Seoul (AFP) – US chip titan Nvidia on Monday announced a large-scale data centre construction project in South Korea with SK Telecom, among a raft of other business deals in the country. Nvidia, the world’s most valuable company, also said it would work with chipmaker SK hynix to develop the advanced memory components that help run AI systems and are currently in short supply. The tie-ups were unveiled after CEO Jensen Huang spent the weekend eating barbecue in Seoul with the country’s tech leaders and appearing on a popular TV show.
SK Telecom and Nvidia plan “to build a gigawatt-scale AI Cloud in Korea…with the first AI factory planned to come online in 2027,” a joint statement said. The project “will support sovereign, physical and agentic AI services for enterprises and industries across Korea, with the vision to expand to greater Asia regions,” it added. No figure was given for how much the two companies will invest in the data centres. SK Telecom operates under the same parent company — SK Group — as SK hynix, which on Monday announced a “multi-year technology partnership” for memory chips with Nvidia.
“The agreement supports supply for advanced memory, addressing the extended development cycles, advanced fabrication and capital investments to sustain the global buildout of AI factories,” their statement said. “Through this partnership, SK hynix will diversify into new markets Nvidia is creating — spanning AI infrastructure, personal AI and physical AI,” through co-developing memory components for Nvidia hardware, it said. As governments and companies pour hundreds of billions of dollars into AI infrastructure, Nvidia’s value has topped $5 trillion, more than the gross domestic product of Japan or India.
The race to build AI data centres has created a global shortage of memory chips — sending profits skyrocketing for manufacturers like SK hynix and rival Samsung Electronics, whose workers’ union recently agreed a deal with management on bonuses, averting a strike. SK Group chair Chey Tae-won last week vowed to double production capacity of silicon wafers used to make memory chips. But he also reiterated his prediction that shortages could persist until 2030, with chip factories taking at least three years to build.
Nvidia’s Huang signed a memory chip display at the SK hynix booth at the Computex trade show in Taipei, writing: “Please make more.” When he landed in South Korea on Friday, Huang said he had “brought a lot of business to Korea,” promising some new “surprises.” On Monday, the California-based company also announced AI-related collaborations with tech giant Naver, and with Doosan Group on robotics.
Nvidia is best known for its GPUs, specialized computer chips originally designed to render video game graphics at high speed. These chips have become the engine behind AI tools from chatbots to image generators and agents that can carry out tasks for users. Nvidia last week unveiled a powerful laptop chip for Windows machines, staking its claim in the market for next-generation consumer PCs integrated with AI.
© 2024 AFP
