South Korea binds the US AI industry chain! Samsung, SK Hynix sign $950 billion chip deal with Nvidia, Broadcom and others
Jin Qilin analyst research report , authoritative, professional, timely and comprehensive, helping you tap potential theme opportunities! Samsung and Broadcom sign $200 billion agreement to deepen cooperation in areas such as memory and foundry technology Nvidia and South Korea's SK Group reach $500 billion cooperation South Korea is rapidly establishing its core position in the global artificial intelligence supply chain by reaching an unprecedented scale of capital and production capacity cooperation with U.S. technology giants, marking the two countries' deep strategic binding in cutting-edge technology industries. According to Reuters, South Korean presidential adviser Kim Yong-beom revealed at a press conference on Saturday that, Samsung Electronics and SK Hynix will enter into memory chip supply cooperation with a total value of up to 950 billion US dollars with major U.S. technology companies including Nvidia. Among them, SK Hynix will provide long-term memory chips worth US$750 billion to U.S. companies, and Samsung Electronics will provide chips worth US$200 billion to Broadcom. At the same time, Nvidia and SK Group further announced an AI infrastructure plan worth more than US$500 billion. This series of historically huge orders and infrastructure investment will not only directly alleviate the current bottleneck of the high-bandwidth memory (HBM) supply chain, but also lock in long-term and highly visible revenue expectations for Korean chip manufacturers, reshaping the production capacity layout and investment pattern of the global semiconductor market. This commercial breakthrough coincides closely with a political push from South Korea's top levels. Against the macro backdrop of AI computing power demand far exceeding global supply, South Korean President Lee officially released the "San Francisco AI Declaration" tomorrow. This cross-border strategic binding directly ensures the stable supply of key hardware for American technology companies, and at the same time pushes South Korea's semiconductor manufacturing base to the forefront of global AI computing power expansion. Huge chip orders lock in long-term supply Ensuring the stability of the underlying supply of computing hardware is the core of this cooperation. According to Reuters, a memory chip supply agreement of up to US$950 billion will build a long-term production capacity moat for US technology giants. As Nvidia's current largest memory chip supplier, SK Hynix plays a key role in this cooperation. At the same time, the cooperation between Samsung Electronics and Broadcom also goes far beyond the traditional supply relationship. According to Bloomberg, the two parties signed a strategic cooperation memorandum of understanding (MOU) and plan to carry out cooperation worth more than US$200 billion in the next five years (to 2030) around memory chips and wafer foundry. Among them, in terms of storage business, Samsung will provide advanced storage solutions for Broadcom's next-generation AI accelerator; in terms of wafer foundry, the two parties will jointly develop Broadcom products around Samsung's 2-nanometer and below advanced process, including wireless broadband communication chips. The cooperation will also extend to 2.3D and 2.5D advanced packaging technologies based on Samsung’s 2nm process to support higher-performance, lower-power AI and network chips. According to reports, SK Hynix’s long-term agreement with Nvidia aims to ensure the supply of advanced memory chips. The two parties will also jointly develop high-bandwidth memory (HBM) for AI training, AI Agents, and Physical AI (Physical AI) applications. HBM, which provides the data transfer speeds required by advanced AI processors by stacking multiple memory chips, has become a key constraint in the current AI supply chain. With the AI server, and supercomputer demand continues to grow, SK Hynix plans to double its wafer production capacity by 2030 to further enhance HBM supply capabilities and enable future memory chip production to achieve closer coordination with Nvidia's chip development schedule. More than $500 billion to reshape AI infrastructure In addition to locking in the supply of underlying chips, the construction of ultra-large-scale data centers has become another investment focus. According to reports, Nvidia and SK Group’s more than $500 billion plan covers the construction of large-scale AI data centers and the development of next-generation HBM. SK Telecom will build an AI data center with a total capacity of 2 GW in South Korea. The first project will use NVIDIA's Vera Rubin platform and SK Hynix's HBM4 memory.
The first facility of this "AI factory" is expected to be officially put into operation in 2027 (i.e. next year). This two-gigawatt scale plan builds on SK Telecom’s previous commitment to build a gigawatt-level AI cloud infrastructure. Funding sources may include global technology companies, overseas capital and other strategic partners. In addition, NVIDIA continues to expand its investment territory in South Korea. According to Bloomberg, Nvidia said late on Friday that it would invest US$1 billion in Naver to provide financial support for its AI data center to be built in South Korea. This capital injection will allow Naver to more than triple the size of the data center using Nvidia's AI computing platform. "San Francisco AI Declaration" promotes national-level strategic binding Behind the comprehensive alliance among the industry is South Korea’s national strategic plan to become a global AI power. On July 24, South Korean President Lee Jae-myung hosted a high-profile AI summit in San Francisco, bringing together the world's leading AI developers and South Korea's top business leaders. At the summit, Lee Jae-myung officially released the San Francisco AI Declaration, clarifying the core vision of South Korea and the United States in AI technology cooperation and promising to open its "vibrant AI ecosystem" to the global market. The lineup for this summit highlights the strategic level of cooperation. Before the summit, Li Zaiming met with Nvidia CEO Jen-Hsun Huang, OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei, and Broadcom CEO Hock Tan. At the same time, attending the summit on behalf of the Korean industry were Samsung Electronics Executive Chairman Jay Y. Lee, SK Group Chairman Chey Tae-won, Hyundai Motor Group Executive Chairman Euisun Chung, and Naver founder Lee Hae-jin. According to Bloomberg, Nvidia CEO Huang Jensen spoke highly of the cooperation between the two parties in an interview. He said that now is the "golden age of South Korea", and South Korea's semiconductor industry and industrial system are booming simultaneously, and emphasized that South Korea is a "country that has the ability to help the world build AI infrastructure."
The first facility of this "AI factory" is expected to be officially put into operation in 2027 (i.e. next year). This two-gigawatt scale plan builds on SK Telecom’s previous commitment to build a gigawatt-level AI cloud infrastructure. Funding sources may include global technology companies, overseas capital and other strategic partners. In addition, NVIDIA continues to expand its investment territory in South Korea. According to Bloomberg, Nvidia said late on Friday that it would invest US$1 billion in Naver to provide financial support for its AI data center to be built in South Korea. This capital injection will allow Naver to more than triple the size of the data center using Nvidia's AI computing platform. "San Francisco AI Declaration" promotes national-level strategic binding Behind the comprehensive alliance among the industry is South Korea’s national strategic plan to become a global AI power. On July 24, South Korean President Lee Jae-myung hosted a high-profile AI summit in San Francisco, bringing together the world's leading AI developers and South Korea's top business leaders. At the summit, Lee Jae-myung officially released the San Francisco AI Declaration, clarifying the core vision of South Korea and the United States in AI technology cooperation and promising to open its "vibrant AI ecosystem" to the global market. The lineup for this summit highlights the strategic level of cooperation. Before the summit, Li Zaiming met with Nvidia CEO Jen-Hsun Huang, OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei, and Broadcom CEO Hock Tan. At the same time, attending the summit on behalf of the Korean industry were Samsung Electronics Executive Chairman Jay Y. Lee, SK Group Chairman Chey Tae-won, Hyundai Motor Group Executive Chairman Euisun Chung, and Naver founder Lee Hae-jin. According to Bloomberg, Nvidia CEO Huang Jensen spoke highly of the cooperation between the two parties in an interview. He said that now is the "golden age of South Korea", and South Korea's semiconductor industry and industrial system are booming simultaneously, and emphasized that South Korea is a "country that has the ability to help the world build AI infrastructure."