Ruchir Sharma warns of AI investment bubble despite technological revolution

Global investment strategist Ruchir Sharma warns that the AI revolution could be accompanied by a massive investment bubble, noting that actual financial burdens for infrastructure could reach trillions of dollars. While 30% of Americans have used AI since ChatGPT emerged, Sharma identifies four "Overs" as symptoms of market risk.

While there is no disagreement that artificial intelligence (AI) technology is changing the world, the flow of capital being poured into that technology may be pointing in a completely different direction. Warnings have emerged that as the AI revolution intensifies, the bubble formed by related stock prices and investment fever could grow even larger.

A video from Professor Lee Hyun-hoon's Economic Forum analyzes the structural risks of the ongoing AI investment craze, based on an article written by global investment strategist Ruchir Sharma for the United States foreign affairs publication Foreign Affairs. Sharma, who worked at Morgan Stanley for 25 years and currently serves as the chairman of Rockefeller International, emphasizes that the success of AI technology does not necessarily translate directly into returns for investors.

The Gap Between Rapid Technological Diffusion and 'Profitability'

Sharma identifies the "speed of diffusion" as what distinguishes AI from past technological innovations. While it took considerable time to build the infrastructure for the steam engine, electricity, or computers to take root across society, AI has spread globally almost instantly via existing computers, smartphones, and high-speed internet networks. Indeed, the speed is unprecedented, with approximately 30% of Americans using AI within less than four years of the emergence of ChatGPT.

However, these powerful expectations widen the gap between the actual value of the technology and the future value calculated by investors. Pointing out that large technology companies with strong performance were at the center of the market during the past internet bubble, Sharma explains that the fact that current AI companies are generating massive profits cannot, by itself, justify rising stock prices. This is because an increase in technology companies' profits and an excessive rise in stock prices can occur simultaneously.

In particular, the capital circulation structure within the AI industry is also identified as a risk factor. A circular revenue model can occur where large cloud companies invest in AI model developers, and those developers then use that money to purchase cloud services. Sharma's analysis suggests that while this can inflate apparent growth, separate verification is needed regarding whether final customers actually have the ability and willingness to pay.

Four Typical Symptoms of a Bubble: The Four 'Overs'

Sharma defines the current AI craze through four "Overs," which are typical signs of a market bubble. The first is Overvaluation, where prices are excessively high; the second is Overholding, where investor assets are concentrated in specific stocks; the third is Overinvestment, referring to competitive facility expansion; and the last is Overborrowing, where debt expands. As these four overlap, the boom becomes stronger, but the market becomes easily shaken by even small changes.

The proportion of stocks in U.S. household assets exceeds 30%, a level that surpasses the peak record during the dot-com bubble. As index-tracking funds concentrate on large technology stocks, both individuals and institutions are exposed to the same price volatility risks. Furthermore, the pace of corporate capital expenditure is much faster than during past technological revolutions. Some analyses suggest that the current ratio of technology capital expenditure to GDP is higher than it was during the internet bubble era.

The most concerning aspect is overborrowing. While AI investment has been reported to be led by cash-rich Big Tech companies, Sharma warns that the actual structure is much more complex. Companies are taking on long-term payment obligations through data center leases or special project finance (PF) structures. While currently disclosed net debt stands at approximately $250 billion, the actual financial burden related to infrastructure investment is likely to reach trillions of dollars over the coming years.

AI Supply Chain Connectivity and National Risk

The AI industry forms a massive interconnected network spanning semiconductors, memory, data centers, power, and core minerals. In this process, companies such as Taiwan's TSMC and South Korea's SK hynix and Samsung Electronics play pivotal roles. Because these companies hold very high weights in their respective domestic stock markets, a tremor in AI capital expenditure could extend beyond the problems of specific companies to shock the entire national financial markets.

While capital flows along the AI supply chain to data centers in Southeast Asia or server production bases in Mexico, not all regions enjoy the benefits. While funds concentrate on AI hardware, manufacturing industries in Europe or certain service-exporting countries may experience investment exclusion, deepening economic imbalances. Even if the technology provides benefits to society as a whole, it remains uncertain whether the companies that invested massive capital will be able to fully recover those profits.

Source: original video (YouTube)