South Korea’s prominent online giant, Naver Corp., inaugurated Asia’s largest data center in Sejong, the country’s administrative city 150 kilometers south of Seoul.
“This is a data center we designed based on future visions for the next ten years or more,” Naver CEO Choi Soo-yeon stated.
The GAK Sejong campus, covering an expansive area equivalent to 41 football fields (294,000 square meters), has Naver’s capabilities in high-tech domains such as AI, cloud computing, robotics, and self-driving technologies. The Gak Sejong campus offers a capacity of 270MW, supporting densities of 20kW and facilitating the hosting up to 600,000 server units.
“It will become a technology innovation engine of not only Naver but also all industries,” Choi added.
GAK Sejong represents Naver’s second data center, complementing the tech giant’s existing facility in Chuncheon, which has been operational since 2013. The initial phase of GAK Sejong is set to be operated at just one-sixth of its total capacity, with Naver strategically planning expansions in six stages.
The company aims to gauge the pace of technology development and data increase before initiating subsequent phases, with plans for the second phase slated to commence in 2025 and the third in 2016.
The GAK Sejong site, known as the “Cloud Ring,” has been in planning since 2019 and was officially announced in 2020. Construction began in 2021, with the completion initially targeted for late 2022.
Future developments are predicted for the Sejong site, with three phases planned as the company aims for gradual, strategic expansions.
Naver Cloud’s CEO, Kim Yu-won, highlighted the site’s current utilization, representing only one-sixth of the total area, with a future-oriented design accommodating scalability to align with technological advancements and anticipated data growth.
The Sejong site incorporates a resilient system to ensure service continuity in the face of potential disasters, featuring an earthquake-resistant design capable of withstanding seismic events up to magnitude 9.0.
The company further plans to advance data center technology, particularly for generative AI requiring substantial data processing. For this, Naver is engaging in discussions with Samsung Electronics, a leading global memory chipmaker. The focus of these talks centers on the joint development of a neural processing unit (NPU), an AI-specific chip designed to mitigate GPU costs.
The collaboration aims to enhance the efficiency and performance of data centers, integral facilities for robust generative AI applications that demand processing large datasets.
Naver Cloud, currently operating four cloud regions in South Korea and additional regions globally, is actively expanding its offerings. The company has introduced a container module with eight racks, providing flexible densities ranging from 5kW to 11kW. With the recent launch of its LLM (Large Language Model) called HyperClova X.
Naver Cloud CEO Kim emphasized the company’s position as the major provider of GPU optimization for high-computing tasks in Korea, especially in large-scale AI applications.
Additionally, Naver is contemplating the establishment of overseas data centers, leveraging its recent deal in Saudi Arabia for constructing a cloud-based digital twin platform as part of the kingdom’s smart city initiative.
- Naver Corporation Shifts to Intel amid Nvidia GPU Prices and Supply Delays
- Naver Cloud Finance Day 2023: Showcasing Financial AI Innovation Cases Using HyperclovaX
- Kakao’s First Self-Owned Data Center: Aims for Disaster-Resilient Operations
- KT to Establish Overseas Internet Data Centers and Facilities with LH
- Naver Seals $100M Deal for Saudi Digital Twin Platform Project