LG CNS completes first overseas AI data centre in Indonesia

A milestone for South Korean tech

LG CNS has officially completed a 30-megawatt AI data centre located in Jakarta, Indonesia. This project represents a significant milestone for the company as it is the first instance of a South Korean firm directly building an overseas client's AI data centre on-site. The facility was commissioned last year by KMG, which is a joint venture between Indonesia's Sinar Mas Group and Korea Investment Real Asset Management.

The company demonstrated strong market confidence by securing customers for the entire capacity of the data centre before the construction was finished. This early commitment suggests a high demand for specialized AI infrastructure in the Southeast Asian market. By establishing this presence, LG CNS aims to prove that the data centre capabilities developed in South Korea are competitive on a global scale.

CEO Hyun Shin-gyun noted that the construction serves as a bridgehead for the company's international ambitions. He stated that the project proves the capabilities accumulated in South Korea can succeed in the global market. The company now intends to use this success to expand its AI data centre and AI factory businesses across the Asia-Pacific region.

Engineering for high-heat AI

To support high-performance computing, LG CNS designed the Jakarta facility to handle next-generation graphics processing units (GPUs). These processors generate significant heat, which requires specialized cooling solutions to maintain stable operations. The company opted for a hybrid cooling structure that combines traditional air cooling with advanced liquid cooling systems.

The current first phase of the data centre utilizes an air-cooling system to manage the immediate needs of the hardware. As the facility moves into later phases, the company will integrate liquid cooling to handle even higher thermal loads. This modular approach allows the infrastructure to scale in complexity as the computing demands of the tenants increase over time.

Power management is another critical component of the design, with a system capable of supplying more than 300 kilowatts per rack. This amount of power is approximately 60 times greater than what a typical data-centre rack can handle. To ensure reliability, LG CNS duplicated the power infrastructure so that backup power remains available even if the main supply is interrupted.

The 'One LG' integrated solution

The project is notable for its use of a combined solution called 'One LG.' This integrated approach brings together various technologies from different affiliates within the LG Group to create a cohesive infrastructure. Instead of relying on third-party vendors for every component, the company utilized its own internal ecosystem to build the facility.

Specifically, the project combined LG CNS's expertise in data-centre design, construction, and operations with other group technologies. LG Electronics provided the HVAC (heating, ventilation, and air conditioning) systems, while LG Energy Solution supplied the battery technology. This collaboration provided a unified framework for power, cooling, and energy infrastructure for the Jakarta site.

By leveraging these internal resources, the company can offer a more streamlined service to its clients. This model allows for better synchronization between the hardware, the cooling systems, and the energy storage. It serves as a blueprint for how the company plans to deliver integrated AI infrastructure to international clients in the future.

Expansion across Asia-Pacific

Following the completion of the Jakarta site, LG CNS is already looking toward the next steps for the project. The company is currently working on the designs for the later phases of construction in Indonesia. Once these phases are finished, the total receiving capacity of the facility is expected to expand to 400 megawatts.

The success in Indonesia provides a foundation for LG CNS to expand its AI data-centre business across the broader Asia-Pacific region. The company has identified several target markets for this expansion, including Malaysia, Thailand, Vietnam, the Philippines, and Japan. These regions are increasingly seeking the infrastructure necessary to support local AI initiatives and data processing.

In its home market of South Korea, the company is also diversifying its service offerings. Beyond its existing colocation and DBO (Design, Build, and Operate) businesses, it is moving into pre-fabricated modular data centres and GPU subscription services, known as GPUaaS. The company also plans to pursue an AI factory business that links directly with AI platforms and services.