Back to News
IndustryAI Understanding briefing

Socomec announces new SiC‑based UPS and expands AI data‑center power investments in South Korea

French power‑solutions firm Socomec said it will deepen its South Korean presence with a new silicon‑carbide UPS and a broader sales and service push aimed at AI data‑center and semiconductor customers.

5 min readRead the linked source
Source-provided image accompanying Socomec announces new SiC‑based UPS and expands AI data‑center power investments in South Korea
Source referenceSource recorded
Publisher
thelec.net
Source type
Linked source — primary-source status has not been established.
ContextUnderstand this in 60 seconds

Key terms

Inference
The runtime phase where a trained model generates predictions or outputs.
Compute
The processing resources required to train and run models, often measured in FLOPS or GPU hours.
Latency
The time between sending a request and receiving the model's output.

What happened

Socomec Korea held a “2026 Designer Day Korea” event on 5 Nov 2026 in Seoul, where it unveiled a next‑generation uninterruptible power supply (UPS) called DELPHYS XL+. The UPS uses silicon‑carbide (SiC) power semiconductors to deliver higher power density and efficiency for AI‑focused data‑center workloads that rely on GPU‑intensive servers. The company also announced that it will increase its investment in South Korea, strengthening its sales, service and technical‑support teams for AI data‑center and semiconductor customers, and expanding its product portfolio into biotechnology and advanced industrial facilities. The move follows the appointment of Yu Gi‑tae as country manager in April and reflects Socomec’s view of South Korea as a strategic market for digital‑infrastructure growth.

On 5 Nov 2026 Socomec Korea hosted a “2026 Designer Day Korea” event at the JW Marriott Hotel in Seoul. The gathering brought together domestic data‑center developers, architectural design firms and engineering companies. During the event, Socomec introduced its DELPHYS XL+ UPS, which incorporates silicon‑carbide (SiC) power semiconductors. According to the company, the SiC‑based design enables higher power density and improved efficiency, addressing the growing demand for reliable power in AI data centers that run high‑performance GPU servers.

The company also outlined a broader strategic plan to expand its South Korean operations. Since appointing Yu Gi‑tae as country manager in April, Socomec has been building a local sales and service organization focused on data‑center and semiconductor customers. The announced expansion includes hiring additional technical‑support staff, establishing longer‑term service contracts, and extending its product line into biotechnology and advanced industrial facilities, which are also expected to require high‑density power solutions.

Wai Tai Yeap, Socomec’s Asia‑Pacific managing director, emphasized that South Korea is a “important strategic market for data centers and digital infrastructure.” He said the company intends to build long‑term partnerships with Korean customers and increase its investment and support in the region. No specific financial figures, pricing details for the DELPHYS XL+, or rollout timelines were provided in the announcement.

Source details: thelec.net

Why it matters

The announcement signals a sharpening of the power‑infrastructure supply chain for AI data centers in a region that is rapidly scaling GPU‑driven workloads. Silicon‑carbide technology can reduce energy loss and cooling requirements, which are major cost drivers for high‑density AI servers. By localising sales, service and technical support, Socomec aims to reduce downtime and improve the reliability of power delivery for Korean data‑center operators, potentially lowering total‑cost‑of‑ownership for AI workloads. The broader expansion into biotech and advanced industrial facilities suggests that Socomec expects similar power‑density challenges in those sectors, indicating a convergence of high‑performance needs across multiple high‑value industries. However, the article does not disclose pricing, order‑lead times, or the scale of the investment, leaving open questions about how quickly the new UPS will be adopted and whether it will be competitive against existing SiC‑based solutions from rivals such as Schneider Electric or Delta.

AI data centers are distinguished by their reliance on GPU clusters that consume significantly more power per rack than traditional CPU‑based servers. As AI model training and workloads expand, operators face mounting challenges in power delivery, heat dissipation and overall energy efficiency. Silicon‑carbide technology, with its lower on‑resistance and higher switching speeds, can reduce conversion losses and enable more compact UPS designs, directly translating into lower operating costs and higher rack‑density.

Socomec’s decision to localise its sales and service functions reduces the of support and maintenance, which is critical for AI workloads that cannot tolerate power interruptions. By positioning itself as a regional partner rather than a distant supplier, Socomec may capture a larger share of the Korean market, which is seeing rapid growth in AI‑focused data‑center construction driven by both domestic tech giants and foreign cloud providers.

The expansion into biotech and advanced industrial facilities suggests that the power‑density challenges seen in AI data centers are spreading to other sectors that rely on high‑performance , such as genomics, drug discovery and advanced manufacturing. This could create a broader market for SiC‑based UPS solutions beyond the traditional data‑center space.

Nevertheless, the lack of disclosed pricing, order‑lead times or the scale of the investment leaves uncertainty about the speed of market adoption. Competitors like Schneider Electric have already launched SiC‑enabled power modules, and the competitive dynamics will hinge on cost, reliability data and the ability to integrate with existing data‑center infrastructure.

Interactive Mechanism

Interactive Mechanism: How It Actually Works

Explore the underlying technology behind this development interactively.

Model Parameter Size:8B Parameters
VRAM Required5.5 GBGPU memory footprint
Target HardwareMacBook / Single GPUDeployment tier
Privacy100% Air-GappedLocal device capability
Core takeaway: Small, quantized models (3B–8B) now run directly inside smartphones and laptops with complete data privacy, while mammoth 400B+ models remain the domain of datacenter clusters.
Interactive Concept Check+10 Points
Future of AI Quiz

What should a useful AI forecast state?

What to watch next

Key indicators to monitor include: (1) uptake of the DELPHYS XL+ by Korean AI data‑center operators and any announced pilot projects; (2) further statements from Socomec on the size of its South Korean investment and any partnership agreements with local semiconductor or biotech firms; (3) competitive responses from other power‑solution vendors, especially regarding SiC‑based UPS offerings; and (4) regulatory or incentive developments in South Korea that could affect data‑center power‑infrastructure spending.

Adoption rates: Watch for announcements from Korean data‑center operators (e.g., Naver, Kakao, or foreign cloud providers) that they have deployed the DELPHYS XL+ or other Socomec SiC solutions.

Investment scale: Follow any follow‑up press releases or regulatory filings that quantify Socomec’s capital commitment in South Korea, including potential joint ventures or local manufacturing plans.

Competitive landscape: Monitor product launches from rivals such as Schneider Electric, Delta, or local Korean power‑solution firms that may introduce comparable SiC‑based UPS offerings.

Policy environment: Track South Korean government incentives or energy‑efficiency standards that could accelerate the adoption of high‑efficiency power infrastructure for AI workloads.

Related guides & quizzes

Found this useful?