AI Investments Continue Even as Interest Rate Bombs Drop

An analysis suggests that the current high-interest rate situation is unlikely to completely halt artificial intelligence (AI) investment. Lee Seon-yeop, CEO of AFW Partners, explained that despite the burden of high interest rates, "as long as it's managed, if it's AI, it probably won't be held back." Big tech companies like Google and OpenAI have sufficient capacity to absorb relatively higher capital raising costs, and there is a gap between financial market concerns and the actual investment environment experienced by companies.

In response to increasing interest burdens, Nvidia and cloud companies are actively exploring new financial structures. A recent market development is Nvidia's $500 billion financial platform, which secures funding by collateralizing GPUs and data centers. Large financial institutions like BlackRock and Goldman Sachs have also entered this market, expanding the methods available for financing AI investments.

Big tech companies believe that the profitability generated from data center operations sufficiently outweighs the cost of raising capital at high interest rates. Lee Seon-yeop emphasized, "If data centers are profitable, and investment is profitable, then who wouldn't invest?" highlighting that the ability to generate returns from investment is more critical than the interest rate itself.

A question arises about whether a 2% point increase in interest rates has a fatal impact on corporate activities. Lee Seon-yeop asked, "What's the problem if Google pays about 2% points more in interest?" pointing out that the current market allows vague fears about macroeconomic indicators to overshadow micro-level analysis. In reality, major big tech companies maintain robust financial health in a high-interest rate environment due to their low debt ratios.

Lee Seon-yeop explained that "when we were young, 5% was a cheap interest rate," clarifying that historically, 5% is a normal level, not an abnormally high one. Market volatility is presumed to be caused by the process of adapting to the current interest rate level.

Lee Seon-yeop explained that "if liquidity keeps flowing in, prices rise, and if liquidity dries up, prices fall; there's no reason it can't rise then." He added that when liquidity is exhausted, the market starts looking for excuses for decline. Therefore, it is necessary to examine liquidity indicators comprehensively, in addition to interest rates, to measure market energy.

Recently, new channels of liquidity supply, such as government fiscal spending and stablecoins, are expanding, which are difficult to capture with traditional interest rate indicators alone. These factors can provide sufficient liquidity to the market even if interest rates remain high, supporting growth drivers like AI investment.

Market participants should not judge the entire market by interest rates alone, but rather consider credit spreads and total liquidity indicators together. It is particularly important to keep in mind the possibility that high interest rates may become a new constant, but the actual market crisis should be judged by a comprehensive assessment of the credit market and liquidity indicators.

Ultimately, market movements are not determined by a single indicator but change according to the flow of liquidity, which is formed by the complex interaction of various factors. Therefore, a comprehensive analysis covering a wide range of indicators is essential to assess market health.

The semiconductor market has experienced repetitive cycles due to mismatches between supply and demand over the past 50 years. Lee Seon-yeop expressed the difficulty of future prediction, stating, "I don't know well what will happen in six months. Because if AI isn't regulated, or if it is regulated, what kind of discussion will emerge?" The rapid pace of AI technological development makes it difficult to definitively forecast an existing supply surplus.

The emergence of new AI models like Astra is expected to cause a surge in additional GPU demand. With the required GPU performance and quantity increasing more than tenfold compared to early generative AI models, it is difficult to predict when the AI investment cycle will end.

Lee Seon-yeop explained, "The effect brought by GPT-6 Astra, AI thinking similarly to humans, is expected to generate massive additional demand for GPUs." He noted that while past generative AI models required units of 100,000 GPUs, the latest models demand over a million GPUs.

Considering this pace of technological advancement, predicting the peak of AI investment is virtually impossible. It is emphasized that continuous verification is needed to ensure that technological progress directly translates into actual demand.

As Korean companies like Samsung Electronics and SK Hynix have grown, they compete in the global market beyond the domestic one. Lee Seon-yeop emphasized, "From Samsung Hynix's perspective, shareholders think they worked hard this time, but our investors' expectations are already at global standards." He stressed the need for governance improvement in line with global standards to attract foreign investors.

Korean companies' shareholder return policies are still considered inadequate compared to global standards. Issues such as corporate governance and circular shareholdings act as stock price discount factors, causing global investors to turn away.

Lee Seon-yeop mentioned, "Micron and SanDisk didn't say they wouldn't invest in facilities, did they? They said they would do facility investment and then give all remaining funds to shareholders." He noted that Korean companies' shareholder return policies are often concentrated in chairman-centric decision-making.

There is a growing call for establishing a corporate culture that balances long-term investment and shareholder returns. Global investors' perception of Korean companies largely depends on management transparency and shareholder-friendly policies, thus proactive improvement efforts are required.

Lee Seon-yeop stated, "They are set up to only look to the chairman. It's a chairman-centric company, not a shareholder-centric one." He noted that there is a perception that this management style has persisted for a lifetime and is difficult to change in the short term.

To change shareholders' perceptions, a substantive change in corporate attitude must precede. For domestic companies to achieve global competitiveness, improving governance is essential, and it is time to enhance transparency and responsible management through it.