SK Telecom (SKT) is partnering with NVIDIA to establish artificial intelligence (AI) infrastructure in South Korea, marking the beginning of a broader initiative across Asia. The collaboration involves joining NVIDIA's global AI infrastructure ecosystem program, known as NVIDIA Cloud Partner (NCP), and developing an AI factory.
During a briefing at the SK Serin Building in Jongno, Seoul, on June 8, Jensen Huang, CEO of NVIDIA, stated, "SKT and NVIDIA are forming a partnership to build an AI factory in Korea. In the future, AI will be utilized in all industries, just like electricity, water, and the internet."
Huang emphasized that "telecommunication networks will not just serve to transmit bits; AI will permeate networks, providing higher frequency efficiency and greater network capabilities."
The two companies plan to collaborate on a "full-stack AI cloud" based on NVIDIA's DSX platform, covering everything from chips to data center operations. The focus is on establishing an AI factory, a data center specialized in AI learning and inference.
Huang noted, "Korea currently needs AI infrastructure, but it is insufficient. Korea is one of the leading AI ecosystems in the world, and AI infrastructure is essential for future development."
The AI factory will serve as a next-generation AI infrastructure that produces tokens, the core units of AI services, on a large scale, based on power and data. It will be built to optimize AI computations, moving beyond traditional data centers focused on general computing and data storage.
The companies plan to launch the first AI factory in South Korea by 2027. This initiative aims to validate the AI cloud operating system and governance before gradually expanding operations across Asia.
To achieve this, SKT will join NVIDIA's NCP, a global partner ecosystem program that provides high-performance AI cloud services using NVIDIA's AI infrastructure and software. The goal is to secure the lowest token costs and the highest performance per watt, enhancing competitiveness in the global AI infrastructure market.
The collaboration with NVIDIA is expected to be a cornerstone of SKT's AI cloud business. SKT plans to utilize NVIDIA's Blackwell graphics processing units (GPUs) to provide AI learning and inference services, with the next-generation platform "Vera Rubin" set to be introduced in the second half of this year.
SKT aims to accelerate the expansion of its AI cloud business through this partnership. Unlike general cloud services, AI cloud services are specialized for AI learning, inference, and data processing, emerging as a key business in the AI supply chain as global tech giants increasingly lease infrastructure.
The company also plans to expand global partnerships, positioning itself as a strategic partner providing the infrastructure and business network needed by NVIDIA across Asia, with the goal of becoming a leading AI cloud provider in the region.
Huang concluded, "Just as the semiconductor industry needed fabs, the AI industry needs AI factories. I expect we can accelerate the establishment of AI infrastructure in Korea with SKT."
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.