South Korea Pioneers Underwater Data Centers to Tackle AI Expansion and Cooling Challenges
August 22, 2026
Ulsan, South Korea, is advancing underwater data centers to address land, power, and cooling constraints as AI computing expands nationwide, with a pilot off the coastline about 20 meters offshore using seawater cooling.
If the pilot succeeds, underwater data centers could serve as a blueprint for expanding coastline-based AI infrastructure across Korea.
The initial project budget totals 51.1 billion KRW (about $37 million), with the central government contributing 40 billion KRW and the remainder coming from local sources.
A longer-term plan contemplates investing roughly 500 billion KRW (around $360 million) to power up to 100,000 servers, contingent on early testing outcomes.
The city is partnering with the Korea Institute of Ocean Science and Technology to design a standardized subsea data center model, targeting a timeline through 2030 and commercialization starting in 2031.
SK Telecom joined the initiative in January 2026 to handle GPU-based infrastructure and server operations, while a broader consortium oversees vessel engineering and software management for underwater AI workloads.
The initial phase focuses on underwater facilities about 20 meters off Ulsan’s coast, using ocean water cooling to reduce reliance on conventional cooling systems.
This project reflects a broader push toward innovative cooling solutions as data center demand grows, with a standard model planned by 2030 and commercial development to begin in 2031.
Officials envision a carbon-neutral, replicable underwater computing template that could be deployed at other sites if the pilot proves successful.
South Korea faces rising demand for computing capacity from chipmakers, cloud providers, and researchers, and ocean water cooling offers a resource that minimizes or eliminates mechanical refrigeration needs.
Summary based on 1 source
Get a daily email with more Tech stories
Source

TechRadar Pro • Aug 22, 2026
South Korean city bets on the ocean to solve its AI data center crisis — but this is only the beginning