The AI Bubble: Are We Living In It, and What Happens When It Bursts?

The world is racing at a furious pace in the field of Artificial Intelligence (AI), with tech giants like Google, OpenAI, and Meta competing to offer their latest solutions. AI investments have reached record numbers, leading to a significant boom in stock markets, and even the White House is participating in this technological development. But the central question here is: Is the world currently experiencing an AI bubble? And will this massive investment wave, characterized by exaggerated expectations, inevitably experience a contraction? The AI bubble is defined as a period of rapid investment and exaggerated expectations in this field, where enthusiasm outweighs practical results, raising concerns that valuations and promises may exceed the technology's actual capabilities and long-term sustainability (Built In). Paul Kedrosky, a partner at SK Ventures and a fellow at the MIT Initiative on the Digital Economy, believes that we are indeed in a bubble, but not the expected kind. While AI is a technology of utmost importance, concern is growing about the enormous amount of money flowing into AI infrastructure, especially data centers.

Kedrosky points out that "there is massive spending on core AI infrastructure with little chance of recovering most costs, and a high probability that most of these assets will lose their value due to their rapid obsolescence."

Repercussions of the AI Bubble Burst

Investments in data centers are expected to reach trillions of dollars, with projected future spending exceeding $2 trillion on these centers. What makes the situation more critical is that an increasing proportion of these funds comes from debt, which imposes unavoidable obligations.

Despite the fundamental importance of AI, this massive influx of funds could lead to a "rational bubble," where all investors believe they are making the right decision. However, when everyone's "right decisions" accumulate, it can lead to enormous waste of resources. This scenario is not new; it has recurred in previous historical bubbles such as the railway bubbles of the 19th century in the UK and the US, where the construction of an excessive number of converging tracks led to significant waste, company failures, and severe market crises. If the AI bubble bursts, companies that have invested heavily in this technology will likely face massive financial losses, leading to layoffs and a decline in innovation rates. The collapse could also result in a loss of confidence in AI, potentially hindering future investments and developments (Built In). Jared Bernstein, a policy fellow at the Stanford Institute for Economic Policy Research, believes that such bubbles can lead to economic recession, noting that the dot-com bubble burst raised unemployment by two points, while the housing bubble led to the collapse of global credit markets and an unemployment increase of over five points (NPR). Bubbles cause severe damage, as a 20 or 30 percent decline in the stock market affects consumer spending, with potential repercussions for an economic recession.

Why Does the Massive Influx of Investments in AI Infrastructure Continue?

Many players in the technology sector are not convinced that we are experiencing an AI bubble. They see AI as more than just a bubble; it represents the most important technology of our era. They look forward to the opportunity to build superintelligence, relying on chips, buildings, and infrastructure dedicated to AI, and consider any slowdown in this development a grave mistake. However, there are voices outside the tech sector warning against this massive spending. For example, the Bank of England warned of the increasing risks of a potential global market correction due to the overvaluation of leading AI technology companies in financial markets. It noted that valuations could be further harmed if the infrastructure requirements needed to sustain the technology are too high and cannot be met. It also added that investors have not been adequately warned of the risks of a stock market collapse if AI does not live up to market expectations (Wikipedia).

In contrast, the chief equity strategist at Goldman Sachs believes that the rapid increase in stock valuations is likely justified by strong and sustainable earnings growth, noting that valuations of large growth stocks in the US were modest compared to the dot-com bubble (Wikipedia). Morgan Stanley also pointed out that the average cash flow of the largest 500 US companies is three times what it was in 1999, and corporate profit margins are much stronger now (Wikipedia). The current Chairman of the US Federal Reserve, Jerome Powell, stated that AI differs from other technology bubbles, such as the dot-com bubble, as the companies behind it generate massive revenues, and investment in AI data centers creates significant economic growth (Wikipedia).

Lessons from Similar Historical Bubbles in the Investment World

The United States and humanity in general have witnessed many economic bubbles throughout history. The Railway Bubble is one of the most prominent examples in the United States, fueled by intense enthusiasm for the idea. A similar scenario recurred in the 1920s during the period of rural electrification, when utility companies proliferated, leading to excessive spending. It can be said that "electricity fever" contributed to the rise of the stock market during the 1920s, which ultimately led to the 1929 crash and accelerated the pace of the Great Depression.

While many are familiar with the telecom and internet (dot-com) bubbles, the closest historical analogy to the current situation lies in the railway and electricity bubbles. Just as there was no need for two converging sets of railway tracks leading to Philadelphia, it is unlikely that we will need the same large number of companies offering large language models or AI models used by the public. These models are expected to naturally contract as the market evolves.

The Destructive Effects of Economic Bubbles

All bubbles cause severe damage, and the intensity of this damage depends on the size of the bubble and the extent of its impact. If you own an index fund and consider yourself a conservative investor, you are actually heavily immersed in the current AI market. If the market reverses course and stocks fall by 20 or 30 percent, you will become significantly less wealthy, which will affect your spending and have potential repercussions for an economic recession.

A common saying in tech circles is that a bubble bursts and then leaves something usable behind, even if not perfect. However, reality shows that almost every financial and technological revolution has caused severe damage, and it can take decades for markets to recover. As the famous economic saying goes: "In the long run, it may work out, but in the long run, we are all dead." Some experts believe that the collapse of the AI bubble will lead to a "natural selection" of companies, where only a few strong ones will survive and thrive (Medium). This means the economy will readjust to be more efficient, which is a natural part of the business cycle, where some companies will exit the market, layoffs will occur, and then the sector will reset, and technology will continue to evolve and improve (Built In).