NVIDIA to Supply 260,000 GPUs to South Korea

NVIDIA CEO Jensen Huang on Friday said that his company will supply 260,000 GPUs (graphics processing units), essential for artificial intelligence (AI) development, to the South Korean government and businesses.

NVIDIA plans to supply these GPUs to the Ministry of Science and ICT, Samsung Electronics, SK Group, Hyundai Motor Group, and Naver. Except Naver, which will receive 60,000 units, the others will each get 50,000, according to South Korean media.

This volume exceeds five times the current number of NVIDIA GPUs in South Korea (45,000 units). Considering the latest GPU, the “Blackwell B200,” costs $30,000–$40,000 per unit, the total value is estimated at $7.8 billion–$10.4 billion.

This collaboration is expected to accelerate South Korea’s AI development and the AI ​​transformation (AX) of its manufacturing sector.

Samsung Electronics, SK, and Hyundai Motor plan to build AI factories by integrating AI into every stage of production—from design to manufacturing for semiconductors, automobiles, robots, and autonomous driving—using the secured NVIDIA GPUs.

To catch up with the US and China in AI, South Korea needs to focus on vertical AI capabilities in manufacturing, where it holds strengths, rather than general-purpose chatbots like OpenAI’s ChatGPT or Google’s Gemini and securing GPUs is critical for this.

With NVIDIA dominating 90% of the global AI accelerator market, GPUs are scarce despite premium prices. Korean companies will now receive stable supplies of NVIDIA GPUs in bulk.

Additionally, increased GPU shipments to South Korea will boost domestic semiconductor firms like Samsung Electronics and SK Hynix, which supply HBM (high-bandwidth memory) for GPUs, creating a win-win structure for both NVIDIA and Korean chipmakers.

AI Foundation Models

The 260,000 GPUs prioritised for South Korea will first be used by the Ministry of Science and ICT to secure 50,000 units for developing independent AI foundation models and building a national AI computing center. This will enable universities, research institutions, and start-ups, which previously struggled to secure GPUs, to access AI infrastructure.

Samsung Electronics, a comprehensive semiconductor company with the world’s largest semiconductor manufacturing facilities, plans to introduce over 50,000 NVIDIA GPUs in the coming years, applying AI to all design and production processes.

SK Group is building a manufacturing AI cloud using NVIDIA GPUs and manufacturing AI platforms. Toward this goal, it is progressing with a 100 MW AI data center project in Ulsan, targeting completion by 2027.

Hyundai Motor Group will develop models in vehicle AI, autonomous driving, and robotics through NVIDIA GPU-based AI factories. To this end, it plans to jointly invest approximately $3 billion.

Naver is planning to expand its AI infrastructure, including its own data centers, to develop a Korean-language specialised large language model (LLM) and Naver Cloud signed an MOU with NVIDIA to jointly develop a next-generation ‘physical AI’ platform that organically connects physical and digital spaces. It also plans to build AI infrastructure centered on national key industries like semiconductors, shipbuilding, and energy to support AI adoption in industrial sites.

LG Electronics also plans to collaborate with NVIDIA to innovate next-generation technologies like physical AI and digital twins. The company is developing its own physical AI model based on NVIDIA’s general-purpose humanoid reasoning model, “Isaac GR00T.”