SK Group has joined forces with global tech giants in Silicon Valley to expand the global AI ecosystem. On July 24, during the K-AI Summit held at the Fairmont Hotel in San Francisco, SK Group reached agreements with Nvidia, Microsoft, and Anthropic to collaborate on AI infrastructure, including AI semiconductors and data centers.
Additionally, in a separate meeting with Amazon Web Services (AWS), SK Group discussed expanding domestic AI infrastructure cooperation based on their existing partnership in building a joint AI data center in Ulsan.
Following the announcement of nationwide AI data center construction and semiconductor belt investment plans at the 'Three Mega Projects National Report Meeting' in June, SK Group aims to strengthen the execution foundation of these projects through collaboration with global tech firms. This initiative is designed to enhance South Korea's AI competitiveness while expanding the global AI ecosystem.
An SK Group official stated, "The expansion of cooperation with global tech giants like Nvidia reaffirms the rapidly growing demand for AI infrastructure, such as AI memory. We expect this collaboration to further energize the recently announced 'Mega Projects' by the government."
Nvidia and SK Group to Collaborate on Comprehensive AI Infrastructure
SK Group will engage in comprehensive cooperation with Nvidia to build AI infrastructure worth over $500 billion. This partnership, which includes the establishment of AI factories and the supply of AI memory, was formalized with a letter of intent (LOI) signed on the same day.
Initially, SK Telecom will receive next-generation GPUs to build a domestic AI factory with a capacity of up to 2 gigawatts (GW) and pursue investment collaboration. This partnership is a concrete step following their agreement in June to establish a GW-class AI factory in South Korea.
SK Telecom plans to introduce the next-generation AI computing system 'Vera Rubin,' equipped with SK Hynix's HBM4, based on Nvidia's full-stack AI factory architecture, the 'DSX' platform. The goal is to gradually establish and operate high-performance AI factories starting in 2027, accelerating the development of large-scale AI infrastructure that encompasses sovereign, physical, and agentic AI, while jointly addressing the growing AI demand in the Asia-Pacific region.
This collaboration is expected to accelerate the spread of AI infrastructure utilizing Nvidia technology and broaden customer access to advanced cloud services. As a result, SK Telecom will embark on large-scale AI infrastructure investments and expansions, making cutting-edge AI infrastructure more accessible to a wider range of customers.
SK Hynix will also engage in long-term cooperation for AI memory supply. Following their previously established long-term technology partnership, Nvidia will provide next-generation AI memory, allowing SK Hynix to expand its growth base in the AI memory business.
The two companies plan to jointly develop and optimize next-generation AI memory, such as HBM, to meet the infrastructure demands ranging from large language model (LLM) training to agentic and physical AI.
Jensen Huang, founder and CEO of Nvidia, stated, "South Korea has all the conditions to become a global AI powerhouse, including world-class networks, data centers, semiconductor technology leadership, and a vast industrial base. Nvidia will work with SK Telecom and SK Hynix to create a new generation of AI factories that will drive the next growth phase in Korea."
Chey Tae-won, chairman of SK Group, remarked, "The competitiveness in the AI era depends not only on how well we utilize AI but also on how much intelligence we can generate. SK aims to build world-class AI factories in collaboration with Nvidia, leveraging SK Telecom's AI infrastructure capabilities and SK Hynix's AI memory, contributing to South Korea's transformation from an AI-consuming nation to a global hub for AI innovation."
SK Hynix and Microsoft to Establish Long-term AI Memory Supply Cooperation
SK Hynix has decided to pursue long-term memory supply cooperation with Microsoft. The two companies aim to supply server memory optimized for AI workloads to Microsoft's data centers in response to the surging demand for AI services.
They plan to explore memory solutions tailored to AI workload requirements and establish new standards for AI server memory through validation in actual server environments.
SK Telecom Expands AI Data Center Cooperation with Anthropic and AWS
SK Telecom and Anthropic signed a comprehensive memorandum of understanding (MOU) to collaborate on building GW-class AI data centers in South Korea. Anthropic is expected to participate in SK Telecom's ongoing AI data center construction project, with plans to discuss specific cooperation methods.
With AWS, SK Telecom will expand the scope of cooperation in AI infrastructure based on their partnership in jointly building the Ulsan AI data center. The two companies aim to combine SK Telecom's domestic AI data center construction and operational capabilities with AWS's global cloud and AI technology and customer base to broaden their business in key domestic locations and secure GW-class AI computing capacity in the long term.
On the same day, Chey Tae-won met with Sam Altman, CEO of OpenAI, to discuss long-term cooperation in the AI infrastructure sector. SK and OpenAI have been continuously discussing collaboration plans for AI infrastructure, including memory, and have been strengthening their strategic partnership since signing a partnership agreement last October regarding memory supply and the establishment and operation of a southwestern AI data center.
Chey Tae-won Proposes Three Key Tasks for Mega Project Success
During the K-AI Summit attended by President Lee Jae-myung, global tech representatives, AI experts, researchers, startups, and investors, Chey Tae-won emphasized the recent announcement of three mega projects, stating, "The investment scale may seem unfamiliar due to its size, but it is a realistic plan based on actual demand in the global AI market."
Chey introduced his meetings with executives from global tech firms like Nvidia, Broadcom, Anthropic, and OpenAI, noting, "The scale of AI chips and computing power demanded by companies is growing much faster than expected. The announced investment scale is realistic when considering future demand."
He added, "The three mega projects are AI initiatives for South Korea's growth and a project for all of us. We must consolidate our capabilities as quickly as possible to ensure success."
Chey outlined three key tasks for project success: 1) Implementing an 'AI Native Nation' where each citizen utilizes at least one AI agent, making South Korea a global AI testbed; 2) Reducing token costs through memory production and expansion of AI data centers; and 3) Establishing AI governance while minimizing AI-related side effects, including job creation and environmental preservation.
He also stated, "AI is still in its early stage. South Korea must leap to become an AI infrastructure hub connecting the U.S., Europe, and Asia, based on more memory and AI data centers."
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.