South Korea bets on AI industrial policy

South Korea bets on industrial policy – this time, AI-focused

A $1 trillion push for chips and data centers

South Korea’s government is launching a massive state-led economic initiative to boost artificial intelligence and semiconductor production. President Lee Jae-myung’s administration has announced a national strategy aimed at making the country one of the world’s top three AI powers and reinforcing its global leadership in semiconductors. The Special Act on Semiconductors provides infrastructure and regulatory support to accelerate new investment, while the Three Mega Projects for a Great Leap Forward aim to develop regional hubs for cutting-edge industries through public-private partnerships in semiconductors, AI data centers, and robotics.

To fund these efforts, the government and major firms have announced plans for new investment exceeding 1,500 trillion South Korean won, which is roughly US$1 trillion. This amount equals about two-thirds of the country’s GDP, marking the largest industrial initiative since the Heavy and Chemical Industry drive in the 1970s. Samsung Electronics and SK Hynix plan to invest 911 trillion won in new semiconductor manufacturing hubs, while other leading firms are investing heavily in AI data centers. For its part, the government is offering new financing vehicles, including a Future Response Fund and a National Growth Fund, to mobilize long-term capital for strategic industries and innovation ecosystems.

The HCI drive was not perfect. It encouraged excessive borrowing, investment duplication, and significant overcapacity that contributed to the 1997 financial crisis. Despite these flaws, the program transformed the nation into a capital-intensive technological powerhouse.

Lessons from the past, but different challenges

South Korea has attempted this type of industrial transformation before. In the 1970s, amid security concerns from North Korea and the Nixon Doctrine, the government launched the Heavy and Chemical Industry (HCI) drive to accelerate industrial upgrading and reduce dependence on imported machinery and defense equipment. The strategy used targeted credit, tax incentives, foreign borrowing guarantees, and industrial complexes to promote exports in six strategic industries: steel, machinery, petrochemicals, shipbuilding, electronics, and non-ferrous metals.

The success of that earlier effort hinged on a capable economic bureaucracy that coordinated finance, infrastructure, and export promotion, alongside sustained political commitment. Skilled engineers and technicians absorbed foreign technologies and improved upon them, while the combination of government support and exposure to international competition supported learning-by-doing. This allowed South Korean firms to ascend global value chains rapidly in the 1980s, aided by favorable external conditions like cheaper oil and lower interest rates.

Today’s environment differs in key ways. The HCI drive sought to absorb and build on foreign technologies, whereas South Korea’s new strategy aims to compete at the global technological frontier, where uncertainty is heightened. Moreover, the practical challenges are substantial. Semiconductor fabrication plants and AI infrastructure require enormous capital investment, highly specialized engineers, and massive, reliable supplies of electricity and water. The process of securing these resources, along with land and permits, could delay projects for years. Additionally, private firms operate within a much denser regulatory regime today than they did in the 1970s, and democratic politics often produce short policy horizons that make decade-long credibility difficult to maintain.

It does not help that today’s democratic politics inevitably produce short policy horizons, making credibility over decades considerably harder than it was under Park. These challenges are compounded by intensifying competition from China, which has already emerged as a leader in AI and many advanced manufacturing industries, and is rapidly closing the gap in semiconductors, partly owing to sustained state support. AI systems are currently struggling to meet these high expectations, creating further pressure on the nation’s industrial plans.