The government's strategy around next-generation power semiconductors has moved into full swing. The Ministry of Trade, Industry and Resources recently officially launched the Next-Generation Power Semiconductor Task Force and appointed Koo Sang-mo (51), a professor in the Department of Electronic Materials Engineering at Kwangwoon University, as Director General. Power semiconductors are essential components that control and convert electricity across core national infrastructure, including electric vehicles, the national power grid, artificial intelligence (AI) data centers, defense and robotics. The government set a goal of raising the technology self-reliance rate to 20% by 2030, up from the current 10%. This is seen as a policy shift to treat power semiconductors as national strategic infrastructure, beyond simple research and development (R&D) support.
Unlike memory and central processing units (CPU), which handle computation, power semiconductors deal with the flow of electricity itself. Recently, beyond conventional silicon (Si), compound power semiconductors such as silicon carbide (SiC) and gallium nitride (GaN) have emerged as key technologies. SiC is strong in high-voltage, high-power environments, making it suitable for EV drivetrains, ultra-high-voltage power grids, and high-power AI data centers, while GaN, with its high-frequency and high-efficiency characteristics, is used in chargers, power supply units, and telecommunications equipment. With the spread of AI aligning with a surge in energy demand, power semiconductors are drawing attention as a core strategic technology.
When we met Koo, the Director General, at Kwangwoon University in Nowon District, Seoul, on the 13th, he likened power semiconductors to "the heart that circulates energy." If power semiconductors do not function properly, not only EVs but the entire national infrastructure—from the national power grid and the energy highway to high-voltage direct current (HVDC) transmission and AI data centers—comes to a halt. He emphasized that "technology self-reliance" in power semiconductors is not about a simple localization ratio, but about building a structure in which domestically made power semiconductors are actually used and can be stably operated in core infrastructure.
Koo, the Director General, is an international researcher in compound power semiconductors and nanoscale devices, and has led research on SiC- and GaN-based power devices. After earning his doctorate from the Royal Institute of Technology in Sweden, he served as a visiting researcher at MIT and worked at NIST, then joined Kwangwoon University's Department of Electronic Materials Engineering as a professor in 2006.
The global competitive landscape of the current power semiconductor industry is not favorable for Korea. Europe has accumulated power semiconductor technology for decades on the back of its automotive industry and is leading SiC commercialization focused on EVs and industrial power systems. The United States also treats high-reliability, high-performance power semiconductors as strategic assets, anchored by demand from AI data centers and defense, maintaining a technological edge. In contrast, Korea is seen as a latecomer in power semiconductors due to an industrial structure focused on memory semiconductors. On top of that, China is rapidly catching up across SiC and GaN materials and devices with large-scale investment and volume offensives, raising concerns that the technology gap could widen further.
The task force's defining feature is a "demand-centered" approach. Whereas past semiconductor policy developed technology first and then searched for use cases, the task force aims to define required performance and specifications from the planning stages of demand industries such as EVs, power grids, AI data centers, and defense. The intent is to resolve, through policy, the reality that domestic customers had little incentive to replace foreign parts already in stable use. Koo summarized this as "demand first, development later," adding that linking public infrastructure with private demand at the same time is what sets the task force apart.
The task force plans to flesh out a power semiconductor technology development roadmap by the first half of this year. It will organize required performance by demand industry and, based on that, set mid- to long-term R&D directions across materials, processes, design, and foundry. In the second half, planning for large research and development projects and discussions on institutionalization will follow. It will also review ways to ensure that domestically made power semiconductors are actually applied in public domains such as power grids, data centers, and weapons systems through legal refinements and institutional improvements. It will focus on workforce development as well. The following is a Q&A with Director General Koo.
— Why "power semiconductors" now? What prompted the government to form a task force at this moment?
"Power semiconductors are not merely a matter of one component for EVs. If power semiconductors do not function properly, the entire infrastructure—from EVs beyond to the national power grid, the energy highway, HVDC, and AI data centers—comes to a standstill. As AI spreads and power demand soars, how efficiently we control and supply electricity has become national competitiveness. Power semiconductors are the heart, blood vessels, and muscles that circulate energy. Relying on the outside for this has limits in terms of energy sovereignty and technology sovereignty."
— In power semiconductors, what should be the criteria for judging "technology self-reliance" and "sovereignty"?
"It is risky to view self-reliance in power semiconductors only as a localization percentage. The key is whether it is actually used. We can call it self-reliance only if domestically made power semiconductors are applied in core infrastructure such as EVs, the national power grid, HVDC, and AI data centers, and can be stably operated over the long term. The standard should be technology that operates within national infrastructure, not merely technology that we can manufacture."
— Unlike memory semiconductors, many say power semiconductors hinge more on tight integration with demand industries.
"Until now, semiconductor policy has developed technology first and then pondered 'where it will be used.' The task force has reversed that order. From the planning stage of demand industries such as EVs, power grids, AI data centers, and defense, we first define what performance and specifications are needed. From the standpoint of domestic customers, there was no reason to swap out foreign parts already in stable use, and we aim to resolve that structural issue through policy. 'Demand first, development later' is the task force's biggest differentiator."
— Which demand customers is the task force prioritizing?
"EVs are important, but structurally larger demand lies in power grids and AI data centers. AI data centers require tremendous power, and high-efficiency, high-reliability power semiconductors are essential to supply it stably. Defense and robotics—so-called physical AI—are also bound to become important demand customers. Power semiconductors cut across specific industries and connect to national infrastructure as a whole."
— In SiC and GaN power semiconductors, where is Korea's weakest link?
"Rather than pointing to one or two spots, we need to look at the entire supply chain. We have some capability across materials, processes, design, and foundry, but there is a gap with Europe and the United States, and China is catching up quickly. Among these, materials must be addressed. It is not that we lack all technology, but dependence on overseas sources is very high. If we fail to internalize now, the industry itself could become barren in the future."
— How do you plan to advance process and foundry capabilities?
"Process and foundry capabilities do exist in Korea. The problem is that they are not being used extensively in practice. Technology advances when we produce a lot, domestic customers absorb it, and experience accumulates. That is why we emphasize public and private demand together. Research and development, mass production, and demand are not separate."
— Is the goal of a 20% technology self-reliance rate by 2030 realistic?
"The figure of 20% is less a goal than a minimum threshold we must surpass. From the current level, we absolutely must more than double. In talking with key players such as Samsung Electronics, SK Siltron, DB HiTek, LS, and Korea Electric Power Corporation, there is already a shared understanding of the need for power semiconductors. But it is not a problem any single corporation can solve alone. That is why a task force is needed."
— What differentiators are only possible under a task force framework?
"It looks not only at research and development but also at institutions and public demand together. We want to build a structure where domestically made power semiconductors are actually used in public infrastructure such as power grids, data centers, and weapons systems. Workforce development is the same. We must build a self-sufficient ecosystem. Power semiconductors are not about market size. They are about infrastructure, security, and sovereignty. With AI spreading and power demand surging, now is a decisive moment."