[Edaily Reporter Shin Yeong-bin ] LG CNS (#LG CNS) is launching a “prefabricated modular data center” business, in which key equipment for artificial intelligence (AI) data centers is prefabricated in a factory and then assembled on-site. The company aims to capture the AI factory market by shortening construction time and expanding infrastructure according to customer demand.

LG Corp. announced on the 30th that it unveiled its Prefab Modular Data Center (PMDC) and “AI Factory Full-Stack Service” at “Data Center World Asia 2026,” held in Singapore on the 29th (local time).

The PMDC approach involves pre-manufacturing the graphics processing units (GPUs), servers, and power and cooling equipment required for an AI factory into standardized modules, which are then installed on-site. This reduces on-site construction work compared to the conventional method of building mechanical, electrical, and plumbing (MEP) systems and AI infrastructure after the data center building is completed.

The infrastructure can also be scaled up by adding modules as the required AI computing capacity grows. LG Corp. plans to supply AI infrastructure to customers when they need it and support them in expanding capacity in phases as their businesses grow.

The company will also integrate relevant technologies from LG Group affiliates. By combining LG CNS’s data center design, build, and operation (DBO) capabilities with LGELECTRONICS’ modular cooling infrastructure and LG Energy Solution’s battery solutions, the company will utilize these resources to build AI factories.

LG Corp. is also pursuing a “modular AI factory” business utilizing PMDC. This approach involves building small, container-based AI data centers without constructing separate large-scale data center buildings, and then adding containers as needed based on required functions and computational scale.

Currently, the company has unveiled modules with an IT load capacity of 4 megawatts (MW) based on NVIDIA’s next-generation “Vera Rubin” GPU, as well as 1.2 MW modules based on the “GB300.” It plans to expand its product lineup to include various specifications in the future.

The company also unveiled its “AI Factory Full-Stack Service” based on NVIDIA’s Vera Rubin. This service integrates GPU-as-a-Service (GPUaaS), which provides Vera Rubin GPU resources via the cloud; intelligence services that support real-time training and inference, such as Physical AI; and the power and cooling technologies that underpin them.

LG Corp. is expanding its capabilities in transforming existing data center infrastructure into AI data centers and in building power and cooling infrastructure tailored to high-performance GPUs. The company plans to use this exhibition as an opportunity to expand its global AI data center business, including in Southeast Asia, North America, and Europe.

Kim Tae-hoon, Executive Vice President and Head of the AI Cloud Business Division at LG Corp., stated, “AI Factory is a comprehensive service that combines data centers, GPU infrastructure, and AI services,” adding, “We will lead the AI Factory ecosystem based on our capabilities in data center design, construction, and operation, as well as our AI infrastructure technology.”