Lee Secures AI Chip Cooperation Deals
President Lee Jae Myung's July 24-25 San Francisco meetings with U.S. technology executives produced AI semiconductor supply and manufacturing cooperation projects valued at $950 billion, according to South Korea's presidential office and Yonhap. The reported agreements include a long-term arrangement under which SK Group would supply $750 billion in high-performance chips to Nvidia and other technology companies. Yonhap also reports that Lee's South American itinerary expanded critical-mineral cooperation.
President Lee Jae Myung's San Francisco meetings with U.S. technology executives produced semiconductor supply and manufacturing cooperation projects valued at a combined $950 billion, according to South Korea's presidential office, Yonhap and NHK World.
The announced projects were discussed during Lee's July 24-25 visit, when he met executives including OpenAI CEO Sam Altman and Nvidia CEO Jensen Huang. NHK World reports that South Korean and U.S. businesses signed the AI-related deals around chip production and supply over the next five years.
Yonhap reports that the package includes a long-term arrangement under which SK Group would supply $750 billion in high-performance chips to Nvidia and other technology companies. The Korea Herald reported that Samsung Electronics and Broadcom signed a memorandum of understanding covering $200 billion. The Herald described the broader $950 billion figure as partnerships that companies agreed to pursue, a distinction relevant when assessing how much of the headline total represents binding orders versus prospective cooperation.
AI chips, data centers and physical AI
At the San Francisco AI Summit, Lee presented South Korea as a prospective production and supply-chain partner for AI semiconductors. According to Yonhap, he said: "Based on the world's highest-level competitiveness and production capacity in memory semiconductors we have, (South Korea) will become a trustworthy AI chip production base and supply chain partner."
Presidential policy chief Kim Yong-beom said the cooperation plans covered three government-designated megaproject areas: semiconductors, AI data centers and physical AI, according to the Korea Herald. NHK World also noted that Lee had announced a government program in July to invest 800 trillion won, or more than $547 billion, in chip-manufacturing hubs in southwestern South Korea.
The reported agreements bring South Korean memory suppliers into discussions with U.S. AI companies. In comparable AI infrastructure buildouts, high-bandwidth memory availability, packaging capacity and long-term supply commitments can materially affect accelerator deployment schedules and system costs. That makes the stated supply arrangements relevant not only to chip vendors but also to teams planning large-scale model training and inference capacity.
Minerals cooperation accompanies the technology agenda
Yonhap reports that Lee also used visits to three South American countries to expand critical-mineral cooperation. The retrieved reporting does not specify the countries, minerals, contractual terms or individual projects.
Critical minerals are an adjacent but consequential input to AI infrastructure because data-center, networking, storage and semiconductor supply chains depend on globally distributed materials and manufacturing capacity. Public details on specific mineral agreements would be needed to assess whether the reported cooperation changes near-term supply availability or is principally a framework for future projects.
Key Points
- 1South Korea and U.S. firms announced $950 billion in chip cooperation, linking memory supply with AI-related projects.
- 2Yonhap reports SK Group's $750 billion high-performance-chip arrangement with Nvidia and other technology companies, subject to disclosed contractual detail.
- 3Comparable infrastructure programs make memory supply, advanced packaging and materials sourcing important constraints for AI training and inference deployment.
Scoring Rationale
The reported $950 billion semiconductor cooperation package is unusually large and directly concerns memory, accelerators and AI data-center supply chains. Its practical impact depends on execution and the binding status of individual agreements, but the named participants make it material for AI infrastructure practitioners.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
