After Taiwan, NVIDIA expands Korea AI partnerships

NVIDIA announced Korea partnerships with SK Hynix, SK Telecom, SK Group, and NAVER across memory, AI cloud, chip design, and sovereign AI.

Image:
Jensen Huang addressing a reception that brought together roughly 200 partners from every part of the Korea AI ecosystem. (Image from NVIDIA)

NVIDIA announced a series of partnerships with South Korean technology companies during Jensen Huang's visit to the country, covering advanced memory, AI cloud infrastructure, semiconductor design and manufacturing, sovereign AI, physical AI, and NVIDIA computing platforms.

The announcements involve SK hynix, SK Telecom, SK Group, and NAVER. The announcements focus on AI factory infrastructure for training, post-training, inference, and AI service delivery.

For Huang Korea is a “world-class leader in heavy industries” — and now in AI.

“Now we’re sitting in a country where you are world-class at manufacturing, world-class at electronics, world-class at software — and you are now world-class at AI,” he said at one of the ecosystem gatherings in Korea.

Here’s a look at some of the big partnerships announced during his visit to Korea.

SK Hynix to work with NVIDIA on memory and fab simulation

NVIDIA and SK Hynix signed a multiyear technology partnership to develop advanced memory for AI computing systems. The agreement builds on earlier co-engineering work between the companies on NVIDIA AI platforms.

Under the partnership, SK Hynix will work with NVIDIA on memory for NVIDIA Vera Rubin AI supercomputers, NVIDIA Vera CPUs, RTX Spark-powered PCs, and Jetson Thor robotic computing platforms. NVIDIA said the agreement also supports supply planning for advanced memory, which has long development cycles.

SK Hynix is using NVIDIA CUDA-X libraries and AI tools to speed up semiconductor simulation. These workflows include technology computer-aided design, computational lithography, in-house simulation codes, and AI physics work using NVIDIA PhysicsNeMo.

The work could involve chipmakers, NVIDIA, and electronic design automation software vendors. It focuses on semiconductor simulation and design workflows.

SK Hynix is also developing digital twins for semiconductor fabs. The company is using NVIDIA Omniverse libraries and OpenUSD pipelines to build 3D factory scenes for visualizing, simulating, and optimizing manufacturing environments.

The digital twin work also involves NVIDIA cuOpt, an open-source GPU-accelerated decision optimization engine, and the NVIDIA Metropolis platform. These tools can support planning for autonomous mobile robots and other fab assets.

SK Telecom plans AI Cloud using NVIDIA DSX

NVIDIA and SK Telecom announced a separate plan for SK Telecom to build a gigawatt-scale AI Cloud in Korea using the NVIDIA DSX platform. The first AI factory under the project is expected to come online in 2027.

SK Telecom said the AI Cloud will support training, inference, and agentic workloads, including sovereign, physical, and enterprise AI services for companies and industries in Korea. The company also described a plan to expand into other parts of Asia.

As part of the collaboration, SK Telecom will become an NVIDIA Cloud Partner. The program includes NVIDIA infrastructure, software, and developer tools for AI cloud services.

The SK Telecom project follows earlier work involving NVIDIA tools in Korea. At GTC Taipei, SK Telecom described its work applying digital twins to SK hynix semiconductor fabs using NVIDIA Omniverse libraries.

In April, SK Telecom also said it had adopted open-source NVIDIA Nemotron datasets to train its A.X K1 model. The model is part of the Korean government's Sovereign AI Foundation Model Project.

NVIDIA and SK Group said they plan to pursue joint research on next-generation AI factory architecture. The work will cover accelerated computing, memory technologies, data center operations, and full-stack AI factory optimization across future infrastructure projects.

NAVER also announced plans to expand its sovereign AI infrastructure using NVIDIA DSX. The project will begin at 55 megawatts, with plans to move toward gigawatt scale.

The DSX-powered cloud will expand NAVER's GAK Sejong data center in Sejong, South Korea. NAVER described the facility as a large-scale site designed for high-density NVIDIA accelerated computing, automation, energy-efficient operations, sustainability features, and disaster-response capabilities.

NAVER said the infrastructure will serve enterprises, industries, government organizations, and AI cloud customers. NAVER also said it supports customers in Europe and the Middle East with services aligned with local regulatory and data-sovereignty requirements.

The infrastructure project builds on NVIDIA and NAVER's model development work. NAVER is advancing HyperCLOVA X by fine-tuning NVIDIA Nemotron 3 Ultra with its own data and training expertise.

NAVER is also the first Korean company to join the NVIDIA Nemotron Coalition. The company said it will contribute to open model development across pretraining, post-training, and reinforcement learning.

NAVER plans to launch an AI Agent Platform in Korea in the second half of the year, powered by NVIDIA NemoClaw blueprints. It is also developing a Seoul World Model using urban street-view data and spatial modeling technology, building on NVIDIA Cosmos world foundation models.