[One Year After the Kkanbu Meeting (3)] Expand Innovation, Business Opportunities and SME Adoption to Become an AI Industrial Powerhouse

Beyond being a major consumer of ChatGPT, South Korea must boost manufacturing competitiveness and foster new industries

Korean-made manufacturing AI cuts fuel costs at a steel plant by 400 million won...and expands overseas

Large corporations are adopting AX, but SMEs are taking a wait-and-see approach...citing costs and staff shortages

The challenge is to develop software tailored to specific industries

South Korea faces the challenge of moving beyond being a major consumer of artificial intelligence (AI) and becoming an industrial powerhouse that uses AI to strengthen its industrial competitiveness. Observers say the country should build on the AI infrastructure it has moved quickly to secure since the “Kkanbu meeting” last year, including graphics processing units (GPUs), expand AI adoption from large corporations to small and medium-sized enterprises (SMEs), boost manufacturing productivity, and create new industries and markets.

Kim Donghwan, CEO of 42Maru, pointed to a gap between companies, saying, “Large corporations are changing, but AI adoption by SMEs is still slow.” Choi Jaesik, a professor at KAIST, stressed that South Korea should turn the strengths of its industries into AI competitiveness: “NVIDIA sells the GPUs, but we make and supply the memory that is essential for them.”

AI runs the plant... Korean-made manufacturing AI also makes inroads overseas=On October 14, KG Steel announced that applying AI to the annealing furnace at its Dangjin plant had cut its liquefied natural gas (LNG) costs by about 400 million won over the past year. The technology was developed with INIZI, where Professor Choi serves as CEO. The AI analyzes the type and thickness of the steel plate, production speed, and other factors to predict the optimal temperature and adjust operating conditions. This has enabled more precise AI-based control of a process previously dependent on workers’ experience.

Korean technology is also beginning to expand overseas. After conducting a pilot test of INIZI’s heating-furnace optimization technology, Japan’s Nakayama Steel Works signed a contract to adopt it in December 2025. In its 2026 earnings presentation materials, the company estimated that the technology could reduce energy consumption per unit of production by about 4.2%.

Professor Choi emphasized that South Korea should expand innovation and business opportunities through AI in manufacturing, an area where the country has strengths. Rather than developing every AI technology independently, South Korea should also make use of advanced technologies from overseas platforms while building the competitiveness needed to solve problems on factory floors. Just as Korean HBM is used in NVIDIA GPUs, South Korea should develop differentiated technologies and services in manufacturing AI that it can supply to global markets.

KAIST Professor Choi Jaesik. Photo by Paek Jongmin, Tech Specialist.

Large corporations have started, but SMEs are taking a wait-and-see approach to AX=The problem is that AI adoption has yet to spread across industrial sites. “Private-sector SMEs are not yet actively pursuing AX,” Kim said. “Many are still hesitant, even after making preparations.” He explained that rapid technological advances and high development and implementation costs make it difficult to assess return on investment (ROI).

Small and medium-sized manufacturers face a heavy burden, from organizing their data and building AI systems to integrating them with existing equipment. They also lack specialized personnel and funding. Simply introducing AI software does not necessarily increase productivity.

In March 2026, the government also issued a joint funding announcement by the Ministry of Science and ICT, the Ministry of Trade, Industry and Energy, and the Ministry of SMEs and Startups, allocating 423 billion won to 11 programs, including AI agents and manufacturing AX. Kim stressed that the public sector should help create initial demand and, in turn, encourage greater private investment.

Kim Donghwan, CEO of 42Maru. Photo by Yoon Dongju

AI consumption ranks among the highest globally...the challenge is to turn it into industrial competitiveness=South Korea ranks among the world’s leaders in AI adoption. OpenAI announced in May 2025 that South Korea had the world’s second-highest number of paid ChatGPT subscribers, after the United States. South Korea also ranked 12th globally, with a generative AI usage rate of 40.6%, according to Microsoft’s “Second Quarter 2026 Global AI Diffusion Report.” That was more than double the global average of 18.8%.

However, high usage rates do not necessarily translate into growth in South Korea’s domestic AI industry. Professor Choi said that U.S. Big Tech companies had strengthened their dominance in South Korea’s consumer AI market since the Kkanbu meeting. The argument is that even as South Korea actively uses foreign AI, it should strengthen its domestic AI industry ecosystem and expand AI-driven innovation and commercialization in areas of strength, including manufacturing.

Kim pointed to the active adoption of AI in China and Taiwan and stressed that domestic companies should also accelerate AI adoption at industrial sites. “Even if you build AI well, you have to use it well,” he said. “AI must not become the goal.”

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