LG Uplus Targets $3.26B in AI Data Center Orders by 2030
LG Uplus is significantly investing in AI data centers, targeting $3.26 billion in orders by 2030, with its Paju AI Data Center designed to handle inference-driven AI workloads. The facility will use prefabricated modular construction and advanced cooling technologies from LG Group allies to accelerate deployment and optimize AI operations. LG Uplus plans to transition into an "AI factory operator," managing GPUs, power, and cooling systems for optimized AI workloads.
Key Takeaways
- LG Uplus projects $3.26 billion in cumulative AIDC orders by 2030 and 15-20% annual sales growth.
- The 200MW Paju AI Data Center will accommodate high-power, inference-driven AI workloads.
- Prefabricated modular construction and 'One LG' alliance technologies will accelerate deployment and optimize operations.
- The Paju facility supports both air and liquid cooling, with co-developed Direct-to-Chip technology improving energy efficiency by 24% over standard air cooling.
- LG Uplus will transition into an "AI factory operator," managing GPUs, power, and cooling for optimized AI workloads, with the first Paju building fully pre-sold for June 2027 completion.
Why It Matters
LG Uplus's significant investment in AI data centers, particularly for inference workloads, signals a shift in infrastructure priorities as AI applications mature. This move positions the telco beyond traditional connectivity, directly addressing the growing demand for specialized, high-density computing environments. The strategy of leveraging in-group alliances and modular construction could establish a replicable blueprint for rapid, integrated AIDC deployment. Observers should track LG Uplus's success in securing follow-on contracts and the operational performance of the Paju facility to assess the scalability and efficiency of this vertically integrated approach for other telcos.
Additional Context
The expansion by LG Uplus arrives as competition for 'sovereign AI' infrastructure in South Korea reaches an inflection point. Per NVIDIA and SK Telecom (June 2026), the two companies have partnered to build a gigawatt-scale AI Cloud in Korea, utilizing NVIDIA's full-stack DSX architecture to support agentic and physical AI workloads. This project, with its first 'AI factory' slated for 2027, mirrors LG Uplus's timeline and reinforces a broader national push to localize critical computing resources. Simultaneously, NAVER announced in June 2026 that it is expanding its GAK Sejong data center facility using NVIDIA platforms to serve as a sovereign AI foundation for government and enterprise sectors.
Industry analysts at ABI Research (February 2026) forecast that AI workloads will overtake legacy compute in active power capacity by 2031, with inference becoming the primary driver of data center demand as early as 2027. This shift is driving radical changes in data center engineering; per JLL (January 2026), rack densities are now approaching 100kW, making liquid cooling a requirement rather than an option. The South Korean Ministry of Science and ICT also reported in April 2026 that the nation's top three telcos—SK Telecom, KT, and LG Uplus—agreed to increase their collective AI infrastructure investment by 15% this year to secure national competitiveness. While SK Telecom holds a dominant 47% mobile market share as of mid-2025 (per MatrixBCC), rivals like LG Uplus are attempting to differentiate via specialized B2B infrastructure like the Paju facility to capture the surging demand for AI token manufacturing.
Read full article at lightreading.com
Enjoy our coverage?
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source