'Father of HBM' Kim Joungho of KAIST Offers Urgent Assessment
Despite Global Leadership in Profits and HBM, Korean Firms Remain Undervalued
The Market Must Abandon the Outdated View That "Memory Equals a Cyclical Industry"
Apple's Use of Ch

"Going forward, the core of AI lies in memory, but it seems the global capital markets are refusing to recognize this."
Professor Kim Joungho of KAIST was interviewed by The Asia Business Daily on the 31st of last month at his research lab in Daejeon. Photo by Paek Jongmin, Tech Specialist

Professor Kim Joungho of KAIST was interviewed by The Asia Business Daily on the 31st of last month at his research lab in Daejeon. Photo by Paek Jongmin, Tech Specialist

View original image

On the 31st of last month, Professor Kim Joungho of KAIST's School of Electrical and Electronic Engineering, whom we met at his research lab in Daejeon, directed his concerns not toward his research topics, but toward other pressing issues.


Although the core of AI has shifted to memory, the global capital markets continue to treat Korean companies as cyclical firms, affected primarily by price and production volume fluctuations. He warned that unless Korea succeeds in changing this perception—and achieves proper valuation despite possessing world-class profitability and High Bandwidth Memory (HBM) technology—even its technological leadership could be threatened.


Professor Kim emphasized that Korean companies must continue to lead AI memory technologies through both technology and management for at least the next ten years to change market perceptions. From an industrial standpoint, he pointed to a more direct threat than China’s own advancement: Apple’s adoption of Chinese memory chips and its support of Chinese memory makers, which is enabling local firms to climb the value chain. To counter this, he proposed both lobbying efforts in the US and the need for investments and production capacity inside the United States. The following is a Q&A with Professor Kim.


- How do you view the recent sharp decline in domestic memory semiconductor stock prices?


△ Samsung Electronics and SK hynix generate some of the highest net profits in the world. Looking toward the future, memory—including HBM and HBF—is at the heart of AI. But it appears that the global capital markets are not acknowledging this reality. Compared to companies like Micron, TSMC, and Nvidia, Korean firms are valued much lower. It's as if the markets do not recognize memory as a core component of AI systems.


- Why do you believe only memory companies are undervalued?


△ The markets continue to see memory as a cyclical business—one where production can be adjusted and prices fluctuate accordingly. Meanwhile, companies such as TSMC and Nvidia are regarded as irreplaceable. Even though Samsung Electronics and SK hynix post large revenues and net profits, their corporate value remains low. The old logic of market cycles is still used to undervalue them, even in the AI era.


- What needs to be done to change these market perceptions?


△ The strength of our domestic capital is not substantial. Ultimately, it is only through technological leadership and business management that we can maintain our edge for another ten years. We need to prove to the market that memory is central to AI systems. There seems to be no other viable approach.


- What do you perceive as the most significant industrial threat?


△ It's Apple’s use of Chinese-made memory chips. Initially, this may be confined to phones sold in China, but if Apple also provides technical support, it could significantly boost the market share of Chinese players such as Changxin Memory or Yangtze Memory. This would deal a direct blow to our companies. Thus, the greater concern isn’t just the rise of Chinese companies themselves, but the fact that Apple is helping nurture and support them.


- How should Korean companies respond?


△ Politically, it may be necessary to mobilize the US Congress to prevent this from happening. To do so, our companies should also invest in the United States. For example, when expanding production capacity in the future, a portion of this increased capacity should be produced in the United States.


- What if China catches up in HBM technology?


△ Over the next three or four years, China will probably be able to manufacture something like HBM3. Once China can meet its own internal demand with domestic products, we will have to move forward again. That means progressing to HBM6 and HBM7, combining those with HBF and HBS, and coupling them with GPUs. We will have to keep innovating relentlessly.



Professor Kim also expressed concern about the situation where young investors, who invested in leveraged financial products during the recent stock price downturn, suffered even greater setbacks. He noted that there is a possibility for the net profits of Samsung Electronics and SK hynix to be maintained or even increase over the next two to three years, highlighting the importance of a patient, long-term approach—focusing on the companies' technology and profitability rather than short-term trading.


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

© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.

Today’s Briefing