Korean Builders Chase AI Data Centre Work As Power Demand Doubles
The Korea Herald reported that Korean contractors are moving deeper into AI data centre construction as JLL expects global data centre power demand to rise from about 103 GW in 2025 to 200 GW by 2030.
Korea’s largest builders are turning AI data centres into a construction pipeline that looks less like conventional property development and more like energy, grid and cooling infrastructure, The Korea Herald reported.
The shift is being pulled by two forces at once: shorter project timelines than housing and contract sizes that can start in the hundreds of billions of won and rise into multitrillion-won territory.
Generative AI demand has made capacity, power procurement and technical delivery central to the business case for contractors that previously depended more heavily on apartments and civil works.
200 GW Power Demand Frames The Market
JLL projects average annual growth of 14 percent for the data centre market through 2030.
Worldwide facility power use is projected to rise from about 103 gigawatts in 2025 to 200 GW by 2030, while the buildout could require as much as $3 trillion in total spending.
Construction costs are rising with that demand.
A megawatt of data centre capacity averaged $7.7 million to build in 2020 and $10.7 million in 2025, a 38 percent increase over the five-year period; the figure is expected to reach $11.3 million per MW this year.
Those figures give Korean contractors a higher-value project class at a time when traditional housing cycles remain exposed to rates, land costs and local demand.
Korean Contractors Move Beyond Build-Only Roles
Hyundai E&C already has four domestic data centre completions on its record.
Its current pipeline includes work on a Saemangeum AI data centre in North Jeolla Province, and June marked the construction start for the 40-MW Ansan Sihwa AI Data Centre.
Daewoo E&C created a dedicated data centre task force in April and signed a July partnership with operator KINX to connect development, construction and operations.
Last year, the company delivered the Empyrion Digital AI Campus in Gangnam at 40 MW.
Its next pipeline includes an investor-constructor role at the Pine Data Centre in Jangseong, which can reach 60 MW, and separate 200-300 MW pursuits in Jangseong and Gangjin.
Samsung C&T is chasing orders from global technology companies while looking toward the Middle East and Australia.
Its 40-MW Ansan Global Cloud Center order was worth 400 billion won, or about $283 million, and a 2.4-MW Chuncheon project uses modular methods intended to shorten construction time and make later expansion easier.
GS E&C is attached to one of the largest domestic plans: an AI data centre campus in Gangwon Province and Donghae with total planned capacity of 2.4 GW.
The first 1.2-GW phase is estimated at around 10 trillion won, with groundbreaking targeted for early October.
Grid And Cooling Work Define The Next Margin
Other builders are adding capacity to the same pipeline.
DL E&C built a 20-MW AI data centre in Seoul and now has an 80-MW Gimpo project under development, while DL Construction won a 126.8-billion-won contract for a 9.8-MW AI computing-focused facility in Bucheon.
SK Ecoplant has a role in the Ulsan SK-Amazon Web Services project, which is sized at 103 MW, and recently created its own AI data centre team.
A construction industry official said the attraction is not only the initial order value.
Builders that expand into land development, equity investment and operations can keep earning rental and operating income after construction ends.
Korea’s policy backdrop adds scale.
The government has laid out support for 8.4 GW of AI data centres through 550 trillion won of public-private investment by 2029.
A Korea Research Institute for Construction Policy assessment treats AI data centres as strategic infrastructure assets, where contractors must coordinate grid connections, substations, backup power, cooling, telecommunications links and approvals.
Lee Gyu-eun, an associate research fellow at the institute, said future competitiveness will depend on shorter power procurement lead times, modular prefabrication, parallel construction and early commissioning.
For contractors, the strongest proof will come when those capabilities turn planned gigawatts into operating campuses rather than construction backlogs.
