"Korea is a country that did not shy away from difficult hardware manufacturing. It has everything from AI chips to steel across the board. If Korea becomes good at AI, enormous commercial opportunities will open up. Please support the startups here."
Jensen Huang, Nvidia's chief executive officer (CEO), sent a clear message at a closed-door roundtable with domestic robotics and AI startups, which he chose as his last official engagement on the 8th in Korea. Korea is a rare stage for physical AI with a semiconductor supply chain, manufacturing sites, and robot hardware all together, and if this potential is to lead to a bigger industry, investment and customers need to come together for robot and AI startups.
At 6:30 p.m. that day, a so-called "marketplace for investment and collaboration" opened in the courtyard of the Youngbingwan at The Shilla Seoul. Not only did leaders of robotics and AI startups collaborating with Nvidia, such as ROBOTIS and AeiROBOT, attend, but also executives from major conglomerates like Samsung Electronics, Hyundai Motor Group, LG Electronics, and Naver, as well as more than 100 people including major domestic and foreign venture capital (VC) firms and government officials gathered in one place. As figures who are usually hard to assemble in one room arrived one after another, invited startup heads said they realized this event was a much bigger networking opportunity than expected.
In front of them, Huang repeatedly stressed that Korea is a "good investment destination" with the conditions to grow a physical AI ecosystem. He cited Korea's ability to execute by rapidly adopting new technologies, its diverse manufacturing base, and its industrial positioning that is not tied solely to either the United States or China as strengths.
Huang then called Bae Kyung-hoon, Vice Prime Minister and Minister of the Ministry of Science and ICT, to the stage and, emphasizing support, said, "There are many excellent startups, so please help them grow well." Riding the momentum, he reportedly called out VCs in the room and encouraged them, saying, "What startups need is money, so please invest."
Organizations that actually hold money and demand attended in force. Major investment firms such as Korea Investment Partners, IMM Investment, and SoftBank Ventures were there, along with representatives from heavy industry and shipbuilding companies with acute labor shortages and strong automation needs, including HD Hyundai, Doosan, and Hanwha Ocean. VCs looked for targets to invest in, conglomerates examined technologies to apply on site, and startups could meet both capital and customers at once—Nvidia set the stage.
This is also what surprised the startup leaders. In Korea, even when conglomerates, startups, and investors gather, hierarchies often surface, but on this day the common denominator of the Nvidia ecosystem created a free-flowing atmosphere for exchange.
AeiROBOT CEO Eom Yun-seol said, "It is not easy for big and small companies to meet and talk without reserve, but this event was significant in that startups met both investors and actual customers at the same time." Angel Robotics CEO Gong Kyung-chul said, "When the government or a conglomerate hosts, people can become self-conscious, but under the umbrella of Nvidia, everyone debated on relatively equal footing," adding, "It was a venue where Nvidia acted as a catalyst."
It is hard to see Huang's push for investment and collaboration in Korea's physical AI ecosystem as mere well-wishing. The startups gathered that day build industrial robots, humanoids, and AI models using Nvidia's graphics processing units (GPUs), robot simulation platforms, and AI model development tools. As they grow, the Nvidia ecosystem grows with them.
Physical AI advances based on data from real factories, shipyards, and logistics centers. The more robot startups enter manufacturing sites using Nvidia's platforms, the more industrial data and use cases accumulate. This, in turn, can drive broader adoption of Nvidia's robot and AI platforms. The reason Korea, with its diverse manufacturing base not only in semiconductors but also autos, shipbuilding, electronics, and steel, is cited as a stage for physical AI collaboration lies here.
In the industry's view, if Korea's primary role in early AI development was focused on the memory semiconductor supply chain, in the physical AI stage its significance is growing as a market that brings together chips, robot hardware, manufacturing sites, and industrial data. Over five days, Huang elevated Korea beyond a chip partner to a collaborative stage to implement physical AI on the industrial front lines. Interest has gathered. The remaining task is to connect this interest to actual investment and on-site demonstrations, and for domestic robot and AI startups to prove commercialization potential on manufacturing floors such as shipyards and factories.