Nvidia reaffirms 260,000 GPU supply plan for South Korea
The hardware giant confirmed it will stick to its massive infrastructure commitment to South Korea, supporting national sovereign AI and industrial projects.
The short version
- Nvidia confirmed the 260,000 GPU supply plan remains on track despite recent leadership changes at the company.
- The government aims to secure 35,000 GPUs directly, with 21,000 already secured and 14,000 expected by early 2027.
- Major companies like Samsung Electronics, SK Group, and Hyundai Motor Group will each build infrastructure using over 50,000 GPUs.
- The plan includes technical cooperation to connect Korean-designed AI accelerators with Nvidia infrastructure via NVLink.
Why it matters
This deal solidifies South Korea's position as a major hub for sovereign AI, ensuring the country has the hardware needed to build its own foundation models. It also integrates South Korea's largest industrial giants into a massive, unified AI infrastructure roadmap.
Confirming the hardware roadmap
Nvidia has officially reaffirmed its commitment to supply more than 260,000 graphics processing units (GPUs) to South Korea. This confirmation comes after a meeting between the company and the South Korean Science and ICT Minister, Bae Kyung-hoon, on November 13. The meeting served to review the supply program and ensure the project remains on track following recent leadership changes at Nvidia.
The commitment covers a broad spectrum of the country's technological needs, spanning both public-sector procurement and private deployments. By maintaining this plan, Nvidia is supporting a massive national infrastructure program designed to bolster South Korea's artificial intelligence capabilities. The agreement ensures that the hardware will be available for a variety of projects, from research to large-scale industrial use.
The broader supply plan was originally first announced in October 2025. By reaffirming these numbers now, Nvidia provides a stable timeline for the South Korean government and its major corporations. This stability is essential for long-term planning in a sector where hardware availability can often be unpredictable or subject to rapid shifts in demand.
Public and private allocations
The distribution of these GPUs is divided between government initiatives and private corporate projects. The South Korean government intends to obtain roughly 35,000 GPUs directly from Nvidia, measured in B200 equivalents. As of the recent update, the ministry has already secured about 21,000 of those units, with the remaining 14,000 expected to arrive by early 2027.
In addition to the direct government procurement, South Korea plans to obtain a quantity of 15,000 additional units via the National AI Computing Center. This brings the total government allocation to more than 50,000 units. These resources are specifically earmarked to support sovereign AI projects, cloud providers, research institutes, startups, and universities.
On the private side, the plan involves several of the country's largest technology and industrial groups. Samsung Electronics, SK Group, and Hyundai Motor Group are each slated to build AI infrastructure using more than 50,000 GPUs. Furthermore, NAVER Cloud has planned to deploy more than 60,000 units to support its own expanding digital services.
Industrial and sovereign AI goals
The hardware is not just intended for general computing but is tied to specific industrial applications. The private deployments are designed to support semiconductor manufacturing, automotive systems, robotics, telecommunications, and enterprise cloud services. By linking the hardware program to these specific sectors, the plan aims to modernize the core industries that drive the South Korean economy.
For the public sector, the focus is on 'sovereign AI,' which refers to a nation's ability to develop and maintain its own independent AI models and infrastructure. The government-procured capacity will support teams working on South Korea’s sovereign foundation-model and 'AI-for-all' initiatives. This ensures that the country's data and AI capabilities remain under national oversight rather than relying solely on external providers.
Nvidia characterized the intended rollout as a combination of sovereign cloud capacity and AI factories run by individual companies. This dual approach allows for both broad public access to computing power and specialized, high-intensity environments for private companies. It creates a comprehensive ecosystem where research and industrial application can happen simultaneously across the country.
Technical cooperation and integration
Beyond the physical delivery of hardware, the two sides intend to continue deep technical cooperation. This includes a focus on connecting Korean-designed AI accelerators with Nvidia's infrastructure through NVLink. NVLink is a high-speed interconnect technology that allows multiple GPUs to work together as a single, unified system for massive workloads.
The cooperation also extends to developing broader AI computing capacity and conducting research into physical AI. Physical AI refers to AI systems that can interact with and move through the physical world, such as in robotics or automated manufacturing. This research is a key component of the long-term strategy to integrate AI into the physical infrastructure of the nation.
To ensure the success of these technologies, the partnership also includes plans for training specialized workers. The goal is to build a workforce capable of managing and developing these complex systems. The ministry will continue to monitor the supply and demand of GPUs across both the public and private sectors to ensure the rollout remains balanced and effective.
Future milestones and expansion
The reaffirmation of the supply plan comes at a time when South Korea is actively expanding several Nvidia-linked infrastructure programs. One such expansion involves a partnership between Nvidia, SK Telecom, and the Korean chip developer Rebellions. This project aims to link neural processing units made in Korea with Nvidia's rack-scale infrastructure for sovereign AI systems targeted for 2028.
These ongoing projects demonstrate a trend toward integrating domestic hardware designs with global standards. By using Nvidia's infrastructure as a foundation, Korean developers can scale their own specialized chips more quickly. This synergy is intended to create a robust domestic AI ecosystem that can compete on a global scale while maintaining local technological interests.
The government's role in this expansion is to act as a monitor and facilitator for these initiatives. By overseeing the distribution of GPU capacity to universities and research institutions, the ministry ensures that the technology reaches the innovators who will build the next generation of applications. The roadmap established in 2025 provides a clear trajectory for South Korea's AI ambitions over the next several years.
What to watch
- The arrival of the remaining 14,000 government GPUs in early 2027.
- Progress on connecting Korean-designed neural processing units with Nvidia's NVLink infrastructure.
- How the 'AI-for-all' initiative translates into specific public services for citizens.
Try it
Look up the 'National AI Computing Center' to see how South Korea plans to distribute its public AI computing resources.
Sources and links
- ReportingTNGlobal: the article we sourced
Liked it? Say why
Add a line for other readers.
No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.
Listen to the AI Wire headlines
Listen to the AI WireListen to our daily AI shows every morning: on your commute, in the shower, wherever you are
Listen to the daily showsFirst reported by TNGlobal · Read the original. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.
