"Samsung Must Be Ready to Change Its Name and Defend Its Foundry" [Paek Jongmin's Shockwave]

Interview with Kim Joungho, Professor at KAIST
"Foundry Is Essential for Maintaining Semiconductor Competitiveness in the HBM Era"
"The Government Must Provide Infrastructure; Companies Need Creativity and Technological Understanding"
"Comp

Professor Kim Joungho of KAIST (center) is discussing with students in the laboratory. TerraLab

Professor Kim Joungho of the School of Electrical Engineering at KAIST opened the interview with a grave assessment. He acknowledged that while the Korean semiconductor industry has gained an edge through high bandwidth memory (HBM), he emphasized that artificial intelligence (AI) marks the beginning of a new order that will determine the fate of the economy, industry, and nation. He urged thorough preparation from the government, corporations, and individuals alike.

Professor Kim offered a candid evaluation of the current state of Korea’s semiconductor industry. “We have always stepped onto platforms created by others. During the era of central processing units (CPUs) and the mobile era, we only contributed as component suppliers?we never created the platform ourselves. Now, we need the creativity to design a new platform.” He warned that if Korea remains fixated on short-term achievements, even survival in the AI era cannot be guaranteed.

He also delivered sharp advice to corporate executives. “If the chief executive officer (CEO) and chief technology officer (CTO) do not truly understand technology, they cannot have insights into the future. Like Jensen Huang, they must study directly and be able to articulate both technology and vision.” He stressed that leaders must go beyond simply managing financial performance and must possess both technological understanding and creative leadership.

Regarding Samsung Electronics, which has fallen behind in the HBM sector, he offered direct advice. He said, “Samsung Electronics must be prepared to change its mission to AI. It cannot survive on memory alone. It must lead with total solutions. In particular, if it gives up on foundry, its overall competitiveness will collapse.”

This is because foundry technology is essential in the HBM development roadmap Professor Kim envisions. He explained that the capabilities of an integrated semiconductor company are the foundation for securing future leadership in HBM. He emphasized, “Even in the HBM era, foundry is key. It must never be abandoned.” He warned that this could become a national issue, not just a problem for Samsung.

Professor Kim also cautioned against excessive dependence on Nvidia. “Alternatives such as AMD, Intel, Google, and Meta must be fostered.” The implication is that competitors should be developed so that AI solutions cannot be created without going through Samsung.

Professor Kim stated, “HBM is not just simple memory; it is a national strategic asset.” He said, “HBM must be managed at the national strategic level. The era of selling memory cheaply and in large quantities is over. From now on, whichever country secures HBM first will directly determine its survival.” He proposed that the nation should strategically utilize HBM.

The core issue is leadership. He noted, “The United States is checking China under the pretext of security, while China is developing its own independent technology in response. Korean companies caught in between are facing pressure from both sides.” He expressed concern that “someday, the United States may even demand the relocation of headquarters and factories.” Therefore, he argued that Korean semiconductor companies must move beyond simply following and secure leadership by devising creative strategies to establish new platforms.

He also clarified the role of the state in the AI era. “The government should not cling to short-term projects but focus on building infrastructure. Only with foundational facilities such as supercomputers, data centers, power grids, and research equipment can companies and laboratories compete.” He emphasized that the government must support talent development and fundamental research while building national strategic infrastructure from a long-term perspective. He also suggested that companies could secure large-scale graphics processing units (GPUs) and share them with universities to support students’ research. He noted that in the United States, as the AI era unfolds, collaboration between companies with abundant GPUs and university researchers is increasing.

He took a critical stance on the recent trend of domestic universities rushing to establish AI departments. Professor Kim stated firmly, “Departments that teach only AI are not the answer.” Instead, he said that interdisciplinary talent capable of integrating existing disciplines with AI is needed. He asserted that individuals who can understand both AI and semiconductors, as well as those with a strong mathematical foundation, will ultimately lead the AI era.

Professor Kim repeatedly emphasized the importance of mathematics education. “The essence of AI and semiconductors is mathematics. Without a mathematical foundation, neither application nor integration is possible.” He diagnosed that mathematical thinking forms the basis of AI algorithms and semiconductor design, and that simple programming education is not enough to compete globally.

Fostering the next generation of scholars is a non-negotiable priority for Professor Kim. He mentioned that he recently received the highest teaching evaluation among his department’s professors. He promised the reporter that if he receives the top evaluation again in the fall semester, he would buy dinner.

At the end of the interview, he briefly shared his plans for the later years of his life, which are also connected to education. “When I am in my seventies, I want to open a small office in Daechi-dong and provide counseling to students about AI and semiconductors.” Even when he can no longer conduct research, he hopes to guide students toward science and engineering fields.