AI investment is a hot topic in the business community and policy authorities these days. As global big tech and governments enter the AI competition worth tens of trillion won, there are also calls in Korea to “flexibly lift regulations on holding companies’ CVC or financial holdings to combat this.” The background is the urgency that it is difficult to endure with its own funds and a shallow domestic capital market alone.
The core of the debate is not the complete abolition of the principle of separation of financial and industrial sectors. It is more of a claim that “Let’s make exceptions only for strategic industries.” But the question is direction. Regulations once relaxed are hard to reverse. This is why it is dangerous to view the separation of financial and industrial capital as simply an “old shackle restricting corporate activities.” This is because this principle is a device designed as a macroeconomic safety plate to prevent a crisis in the financial system.
Concern is based on theory and history, not on vague fear. As the “diamond-deepvig model” shows, banks have a weak structure of managing long-term assets with short-term deposits. Even if the assets are sound, if trust is shaken, a bank run is inevitable. The Silicon Valley Bank (SVB) crisis in the United States shows the ripple effect when industry bias is added to this. With interest rate shocks applied to the asset structure focused on tech companies and long-term bonds, 36 hours of deposit withdrawal in the era of smartphone banking was sufficient. When the separation of financial and industrial risks is eased and transferred to the financial system, the shock can turn into an uncontrollable crisis.
On the other hand, the situation in the East shows what happens when the boundaries between finance and industry are blurred. At that time, affiliated financial companies acted as “internal sources of funds,” resulting in the transfer of insolvency to the market. Discussions on CVC deregulation should also be cautious. There should be no deregulation such as opening a ‘special passage’ without governance and risk management devices. This can repeat the limitations of the ‘internal capital market’ in which resources are inefficiently allocated by the group’s internal logic rather than being the priming water for innovation investment.
The more competition over AI and semiconductors intensified, the greater the investment risk. The size of the initial investment is large, the technology path changes from time to time, and the market is winner-take-all. The greater the risk, the more sophisticated “Market Validation” is needed, not “removing safeguards.” If you are confident of passing the market’s cold verification, there is no reason to rely on easy internal funds. Deregulation with less transparency will only deepen the “Korea Discount” by raising the perception that “Korea is a country that changes principles whenever necessary.”
In the end, the key is “what kind of money will be mobilized?” Safe assets such as bank deposits have a share to bear, and “adventure capital” such as stocks, corporate bonds, private equity, and infrastructure funds have a share to bear. The principle of finance is that the latter takes charge of AI infrastructure, which is a high-risk field. The government’s role is not to unbar banks, but to open banks that are preventing market money from flowing into AI. The order is to repair regulations, governance, and tax incentives so that pension funds, insurance, and infrastructure funds can actively invest in AI infrastructure. Competition for AI infrastructure has already exceeded the scale that a single company can afford. Instead of the old solution that breaks down the financial and industrial firewalls, it should focus on diplomatic and tax support so that Korean companies can form an “infrastructure consortium” with global pension funds and government funds.
I understand the urgency of the need for astronomical ‘live ammunition’ for investment in AI infrastructure. However, the argument to tear down the ally’s “sanseongbyeok (financial safety net) first to win the “war” can be a dangerous gamble that will drive companies and the financial ecosystem to annihilate. As AI is the future of the country, the financial principles underlying it must be more sophisticated and solid.
[Lee Yoon Soo, professor of economics at Sogang University]