Nvidia’s Jensen Huang Says South Korea, Taiwan Both ‘Special’ and Not Meant for Comparison

‘Agentic AI’ Era Has Arrived and Will Be Everywhere

Nvidia Chief Executive Officer Jensen Huang said both South Korea and Taiwan are special when asked to compare the two.

Huang made the remarks in Taipei on June 1 after a dinner with representatives of major South Korean companies, Taiwan’s Central News Agency and other outlets reported on June 2. Nvidia’s annual artificial intelligence conference, GTC Taipei, was being held in the city.

In the semiconductor industry, there is no need to choose one over the other, Huang said, adding that comparison is not important. Both South Korea and Taiwan can be very special at the same time, he said.

Huang arrived in Taiwan on May 23 and has been meeting figures in the island’s semiconductor industry, including TSMC Chairman C.C. Wei over dinner.

He is also attending Computex, Asia’s largest information-technology trade show, running from June 2 to June 5. He is then scheduled to visit South Korea around June 5 for meetings with the heads of major business groups and other events.

He is expected to hold a samgyeopsal dinner in Seoul with SK Group Chairman Chey Tae-won, Hyundai Motor Group Executive Chair Euisun Chung, LG Group Chairman Koo Kwang-mo and Naver Chairman Lee Hae-jin.

Huang said the most important purpose of his South Korea trip is to personally express thanks and congratulations to Nvidia’s Korean partners. He added that he likes Korean food and hopes to visit restaurants serving fried chicken, samgyetang and samgyeopsal.

He called South Korea the birthplace of esports and PC bangs, or gaming cafes, and said the country has remained close to his heart since the early days of Nvidia’s GeForce graphics cards.

In his GTC Taipei keynote on June 1, Huang also stressed the arrival of the agentic AI era, in which AI moves beyond simple chatbot functions to carrying out work.

He said the emergence of AI agents is the main driver behind the transformation in computing models. Everything has changed over the past six months, he said, and with agents now realized, AI can finally perform truly useful tasks.

He also said all computing infrastructure will eventually adopt the same agentic AI computing architecture. That architecture will extend beyond data centers to personal computers, cars, humanoids, factory equipment, base stations and satellites.

Nvidia has spent the past two years focused on building AI factories, Huang said, and the next step will be to bring agentic AI from the cloud to edge devices. In the future, AI agents will be everywhere, he said.

He also said AI has become a machine for producing profits and gross domestic product, or GDP. Tokens, the basic unit of computing, can already be converted into revenue and profit, he said.

Go Jeong-sam, Hankyung.com reporter, jsk@hankyung.com