NVIDIA to Deploy 260,000 GPUs in South Korea

Gyeongju, South Korea — NVIDIA Corporation plans to deploy 260,000 graphics processing units in South Korea through partnerships with four domestic companies.

Jensen Huang, NVIDIA’s chief executive, announced the figure at APEC CEO Summit 2025 Korea in Gyeongju.

The planned deployments include arrangements with Naver Corporation, Samsung Electronics, SK Group, and Hyundai Motor, according to Huang.

Naver is expected to add 60,000 GPUs. Samsung Electronics plans to install 50,000 GPUs at its facilities, which the company refers to as AI factories. SK Group will also deploy 50,000 GPUs. Hyundai Motor is expected to utilize its existing infrastructure for autonomous vehicle development and factory operations.

The partnerships are expected to support digital manufacturing systems and robotics applications. NVIDIA already collaborates with Samsung on high-bandwidth memory technology and chip manufacturing.

NVIDIA’s Grace Blackwell system weighs two tons, consists of 1.5 million parts, and consumes 120 kilowatts of power, according to Huang.

Huang said during the speech that AI applications became profitable in 2025.

He said this trend drives infrastructure investment and attributed growth to advances in reasoning and problem-solving capabilities.

NVIDIA has operated in South Korea for 30 years, according to Huang. The company introduced its GeForce product line to the country 25 years ago.

NVIDIA collaborated with Naver on a Korean language model, which Huang described as the world’s third-largest.

NVIDIA also announced plans to collaborate with Korean research institutions, including the Korea Advanced Institute of Science and Technology, as well as government entities and startups.

Samsung manufactures memory components for NVIDIA’s AI systems and produces chips for the company. SK Group supplies high-bandwidth memory, according to Huang.

Software capabilities, scientific expertise, and manufacturing capacity are relevant to South Korea’s development of artificial intelligence, Huang said.

He identified robotics and industrial automation as areas of application.

The GPU deployments are expected to be used for language processing, autonomous vehicle systems, industrial operations, and manufacturing applications, the company said.