In brief
LG CNS announced the completion of a 30‑megawatt hyperscale AI data center in Jakarta, the first of its kind built by a South Korean company for an overseas client. The facility will support high‑performance GPU computing and is designed for future expansion to 400 MW. The company now aims to extend its AI data‑center operations to other Asia‑Pacific markets and broaden its portfolio with subscription‑based GPU services and liquid cooling infrastructure.
On Monday, the Korean technology group LG CNS announced that it had finished building a 30‑megawatt (MW) hyperscale (a large‑scale, cloud‑based facility capable of handling vast computing loads) artificial‑intelligence data center in Jakarta, Indonesia. The project was commissioned by KMG, a joint venture between Indonesia’s Sinar Mas Group and Korea Investment Real Asset Management Co., and marks the first time a South Korean firm has constructed an AI data center for an overseas client.
LG CNS designed the Jakarta facility to house high‑performance GPUs—graphics processing units that accelerate machine‑learning workloads by performing many parallel calculations. The 30‑MW installation includes a cluster of next‑generation GPU servers, advanced cooling infrastructure and redundant power supply, enabling continuous operation for AI training and inference tasks.
The company said the data center will be expanded in the near future so that its total power capacity could reach up to 400 MW, a figure that would place it among the largest AI facilities in the region. The expansion is slated to increase the number of GPU nodes, enhance storage bandwidth, and add liquid‑cooling systems that cool server racks by circulating chilled water instead of using conventional air‑conditioning units.
LG CNS noted that the Jakarta project is part of a broader strategy to extend its AI data‑center business across the Asia‑Pacific. After Indonesia, the group plans to establish facilities in Malaysia, Thailand, Vietnam and Japan, citing growing demand for cloud‑based AI services from regional enterprises.
Key to the center’s design is the use of automated control systems that monitor temperatures, power usage effectiveness and server health in real time. These systems reduce human intervention, lower operating costs and improve reliability, which is essential for meeting the strict uptime requirements of AI workloads.
According to a company statement, the new data center will also provide subscription‑based GPU services, allowing clients to pay for compute time on a flexible basis rather than purchasing their own hardware. This model mirrors those used by global cloud providers, but with a focus on specialized, high‑density GPU workloads.
In addition to compute services, LG CNS plans to expand into the AI factory sector, offering end‑to‑end AI solutions for manufacturing and logistics. The company claims that the Jakarta facility’s modular design will facilitate the rapid deployment of new services and specialized AI workloads tailored to local market needs.
Industry observers see the Jakarta project as a sign that Southeast Asia is becoming a new hub for AI infrastructure. The ability to localize high‑performance computing brings data sovereignty benefits for Indonesian companies that face strict data‑protection regulations.
LG CNS will also broaden its portfolio to include liquid‑cooling infrastructure, a technology that reduces power consumption and increases server density. Liquid cooling is becoming increasingly important in hyperscale facilities where heat generation can limit the number of compute nodes that can be housed in a given footprint.
Looking ahead, the company anticipates that the Jakarta data center will become a key node in a regional network of AI facilities that will support the growing talent pool of data scientists and AI researchers operating in Southeast Asia.
With the completion of this project, LG CNS demonstrates its capability to build and operate large‑scale, high‑performance AI infrastructure abroad, positioning it as a competitive player in the global AI data‑center market.
