Jensen Huang is in South Korea for the second time in less than a year to build large-scale AI infrastructure in Asia. SK Telecom is advancing plans for a first gigawatt-scale AI cloud service in South Korea, expanding to other parts of Asia. This combines Nvidia's GPUs with SK Telecom's network to create AI factories generating tokens, first expected online by 2027.

Jensen Huang said without SK partnership, today's AI industry would not have developed as wonderfully as it has, and we're at the beginning of the AI infrastructure build out.

SK Hynix has a multi-year technology agreement to develop advanced memory for global AI data centers, enabling them to keep pace with Nvidia's plans into robotics, PCs, and AI supercomputers. Jensen Huang said SK Hynix's plan to double memory wafer capacity by 2030 would not be enough to meet surging AI demand, but they will remain Nvidia's largest memory partner. Nvidia's new Vera CPU will use SK Hynix memory.

Memory chip makers have been straining to keep up with demand.

Naver will collaborate with Nvidia to build AI factories as part of the gigawatt-scale push. They discussed a concrete roadmap to jointly enter AI markets in Europe, the Middle East, and Asia Pacific. Naver has its own homegrown LLM in Korea, central to South Korea's sovereign AI efforts.

None of the partners have disclosed the value of these deals. Jensen's meetings continue with Doosan, LG, and Hyundai.