NVIDIA and NAVER said they will expand AI infrastructure for sovereign AI workloads, beginning with a 55-megawatt project planned for 2027 and aiming to scale toward gigawatt-level AI factories across multiple regions.
NVIDIA and NAVER announced an agreement to expand AI infrastructure for sovereign AI workloads, with an initial 55-megawatt deployment planned as the first step toward larger AI factory projects.
NVIDIA Newsroom said NAVER will use NVIDIA DSX to expand sovereign AI infrastructure, beginning at 55 megawatts and scaling toward gigawatt-scale AI factories intended to support agents, physical AI and inference workloads.
NAVER Corp., in a Korean-language press release, described the agreement as a partnership to build gigawatt-class global AI factories. NAVER said the first 55MW operation is expected to begin in 2027 and that the companies plan to expand capacity to at least four times the scale of GAK Sejong, NAVER’s large data center in South Korea.
NAVER framed the project as an AI infrastructure alliance spanning Asia, the Middle East and Europe. NVIDIA’s announcement similarly positioned the effort around rising global demand for AI computing and the need for infrastructure that can support local and regional AI requirements.
Both companies emphasized sovereign AI, a term used in the industry to describe AI infrastructure and systems designed to support local control over data, computing resources and AI services. NVIDIA said NAVER’s expansion is intended to serve increasing demand for AI infrastructure across areas including AI agents, physical AI and inference.
The companies did not provide a full site-by-site buildout schedule in the excerpts provided. NAVER’s release, however, stated that the first 55MW operation is targeted for 2027 and that the broader plan aims for gigawatt-class scale.
The use of “AI factories” reflects NVIDIA’s preferred term for large-scale data center infrastructure optimized for AI model training, deployment and inference. In this case, NVIDIA said the planned infrastructure will use its DSX platform.
Investing.com reported that NAVER shares rose after the company announced the NVIDIA partnership. The outlet said investors reacted to plans for an initial 55MW project and a longer-term path toward gigawatt-scale infrastructure using NVIDIA’s DSX platform.
The market reaction report adds financial context, but the core details of the infrastructure plan come from NVIDIA Newsroom and NAVER Corp.’s press release. Those company announcements are the primary sources for the stated capacity targets, timeline and strategic framing.
The announcement shows how major internet platforms and AI hardware suppliers are moving from individual data center expansions toward larger regional infrastructure programs. NAVER, best known for its search, commerce, cloud and consumer internet businesses in South Korea, is positioning the NVIDIA partnership as part of a broader global AI infrastructure strategy.
For NVIDIA, the agreement reinforces its push to supply AI computing systems for sovereign and regional AI deployments. For NAVER, the planned 55MW start and eventual gigawatt-class ambition signal a significant long-term infrastructure commitment, though the companies have not yet disclosed all deployment locations, capital costs or detailed capacity milestones beyond the 2027 first operation described by NAVER.
The practical impact will depend on execution: whether the companies can secure power, sites, equipment and customers at the scale described in their announcements. For now, the sourced facts show a substantial planned expansion, anchored by a 55MW first phase and a stated goal of scaling much further.
NVIDIA and NAVER announced an agreement to expand AI infrastructure for sovereign AI workloads, with an initial 55 megawatt deployment planned as the first step toward larger AI factory projects.
NAVER Corp., in a Korean language press release, described the agreement as a partnership to build gigawatt class global AI factories.
NAVER said the first 55MW operation is expected to begin in 2027 and that the companies plan to expand capacity to at least four times the scale of GAK Sejong, NAVER’s large data center in South Korea.
Continue reading
