Facility will feature 50,000 GPUs and advance semiconductor R&D, digital twins and AI agent development.<\/p>\n
NVIDIA announced that it is working with SK Group to build an AI factory to advance semiconductor research, development and production, as well as cloud infrastructure to support digital twin and AI agent development.<\/p>\n
SK Group is building an AI factory featuring more than 50,000 NVIDIA GPUs, with the first phase planned for completion by late 2027. Once complete, the system is expected to be one of Korea\u2019s largest AI factories.<\/p>\n
The new factory will serve SK subsidiaries \u2014 including SK hynix and SK Telecom (SKT) \u2014 as well as external organizations through a GPU-as-a-service model, accelerating digital transformation and industrial innovation for Korea\u2019s industries.<\/p>\n
Further expanding the NVIDIA and SK Group partnership, the companies are collaborating on the development of SK hynix high-bandwidth memory (HBM) and next-generation advanced memory solutions for NVIDIA GPUs, semiconductor manufacturing and telecommunications infrastructure.<\/p>\n
The SK Group AI factory infrastructure will be available to sovereign model developers participating in the Korea government\u2019s Sovereign AI Foundation Models project. SKT is also taking part in the program as a model developer to provide foundation models for enterprises, industries and researchers working on AI agent and application development.<\/p>\n
SK Telecom advances physical AI and robotics with NVIDIA Omniverse<\/p>\n
NVIDIA Cloud Provider SKT plans to build an industrial AI cloud with\u00a0NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs<\/a>\u00a0in Asia, empowering the nation\u2019s manufacturing leaders to advance physical AI and robotics development.<\/p>\n The initial deployment will include over 2,000 NVIDIA RTX PRO 6000 Blackwell GPUs, and will run\u00a0NVIDIA Omniverse<\/a> workloads to support SK hynix\u2019s semiconductor manufacturing, fab digital twins and internal AI agents.<\/p>\n SK Hynix boosts productivity through AI physics, digital twins, robotics and AI agents<\/p>\n SK hynix is using the accelerated cloud infrastructure from SKT to scale its operations.<\/p>\n SK hynix is tapping\u00a0NVIDIA CUDA-X<\/a> technologies, including through the\u00a0NVIDIA PhysicsNeMo<\/a> framework, to accelerate chip design with AI physics. This accelerates technology computer-aided design simulations, enabling SK hynix to deliver next-generation semiconductor products faster and with greater precision. SK hynix is also testing opportunities to accelerate its current use of Synopsys software with NVIDIA Blackwell GPUs for technology computer-aided design and circuit simulation.<\/p>\n To boost efficiency, SK hynix is building autonomous fab digital twins using NVIDIA Omniverse libraries and\u00a0NVIDIA RTX PRO Servers<\/a>\u00a0with NVIDIA RTX PRO 6000 Blackwell GPUs. These digital twins enable real-time simulation, monitoring and optimization of fab operations, resulting in faster production ramp-up, improved operational agility and progress toward fully self-optimizing manufacturing facilities.<\/p>\n SKT is developing a foundation model called A.X. that will power SK hynix\u2019s AI agents, built with\u00a0NVIDIA NIM<\/a> microservices and the\u00a0NVIDIA AI Enterprise<\/a>\u00a0software platform. These agents will help more than 40,000 employees, production staff and office workers collaborate and solve problems faster, accelerating productivity across chip development and fabrication.<\/p>\n Learn more about these collaborations by joining the NVIDIA\u00a0keynote and sessions<\/a>\u00a0at the SK AI Summit, running Nov. 3-4.<\/p>\n