en.Wedoany.com Reported - South Korean telecom operator LG Uplus plans to achieve cumulative orders of 5 trillion won (approximately $3.26 billion) in its AI data center (AIDC) infrastructure business by 2030, with annual sales growth of 15% to 20%. To achieve this goal, the company is building a 200MW hyperscale data center in Paju, Gyeonggi Province.
The Paju data center will consist of five buildings with a total area of approximately 150,000 square meters. LG Uplus stated that this facility is the only data center in the Seoul metropolitan area capable of supporting high power levels, and is expected to become a key hub for inference-driven AI workloads.
LG Uplus noted that the shift in AI operations from training to "inference" has led to increased power consumption, greater volatility, and a sharp rise in heat generation in AI data centers. The construction cycle for data centers spans several years, creating a gap with the rapidly growing demand for AI infrastructure. The company explained that GPUs can be acquired within months, but building the AIDC to house computing power takes three to four years. Additionally, AI token usage more than doubles every quarter, making data center construction delays a key risk to business progress.
To accelerate deployment, LG Uplus plans to adopt a prefabricated modular data center construction method. This approach involves manufacturing standardized components off-site and assembling them on-site, reducing construction time by months while supporting flexible scaling from proof-of-concept projects to hyperscale facilities.
The company will build the Paju AIDC under the "One LG" alliance, integrating localized technologies from the LG Group ecosystem. LG Electronics provides advanced thermal cooling units and free-cooling chillers; LG Energy Solution supplies high-performance uninterruptible power supply (UPS) batteries equipped with multi-layer safety cells; LG Uplus collaborates with LS Electric to develop a high-capacity DC 800V power distribution system. The Paju facility is South Korea's first data center compatible with both air-cooling and liquid-cooling systems. The chip-direct cooling technology, co-developed with LG Electronics, circulates cooling water through a dedicated metal plate attached to the GPU processor. Internal verification shows this technology improves energy efficiency by 24% compared to standard air cooling. LG Uplus stated that from initial construction, the building's load, waterproofing, and piping systems have been optimized for liquid cooling to handle the heat generated by inference-focused GPU servers, enabling efficient cooling for all types of AI chips, including GPUs, CPUs, and NPUs.
LG Uplus is developing an AI-based data center infrastructure management system (DCIM), leveraging predictive AI analysis to anticipate potential power shortages, detect micro-overheating spikes, and strive for 99.99% infrastructure uptime. The company will also deploy autonomous inspection robots to patrol the data center 24/7, monitoring humidity, temperature fluctuations, dust accumulation, and water leaks.
Once the Paju AI data center is operational, LG Uplus plans to transform into an AI factory operator, managing GPUs, power, and cooling systems in an integrated manner to ensure AI workloads run under optimized conditions. Ahn Hyung-kyun, head and executive director of the company's enterprise AI business group, stated that the competitiveness of an AI data center depends not on the scale of the facility, but on the stable operation of the entire infrastructure. The first building of the Paju AI data center is scheduled for completion in June 2027, and it is already sold out, with all contracts signed.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
