Nvidia and South Korea's SK Group reach $500 billion cooperation
On July 25, Beijing time, Nvidia (NVDA.O) published a blog announcing a US$500 billion cooperation with South Korea's SK Group. The cooperation covers the construction of AI factories and the supply of next-generation memory. NVIDIA said that the two parties signed a letter of intent, and the cooperation between the two parties aims to establish AI infrastructure to meet the world's growing computing needs. It was previously reported that South Korean President Lee Jae-myung met with Nvidia CEO Huang Renxun on July 24. Huang Renxun promised to cooperate with Korean companies, including reaching a US$500 billion contract with SK Group. According to NVIDIA, relevant cooperation includes the establishment of a long-term partnership between NVIDIA and SK Hynix (000660.KS) to jointly develop next-generation artificial intelligence memory including HBM (high-bandwidth memory). This will ensure that Nvidia has a stable supply of next-generation AI memory. SK Telecom, a subsidiary of SK Group, will build an AI cloud computing center with a capacity of 2 GW, using NVIDIA Vera Rubin DSX and deploying a Vera Rubin accelerated computing system based on SK Hynix HBM4. The first facility is scheduled to go online in 2027. SK Telecom will use this to expand its AI infrastructure. NVIDIA said the two companies aim to accelerate the development of large-scale AI infrastructure, including sovereign AI, physical AI, agent-based AI and enterprise AI services, and jointly respond to the growing AI demand in the Asia-Pacific region. SK Group Chairman Chey Tae-won said that competitiveness in the AI era not only depends on the efficiency of AI utilization, but also how much intelligence can be produced. SK Group will join hands with NVIDIA to build a world-class AI factory. NVIDIA has previously reached cooperation with the Korean government and enterprises, which mainly involves Korean enterprises purchasing NVIDIA products and NVIDIA purchasing storage products from Korean enterprises. In October last year, NVIDIA announced at the APEC (Asia-Pacific Economic Cooperation) Summit in South Korea that it would cooperate with the Korean government and Korean companies to build AI infrastructure and ecosystem, involving the deployment of more than 260,000 NVIDIA GPUs for physical AI and agent AI. This includes the Korean government’s investment in sovereign AI through relevant departments, involving the deployment of more than 50,000 new NVIDIA GPUs. These GPUs will be deployed on the Korean National AI Computing Center, Korean cloud service and IT providers NHN Cloud, Kakao Corp and NAVER Cloud to accelerate the development of sovereign AI in enterprises and industries. Samsung Electronics, SK Group, and Hyundai Motor Group each plan to build AI factories containing more than 50,000 NVIDIA GPUs. When Huang Renxun visited South Korea in June this year, Nvidia also announced a multi-year technical cooperation agreement with SK Hynix to cooperate in promoting the development of next-generation memory in global AI data centers. SK Hynix planned to jointly develop the Vera Rubin AI supercomputer, Vera CPU, RTX Spark-powered personal computers, and Jetson Thor with Nvidia. The memory used by the computing platform. NVIDIA said that the relevant agreements will ensure the supply of advanced memory and address the challenges of long development cycles, advanced manufacturing, and capital investment. (This article comes from China Business News)