NAVER Corp. (KOSPI: 035420), NVIDIA Corporation (Nasdaq: NVDA) and global investment firm Brookfield Corporation (NYSE: BN) have unveiled plans to significantly expand South Korea’s sovereign artificial intelligence infrastructure, proposing a 200-megawatt AI factory that would more than triple the scale of a previously announced deployment and lay the foundation for a future 1-gigawatt national AI computing network.
The announcement was made during South Korean President Jae Myung Lee’s visit to San Francisco for the AI Summit, underscoring the country’s growing efforts to strengthen domestic AI capabilities amid increasing global competition for computing infrastructure.
The proposed expansion builds on plans announced in June, when NAVER committed to deploying a 55-megawatt AI factory based on NVIDIA’s DSX platform at its GAK Sejong hyperscale data centre in Sejong, South Korea. Under the revised plan, the facility will be expanded to 200 megawatts, while NAVER said it ultimately intends to scale its NVIDIA AI infrastructure deployment to 1 gigawatt.
The project is backed by a multi-billion-dollar financing structure. NVIDIA plans to invest $1 billion in NAVER, while Brookfield has signed a non-binding term sheet to provide up to $9 billion in funding. NAVER will finance the remaining project costs.
NVIDIA said its planned investment remains subject to customary closing conditions, including NAVER securing at least $9 billion in committed financing independent of NVIDIA’s investment.
Once completed, the AI factory is expected to provide production-scale computing resources for AI developers in South Korea and the United States, supporting the development of next-generation foundation models, AI agents and enterprise AI applications.
The announcement reflects a broader global push by governments and technology companies to establish sovereign AI infrastructure—domestically controlled computing resources designed to reduce reliance on overseas cloud providers while supporting national AI strategies.
The expanded facility will be powered by the NVIDIA DSX platform and is expected to incorporate the company’s latest NVIDIA Vera Rubin and NVIDIA Blackwell AI computing architectures. Together, the systems are designed to provide high-performance computing, networking, software and operational tools required to train and deploy large-scale AI models.
As an NVIDIA Cloud Partner, NAVER operates hyperscale AI infrastructure using NVIDIA’s full-stack AI platform and has been expanding its investments in AI services and cloud computing in recent years.
“Our infrastructure supply agreement with Brookfield and NVIDIA’s planned strategic investment have accelerated NAVER’s vision for the AI factory business into a robust execution phase,” said Haejin Lee, founder and chairman of NAVER.
Lee said the company aims to strengthen South Korea’s sovereign AI ecosystem through partnerships with global technology providers while enhancing the country’s long-term competitiveness in artificial intelligence.
Brookfield said the investment aligns with its strategy of expanding AI infrastructure globally as demand for computing capacity continues to rise.
“The partnership combines Brookfield’s global AI infrastructure investment capabilities, NAVER’s full-stack AI expertise and NVIDIA’s accelerated computing platform,” said Sikander Rashid, Global Head of AI Infrastructure at Brookfield.
He added that trusted and scalable AI infrastructure is becoming an increasingly important strategic priority for both governments and enterprises as AI adoption accelerates worldwide.
NVIDIA founder and Chief Executive Officer Jensen Huang described AI factories as critical national infrastructure for the emerging AI economy.
“AI factories are the infrastructure nations need to compete and innovate in the intelligence era,” Huang said, adding that the collaboration would help support South Korea’s startup ecosystem and industrial AI development.
Beyond infrastructure deployment, the companies also outlined plans to deepen collaboration on open AI model development.
NAVER said it is advancing its HyperCLOVA X large language models using NVIDIA Nemotron 3 Ultra open models combined with its proprietary datasets and training techniques. The company also became the first Korean member of the NVIDIA Nemotron Coalition, contributing to open model development across pre-training, post-training and reinforcement learning.
The collaboration also extends to AI applications.
NAVER plans to launch an AI agent platform in South Korea during the second half of the year using NVIDIA Agent Toolkit software, including NVIDIA NemoClaw blueprints. The company is also developing a “Seoul World Model” built on NVIDIA Cosmos world foundation models, using proprietary street-view imagery and spatial data to create AI models capable of understanding real-world urban environments.
The proposed AI factory will also deploy NVIDIA’s latest software stack designed specifically for large-scale AI infrastructure.
The NVIDIA DSX platform integrates AI chips, computing systems, networking, software and facility management into a unified architecture optimised for AI factories. NVIDIA said its DSX MaxLPS software is designed to maximise AI token throughput per megawatt of power consumption, while DSX OS provides lifecycle management, health monitoring, resiliency and multi-tenant operations across large AI deployments.
The project also strengthens Brookfield’s growing presence in AI infrastructure. The investment firm said it manages approximately $100 billion in assets across AI-related sectors, including data centres, semiconductor manufacturing, compute infrastructure and power generation.
Brookfield has operated in South Korea since 2014 and currently manages around $12 billion in infrastructure, real estate and energy assets in the country.
The proposed investment also builds on NVIDIA’s role as a founding partner in Brookfield’s Global AI Infrastructure Program, highlighting closer collaboration between technology providers and infrastructure investors as demand for AI computing accelerates.
Industry analysts view sovereign AI infrastructure as an increasingly important competitive differentiator, with countries investing heavily in domestic compute capacity to support national AI ecosystems, data sovereignty requirements and strategic industries.
If completed, the expanded GAK Sejong deployment would rank among Asia’s larger AI infrastructure projects and further position South Korea as one of the region’s key hubs for advanced AI development, cloud services and next-generation computing.
