Republic of Korea announced $950bn in new AI initiatives involving Samsung Electronics, SK Group and US tech firms as global AI leaders rush to fill a shortage of faster chips to power and develop more advanced systems.
SK Group has signed deals worth $750bn, including SK Hynix’s partnership with Nvidia valued at more than $500bn, Korea's presidential Blue House said after an AI summit hosted by President Lee Jae Myung in San Francisco.
Samsung Electronics said it had signed a memorandum of understanding with Broadcom covering up to $200bn across memory chips, foundry services and advanced packaging.
The deals are the latest moves by US AI firms hoping to lock in supply as demand for chips and computing infrastructure outpaces supply, driving tech companies into closer partnerships with Korea's semiconductor and memory leaders.
Speaking at the summit, Lee said Korea would lead the way to a new era of AI in partnership with global technology companies, making available its "dynamic AI ecosystem" to expand the global market for industrial and personal AI use.
He unveiled a "San Francisco AI Declaration" that outlined Republic of Korea's AI ambitions for technological cooperation with the US.
"I'd like to present a vision for the future of AI and propose global cooperation to make it happen," Lee said.
Lee separately met with Nvidia CEO Jensen Huang, OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei and Broadcom CEO Hock Tan, before joining the summit.
Samsung Electronics Executive Chairman Jay Y Lee, SK Group Chairman Chey Tae-won, Hyundai Motor Group Executive Chair Euisun Chung and Naver founder Lee Hae-jin also attended.
Nvidia and Korea's SK Group unveiled a new AI initiative valued at $500bn spanning large-scale AI data centers and next-generation memory.
The initiative includes a long-term partnership with SK Hynix to secure next-generation memory supply for Nvidia and jointly develop high-bandwidth memory for AI training, AI agents and physical AI applications.
As part of the agreement, SK Telecom plans to build a 2-gigawatt data center powered by Vera Rubin chips and SK Hynix's HBM4 high-bandwidth (HBM) memory chips. The data center is due online in 2027.
Samsung Electronics said it will work with Broadcom on the development of its next-generation AI accelerator based on Samsung's HBM technology and provide it with sub-2-nanometer foundry process and advanced packaging solutions.
SK Chairman Chey said AI customers were seeking far more memory than previously anticipated, pointing to recent discussions with Nvidia, Anthropic, OpenAI and Broadcom.
Chey said Nvidia's projected memory requirement over the next five years could soon prove to be higher.
"I wouldn't be surprised if, the next time I see (Huang), he'll tell me those figures were too low and say, 'I want more,'" Chey said.
He said Broadcom CEO Hock Tan also told him the company's memory demand was likely to surpass available supply, and that Anthropic was regularly checking on future chip availability as it moves beyond developing AI models to building computing capacity.
Chey said OpenAI would require enormous computing power and had urged SK to supply chips directly to the ChatGPT maker.
The SK Group chairman also said Anthropic was making its own "computing power" without elaborating, potentially referring to a plan reported by Reuters that the AI lab is exploring the possibility of designing its own chips.
An Anthropic spokesperson said the startup was always exploring options and drawing on a range of chips for a diversified approach to computing.
Anthropic CEO Amodei made no comment on the reported plan, saying the company had signed supply deals with Samsung Electronics and SK Hynix, without providing further details.
In May, Anthropic raised funding at a post-money valuation of $965bn, naming Samsung Electronics, SK Hynix and Micron as partners whose technologies play a critical role in the supply of memory, storage and logic chips.
