South Korea Minister Discusses AI Partnerships With Nvidia

Science Minister Bae Kyung-hoon met with Nvidia Chief Executive Officer Jensen Huang in Seoul on June 8, 2026, to discuss expanded cooperation across the domestic AI ecosystem, according to the Ministry of Science and ICT and reporting by Yonhap News Agency. The meeting took place during Huang's four-day visit to South Korea, during which he met executives from major conglomerates, researchers and startup representatives, and addressed a Korea AI Ecosystem Reception attended by representatives from 18 companies, including Samsung Electronics, SK hynix, Hyundai Motor Group, LG Electronics, Naver and Krafton (Yonhap; Korea Herald).

Yonhap reports that Minister Bae asked Huang for the smooth delivery of about 260,000 advanced chips that were discussed during the Asia-Pacific Economic Cooperation meeting last year. Korea Herald reports that talks focused on the planned delivery of 260,000 Nvidia GPUs, the introduction of Vera Rubin-based AI factories in Korea, and closer collaboration with Korean universities, research institutes and companies. Yonhap additionally records that Bae expressed hope Nvidia's upcoming AI R&D center in Seoul could become a partnership base.

UPI quotes Huang saying that he views robotics as the next major growth sector for South Korea and that Nvidia intends to partner with domestic manufacturers in robotics and AI, remarks he made after arriving in Seoul.

Reporting by Korea Herald names Nvidia's Vera Rubin NVL72 as the platform discussed for so-called "AI factories," described in coverage as specialized facilities for training and running large-scale AI workloads. Yonhap and Korea Herald both point to the 260,000 chips/GPUs as a central operational item in talks; Korea Herald frames Vera Rubin-based AI factories and NVL72 specifically as part of the delivery and deployment conversation.

Governments and hyperscalers increasingly combine hardware supply commitments with local R&D and deployment partnerships to accelerate adoption. Comparable public agreements typically include multi-hundred-thousand chip deliveries, commitments to build local training or inference facilities, and engagement with universities and system integrators to create use cases. For practitioners, such arrangements usually lower procurement friction while increasing demand for local systems integration, cluster management, and GPU-accelerated software stacks.

For practitioners: Expect implementation work to center on three technical streams: integration of large-scale GPU clusters with existing on-prem and cloud infrastructure; adaptation of orchestration, storage and networking to support Vera Rubin-class platforms; and development of production-grade "physical AI" applications that marry robotics, controls and perception models. These are industry patterns observed when chipmakers and national governments coordinate deployments.

Editorial analysis: The combination of a large-volume GPU delivery figure (260,000) and discussion of Vera Rubin AI factories indicates a push toward capacity growth and on-prem AI compute in South Korea. Public-private conversations that also reference R&D centers and university links tend to accelerate ecosystem development by channeling hardware, talent and pilot projects into manufacturing, automotive and robotics verticals. For the regional AI supply chain, early local deployment of Nvidia Vera Rubin-class systems would shift some training and inference workloads toward South Korean campuses and enterprise sites, affecting procurement, operations and talent demand.

• •Whether Nvidia publicly confirms a binding timeline for the 260,000 GPUs and any delivery schedule, and whether those GPUs are specified as particular NVL or H100-class SKUs (Yonhap; Korea Herald).
• •Announcements about an Nvidia R&D center in Seoul that include staffing, research focus or formal partnerships with universities (Korea Herald; Yonhap).
• •Early pilot projects or "AI factory" proofs of concept with major conglomerates named at the reception, which would indicate how the Vera Rubin platform is being applied to manufacturing or robotics (Yonhap; Korea Herald).
• •Regulatory or infrastructure steps by South Korean ministries to host high-density GPU clusters, including power and data center investments (reported demonstrations at domestic forums, Biz Chosun).

This account restricts itself to reported events and public remarks; no internal company plans beyond direct quotations or reporting are asserted.