SK to Launch USD 750 Billion AI Infrastructure Partnership with U.S. Big Tech Firms

Chey Tae-won and Jensen Huang to Build AI Infrastructure Worth Over $500 Billion

Microsoft to Pursue Long-Term AI Memory Supply Partnership

Anthropic and AWS to Expand Cooperation for AI Data Center Construction

Enhancing National AI Competitiveness and the Global AI Ecosystem

Chey Tae-won: "The Mega Projects Are for Korea’s Growth"

SK has agreed to strengthen collaboration with global big tech companies to expand the artificial intelligence (AI) ecosystem, assembling a partnership valued at USD 750 billion (approximately KRW 1,085 trillion). Building on its previously announced "Three Major Mega Projects National Briefing" last month, SK plans to broaden its execution base by working alongside big tech firms.

SK Chairman Chey Tae-won is speaking at the "San Francisco AI Summit" held on the 24th (local time) at The Midway in San Francisco, USA, attended by President Lee Jae-myung. Photo by Yonhap News

On the 24th (local time), at the "K-AI Summit" held at the Fairmont Hotel in San Francisco, USA, SK announced that it reached agreements with NVIDIA, Microsoft, and Anthropic to cooperate in AI infrastructure areas such as AI semiconductors and AI data centers. In a separate meeting with Amazon Web Services (AWS), the companies discussed expanding cooperation on domestic AI infrastructure, leveraging their existing partnership built through their joint construction of the Ulsan AI Data Center.

◆ SK Group to Lead Globally as a Full-Stack AI Factory= Following last month’s announcement of its investment plans for nationwide AI data center construction and the development of a semiconductor belt, SK’s strategy is to further solidify the execution of its mega projects through cooperation with global big tech firms.

SK Group will embark on a comprehensive partnership with NVIDIA for the construction of AI infrastructure projects worth more than USD 500 billion in total. This includes everything from building AI factories to supplying AI memory. The two companies signed a letter of intent (LOI) formalizing this scope of collaboration on the same day.

Initially, SK Telecom will secure next-generation GPUs and pursue investment cooperation to build an AI factory with a maximum capacity of 2 gigawatts (GW) in South Korea. This cooperation concretizes the plan—agreed upon last month—to establish a GW-scale AI factory in the country.

SK Telecom plans to gradually build and operate high-performance AI factories from next year by adopting “Vera Rubin,” a next-generation AI computing system equipped with SK hynix’s HBM4, based on NVIDIA’s full-stack AI factory architecture “DSX” platform.

SK Chairman Chey Tae-won is speaking at the "San Francisco AI Summit" held on the 24th (local time) at The Midway in San Francisco, USA, attended by President Lee Jae-myung. Photo by Yonhap News

SK hynix will embark on a long-term collaboration in AI memory. NVIDIA will gain stable access to next-generation AI memory while SK hynix will expand the foundation for its AI memory business growth.

SK hynix also plans to pursue a long-term memory supply partnership with Microsoft. The two companies have agreed to supply server memory optimized for AI workloads to Microsoft’s data centers in the mid-to-long term to respond to the surging demand for AI services.

SK Telecom and Anthropic have signed a comprehensive memorandum of understanding (MOU) for cooperation to build GW-class AI data centers in Korea. Anthropic intends to participate in SK Telecom’s ongoing domestic AI data center construction projects.

With AWS, SK is expanding the scope of cooperation in AI infrastructure by building on their partnership forged through the joint construction of the Ulsan AI Data Center. By combining SK Telecom’s domestic AI data center construction and operation capabilities with AWS’s global cloud and AI technology and customer base, the companies plan to expand business operations to key domestic hubs and, in the long term, secure GW-class AI computing capacity.

◆ “One AI Agent per Person”—SK Chairman Chey Tae-won: Korea Must Evolve into an AI-Native Country= At the summit held on the same day with President Lee Jae-myung and big tech leaders, SK Group Chairman Chey Tae-won stressed regarding the recently unveiled “Three Major Mega Projects” that “Although the scale of investment may seem overwhelming, this is a realistic plan based on actual demand in the global AI market.”

Chairman Chey introduced his meetings with executives from global big tech companies like NVIDIA, Broadcom, Anthropic, and OpenAI, noting that “the pace and scale of demand for AI chips and computing power are growing much faster than anticipated,” and “Even the current investment scale is, in light of future demand, a realistic figure.”

He went on to say, “The three major mega projects are AI projects for Korea’s future and are projects for all of us,” emphasizing, “We must concentrate our capabilities as quickly as possible to ensure their success.”

Chairman Chey presented three key objectives necessary for project success: implementing a “one AI agent per citizen” vision to make Korea a global AI testbed and AI-native nation; reducing token costs by expanding memory production and AI data centers; and minimizing the side effects of AI, such as by establishing AI governance, creating jobs, and preserving the environment.

He also remarked, “AI is still in its early stages,” adding, “Korea should leap forward as an AI infrastructure hub connecting the U.S., Europe, and Asia, based on more memory and AI data centers.” He concluded, “I look forward to Korea and the U.S. joining forces to build better AI infrastructure and ecosystems together.”

This content was produced with the assistance of AI translation services.