Nvidia has announced that it will deliver more than 260,000 Blackwell AI accelerator chips to South Korea after an agreement brokered between the company and the Ministry of Science. Others involved in the deal include Samsung, Hyundai, and SK Group, according to a report by Bloomberg.
The deal is in line with a larger effort by Nvidia to expand AI computing infrastructure across major technology economies worldwide. Nvidia CEO Jensen Huang is presently in South Korea attending the Asia-Pacific Economic Cooperation (APEC) CEO Summit. During the course of the event, he has been promoting the firm’s hardware as the backbone of national-scale AI.
According to reports, the South Korean government plans to build a sovereign AI infrastructure, which will include installing more than 5,000 Nvidia accelerator chips in public and private data centers. Centers involved include the National AI Computing Center and facilities operated by Kakao, Naver, and NHN Cloud. The plan is expected to focus on maintaining domestic control over high-level AI computing rather than outsourcing AI workloads to foreign cloud networks.
Nvidia set to ship AI chips to South Korea as firms deploy large GPU clusters
According to reports, Samsung has plans to build a facility that is described internally as an AI factory, where it will host more than 50,000 Nvidia chips. Samsung is expected to use the chips to train and deploy its own in-house AI models for devices and internal systems. Nvidia representatives also added that Samsung will modify its chipmaking lithography platform to work with Nvidia GPUs, a change that the chip maker said could improve performance efficiency by up to 20 times.
Samsung is also expected to adopt Nvidia’s Omniverse simulation software. In addition to being a customer, Samsung is also a supplier to Nvidia, producing high-bandwidth memory (HBM) used alongside the GPUs. Samsung also confirmed that it will collaborate with Nvidia to adjust its fourth-generation HMB memory for improved AI chip compatibility.
Hyundai Motor Group will also use a similar number of Nvidia’s Blackwell-based processors that will support the company’s development of AI models for manufacturing automation and autonomous driving systems. In addition, it will also be used in training cycles involving telemetry, robotics workflows, and real-world navigation datasets. The allocation is in line with Hyundai’s move towards an AI-driven car development and software-defined vehicle platforms.
Meanwhile, SK Group is deploying Nvidia RTX Pro 6000 Blackwell server chips in what is described as Asia’s first industrial AI cloud, intended for robotics operations and AI-assisted physical process workflows. By adding Blackwell GPUs in industrial settings, SK Group is trying to apply AI inside active production systems rather than only digital systems, a move that was confirmed by Huang.
