AI Bubble to Burst? What Matters Is What Comes Next

While railway, power, and internet bubbles have burst in history, the infrastructure remained. The true winners of the AI era will be determined after the collapse. Artificial intelligence is widely expected to transform the world, yet whether the massive investments poured into it will yield returns remains uncertain. Technology's success does not necessarily mean investors' success.
Jeffrey's strategist Chris Wood recently warned that the current AI investment frenzy could lead to significant capital losses. U.S. companies are spending billions on AI data centers and semiconductors, but if low-cost Chinese open-source AI spreads, it may become difficult to recoup investments. In short, AI will advance, but companies investing heavily in it might face substantial losses.
Similar patterns have occurred before. In the 1870s, a railway construction boom saw investment flood into areas with no demand, leading to financial panic and bankruptcy of many railway companies by 1873. Yet rail networks remained and later became central to U.S. industrialization. Similarly, in the 1920s, massive capital flowed into electricity infrastructure based on the belief that power was the future. When the Great Depression hit, large utilities collapsed, but the grid endured and transformed American factories and homes.
The most recent example is the dot-com bubble of the early 2000s. Telecom companies built extensive fiber networks expecting endless internet usage growth. After the bubble burst, many firms failed, leaving unused cables. However, these assets later sold cheaply, enabling Google, YouTube, Facebook, and Netflix to grow on affordable infrastructure.
The common thread is clear: investors lost money, but the infrastructure they built became assets for future eras. The same pattern may unfold in today's AI market. Major tech firms are investing heavily in data centers, semiconductors, and power facilities to develop better AI. However, whether they can generate sufficient returns remains uncertain as competition intensifies and pricing pressures grow.
Particularly notable is the role of debt. Past bubbles didn't collapse solely due to cooling investment enthusiasm; they fell when financial institutions stopped providing capital. In AI, if investments and borrowing outpace revenue growth, a similar crisis could emerge. Not all AI investment should be dismissed as a bubble—large tech firms with strong cash reserves differ from startups reliant on debt.
AI semiconductors may also become obsolete faster than rail or power infrastructure due to rapid technological change. Past history doesn't guarantee repetition, but the gap between investment scale and actual returns must be closely monitored. This is especially relevant for Korea, where the stock market heavily depends on semiconductor and AI investments. A slowdown in U.S. AI capital expenditure could significantly impact Korean firms and financial markets.
Investors should look beyond share price increases to assess cash flow, debt levels, and actual profitability. However, dismissing the AI investment boom as a technological failure would be wrong. Instead, preparation for what comes after is key. As data centers, power infrastructure, and computing resources become cheaper, small businesses and startups that currently cannot afford AI may gain opportunities. Companies leveraging affordable technology to create new services could dominate future markets.
The cryptocurrency industry has experienced similar dynamics. After the 2017 ICO boom and 2022 market collapse, numerous projects vanished, and investors suffered losses. Yet blockchain technology endured, with stablecoins and asset tokenization expanding into mainstream finance.
Importantly, surviving technology doesn't justify past investments. The key is not whether a bubble existed but what remains after it bursts and who leverages that infrastructure. While no one knows when or how the AI investment boom will end, history offers a lesson: excessive expectations for new technologies can waste capital, but the infrastructure built during such periods often fuels future industries.
What's needed now is neither blind optimism nor unwarranted pessimism about AI. A balanced approach—remaining cautious about risks while not missing technology's potential—is essential. The true winners of the next era will be those who best utilize the assets left behind by the current AI investment wave, not those that spent the most.
Korean Source
This article is an English localization of a Korean-language crypto news report. Original headline: [사설] AI 거품은 꺼진다…더 중요한 것은 그다음이다