NAVER, NVIDIA, and Brookfield are joining forces to dramatically expand the capacity of Korea's AI factory. The current facility runs at 55 megawatts. The partners aim to push that to 200 MW by 2028, with a longer-term goal of building gigawatt-class AI infrastructure.
Why the capacity jump matters
AI training and inference workloads are power-hungry. A 55 MW facility can handle a limited number of high-end GPU clusters. Scaling to 200 MW means roughly four times the compute density. The gigawatt target suggests the consortium is betting on sustained demand from large language models, autonomous driving, and industrial AI.
Who's doing what
NAVER brings cloud and AI platform expertise. NVIDIA supplies the GPU hardware and software stack. Brookfield, a global infrastructure investor, provides the capital and energy management know-how. The three companies didn't disclose financial terms or a specific site for the expansion.
Construction timelines and regulatory approvals haven't been announced. The partners said they'll release more details in the coming months. For now, the 200 MW target is a public milestone that signals Korea's push to become a regional AI hub.
