NVIDIA has announced a $1 billion investment in South Korean internet and cloud services provider Naver, allocated for the latter's AI data center project currently under construction in South Korea. This represents the latest in a series of significant transactions recently finalized by the chipmaker.
The company stated on Friday evening that the funding will more than triple the capacity of Naver's data center. The project is a joint development between Naver and the U.S. private equity firm Brookfield Group, and will be fully equipped with NVIDIA AI computing hardware.
In addition to the Naver deal, NVIDIA has partnered with SK Group to plan the construction of an AI data center cluster in South Korea with a total computing power exceeding 2 gigawatts, a power load sufficient to meet the electricity needs of approximately 1.5 million households. NVIDIA revealed that the first AI computing factory, to be built by SK Telecom, will begin operations next year.
NVIDIA CEO Jensen Huang stated in an interview, "South Korea is entering a golden period. Its semiconductor and industrial sectors are both thriving, providing the complete capability to build AI infrastructure for the world."
Huang continues to position the company globally in the AI sector through large-scale investments and major collaboration agreements. He has invested billions of dollars cumulatively in companies across the entire AI supply chain, with the core goal of accelerating global customer adoption of NVIDIA products and removing various bottlenecks for AI industry deployment.
This announcement deepens the existing collaboration between NVIDIA and the two leading South Korean conglomerates. The news came during a visit by South Korean President Lee Jae-myung to the United States, where he is attending a series of industry events in Silicon Valley, with more technology cooperation agreements expected to follow.
For NVIDIA, strengthening ties with SK Group helps secure a stable supply of memory chips. SK Group's affiliate, SK Hynix, along with Samsung Electronics, are the world's two major suppliers of high-end memory chips. With the global large-scale construction of new AI data centers, high-bandwidth memory chips remain in persistently short supply.
NVIDIA stated that it will assist SK Hynix in developing next-generation High Bandwidth Memory (HBM) chips to lock in a stable supply channel. Huang revealed in the interview that the total transaction value of the overall AI collaboration between NVIDIA and SK Group will exceed $500 billion. This figure encompasses both NVIDIA's procurement of memory chips and SK Group's procurement of supercomputing equipment from NVIDIA. "The combined business volume of our cooperation will reach $500 billion," he said.
Regarding the Naver data center project, Brookfield has signed a non-binding framework agreement that could provide up to $9 billion in financing.
Huang believes that investments in AI companies like Anthropic and OpenAI can both drive growth for NVIDIA's core business and generate investment returns. However, market critics argue that such investments are a form of artificially creating demand and represent circular transactions.
Earlier this week, NVIDIA signed a $1.5 billion agreement with Amkor Technology to help expand the latter's chip packaging factory. This followed Huang's recent trip to Japan, where he reached agreements with several robotics companies and Toyota Motor Corporation.
Huang is currently pushing NVIDIA to expand its customer base beyond major cloud providers (hyperscalers) and increase AI hardware investments for other enterprises. He has proposed that countries should build their own sovereign and controllable domestic AI computing systems to maintain data sovereignty, while also ensuring that small and medium-sized enterprises and various institutions can easily access AI computing resources.
Comments