[Editorial] To Nurture a Korean Nvidia, Strengthen Semiconductor Fundamentals First

The government unveiled its plan on the 18th for distributing advanced graphics processing units (GPU) to be supplied by Nvidia Corporation. Starting in February next year, the GPUs will be prioritized for small and medium-sized enterprises (SMEs), startups, academia, and research institutes. Schools and research centers will receive them free of charge, while companies will pay 5–10% of the market price. The government has agreed to receive 50,000 GPUs from Nvidia Corporation by 2030, and next year, it will deploy 10,000 units purchased this year for 1.46 trillion won to the field.
Nvidia Corporation's advanced GPUs can significantly enhance the speed and capacity of artificial intelligence (AI) computations, serving as a catalyst for AI research and development in Korea’s academic, industrial, and research sectors. The government stated that it would use this as a foundation to build an 'AI Highway' and encourage the development of proprietary AI Foundation Models. Furthermore, it announced plans to accelerate the development and promotion of domestic Neural Processing Units (NPU) to compete with Nvidia Corporation’s GPUs.
To lead the AI transformation, the government must exhaust every possible measure. Optimizing the allocation of government-assigned GPUs to maximize their utility and using them to strengthen the AI infrastructure is a fundamental responsibility. Becoming one of the world’s top three AI powers is an attainable goal through this process. However, there is another crucial aspect that must not be overlooked: assessing whether Korea’s advanced industries possess sufficient strength and competitiveness.
Without semiconductors and advanced products, Korea’s economy would struggle to grow or export. Many of the country’s leading manufacturing sectors have already lost ground to China’s aggressive pursuit. Even the semiconductor industry is no longer secure, raising further alarm. Leading semiconductor scholars and experts have expressed concerns about Korea’s ability to maintain a strong position in future markets. The Semiconductor Special Committee of the National Academy of Engineering of Korea (NAEK) recently held a forum and bluntly stated, 'Korean AI semiconductors have no presence except for High Bandwidth Memory (HBM)'.
The AI semiconductor market is rapidly expanding. According to Gartner, Inc., the GPU market will grow from $70 billion this year to $180 billion by 2030. The Application Specific Integrated Circuit (ASIC) market is expected to increase from $21 billion to $70 billion, and HBM from $20 billion to $60 billion. Aside from HBM, Korea’s competitiveness is considered minimal, making it essential for both the public and private sectors to enhance their capabilities. NAEK pointed out that the domestic AI ecosystem is fragmented, with each sector advancing independently. It proposed a national-level 'K-Big Tech Alliance' involving hardware companies such as memory manufacturers, cloud service providers, and software firms. This is an issue worth serious consideration.
A robust, government-led AI ecosystem is necessary to attract global investment. According to an analysis by the Korea Chamber of Commerce and Industry (KCCI) of Organization for Economic Cooperation and Development (OECD) venture investment statistics, Korea’s share of AI venture investment through the third quarter of this year accounted for only 1% of total global funding. Compared to major countries, this is just one-seventy-third of the USA, one-seventh of the UK, and one-sixth of China, ranking Korea around ninth in the world.
A complex and unpredictable regulatory environment is a key factor deterring major global investors. Unlike the USA and the European Union (EU), which implement flexible policies tailored to market conditions, Korea’s labor and industrial policies are often seen as rigid. Requests to allow flexible working hours for advanced research positions have made no progress due to opposition from the labor union. While advanced GPUs are important, fostering an innovative mindset to enhance corporate and industrial competitiveness is equally essential.