An investment group led by Brian Koo, grandson of the LG Group founder, and Amin Badr El-Din, founder and CEO of BADR Investments, is spearheading a project to construct the world's largest AI data center in South Korea, as reported by the Wall Street Journal. This ambitious undertaking, initiated by the investment group 'Stock Farm Road,' is projected to cost approximately $35 billion (around 50 trillion KRW) with an initial investment of$10 billion, and is targeted for completion in 2028 after breaking ground in early 2025.

The proposed AI data center is designed for an unprecedented power consumption of 3 gigawatts (GW). This scale significantly surpasses any existing data center globally, with the WSJ noting that no current facility worldwide aggregates AI semiconductors requiring more than 1 GW. To contextualize, 1 GW is equivalent to the electricity consumption of 750,000 to 1 million households. If realized as planned, the South Korean facility would be nearly three times the size of the Texas data center complex being developed under the U.S. Stargate project, which involves major players like OpenAI, Oracle, and SoftBank.

The core methodology for enabling such an immense facility directly addresses the primary challenge in hyper-scale AI data center development: massive power consumption. Recognizing that the lack of sufficient and stable power supply has been the main impediment to building ultra-large AI data centers, this project strategically targets Jeollanam-do (South Jeolla Province) in South Korea, a location far south of Seoul. This province has committed to supporting access to critical resources, including energy and water. This contrasts with the increasing scarcity of suitable sites in the U.S. that offer cheap land, ample power, and robust data connectivity, exacerbated by an aging power grid. Brian Koo emphasized South Korea's inherent suitability for hosting large-scale facilities capable of serving global demand. The project's success, particularly its timely and cost-effective completion, is viewed as a significant opportunity for the global AI infrastructure landscape.