What happened
Schneider Electric introduced three new power solutions targeting AI‑intensive data centers: a software‑defined medium‑voltage (MV) switchgear demonstrated in a live Equinix colocation pilot, the RM AirSeT 38 SF₆‑free GIS extending its platform to 38 kV, and the Galaxy VXL high‑density modular UPS supporting both lithium‑ion and VRLA batteries.
During the YOTTA 2026 conference in Las Vegas, Schneider Electric unveiled its Software‑Defined Medium‑Voltage (SD‑MV) switchgear, describing it as the world’s first fully software‑defined MV solution. The system has already been installed in a live Equinix colocation data center, where it is used to manage medium‑voltage distribution for AI workloads.
The company also announced the RM AirSeT 38, an extension of its SF₆‑free AirSeT GIS platform from 24 kV to 38 kV. Designed for next‑generation AI data centers, the product aims to simplify MV backbone design and support higher power capacities required by dense AI racks.
Finally, Schneider introduced the Galaxy VXL UPS, a modular, high‑density uninterruptible power supply capable of delivering up to 1.25 MW per frame. The UPS supports both lithium‑ion and VRLA battery technologies, offering flexibility for operators seeking to balance performance, resiliency, and sustainability.
Source details: engineering.com ↗
Why it matters
The offerings promise faster deployment cycles, reduced material use, and higher power density—key factors as AI workloads push data‑center power densities beyond traditional limits. By decoupling intelligence from hardware, Schneider claims up to three‑fold faster ordering and manufacturing and up to 85 % less copper in control wiring, potentially lowering capex and operational complexity for hyperscalers and colocation providers.
AI and high‑performance computing workloads are driving a shift toward higher rack densities and more dynamic power profiles. Traditional MV switchgear and UPS solutions often lag behind these demands, leading to longer lead times and higher material costs.
Schneider’s SD‑MV approach claims up to three times faster ordering and manufacturing, and up to two times faster commissioning compared with conventional engineered‑to‑order MV switchgear. The reduction in control wiring (87 % less), terminals (91 % fewer), and copper (85 % less) could translate into significant capex savings for data‑center operators.
The RM AirSeT 38’s higher voltage rating aligns with industry trends toward 36‑38 kV MV architectures, enabling more compact backbone designs and reducing the number of distribution points needed for large AI clusters.
Galaxy VXL’s high power density (1.25 MW per frame) and compact footprint address the need for space‑efficient power delivery in AI‑dense environments, while offering battery flexibility to meet varying resiliency and sustainability goals.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
What did neural scaling law research (e.g. Kaplan et al., 2020) observe?
What to watch next
Watch for the rollout timeline—order intake for RM AirSeT 38 begins in 2027 with first shipments in 2028, and broader availability of the software‑defined MV switchgear is slated for 2028. Pricing, licensing models for the software‑defined features, and real‑world performance data from the Equinix pilot will determine market adoption.
The timeline for commercial availability: Schneider plans to open order intake for RM AirSeT 38 in 2027 with first deliveries in 2028, and broader rollout of the SD‑MV switchgear in 2028 after pilot validation.
Pricing and licensing: No pricing information was disclosed. The cost structure for the software‑defined features, including any subscription or licensing fees, will be critical for adoption decisions.
Performance validation: Independent benchmarks of the Galaxy VXL’s efficiency and the SD‑MV switchgear’s reliability in AI workloads will be needed to confirm the claimed benefits.
Competitive response: Other power‑infrastructure vendors may accelerate similar software‑defined or high‑voltage offerings, influencing market dynamics for AI data‑center power solutions.