SK Group will build a "manufacturing AI cloud" using NVIDIA's graphics processing units (GPUs) and its manufacturing AI platform Omniverse to drive artificial intelligence (AI) innovation in Korea's manufacturing ecosystem. SK Group plans to open it not only to manufacturing affiliates such as SK hynix but also to external parties including manufacturing-related public institutions and domestic Start - Up companies, helping Korea's manufacturing ecosystem boost productivity and efficiency on an AI basis.
Chey Tae-won, chairman of SK Group, and Jensen Huang, NVIDIA founder and chief executive officer (CEO), met at the CEO Summit of the 2025 Asia-Pacific Economic Cooperation (APEC) leaders' meeting in Gyeongju, North Gyeongsang, on the 31st and discussed cooperation measures for the "Manufacturing AI Startup Alliance (Physical AI Startup Alliance)," semiconductor collaboration, and the direction of development for the domestic manufacturing AI ecosystem. Manufacturing AI is AI technology that can be applied to physical devices in tangible form such as automobiles and robots, or used in manufacturing plants that produce goods.
Omniverse is NVIDIA's virtual simulation–based Digital Twin (a technology that virtually reflects and builds manufacturing process designs in a digital space to verify production yield, errors, and more in advance of the actual manufacturing process) platform. It allows building and simulating manufacturing production processes in an online three-dimensional (3D) virtual space.
The manufacturing AI cloud will be based on more than 2,000 units of NVIDIA's latest GPUs (RTX Pro 6000 Blackwell Server Edition) adopted by SK hynix, and SK Telecom will handle construction, operation, and service so it can be used at SK hynix's Icheon campus and the Yongin Semiconductor Cluster.
As the only domestic operator of a manufacturing AI cloud, SK plans to create an environment in which users can access Omniverse directly without relying on overseas data centers, ensuring performance and data security optimized for Korea's manufacturing sector.
In addition to supplying GPUs, NVIDIA plans to develop, together with SK, AI models tailored to Korea's manufacturing industry based on Omniverse and to provide technical cooperation in software optimization, AI model training and inference, cloud operations automation, and simulation tuning.
The two companies expect this collaboration to enable domestic manufacturing corporations, which have faced difficulties adopting AI due to high expense and equipment supply issues, to realize manufacturing AI. The two companies also agreed to jointly foster and support AI Start - Up companies in the manufacturing field with venture capital (VC) firms such as IMM Investment, Korea Investment Partners, and SBVA.
Along with this, SK Group will build an AI infrastructure, "AI Factory," with more than 50,000 GPUs in Korea with NVIDIA. The AI Factory is an "AI industry cluster based on NVIDIA GPUs" that includes the manufacturing AI cloud and an AI data center project in Ulsan. SK Group is pursuing a 100-megawatt (MW) AI data center project in Ulsan with a target of 2027.
Chey Tae-won, chairman of SK Group, said, "SK Group is working with NVIDIA to make AI the engine driving innovation across domestic industries. This will allow industries across the board to break through limits of scale, speed and precision," and added, "Based on the NVIDIA AI Factory, SK Group plans to build infrastructure to power next-generation memory, Robotics, Digital Twin, and intelligent AI agents."
Jensen Huang, NVIDIA founder and CEO, said, "In the AI era, a new kind of manufacturing plant called an AI factory has emerged. SK Group is NVIDIA's key memory technology partner, helping NVIDIA develop the state-of-the-art GPU computing platform that is driving AI progress worldwide," and added, "By building AI infrastructure based on NVIDIA's accelerated computing and software, we are very pleased to be jointly creating an AI factory that will fuel SK Group's innovation and invigorate the Korean AI ecosystem."