[Market Pulse] Why a Korean AI-Dedicated Supercomputer Must Be a National Design Project

World-Class Semiconductor Competitiveness
Yet to Be Linked to System Design
Government-Led Consortium Is Essential

Youngsun Park, former Minister of SMEs and Startups

Is it possible for South Korea to realize a domestically designed supercomputer dedicated to artificial intelligence (AI)? While this question may seem to be a technical issue, it is, in fact, a matter of national policy choices. Japan and Taiwan have already pursued or are currently pursuing such initiatives. Only South Korea has yet to begin.

Japan has defined the supercomputer as a national scientific infrastructure. The supercomputer Fugaku, which Japan proudly claims, was conceived not simply to achieve the highest performance, but to accumulate the capability for Japan to independently design, operate, and improve computing architectures. As a result, this initiative propelled the Fugaku supercomputer to the ranks of the world's top performers.

Taiwan has chosen a different path. Rather than insisting on design sovereignty, Taiwan aims to become the country best at manufacturing and integrating global AI systems. Within the NVIDIA GPU-centered ecosystem, Taiwan has secured an irreplaceable position in packaging, power efficiency, and system integration. Currently, Taiwan is working with NVIDIA to build an AI supercomputer based on 10,000 GPUs, with the goal of becoming the world's leading AI manufacturing nation. If all goes according to plan, it will be operational this year for use in AI factories.

South Korea's Issue Is Not 'Capability,' but 'Connection'

South Korea possesses world-class competitiveness in memory semiconductors, including high-bandwidth memory (HBM). In AI supercomputers, memory is not a subsidiary component but a crucial factor determining performance and power efficiency. Nevertheless, South Korea has failed to connect this strength to system design. This is not due to a lack of technology, but rather a lack of policy.

South Korea's industrial policy has succeeded in strengthening competitiveness at the component level, but it has failed to address HBM, neural processing units (NPUs), systems, and software as an integrated design challenge. For South Korea to become one of the top three AI powerhouses, it must develop a "Korean-style AI supercomputer." There is a significant difference in industrial impact between countries that have experience designing supercomputers themselves and those that do not.

Therefore, the government must adopt the development of purpose-built AI supercomputers for national strategic industries as a national project. Rather than focusing solely on computation, South Korea needs AI supercomputers with architectures optimized for memory and data movement, specialized for large language model (LLM) training, manufacturing and process AI, and defense and energy simulations. This is the most direct way to convert South Korea's HBM competitiveness into system sovereignty. It also provides a valuable opportunity to create global reference cases for domestic fabless NPU companies such as FuriosaAI, which are just beginning to blossom.

AI Supercomputers Must Be a National Project

An AI supercomputer is not a project for a single company. The government should lead the formation of a permanent design consortium. The participants should include memory, fabless, and system companies, government-funded research institutes, universities, and researchers specializing in compilers and AI frameworks. This cannot be managed under a short-term performance evaluation system. It must be managed as a long-term national infrastructure project spanning a decade.

Most importantly, the strategy linking HBM, NPU, and AI supercomputers must be integrated as a national project. This is South Korea's critical differentiator. South Korea is almost the only country capable of simultaneously possessing and nurturing both HBM and next-generation NPUs. Only by doing so can South Korea move beyond being a mere "component supplier" and become a "country that designs AI computing architectures."

It is also essential to design a budget structure that assumes the possibility of failure. Designing an AI supercomputer may involve failures. The problem lies in a system that does not tolerate failure. The scheme that defines audit standards solely by "results" must also change. Without such institutional reforms, the birth of a "domestically developed supercomputer" will not be easy. The reason a South Korean-designed AI-dedicated supercomputer has yet to emerge is not due to technological difficulty, but rather fear of failure, an institutional and cultural unwillingness to accept failure, and issues of technological integration and decisive action.

Will South Korea Become a Design Nation or Remain a Component Supplier?

Japan chose design, while Taiwan chose optimization. South Korea must now decide whether to become a design nation by integrating HBM, NPU, and systems, or to remain a country that continues to supply the world's best components.

An AI supercomputer is not a calculator. It is the blueprint that determines the national industrial structure in the AI era. If the Ministry of Science and ICT cannot make this decision, then the National AI Strategy Committee must set this long-term agenda. The National AI Strategy Committee should serve as the flagship that leads the entire AI industry with a broader perspective and establish this as a symbolic agenda of the Lee Jaemyung administration.

Youngsun Park, former Minister of SMEs and Startups