South Korea has realized that pretty speeches about future technologies may make headlines, but they do not build new factories, power lines, or HBM packages. On November 2, President Lee Jae Myung defended a record budget for 2027 in parliament and announced plans to invest 21.3 trillion won, equivalent to around 16 billion US dollars, in AI and three key areas: semiconductors, Physical AI, and AI data centers. According to Reuters, that would be nearly twice as much as in the current year. This is not yet a finalized spending program, but part of the draft budget. But the direction is clear: South Korea wants not only to maintain its role in the global AI and memory industry, but to expand it.
The answer is relatively simple. The global expansion of AI data centers is driving demand for high-performance memory chips, particularly HBM and server DRAM. Two of South Korea’s most important manufacturers, Samsung Electronics and SK hynix, are positioned right there. If the government wants to secure this industrial momentum, supporting research alone is not enough. Production capacity, infrastructure, power and water supplies, and the smoothest possible expansion of new semiconductor and data center clusters are needed. Reuters quotes Lee as saying that the government intends to invest 21.3 trillion won in AI and the three megaprojects: semiconductors, Physical AI, and AI data centers. Additional reports from South Korea identify specific priorities. Among other things, around 2 trillion won is to go toward expanding semiconductor production capacity. Another 2.1 trillion won is reportedly earmarked for power, water, and industrial infrastructure, so that investment projects do not founder on a lack of utilities or overly slow site preparation.
This point is particularly important because discussions of new semiconductor sites often proceed as if all you have to do is put up a factory and the market will somehow take care of the rest. In practice, modern fabs depend heavily on local infrastructure. Without sufficient power, water treatment, transport, and logistics connections, a state-of-the-art production facility can quickly become an expensive waiting game. Seoul’s apparent intention to provide targeted advance funding here is therefore not a side issue, but a central component of its industrial policy. In addition, South Korea is looking beyond conventional semiconductor manufacturing. Physical AI, meaning AI systems with a direct connection to the physical world, is also set to receive support. This includes, for example, robotics, autonomous systems, smart factories, and applications in which AI works with sensors, actuators, and industrial processes. According to South Korean reports, around 500 billion won is earmarked for related demonstration and application projects.
The announcement is particularly relevant to the international market because South Korea remains a heavyweight in memory. HBM is a key component for modern AI accelerators, because large training and inference models need not only computing power but also extremely high memory bandwidth. If South Korea accelerates its semiconductor and infrastructure policies, that will also indirectly strengthen the position of its memory manufacturers in precisely the markets that are currently growing fastest. At the same time, this sends an industrial policy signal to competitors. The US is supporting chip manufacturing through its own programs, Taiwan remains central to contract manufacturing, and China is investing heavily in technological self-reliance. South Korea is now trying not only to defend its core strengths, but to link them more closely to the next growth surge through government investment in infrastructure and AI.
Despite the scale of the announcement, it is important to put the current status in perspective. Lee presented the spending plans as part of the 2027 budget, but parliament still has to approve it. So these are not funds that have already been fully authorized. It remains to be seen how much individual line items will still change in the coming months. Nevertheless, the political direction is clear, and the size of the proposed package alone shows that Seoul sees competition for AI infrastructure and semiconductor manufacturing as a strategic issue.
Conclusion
What I find particularly interesting about this news is that South Korea is not just talking about AI, but consistently linking it to its existing industrial base. That is precisely what makes the announcement more relevant than many other national AI programs. This is not just about data centers or research labs, but about memory, manufacturing, infrastructure, and applications in industrial settings. Of course, the draft budget has not yet been finalized. But if even a large part of these plans is implemented, it will strengthen South Korea’s role in a market where, in the future, it will not be just chips that matter, but entire industrial ecosystems.
