SK Telecom is joining hands with Nvidia to target the global AI infrastructure market. It plans to combine its network and data center operations capabilities with Nvidia's AI Semiconductor and software platform to leap forward as Asia's largest AI cloud provider.
SK Telecom said it will pursue a "full-stack AI cloud" collaboration based on the Nvidia DSX platform, covering everything from chips to data center operations. Nvidia DSX is a platform that defines how to design, build, and optimize AI data centers, spanning semiconductors and systems, infrastructure software, facilities, and partner technologies.
The two companies plan to gradually expand AI Factories, data centers specialized for AI workloads, into gigawatt-level infrastructure. An AI Factory is a next-generation data center concept that inputs power and data to produce tokens, the core unit of AI. While traditional data centers have focused on general-purpose computing and data storage, AI Factories are designed with an architecture optimized for AI computation, including training, inference, and data processing.
The first AI Factory is set to begin operations in 2027 in Korea. SK Telecom plans to use it as the first case to validate the operating structure and business model of the two companies' AI cloud, and then expand the infrastructure across Asia. SK Telecom will also join Nvidia's global partner ecosystem program, Nvidia Cloud Partner.
The core of this collaboration is to secure both low token production expense and high power efficiency. SK Telecom will start by providing AI training and inference services with Nvidia's Blackwell GPUs and will sequentially leverage the Vera Rubin platform, scheduled for supply in the second half of this year.
Groupwide cooperation within SK Group will also expand. SK Group Chairman Chey Tae-won and Nvidia CEO Jensen Huang met on the 1st in Taiwan to review the AI infrastructure roadmap and agreed to push for collaboration at the group level. While previous cooperation focused on semiconductors such as SK hynix's HBM, the scope will broaden to joint research on AI Factory design and operations and next-generation computing architectures.
The two companies plan to form a joint council for research to boost GPU and memory performance together. SK Telecom will tap Nvidia's computing and software capabilities, while Nvidia will be able to leverage SK hynix's memory technology and SK Telecom's experience in building and operating AI Factories.
They will also continue cooperation in Physical AI and Robotics. SK Telecom is building large-scale Digital Twin technology based on Nvidia Omniverse and applying it to SK hynix's semiconductor manufacturing process. They are also pushing to advance a robot simulation and training platform based on Nvidia Cosmos and Isaac GROOT.
Chairman Chey said, "Based on our close partnership with Nvidia, we have secured full-stack AI infrastructure competitiveness that spans from chips to data center operations," and added, "By jointly addressing GPU, memory, and energy issues, we will reinvent ourselves as a leading AI cloud provider driving the development of the AI ecosystem across Asia."
CEO Huang said, "Communications networks are evolving into national AI infrastructure," and noted, "Through the Nvidia DSX platform, SK Telecom will be able to build large-scale AI clouds and provide agent AI, enterprise AI, and physical AI to corporations and industry."