDC on the source side and 840 VDC on the load side.
Is the Nexans AmpaCity project a factory or a building?
A building. The first commissioned phase covers part of the AmpaCity R&D campus in Lyon and runs USB-C-powered workstations, lighting, a 14.4 kWh battery and two 30 kW EV chargers from a 700 VDC bus, with solar PV planned for phase two. Schneider Electric is credited with the integrated interlink converter and VINCI Energies Building Solutions with the microgrid engineering and design. Industrial adoption is argued by extension from this and other pilots, not demonstrated at AmpaCity.
How is a DC microgrid controlled if the network goes down?
It keeps controlling itself, because the control signal is physical. Sources and loads follow droop curves defined against the bus voltage band, so oversupply pushes loads to maximum capacity and into storage, while undersupply sheds loads by priority and discharges the battery. Communications are used where they add value, but the rules are designed so that source and load management do not depend on them. That resilience is the main architectural difference from a monitoring-platform approach.
Can I use my existing AC circuit breakers on a 700 VDC bus?
No. Direct current has no natural current zero crossing, so arc interruption is a different problem and the required speed is defined by the installation’s DC zone: under 50 ms for electromechanical breakers and fuses in the high-fault-current zones, under 1 ms for superconducting hybrid breakers, and under 10 µs for solid-state breakers in the electronic-source zones. AC-rated breakers such as a standard miniature circuit breaker belong on the AC side of the converter, where they continue to do their normal job.
What does a DC retrofit change about my spare parts?
It raises the criticality of a few categories rather than replacing the catalogue. Control power supplies, battery-buffered and redundant 24 V rails, DC-rated protection, DC bus cable and intertrip circuits, and the converters themselves become the parts that can stop a process. Drives, PLCs, remote I/O, sensors, condition monitoring and everything on the AC side stay in service, and during a phased retrofit they remain most of what you order.
Is DC distribution only a data-centre story?
Data centres are where the economics are clearest — next-generation AI racks are moving above 1 MW against roughly 200 kW today, and ABB estimates that 25–40% of new data-centre capacity in 2030 could use DC distribution. Buildings are where the pilots are running now. Factories are the third case, supported by industrial low-voltage DC standardisation at IEC and through the German VDE-SPEC 90037 work, but still ahead of demonstrated large-scale deployments.
Related reading on this site
Two adjacent pieces explain how the compute side and the component side are moving at the same time: what the Siemens and NVIDIA Erlangen AI factory blueprint changes for ramp-up engineering, and EPC’s new 100 V GaN motor-drive evaluation boards, which are the same power-electronics trend at actuator scale. On the supply side, what ABB’s acquisition of Rotork means for actuator spares shows how electrification portfolios are consolidating, and the Rockwell CompactLogix 5390 launch analysis covers the controller layer that executes load-shedding priorities. For ordering practice, see how to request a quote and get a fast response, and the guide to confirming an exact replacement.
Sources
- Nexans, “Nexans turns AmpaCity into a cutting-edge DC microgrid”, 11 May 2026 — nexans.com
- Nexans, innovation programme page describing AmpaCity as a DC microgrid demonstrator — nexans.com
- Current/OS Technical Rules: voltage bands, droop control, DC zones 0–4, intertripping wire, earthing — currentos.org
- Current/OS project references (Wave Lille, MiRaCCLE, Shift2DC, N470, Pulse) and impact figures — currentos.org
- ABB, “ABB’s new direct current portfolio aims to rewire AI data center energy infrastructure”, group press release, Zurich, 21 September 2026 — abb.com
- NVIDIA, “European ecosystem builds world’s most efficient AI factories with NVIDIA DSX”, 21 October 2026 — nvidianews.nvidia.com
- NVIDIA, 800 VDC architecture overview — nvidia.com
Standards references: IEC TR 63282 (DC voltage bands), IEC SRD 63317:2025 (industrial low-voltage DC standardisation study), VDE-SPEC 90037 (industrial LVDC, DC-INDUSTRIE / DC-INDUSTRIE2), IEC 60364 series as applied by the Current/OS zone model. Autronexa is an independent supplier of industrial automation spare parts; it is not affiliated with Nexans, Schneider Electric, ABB, VINCI Energies or Current/OS, and nothing here is engineering advice for a specific installation. For a parts question on a live retrofit, contact us or see about Autronexa, resources, FAQ and the news and blog index.
