(21) Son Jaegwon, CEO of The Milk
The Hidden Engine of a City That Attracts "Crazy-Talented People"
The Essence of the AI Power Race Lies in "Structure," Not Technology
A World-Class Innovation Hub Built on Talent, Capital, and Environment
Son Jaegwon, CEO of The Milk, is a veteran journalist with over 20 years of experience, including a tenure as a Silicon Valley correspondent. After serving as a visiting scholar at Stanford, he founded The Milk in 2019, operating a cross-border media platform connecting Korea and the United States. As a reporter specializing in IT, he covered companies like Intel and Samsung Electronics, immersing himself in the frontlines of technology and innovation. Why did he leave behind a stable career in journalism to pursue entrepreneurship? From Jensen Huang's recent visit to Korea and the 260,000 GPU investment, to the raw realities of Silicon Valley in the AI era, we listened to his firsthand accounts from the forefront of innovation.
Son Jaekwon, CEO of The Milk, is having a conversation with Kim Daesik, a professor at KAIST, on the 3rd at the Artist's House Lounge in Jongno-gu, Seoul. Photo by Kang Jinhyung
-What motivated you to leave behind 20 years as a journalist and embark on entrepreneurship?
▲As a journalist, I always believed I should be a 'practical intellectual.' Knowledge alone isn't enough-you have to put it into action. Over 20 years of covering Silicon Valley, I became curious about why people like Mark Zuckerberg and Sam Altman keep creating new things. Through in-depth reporting, I realized the secret lies in entrepreneurship.
Especially when faced with massive changes like AI, I felt, "This isn't just something to report on-it's an issue that demands action." How could I not care when my own livelihood was at stake? The internet, PCs, mobile, and social media are all tools for media communication. I believed that as a journalist, I had to deeply explore and respond to these changes. With the advent of YouTube and podcasts, it was time for the journalist's role to evolve, which is why I founded The Milk in 2019.
The San Francisco and Silicon Valley areas are the world's only innovation hubs that concentrate technology, capital, talent, and living environment. CEO Sohn explained that this region has been the core of the IT industry for decades due to the synergy of its compensation system, environment, and first-mover strategy. Photo of the San Francisco skyline. Photo by Getty Image
-What makes Silicon Valley still so special? Many places around the world have tried to emulate it but have failed.
▲The core is a system that attracts 'crazy talent.' But it's not just about having a lot of talent; there's a force that drives them to excel. The first is money-the compensation system. Talented people aren't afraid of competition; in fact, they want to be around those who are better than them so they can learn and grow. Silicon Valley is the place that has optimized incentives, which are a fundamental human drive.
The second factor is the environment. The weather is great, there are top universities like UC Berkeley, the world's best golf courses, beautiful beaches, natural wonders like Yosemite, and tourist attractions like San Francisco. It satisfies not only intellectual curiosity but also quality of life, making people want to take on challenges there.
Son Jaegwon, CEO of The Milk, is having a conversation with Kim Daesik, a professor at KAIST, on the 3rd at the Artist's House Lounge in Jongno-gu, Seoul. Photo by Kang Jinhyung
The third is the 'first mover' effect. The gap was created by starting 20 years ahead. When Amazon launched its online bookstore in 1996, hundreds of companies tried to compete, but the first-place finisher secured the market and talent. People in the IT industry understand the importance of being a first mover, which is why there is such frenzied investment in AI now. And the essence that makes all this possible is the 'power of determination.' Once they decide to do something, they make it happen. Whether AI will change the world or not-no one really knows. But they have made up their minds, and once they commit, they make it a reality.
-You've met key figures leading the AI era, such as Sam Altman, Jensen Huang, and Elon Musk. What are they like?
▲Sam Altman is extremely introverted. He speaks quickly and is an introvert. He has a strong sense of identity and accepts the flow of his own DNA. What stands out most is his learning ability and capacity for rapid adaptation. As president of Y Combinator, he met countless startup founders and listened to their stories. Even when he thought, "Why are there so many pointless stories?" he still listened to them all. And after a few days, those stories became part of him.
Sam also has a strong belief in AGI (artificial general intelligence). What he writes on his blog is exactly what he says at conferences. He truly believes in it. He has a strong faith in a technological utopia. Mark Zuckerberg is the same. That's why they named it 'OpenAI.' Even though it has become closed AI now, they still try to keep the product price at $10. They believe that's what gives the name its value.
Sam Altman, CEO of OpenAI, the developer of ChatGPT, is visiting Korea on the 1st through the Seoul Gimpo Business Aviation Center (SGBAC) in Gangseo-gu, Seoul. Photo by Yonhap News
On the other hand, Jensen Huang is the complete opposite. He is warm-hearted and highly attuned to social context. He has conviction, faith, the strength to overcome adversity, and knows how to care for others while also letting go. This isn't just my opinion; it's what NVIDIA employees commonly say. Born in Taiwan, he moved to the United States, graduated from Washington State University, and even worked as a server at Dennis' Restaurant in Silicon Valley. He came from a humble background, yet possesses all the qualities of an entrepreneur without losing his humanity. As for Elon Musk... I haven't met him in person.
-Jensen Huang recently invested 260,000 GPUs in Korea. What does this mean?
▲This is not just a business transaction; it's a matter of national competitiveness. In today's world, national strength is determined by compute power. AI now dictates military power, economic power, and productivity. Just look at the AI drones in the Ukraine war. The core of compute power is the GPU, and 260,000 of the latest chips have come to Korea. It's an enormous gift.
There are several reasons for this. First, China is currently under export restrictions, so they can't sell there. That means a country in the Asia-Pacific region is needed to serve as a GPU hub. Taiwan has already decided to build an R&D center but faces geopolitical risks. Japan has legacy systems and is slow in digital transformation. Jensen evaluates Korea as "a country strong in hardware, software, and services." Plus, Korea is strong in manufacturing.
Samsung Electronics Chairman Jay Y. Lee (left), Jensen Huang, CEO of Nvidia, and Hyundai Motor Group Chairman Chung Euisun are greeting each other after finishing a "chimaek" gathering at a chicken restaurant in Samseong-dong, Seoul on the evening of the 30th. 2025.10.30 Photo by Jo Yongjun
In particular, NVIDIA's roadmap has shifted to physical AI. After agents, the next step is physical AI. Which country is best positioned to move into physical AI? Germany, Japan, and Korea. But Japan is weak in software and isn't a trendsetter. Korea, on the other hand, is a country that sets trends, like K-POP and K-fashion. Korea also has robot-based technology, has developed platforms like Naver, and has created games. From Jensen's perspective, Korea is the country where Chairman Lee Kunhee proposed the Game Olympics in the 1990s-the country where GPUs were used in PC gaming rooms. He saw that potential.
Of the 260,000 GPUs, 50,000 will go to Hyundai Motor, 50,000 to SK, 50,000 to Samsung Electronics, 60,000 to Naver, and the remaining 50,000 will be used by the government.
Now the question is what we will do with them.
-Where should Korea focus to remain competitive in the AI era?
▲Physical AI. Competing with the United States or China in software AI, like large language models (LLMs), puts us at a disadvantage in terms of scale. But there are areas where we can excel. Just as Israel found its strength in cybersecurity, our strength is in physical AI-the convergence of robotics and manufacturing.
Son Jaegwon, CEO of The Milk, is having a conversation with Kim Daesik, a professor at KAIST, on the 3rd at the Artist's House Lounge in Jongno-gu, Seoul. Photo by Kang Jinhyung
There are only four countries with robot-based technology: China, Japan, Germany, and Korea. Among them, Korea excels at both hardware and software. Korea is better than Japan in software and has a sense for services. Even Americans perceive Korea as being good at creating "physical things." In Silicon Valley, you often see Hyundai and Kia cars like the Ioniq, Palisade, and Telluride.
Americans say that only Korea can create "charismatic and attractive products" like K-POP and Korean dramas. If we combine that tech sensibility with manufacturing capabilities, we can achieve global competitiveness in physical AI.
-How do you assess China's level of AI technology?
▲In terms of technology, talent, and government policy, China is world-class in all three areas. The emergence of models like DeepSeek and Qwen proves their technological prowess. Even under AI export bans, they produce world-class products in every area, from Huawei smartphones to humanoid robots.
The talent level is beyond question. In Silicon Valley, more than half of the engineers at Facebook and Google are Chinese. Chinese researchers publish the most AI papers. Government policy is also overwhelming. Only China and North Korea have five-year plans now. Korea used to have economic development five-year plans, but not anymore. China still maintains them, picks winners, and invests heavily. They are fully integrating technology and the market.
In almost every future technology field-quantum computing, space, robotics, nuclear fusion-China is ahead of the United States. The same goes for biomedical. Visit a biology department in the U.S. and you'll see how many Chinese students there are. China and the United States are the two top players.
Son Jaekwon, CEO of The Milk, is having a conversation with Kim Daesik, a professor at KAIST, on the 3rd at the Artist's House Lounge in Jongno-gu, Seoul. Photo by Kang Jinhyung
-What is your view on the AI bubble controversy?
▲I think there is some wishful thinking mixed in. On Wall Street and in Europe, there are voices hoping "it is a bubble." They can't keep up with the pace, so they wish for a correction, hoping to get a chance to enter the market if it happens. Opportunities arise in times of transition, and since they missed the initial wave, they're asking for another chance.
But historically, bubbles burst when people least expect it. They pop when everyone is complacent, not when people are being cautious. The very fact that wishful thinking is surfacing now is, in my view, a sign that the bubble may persist even longer.
The key is demand and productivity. The real issue is whether it will actually boost GDP. Both Satya Nadella of Microsoft and Sam Altman say that demand will not decrease. Right now, the problem is a shortage of GPUs. Of course, it's true that some of the latest chips, like Grace and Blackwell, are left unused. There aren't enough applications yet to utilize that much computing power. Things are moving too fast. Currently, they're used for inference, but that doesn't require such high specs.
But only countries like Korea, Japan, Germany, the UK, and France can use such high-end chips. Only a limited number of countries can provide the necessary energy infrastructure. I think the bubble won't burst for at least another two to three years, especially until OpenAI's Stargate project is completed in three years. Until then, the momentum will continue one way or another.
Son Jaegwon, CEO of The Milk, is having a conversation with Kim Daesik, a professor at KAIST, on the 3rd at the Artist's House Lounge in Jongno-gu, Seoul. Photo by Kang Jinhyung
-Finally, if you could go back to that moment, would you choose entrepreneurship again? And do you have any advice for young journalists?
▲Even if I could go back, I think I would choose entrepreneurship again. I have never regretted the decisions I've made. I have survived for six years already, and I've learned so much along the way. Although I am not a perfect entrepreneur, the insights I have gained by transitioning from journalist to founder are invaluable. By practicing the entrepreneurship I learned in Silicon Valley, I am proving my belief that 'an intellectual should not just know, but also act.'
What I want to tell young journalists is that the role of journalists is fundamentally changing. Simply relaying what happens on the ground is no longer enough. In a world where anyone can speak directly through YouTube, podcasts, and social media, journalists must provide deeper context and insight. Entrepreneurship can be one way to put this into practice.
However, if you are considering entrepreneurship, I hope you won't just copy existing models. There's no point in creating another similar media outlet. You should start a business only when you have discovered a 'real problem' as a journalist and have an original way to solve it. The goal is not to divide up the market, but to create new value.
If you have the ability to solve problems with a results-oriented approach, execution capability, and your own unique solution, then entrepreneurship is well worth the challenge. The important thing is creation, not imitation. That is the greatest lesson I learned in Silicon Valley and the principle I am still striving to practice today.
Kim Daesik, Professor at the School of Electrical Engineering, KAIST · Kim Hyeyeon, Choreographer and CEO of Yeonist
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.

