LG CNS 印尼 AI 數據中心完工前租出全部容量，GPU 投資採審慎策略

LG CNS has secured lease agreements for 100% of its artificial intelligence (AI) data center capacity in Indonesia ahead of completion. Meanwhile, the company said it would invest gradually in its graphics processing unit (GPU) services business after assessing demand and commercial viability.

During its third-quarter earnings conference call on Oct. 30, LG CNS announced that its AI data center in Jakarta, Indonesia, is scheduled for completion in early November. The company added that the facility has already secured tenants for its entire capacity.

“We secured customers to lease the entire facility before completion, establishing a stable revenue base from the start of operations,” said Yoo Ki-woong, vice president of LG CNS’ AI and Cloud Business Division, during the conference call. He explained that the company is pursuing additional data center construction projects in Indonesia and exploring expansion into North America and Europe.

The Indonesian data center is LG CNS’ first overseas AI data center construction project.

In 2024, the company signed a data center consulting and design contract with local operator KMG. It subsequently secured a 100 billion won infrastructure construction contract in August 2025. KMG is a joint venture established by Indonesia’s Sinar Mas Group and Korea Investment Real Asset Management. LG CNS participated in the project through LG Sinar Mas Technology Solutions, its joint venture with Sinar Mas Group.

The first-phase data center has 11 above-ground floors, a total floor area of 46,281 square meters and a power intake capacity of 30 megawatts (MW). It can accommodate more than 100,000 servers. Project owner KMG plans to expand total power intake capacity to 220 MW.

In North America and Europe, LG CNS plans to focus on prefabricated modular data centers.

“We are pursuing business cooperation through our U.S. subsidiary and local partners,” Yoo said. “We plan to enter the market using modular data center technology designed to support Nvidia’s next-generation AI platform, Vera Rubin.”

According to the company, the modular approach reduces on-site construction work compared with conventional methods and allows facilities to be expanded in stages according to customer demand. However, LG CNS intends to manage investment risks in its GPU services business separately from its broader AI infrastructure expansion.

LG CNS chief financial officer (CFO) Song Kwang-ryun expressed a positive view of the medium- to long-term growth potential of GPU-as-a-Service (GPUaaS) during the conference call. He said the company is reviewing business opportunities involving LG Group affiliates and external enterprises but must consider the substantial upfront investment costs and rapid pace of technological change.

“We will proceed cautiously, focusing on projects with established commercial viability,” Song said of future investments.

LG CNS previously decided at a Sept. 11 board meeting to invest 381.4 billion won in GPUs and infrastructure equipment, including Nvidia’s Vera Rubin. The equipment acquisition is scheduled for June 30, 2027.

The company appears to intend to determine the scale of future GPUaaS investments based on actual demand and profitability, separately from the GPU investment already approved. GPUs require substantial upfront capital, and rapid equipment replacement cycles mean customer acquisition and utilization rates can significantly affect investment returns.

Meanwhile, AI and cloud revenue reached 961.5 billion won in the third quarter, accounting for 57.8% of LG CNS’ total revenue of 1.6643 trillion won.