What Will We Do in the AI Era?... Polarization in Consumption, Investment, and Employment [Weekend Money]

AI Investment Concentrated Among Big Tech... Only Highly Skilled Workers Remain
Hardware Exporters Like South Korea and Taiwan to Benefit
Traditional Manufacturing-Centered Countries Like Europe Left Behind

Artificial intelligence (AI) has emerged as a key driving force underpinning the U.S. economy. As investments in data centers and infrastructure increase, the United States continues to experience growth that exceeds its potential growth rate. However, the benefits of AI-era growth are not distributed evenly. Consumption, investment, and employment are increasingly diverging in a 'K-shaped' pattern.

According to a recent report by Hana Securities, while the U.S. economy is expected to maintain solid momentum driven by domestic demand, the polarization of growth will likely intensify. Low-income households are reducing discretionary spending due to the burden of energy costs and rising prices, whereas high-income households maintain relatively strong consumer sentiment. Even within U.S. consumption, economic conditions are increasingly felt differently depending on income level.

The same is true for investment. Recently, the contribution of non-residential investment to U.S. GDP has grown, with a particular focus on AI data centers and infrastructure. The challenge is that AI investment requires massive upfront costs. Naturally, investment is being concentrated among big tech companies with significant capital power. According to analysis by the Federal Reserve Bank of San Francisco, among companies with a positive outlook on AI, the top 1% of large firms by asset size accounted for most of the increase in physical capital investment projected for 2025.

Hana Securities predicts that a period of structural change will unfold during the transitional phase before the AI cycle matures. The first factor is the ability to secure long-term funding and maintain a leading position. The five major hyperscalers are projected to see their free cash flow fall to the lowest level in a decade in the second half of this year. Big tech companies are increasingly turning to external borrowing, such as issuing corporate bonds and utilizing structured finance. However, the economic lifespan of AI servers is only 2 to 3 years, meaning that even after building data centers, ongoing replacement costs are inevitable. Companies that are slow to monetize may face growing financial burdens and could be forced out of the market.

Changes are also expected in the labor market. As data center construction increases, demand for skilled construction workers is rising. In contrast, industries with a high rate of AI adoption, such as information services, professional services, and finance, face the risk of certain roles being automated. This could widen the gap between workers who can leverage AI and those whose roles can be replaced by it.

Disparities between countries based on industrial structure are also expected to widen. While the United States is achieving growth through AI investment, it remains highly dependent on hardware imports. This means countries exporting AI hardware could benefit in turn. In particular, the growth engines of the top four net exporters of AI hardware—South Korea, Taiwan, Thailand, and Malaysia—are expected to strengthen. The International Monetary Fund (IMF) has already significantly raised its growth outlook for South Korea from 1.9% to 2.6% in its revised economic forecast in July. This stands in stark contrast to its downward revision of the global economic growth rate to 3%. Jun Kyuyeon, an economist at Hana Securities, explained, "Europe and emerging countries focused on traditional manufacturing, which are falling behind in AI investment and the discovery of leading companies, will become relatively more vulnerable."

This content was produced with the assistance of AI translation services.