The government is pursuing an artificial intelligence transformation (AX) initiative to strengthen the competitiveness of small and medium-sized foundational manufacturers, but companies are abandoning AI adoption due to the burden of high self-funding requirements. As the cost burden is too heavy for individual firms to adopt AX on their own, calls are growing to shift policy toward supporting joint development, verification, and maintenance centered on cooperatives whose industry characteristics and processes are similar.
According to a survey conducted by the Korea Federation of SMEs on 14 cooperatives and 25 companies that had considered the "AX Sprint Support Project," which helps small and medium-sized enterprises transition to AI, 18 respondents (72%) said the self-funding requirement for adopting companies—amounting to 30% of total project costs—was excessive, the federation said on the 14th.
Thirteen respondents (52%) said the current standard, which recognizes only half of the self-funding amount as in-kind contributions, was burdensome. The in-kind recognition ratio refers to the rate at which a company's self-funding obligation in a government support project is recognized by converting labor, equipment, facilities, and existing technology into monetary value instead of cash. Ten respondents (40%) also said they found the structure burdensome, as companies must first execute project costs before receiving support.
Of the 13 cooperatives that abandoned the AX Sprint project, more than half—7 (about 54%)—cited the excessive burden on adopting companies as the reason.
Some companies also felt burdened by regulations such as permits and licenses during the AI adoption process.
An official from a surface treatment cooperative said, "In the surface treatment sector, if changes occur in processes during AI adoption and verification, permits must be reobtained due to chemical and environmental regulations, which can take more than three months. Since it is impossible to operate the production process during that period, the permit costs and production disruption burdens that companies must bear to adopt AI are very large."
Still, the survey confirmed strong interest in AI adoption on the ground to improve productivity. Seventy-two percent of responding companies (18) said they had not yet adopted AI but were considering it. Among companies that had already adopted AI, 24% (6) wished to expand further. The most-selected necessary policies were direct support for adoption costs such as AI vouchers (64%, 16 companies) and support for building data infrastructure (48%, 12 companies).
The foundational manufacturing cooperatives that participated in the survey—covering fields such as molds, surface treatment, welding, and machinery—emphasized that AI is not an issue for individual companies alone but infrastructure that raises productivity and competitiveness across the entire industry. They argued that AI could spread rapidly across industries only if cooperatives with industry networks handle project planning, gather member demand, and take charge of joint verification and dissemination.
Sim Woo-pil, head of the planning and policy team at the Korea Die & Mold Industry Cooperative, explained, "Foundational manufacturers currently have a structure in which individual companies find it difficult to compete on cost with Chinese firms due to an aging workforce, labor shortages, and high electricity rates. In the case of cooperatives where similar foundational manufacturers gather, since they share similar manufacturing processes and business processes within the sector, they can be effective in responding to AX."
