NAVER, NVIDIA and Brookfield announced plans on July 24 to expand Korea AI infrastructure at the GAK Sejong data center in South Korea, increasing the proposed AI factory from 55 megawatts to 200 megawatts by 2028. The project is intended to provide production-scale computing capacity for AI developers, businesses and public-sector organizations in Korea and the United States.
The expansion was announced during South Korean President Jae Myung Lee’s AI Summit visit to San Francisco. It combines NAVER’s data center operations, NVIDIA’s computing platform and planned investments from NVIDIA and Brookfield, subject to financing commitments and other closing conditions.
Korea AI Infrastructure Expansion
NAVER, NVIDIA and Brookfield said the proposed expansion would more than triple the initial 55-megawatt NVIDIA DSX deployment. The companies are targeting 200 megawatts of capacity at NAVER’s GAK Sejong hyperscale data center by 2028, creating a larger computing resource for commercial and government AI workloads.
The July announcement builds on a June agreement between NAVER and NVIDIA covering an initial 55-megawatt deployment. According to NAVER, operations are expected to begin during the first half of 2027 before capacity increases to 100 megawatts later that year and 200 megawatts in 2028.
Additionally, NAVER intends to pursue a longer-term expansion toward one gigawatt of NVIDIA-powered infrastructure. The immediate plan is therefore a phased increase in computing capacity rather than completion of a gigawatt-scale facility.
Capacity and Development Timeline
| Indicator | Development Stage | Context |
|---|---|---|
| Initial AI factory capacity | 55 megawatts | NAVER said the first deployment is scheduled to begin operating during the first half of 2027. |
| Intermediate capacity | 100 megawatts | NAVER’s June development roadmap calls for capacity to reach 100 megawatts during 2027. |
| Planned expanded capacity | 200 megawatts | NVIDIA and NAVER said the GAK Sejong deployment is expected to reach 200 megawatts by 2028. |
| Long-term objective | Up to one gigawatt | NVIDIA’s July 24 release said NAVER intends to continue expanding toward gigawatt-scale sovereign AI infrastructure. |
Investment and Financing Plans
NVIDIA plans to invest $1 billion in NAVER, while Brookfield has entered into a nonbinding term sheet to provide up to $9 billion. NAVER would fund the remaining amounts required for the infrastructure project, according to the joint NVIDIA announcement.
However, the planned financing has not been presented as fully completed. NVIDIA said its investment remains subject to customary closing conditions and NAVER securing at least $9 billion in committed project financing, separate from NVIDIA’s proposed contribution.
NAVER described the announced development as a $10 billion project, although the final capital structure and any additional NAVER contribution remain dependent on the financing arrangements.
Planned Funding Structure
| Indicator | Planned Amount | Context |
|---|---|---|
| NVIDIA investment | $1 billion planned | NVIDIA said the investment in NAVER remains subject to closing conditions and completion of the required financing arrangements. |
| Brookfield funding | Up to $9 billion | NVIDIA described Brookfield’s arrangement as a nonbinding term sheet, while NAVER identified Brookfield as the project’s capital partner. |
| NAVER contribution | Remaining project costs | The joint announcement said NAVER would finance amounts not covered by the planned NVIDIA and Brookfield commitments. |
| Financing condition | At least $9 billion committed | NVIDIA said NAVER must finalize at least $9 billion of committed financing before NVIDIA’s planned investment can close. |
GAK Sejong AI Factory Development
The expanded infrastructure is planned for NAVER’s GAK Sejong hyperscale data center in Sejong, South Korea. NVIDIA said the facility would use its DSX platform to combine computing hardware, systems, software, facilities and partner technologies within a multi-tenant AI factory.
The proposed 200-megawatt installation is expected to include NVIDIA Vera Rubin and Blackwell platforms. These systems would provide computing and software resources for organizations developing AI models, autonomous agents and AI-enabled services.
Meanwhile, NAVER would operate the infrastructure as an NVIDIA Cloud Partner. The practical effect would be to make large-scale computing capacity available to multiple customers rather than reserving the facility exclusively for NAVER’s internal products.
Technology and Operating Model
NVIDIA said its DSX platform is designed as an end-to-end system for constructing and operating AI factories. The platform includes DSX MaxLPS software for increasing token throughput per megawatt and DSX OS for lifecycle management, system health, resilience and multi-tenant operations.
According to the companies, the infrastructure will serve enterprises, industrial users, government organizations and AI cloud customers. The 200-megawatt target provides a measurable capacity objective, while its real-world purpose is to support the training and deployment of larger AI applications across Korea and the United States.
Support for AI Models and Applications
NAVER and NVIDIA are also extending their cooperation beyond data center capacity into open-model development. NAVER is adapting its HyperCLOVA X models using NVIDIA Nemotron 3 Ultra open models alongside NAVER’s proprietary data and training methods.
NAVER has become the first Korean company to join the NVIDIA Nemotron Coalition. The company said its participation covers pretraining, post-training and reinforcement learning, allowing its researchers to contribute to shared model-development work while continuing to develop NAVER-specific services.
Additionally, NAVER plans to launch an AI agent platform in Korea during the second half of 2026. The platform is expected to use NVIDIA Agent Toolkit software, while a separate Seoul World Model project will combine NAVER’s urban street-view and spatial data with NVIDIA Cosmos world foundation models.
NAVER and NVIDIA Collaboration
The model-development program connects the infrastructure investment with specific services. HyperCLOVA X provides the model component, the planned agent platform provides a commercial application layer, and the Seoul World Model extends the work into spatial and physical AI.
The companies have not announced performance results for these planned services. Nevertheless, the projects identify how the expanded computing capacity could be used in practice, including model training, AI agents and digital representations of urban environments.
Strategic Importance for Korea
The companies describe the project as sovereign AI infrastructure because the planned computing capacity would be operated in South Korea and made available to domestic institutions and industries. The 200-megawatt expansion would create a larger national resource while also making capacity available to customers based in the United States.
Brookfield said it manages approximately $100 billion in assets across the AI infrastructure value chain, including data centers, semiconductor manufacturing, computing and dedicated power generation. It also manages approximately $12 billion in Korean infrastructure, real estate and energy assets after establishing operations in the country in 2014.
Therefore, the partnership combines an existing Korean data center operator, a computing technology provider and a large infrastructure investor. The announced capacity and financing commitments give the project a defined scale, while delivery remains subject to funding, construction and technical execution.
Stakeholder Comments
Haejin Lee, NAVER Founder and Chairman, said:
“NVIDIA’s planned strategic investment and our infrastructure supply agreement with Brookfield have propelled NAVER’s vision for the AI factory business into a robust execution phase. Leveraging the solid partnerships with our global partners, we will drive technological innovation, foster a sovereign AI ecosystem and spearhead efforts to strengthen South Korea’s AI competitiveness.”
Sikander Rashid, Brookfield Global Head of AI Infrastructure, said:
“The partnership will combine Brookfield’s global AI infrastructure investment capabilities, NAVER’s full-stack AI and data center operating expertise, and NVIDIA’s accelerated computing platform to advance Korea’s AI capabilities. As AI adoption accelerates across the global economy, access to trusted, sovereign and scalable AI infrastructure is becoming an increasingly important strategic priority for companies and countries.”
NAVER, NVIDIA and Brookfield have outlined a phased expansion of Korea AI infrastructure from 55 megawatts to 200 megawatts by 2028, supported by a planned $1 billion NVIDIA investment and a nonbinding Brookfield financing proposal of up to $9 billion. The GAK Sejong project would provide multi-tenant computing capacity for AI models, agents and services in Korea and the United States.
The announced scale is substantial, but the financing and NVIDIA investment remain subject to stated conditions. Progress will depend on NAVER completing the required funding arrangements and delivering the planned data center expansion.
Sources: NVIDIA Newsroom, NAVER Corporation, NAVER June Infrastructure Announcement.
Prepared by Ivan Alexander Golden, Founder of THX News, an independent news organization delivering timely insights from global official sources.
Research combines AI-assisted analysis with human-edited accuracy and context.
