[Column] "In the AI Era, What Is Scarier Than a Bubble Is Institutions That Cannot Keep Up"

(24) Lee Junpyo, CEO of SBVA
Venture capital investing in tech startups
Feeling AI-driven changes faster than most
Need to focus on areas where small elite teams can generate multiplicative impact

Editor's noteThe Asia Business Daily has decided to explore, from the perspectives of both engineers and artists, what changes ever-evolving generative artificial intelligence (AI) will bring to the field of artistic creation, and what kinds of questions "people" should be asking. To this end, we have created a monthly feature in which Kim Daesik, professor at the School of Electrical Engineering at KAIST, and choreographer Kim Hyeyeon (CEO of Yeonist) hold conversations with artists or discuss works. The "AHA" in the title of the series stands for "AI, Human & Art." We hope that, through Professor Kim Daesik, who is exploring the future of generative AI more passionately than anyone else, and choreographer Kim Hyeyeon, who is boldly combining generative AI with dance, you can take one step closer to the weighty question of AI, people, and art.

CEO Lee Junpyo leads SBVA (formerly SoftBank Ventures), one of Korea's leading venture capital (VC) firms with a 26-year history. He is someone who has experience both as a founder who once received investment and as an investor who has witnessed the rise and fall of numerous tech companies. Today, standing before the giant wave of artificial intelligence (AI), he says he feels "more tension than excitement." In this transition period from the broadband era to generative AI and physical AI, he is sensing the temperature of change earlier than most. As the final interviewee of The Asia Business Daily's "AHA" series, we asked him about our jobs and social structures that may be swallowed by technology, and about the opportunities that remain.

Lee Junpyo, CEO of SBVA, is in discussion with Kim Daesik, a professor at KAIST, on the 5th at SBVA headquarters in Gangnam-gu, Seoul. Photo by Kang Jinhyung

-As CEO of SBVA, what role are you playing, and from what vantage point are you observing the AI industry?

▲SBVA was founded in 2000 as a venture capital firm under the SoftBank Group and has been investing in technology startups in Korea and abroad. Since then, we have supported the growth of various tech companies, mainly in Korea and across Asia, and in 2023 we transitioned into an independent investment firm. I myself am a former founder who received investment from SoftBank in the past, and now, as CEO, I am responsible for identifying the potential of early-stage and growth-stage companies and working with them on their global expansion strategies.

Having sat in both seats, as founder and as investor, I see the current AI industry not as a simple technology trend but as a structural transformation. These days, I feel more tension than excitement. Watching this pace from the very front lines of industry, the feeling is less that we are "choosing" and more that we are being "swept along."

-How are you personally experiencing the recent speed of the AI industry?

▲The unit of change has become completely different. In the past, technology needed time to take root in the market, but now, once a model is announced, corporate strategies and investment flows move immediately. Changes occur not on a yearly basis, but by quarter, or even by week. The speed at which capital is being deployed is incomparably faster than before. Technology development, infrastructure expansion, and global diffusion are happening simultaneously. It is no longer a structure where one stage stabilizes before moving to the next; multiple stages are progressing in parallel. This is why the entire industry is in a constant state of tension. The pace is so fast that even just keeping up is daunting.

-Do you think the transformation of employment structures due to AI has already begun?

▲The signals of change are already visible. In the recruitment market, entry-level positions are rapidly disappearing, and a substantial portion of the analysis, research, and reporting work that used to be handled by juniors is now being done by AI. Repetitive work such as customer service is being automated first. On the surface, it may look like efficiency is improving, but the more fundamental change is that the "learning stage" is being weakened.

In the past, people grew into experts by accumulating experience in an apprenticeship-like manner. They learned the field while performing small roles, and gradually took on more complex work, building up their capabilities. But if that entire middle stage is compressed, the pathway through which expertise is formed itself may change.

Once the value structure of knowledge work changes, it is likely to affect channels of social mobility, the so-called pathways for moving between social classes. If the belief that "you can move up if you work hard" erodes, social tension will inevitably rise. This is not a far-off hypothetical future; it is a phenomenon already being observed in some industries.

-Some say the AI investment boom is a "bubble." What is your view?

▲There are clearly bubble-like characteristics. Capital is flooding in quickly, and it is true that the valuations of some companies are overheated. It does evoke memories of the dot-com bubble. But there is also a decisive difference. Back then, there was still doubt about whether the internet could truly transform industries, whereas today AI is already changing corporate productivity and cost structures. We are not in an experimental stage; AI has deeply penetrated industry.

Even after the dot-com bubble burst, fiber-optic cables and data infrastructure remained, and they became the foundation of today's digital ecosystem. This time, too, some companies will disappear, but the models, data, and computing infrastructure are highly likely to remain. That is why, to me, the more important question is not whether there is a bubble, but what will remain afterward and what we are preparing for.

-Do you think our society is sufficiently preparing for this speed of change?

▲There is active discourse about technological progress, but discussions about how that change will affect employment structures, the education system, and the social safety net are not yet deep. AI is not an issue for specific industries alone; it is a technology that directly impacts labor and income structures. Yet our institutions are still designed on the premise of past industrial structures.

Technology advances exponentially, while policy and institutions move linearly. The wider this gap becomes, the greater the shock will be. I worry less that AI will directly threaten humans, and more that our social systems may fail to absorb its speed. If labor and income structures are rapidly reorganized, it can lead to political and social turmoil. We could move in a more prosperous direction, but that path will not open automatically; it requires preparation and deliberate design.

-How are leaders in Silicon Valley viewing this transformation?

▲When I met Sam Altman, what struck me was that he regarded this transformation not as a mere possibility, but as a flow that is already unfolding in reality. There was a shared premise that AI will cause a sharp increase in productivity, and the discussion naturally extended to the changing meaning of work and shifts in income structures. The reason universal basic income is being seriously mentioned also lies in that context.

However, they do not believe there will be no shock. Technology moves quickly, but society does not change at the same pace. In the end, the crucial question is not whether to stop technology, but how to design structures that can absorb its speed.

Lee Junpyo, CEO of SBVA (center), is speaking with Kim Daesik, professor at KAIST (left), and choreographer Kim Hyeyeon at SBVA's headquarters in Gangnam-gu, Seoul on the 5th. Photo by Kang Jinhyung

-In that case, where do you think Korea stands right now?

▲If you look only at the speed of technology adoption, Korea is an extremely fast-moving country. When new services appear, we test, apply, and modify them with great agility. Even in the broadband era, we did not possess all the core technologies, but we were among the fastest to develop services and content on top of them.

However, adoption and leadership are different. We are highly capable of quickly adapting on top of global models and infrastructure, but we are not in a structure where we directly control the core technologies and computing infrastructure. That gap may widen over time. I see this as a critical juncture: will we remain a testbed, or will we secure at least part of the technological initiative?

-What do you see as the next mega-trends?

▲It is difficult to predict where AI will be in five years. However, the spread of physical AI may be faster than expected. I recently witnessed firsthand in China the scale of government-led investment and its execution power. In robotics, manufacturing infrastructure and data are combining to drive rapid advancement. At a U.S. robotics company, I also saw technology that trains on the hand movements of master craftsmen through sensor gloves. It was at a level that could reproduce not just simple repetition, but also delicate, intricate tasks.

Automation is likely to accelerate not only in knowledge work but also in physical labor. The flow of industry will ultimately converge on the energy question. Nuclear fusion, high-efficiency batteries, and energy storage technologies may become the foundational infrastructure of the AI era. As Sam Altman puts it, AI is a technology that converts electricity into intelligence. The core axis of competition will be how reliably you can convert electricity into computational power.

What is interesting is that AI is also being used in research on these very energy technologies. We are seeing a structure in which AI is exploring the technologies that will support the AI era itself.

-What advice would you give to people in their twenties, especially young people considering starting a business?

▲Paradoxically, I think this is still a period when the window of opportunity remains open. With AI, five people can now do what once required a hundred. Even with limited capital and manpower, you can build products if you use the tools well. But this window will not remain open forever. As technology and infrastructure become more advanced, the barriers to entry may rise again. That is why I think this may be the last window.

Rather than betting on large headcounts, you need to focus on areas where a small team can generate multiplicative effects. Technology is a tool, and the outcome will depend on how strategically you use that tool.

The conversation with CEO Lee Junpyo left a deep resonance about the "human domain" we must protect amid the runaway advance of technology. Over the past two years, "AHA" with Professor Kim Daesik of KAIST and choreographer Kim Hyeyeon has been an intense journey to find that point of contact. When this series began, AI was still the "future"; now it has become the "reality" in which we breathe. What matters now is not the speed of technology, but the speed of the people who must bear it.

What we need today is not hasty answers, but more questions, and ceaseless dialogue and communication among human beings. We sincerely hope that our reflections on connecting technology and people have become small "Aha" moments in the lives of our readers. We express our deepest gratitude to all the readers who have cherished "AHA" thus far.

Kim Daesik, Professor at the School of Electrical Engineering, KAIST · Choreographer Kim Hyeyeon, CEO of Yeonist

This content was produced with the assistance of AI translation services.