How Nvidia’s South Korean AI deals could fuel ‘the next industrial revolution’

A spate of agreements Nvidia has reached with South Korean companies presents a long-term road map for the country to expand its role in the semiconductor and physical AI industries, observers say. The most significant deal Nvidia CEO Jensen Huang made during his whirlwind three-day trip to Seoul that ended on Sunday was with SK Telecom to build AI infrastructure, including factories. An AI factory is a data centre running an end-to-end operational system that continuously collects and processes data, and refines algorithms to generate intelligence on a massive scale. The two companies agreed to pursue a “full-stack AI cloud” collaboration spanning everything from chips to data centre operations, based on Nvidia’s DSX platform. DSX integrates high-performance Graphics Processing Unit (GPU) hardware with an optimised software stack – enabling enterprises, researchers and developers to train and deploy massive AI models. SK Telecom plans to build a gigawatt-scale AI cloud and AI factory in South Korea by 2027 using Nvidia’s DSX platform. Separately, Naver, South Korea’s largest internet portal and search engine, has agreed to deploy a Nvidia DSX-powered AI factory hub, starting with a 55-megawatt expansion in 2027 and ultimately scaling to a gigawatt level. A data centre consuming 55MW of electricity can accommodate thousands of high-end AI servers. Nvidia has also agreed with LG Group and Hyundai to expand collaborations to build AI factories for physical AI, autonomous driving and robotics. A multi-year deal with semiconductor chip giant SK Hynix will see the co-development of next-generation high-bandwidth memory crucial for Nvidia’s AI systems and supercomputers. In addition, Nvidia and Doosan announced plans to advance physical AI and AI factory infrastructure and expand cooperation across Doosan’s key businesses, including energy, electronic materials and robotics. AI road map In January, Nvidia announced plans to establish a research and development centre in South Korea. After arriving in the country on Monday, Huang said the company had begun hiring staff to build it, praising South Korea as a “truly excellent place” for research and development. “This is a vote of confidence for South Korea as a potential AI hub in the region by the world’s leading tech giant,” said Hur Jung, an economics professor at Sogang University. “Nvidia has also presented a long-term road map for South Korean government and companies to develop data centres, AI factories and physical AIs.” The deals are also significant in terms of securing independence in AI and telecommunications. “These agreements would help lay the foundation for South Korea to secure its sovereignty in artificial intelligence and online communications amid mounting geopolitical uncertainties,” Hur said. However, he cautioned that there was no guarantee that these agreements would be realised as they were mostly memorandums of understanding with no legal obligations. “I am afraid they are painting a rosy picture, adding fuel to the semiconductor boom in the stock market,” he said, pointing to the exorbitant amounts of capital required for the projects. Huang’s trip signalled Nvidia’s trust in South Korea as a crucial partner in realising its AI vision, said Kim Gwang-suk, an economics professor at Hanyang University. “South Korea is a rare country that has a full AI value chain for physical AI and Nvidia needs it for its business,” he said. Physical AI refers to AI-equipped robots, autonomous vehicles, industrial systems and other embodied AI agents. The AI value chain includes AI infrastructure, models, hardware and real-world operations. “Of the AI full stack, the GPU is the only thing that South Korea doesn’t have, while Nvidia has GPUs but needs the rest of the value chain. Their cooperation is inevitable in the physical AI sector,” he said. An AI full stack is a complete set of integrated technologies needed to build, run and manage AI applications. Nvidia’s GPUs will be in great demand as South Korean companies step up global physical AI production, according to Kim. Samsung and LG produce home appliances, while Samsung also manufactures IT devices such as mobile handsets and personal computers. Doosan builds robotics systems and Hyundai produces automobiles. Nvidia’s research and development centre would also help turn South Korea into a global AI research hub, as the country’s regulation-free zones could provide test beds for autonomous vehicles, drones and other emerging technologies, Kim said. Shaping the future The agreement with SK Hynix built on years of deep co-engineering collaboration that had powered some of the world’s most advanced AI computing platforms, Nvidia said in a news release. “AI factories are the engines of the next industrial revolution, and advanced memory is essential to their performance,” Huang said. SK Hynix had been an “extraordinary partner”, playing a central role in delivering advanced memory technologies for Nvidia AI computing platforms, he added. “We will co-develop the next generation of memory for AI factories and support the accelerating global expansion of AI infrastructure – from frontier model training to agentic and physical AI.” Agentic AI is an advanced system that moves beyond simply answering questions or generating content and acts as an autonomous digital employee. “SK Hynix and Nvidia have been building towards this for years, and this partnership reflects the depth of that collaboration,” SK Group chairman Chey Tae-won said, calling the collaboration “work that will shape the future of AI infrastructure”. Huang also met LG Group chairman Koo Kwang-mo and agreed to strengthen cooperation in mobility, AI data centres, foundation models and physical AI, according to The Korea Times. With Nvidia AI platforms, LG could move from home appliances to AI factories and infrastructure, Huang said. LG said it would adopt Nvidia’s AI technologies to build an intelligent autonomous manufacturing ecosystem and that the companies would jointly develop general-purpose robotics models. It also plans to build an AI factory and integrate Nvidia’s autonomous driving platform into next-generation driver-assistance systems and other AI-powered mobility solutions. At the meeting with Hyundai Motor Group executive chair Chung Eui-sun, Huang said they discussed ongoing projects to “transform the future of mobility and the future of robotics”. “The next wave of AI is mobility and physical AI. This is the future where AI understands the world and can come out and do good work, productive work for the world,” he said. Hyundai plans to invest 9 trillion won (US$5.9 billion) to transform a vast tract of reclaimed land in Saemangeum, North Jeolla province, into what it calls an “AI valley” for AI and robotics development. For its partnership with Nvidia, Naver plans to build a 55MW AI factory in 2027 and expand its overseas AI infrastructure capacity to 200MW by 2028 to meet growing global demand. “This alliance is highly encouraging because it allows us to offer a tangible blueprint for countries and regions around the world seeking to build their own sovereign AI capabilities,” Naver chairman Lee Hae-jin said.

How Nvidia’s South Korean AI deals could fuel ‘the next industrial revolution’

.css-1c6uqr6{color:inherit;font-weight:inherit;font-size:inherit;font-family:inherit;line-height:inherit;overflow-wrap:break-word;}Jensen Huang’s dealings in Seoul may provide a road map for South Korea’s AI ambitions but experts warn against painting too rosy a picture

Jensen Huang’s dealings in Seoul may provide a road map for South Korea’s AI ambitions but experts warn against painting too rosy a picture