South Korea's LG to deploy Nvidia AI factory - Data Center Dynamics

South Korean conglomerate LG is set to deploy Nvidia-based GPU infrastructure to support its artificial intelligence initiatives. As part of a broader strategic move, LG Uplus plans to build a large-s... Read original →

LG Group’s full-scale adoption of NVIDIA’s AI Factory architecture signals a strategic pivot from end-product manufacturing to foundational AI infrastructure. Technically, integrating 800V DC power with the DSX platform will force upgrades across power delivery, liquid cooling, and high-density interconnect supply chains—boosting LG Innotek’s advanced packaging prospects. On compliance, tightening U.S. GPU export controls compel LG Uplus to localize deployments, yet its 200MW Paju data center remains exposed to American chip dependency. While Samsung pushes vertical integration with in-house AI accelerators and HBM memory, LG’s NVIDIA alignment reflects a pragmatic bet on speed over sovereignty. Over the next 18 months, LG’s containerized data centers could set an edge-AI template across Asia-Pacific—but persistent TSMC CoWoS capacity constraints may throttle GPU availability and delay monetization.