Iris Coleman
Jul 27, 2026 01:08
NAVER, NVIDIA, and Brookfield will scale Korea’s AI manufacturing unit from 55 MW to 200 MW by 2028, aiming for gigawatt-class AI infrastructure.
NAVER, NVIDIA, and Brookfield have unveiled an bold plan to increase Korea’s sovereign AI manufacturing unit infrastructure, boosting capability on the GAK Sejong knowledge heart from 55 megawatts (MW) to 200 MW by 2028. Introduced throughout Korea’s AI Summit in San Francisco on July 25, this venture underscores the nation’s drive to place itself as a worldwide AI powerhouse.
The enlargement builds on NAVER’s June 2026 dedication to develop gigawatt-scale AI infrastructure utilizing NVIDIA’s DSX platform. The preliminary 55 MW deployment, slated to go stay in 2027, will now scale to 200 MW inside a yr of operation, with long-term plans to succeed in 1 gigawatt (GW).
To fund this development, NVIDIA is committing $1 billion, whereas Brookfield plans to offer as much as $9 billion below a nonbinding time period sheet. NAVER will safe the remaining financing. This collaboration integrates NAVER’s experience in hyperscale knowledge facilities, NVIDIA’s full-stack AI platform, and Brookfield’s funding capabilities to create a scalable, multi-tenant AI cloud infrastructure able to serving enterprises, industries, and authorities organizations.
Strategic Implications for AI Improvement
The expanded AI manufacturing unit will present essential compute energy for creating superior AI fashions and purposes, together with NAVER’s proprietary HyperCLOVA X basis fashions. NAVER goals to boost its AI providers with NVIDIA’s cutting-edge platforms, such because the NVIDIA Vera Rubin and Blackwell architectures, which optimize token throughput whereas decreasing prices. This infrastructure may also assist open mannequin initiatives like NVIDIA’s Nemotron 3 Extremely.
“AI factories are the infrastructure nations must compete within the intelligence period,” mentioned NVIDIA CEO Jensen Huang. “This partnership will drive Korea’s AI competitiveness by offering the size and sovereignty essential to gasoline innovation.”
NAVER plans to launch an AI agent platform in late 2026, integrating NVIDIA’s Agent Toolkit software program. Moreover, the corporate is creating a Seoul World Mannequin utilizing NVIDIA Cosmos know-how, which might have implications for city planning and autonomous methods.
A Broader Imaginative and prescient for Sovereign AI
NAVER’s imaginative and prescient extends past home infrastructure to incorporate a worldwide AI footprint. Earlier this yr, NAVER Cloud CEO Kim Yu-won outlined plans for a “World AI Manufacturing facility” spanning Asia, the Center East, and Europe. By leveraging partnerships with NVIDIA and Brookfield, NAVER goals to evolve into a number one operator of world AI infrastructure, a transfer that aligns with Korea’s broader digital transformation technique.
Since January 2026, NAVER has operated Korea’s largest AI cluster, that includes 4,000 NVIDIA B200 GPUs, which lower coaching instances for large-scale fashions by over 90%. The expanded GAK Sejong facility will construct on this basis, providing production-scale AI compute to each home and U.S.-based innovators.
Market and Trade Affect
With AI adoption accelerating throughout industries, sovereign infrastructure initiatives like NAVER’s are more and more seen as strategic property. For Korea, this enlargement represents not solely an funding in its technological future but in addition a bid to scale back dependence on overseas AI infrastructure. The collaboration additionally positions Brookfield as a key participant within the world AI infrastructure sector, constructing on its $100 billion portfolio of knowledge facilities, semiconductors, and energy era property.
Merchants and buyers monitoring the AI sector ought to word the size and ambition of this venture. The $10 billion complete funding alerts sturdy confidence in AI’s transformative potential, with NAVER on the forefront of Korea’s efforts to guide on this house.
Because the infrastructure comes on-line beginning in 2027, it’ll doubtless unlock new alternatives for AI startups and established gamers alike, each in Korea and overseas. This might have ripple results throughout industries, from autonomous autos to superior well being care options, as companies leverage the expanded compute capability to construct next-generation AI purposes.
Picture supply: Shutterstock