Planned

SK Telecom — Ulsan AI Mega Campus

Ulsan, South Korea

Capacity
1000MW
Operator
SK Telecom

SK Telecom’s Ulsan AI Mega Campus is a flagship 1 GW hyperscale project designed to anchor South Korea’s sovereign AI computing ambitions. Announced in early 2026, the facility is one of the most ambitious single-site AI data center announcements in Asia — if completed at full capacity, it would be among the five largest AI data center campuses globally.

Ulsan was selected for several strategic reasons: the city has substantial industrial power infrastructure (it is home to Hyundai Motor’s primary assembly plants and HD Hyundai’s shipyards, requiring heavy industrial electrical supply), available industrial land, and a position on South Korea’s eastern coast with good connectivity. The SK-AWS collaboration at the adjacent Ulsan site (tracked separately) demonstrates the city’s emergence as South Korea’s AI infrastructure hub outside Seoul.

SK Telecom is positioning the Ulsan campus as a GPU-as-a-Service (GPUaaS) offering — selling compute time to Korean enterprises, government research programs, and eventually international customers seeking South Korean AI compute capacity. The facility targets South Korea’s domestic LLM developers (including KT, Kakao, NAVER, and SK’s own T3K AI unit), semiconductor firms (Samsung, SK Hynix), defense AI programs, and industrial automation developers in the automotive and shipbuilding sectors concentrated in the Ulsan-Busan corridor.

South Korea’s government has explicitly targeted AI infrastructure buildout as a national priority, with the Ministry of Science and ICT coordinating investment through the “AI Computing Infrastructure Strategy.” The Ulsan campus is a centerpiece of that strategy. Renewable energy for the campus is planned through SK’s partnerships with offshore wind developers in the Yellow Sea.

Reference Metadata

Country
South Korea
Region
Ulsan
City
Ulsan
Status
Planned
Machine-readable record
JSON GeoJSON

Browse This Facility Through The Index

Entry pages are the atomic records in the index. Use the JSON endpoint for programmatic access, then move through the taxonomy links to the broader operator, country, energy, and timeline views that contextualize this facility.

Related Facilities

Nearby slices in the index that share the same operator, geography, build phase, or AI profile. Use these to move laterally through the buildout instead of jumping back to the full directory.

Follow the buildout feed