Event: San Francisco AI Summit 2026
Date: Friday, July 24, 2026
Location: San Francisco, California, United States
Host: Government of the Republic of Korea, convened by President Lee Jae Myung
Format: Invitation-only government-backed summit, approximately 150 attendees
Event Description
The San Francisco AI Summit was a one-day, invitation-only gathering convened by South Korean President Lee Jae Myung during his visit to the United States, bringing together the leadership of the American artificial intelligence industry and the heads of South Korea’s largest technology and industrial groups. Approximately 150 executives, investors, researchers and startup founders attended.
The summit was organised around two themes that define the current constraint on global AI infrastructure: high-bandwidth memory supply and large-scale data centre construction. South Korea occupies an unusual position in both. Samsung Electronics and SK Hynix together account for roughly eighty per cent of global HBM production, the memory technology that every advanced AI accelerator depends on, while the Korean government has committed to a large-scale domestic data centre buildout as part of a semiconductor and physical AI investment programme announced in June 2026.
Speakers and Attendees
President Lee delivered the keynote address. Ahead of the main session he held individual meetings with Nvidia chief executive Jensen Huang, OpenAI chief executive Sam Altman, Anthropic chief executive Dario Amodei and Broadcom chief executive Hock Tan.
Korean industry was represented at the highest level. 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 all attended, alongside Deputy Prime Minister and Minister of Science and ICT Bae Kyung-hoon. Microsoft was represented by Azure hardware president Rani Borkar. A dinner for the delegation and visiting chief executives followed the main programme.
The San Francisco AI Declaration
The centrepiece of the keynote was the San Francisco AI Declaration, a statement of South Korea’s ambition to establish itself as a core country in the global AI supply chain. The declaration set out four pillars: building an AI semiconductor production base, establishing a world-class AI hub, developing national AI testbeds, and expanding international AI cooperation. President Lee framed the programme as a partnership model rather than a competitive one, calling for closer cooperation between Korean and American companies across semiconductors, data centres, sovereign AI and large language models.
Announcements and Agreements
The Korean presidential office stated after the summit that agreements worth more than 950 billion dollars in combined value had emerged from discussions at the event. The largest components were a set of SK Group agreements valued at 750 billion dollars, including an SK Hynix partnership with Nvidia valued above 500 billion, and a Samsung Electronics memorandum of understanding with Broadcom covering up to 200 billion dollars across memory chips, foundry services and advanced packaging.
A separately disclosed letter of intent between SK Group and Nvidia, also valued above 500 billion dollars and spanning large-scale AI data centres and next-generation memory, may overlap with the SK figure; the degree of overlap has not been made public, and at least one independent calculation of the announced agreements arrived at a materially lower total. The instruments announced are memoranda of understanding and letters of intent rather than executed purchase contracts.
The most concretely specified outcome was an SK Telecom commitment to build a two-gigawatt AI data centre running Nvidia’s Vera Rubin platform on SK Hynix HBM4 memory, with the facility scheduled to come online in 2027. Naver and Hyundai Motor Group also participated in partnership discussions covering AI infrastructure and autonomous driving.
Context
The summit took place during a period of acute competition for AI compute supply, with memory and packaging capacity rather than logic fabrication acting as the binding constraint on the global buildout. It followed a series of earlier Korean engagements with Nvidia and other American technology companies, and formed part of President Lee’s broader strategy of positioning South Korea as a trusted production base within an American-aligned AI supply chain.