Nvidia locks down memory supply from SK Hynix as part of $500 billion AI deal

Nvidia said it's secured AI memory supply from South Korea's SK Hynix, as the chipmaker tries to lock in a key component for its advanced processors and systems. The agreement, announced late Friday in San Francisco, could be worth $500 billion over a number of years, and includes the construction of large-scale data centers expected to come online in 2027, Nvidia said. SK Hynix affiliate SK Telecom will build a cloud business using Nvidia's Vera Rubin systems. Nvidia said it's targeting enough capacity to require 2 gigawatts of power, which indicates a massive buildout with hundreds of thousands of graphics processing units. Nvidia is aggressively securing supply of high-bandwidth memory, which is essential for its GPUs and systems, as the AI boom has created a global memory shortage. SK Hynix is the leader in HBM production, according to analysts. "The expansion will include a co-develop opportunity for us on the next-generation SK Hynix AI memory, and this will help us secure a stable supply of HBM memory," Raj Mirpuri, Nvidia enterprise vice president, said on a call with reporters. The agreement is also a sign that massive AI infrastructure buildouts are moving beyond a handful of hyperscalers, with foreign governments and massive conglomerates starting to get involved. The deal was announced at an AI summit in San Francisco with South Korean officials including President Lee Jae Myung. Nvidia also said Friday that it would invest $1 billion into Naver, a Korean cloud company building data centers around its GPUs. Nvidia said the project would provide an opportunity for potential customers in South Korea and around the world to secure AI computing capacity before the 200 megawatts of capacity is completed. SK Hynix, South Korea's second most valuable company, listed on the Nasdaq earlier this month as part of an effort to finance infrastructure developments.