
A rendering of LG Uplus' planned artificial intelligence (AI) data center in Paju, Gyeonggi Province / Courtesy of LG Uplus
LG Uplus unveiled its artificial intelligence (AI) data center infrastructure strategy at its facility under construction in Paju, Gyeonggi Province, targeting 5 trillion won ($3.2 billion) in cumulative orders by 2030 as it pushes to scale its AI infrastructure business.
The company said Friday it aims to establish a new standard for AI infrastructure by leveraging groupwide One LG synergies and showcasing the global competitiveness of its in-house technologies.
As AI workloads shift from model training to inference, data centers are contending with rising power demands, more variable workloads and intensifying thermal constraints.
To address these challenges, LG Uplus introduced its “ACE on Trust” strategy for the Paju AI data center, which is set to span roughly 150,000 square meters. The initiative centers on agility, capacity and efficiency built on a foundation of operational reliability, aimed at defining a new standard for AI infrastructure.
For agility, the company will adopt a prefabricated modular data center construction method to accelerate deployment timelines. The approach allows for standardized components to be manufactured off-site and assembled on location, reducing build times by several months while enabling flexible scaling from proof-of-concept projects to hyperscale facilities.
Its Paju AI data center, with its first phase slated for completion in June 2027, will use preassembled modular components to accelerate construction and respond more quickly to demand.
With a planned capacity of 200 megawatts, the upcoming facility is currently the only data center in the Seoul metropolitan area capable of supporting that level of power, positioning it as a key hub for inference-driven AI workloads.

A poster for LG Uplus' “ACE on Trust” strategy / Courtesy of LG Uplus
Beyond directly building facilities, the company plans to partner with global firms and asset managers to deliver customized design-build-operate services to meet rising demand.
Efficiency is another core focus. The Paju center is being built with a hybrid cooling architecture that integrates both air and liquid cooling systems at hyperscale for the first time in Korea.
The liquid cooling solution, developed in collaboration with LG Electronics, uses a direct-to-chip method with coolant distribution units, improving energy efficiency by approximately 24 percent compared to traditional air cooling, according to the company.
To improve operational stability, the company said the Paju facility will use robotics systems to monitor temperature, humidity, water leaks and dust around the clock while continuously surveying the surrounding premises.
“AI data center competitiveness is no longer determined by facility size alone, but by how reliably operators can manage the entire infrastructure stack,” the company’s official said. “The Paju AI data center will serve as a flagship example of that capability.”
LG Uplus said it ultimately aims to evolve into what it calls an AI factory operator, moving beyond simply building data centers to managing resources for graphics processing units, power and cooling systems in an integrated manner so AI workloads can operate under optimized conditions.
According to the company, the facility’s first building has already reached full capacity, with all contracts signed.