NVIDIA and Korean partners outline major AI factory buildout plans
NVIDIA, NAVER, Brookfield and SK Group detailed Korea AI factory plans spanning data centers, GPUs and HBM memory.
NVIDIA used a South Korea-focused AI summit in San Francisco to outline a much larger infrastructure push with Korean technology and investment partners, including planned AI factory capacity tied to NAVER, SK Group, SK Telecom, SK hynix and Hyundai Motor Group.
The clearest near-term project is a proposed expansion of NAVER's NVIDIA DSX AI factory at the GAK Sejong data center. NVIDIA says NAVER, Brookfield and NVIDIA plan to grow the initial 55-megawatt deployment to 200 megawatts by 2028, with NAVER describing a longer path toward gigawatt-scale sovereign AI infrastructure. The companies say the expanded site is intended to support Korean and U.S.-based AI builders with production-scale compute for models, agents and AI services.
Why the announcement matters
The announcement is another sign that AI competition is moving from model releases alone to national-scale compute, power and memory supply. NVIDIA said the expanded NAVER deployment is expected to use Vera Rubin and Blackwell platforms, while the DSX stack is positioned as a full system for operating AI factories across chips, systems, software, facilities and partners.
Financing remains conditional. NVIDIA said it plans to invest $1 billion in NAVER, while Brookfield has entered a nonbinding term sheet to fund up to $9 billion. NVIDIA's investment depends on customary closing conditions and NAVER finalizing at least $9 billion of committed financing separate from NVIDIA's planned investment. That means the headline capacity is a roadmap, not completed infrastructure.
A second NVIDIA newsroom announcement broadened the picture: SK Group and NVIDIA said they are expanding a partnership worth more than $500 billion across AI factories and next-generation memory. The plan includes an SK Telecom NVIDIA Vera Rubin DSX AI factory of up to 2 gigawatts and a longer-term NVIDIA-SK hynix memory partnership covering HBM and future AI memory technologies.
Broader AI stack
NVIDIA's blog also tied the infrastructure announcements to research and industrial AI work. The company said it and KAIST launched a joint AI research lab focused on agentic AI in Korea, and that NVIDIA and Seoul National University are planning an NVIDIA AI Technology Center for AI research, education and talent development.
Hyundai Motor Group's role points to physical AI as another target. NVIDIA said Hyundai outlined work on a robot reference platform and integration of NVIDIA DRIVE Hyperion with vehicle platforms, extending a 50,000 Blackwell GPU buildout. Together, the announcements show Korea's AI strategy being framed as a full-stack effort: data centers, memory, robotics, automotive systems, cloud services and research capacity.
Sources
Cover photo by Tima Miroshnichenko on Pexels, used under the Pexels License.
Comments (0)
Leave a Comment