Looking back at history, when will the capital expenditure boom turn into a bubble burst?

Written by: Dong Jing

Source: Wall Street Journal

From the railroads of the 19th century to the artificial intelligence of the 21st century, every major technological innovation in history has sparked a wave of capital expenditure, but the frenzy often ends in a bubble bursting.

BCA Research released a special report in November this year titled “When Capex Booms Turn Into Busts: Lessons From History,” which reviews four typical capital expenditure booms, revealing the core logic of the transition from prosperity to collapse, and issues a warning regarding the current AI boom.

The report summarizes five common patterns: investors overlook the S-curve of technology adoption, revenue forecasts underestimate the extent of price declines, debt becomes a core reliance for financing, asset price peaks occur before investment declines, and the collapse of capital expenditure exacerbates economic recession. These patterns are already emerging in the current AI field—technology adoption rates are stagnating, Token prices have plummeted over 99%, corporate debt has surged, and GPU rental costs are decreasing.

Based on historical comparative analysis, BCA Research concludes that the AI boom is following the historical bubble path and is expected to end within the next 6 to 12 months. The report suggests that investors maintain a neutral allocation to stocks in the short term, moderately underweight stocks in the medium term, and closely monitor forward-looking indicators such as analyst expectation revisions, GPU leasing costs, and corporate free cash flow.

The report specifically points out that the current economic environment has more concerns, as U.S. job vacancies have fallen to a five-year low. If the AI boom fades away and no new bubble offsets the impact, the potential economic recession may be more severe than during the 2001 internet bubble burst.

Historical Reflection: The Collapse Trajectory of Four Capital Frenzies

BCA stated that the essence of the capital expenditure boom is a collective optimistic expectation of capital regarding the commercialization prospects of new technologies. However, history repeatedly proves that this optimism often deviates from the objective laws of technology implementation, ultimately leading to a collapse due to supply and demand imbalance, debt accumulation, and high valuations.

The railway boom in 19th century Britain and America demonstrated the destructive power of overcapacity.

The report pointed out that the success of the Liverpool-Manchester railway in 1830 ignited a frenzy of investment in Britain, with railway stock prices nearly doubling between 1843 and 1845.

By 1847, railway construction spending accounted for a record 7% of the UK's GDP. The tightening of liquidity eventually triggered the financial crisis in October 1847, with the railway index plummeting 65% from its peak.

The report states that the railroad boom in the United States reached its peak during the Panic of 1873, when the New York Stock Exchange was forced to close for ten days, and corporate bond default losses amounted to 36% of face value between 1873 and 1875.

After the total mileage of railways in the United States reached a peak of over 13,000 miles in 1887, overcapacity led to a collapse in transportation prices, and by 1894, about 20% of the U.S. railway mileage was under bankruptcy management.

The electrification boom of the 1920s exposed the vulnerabilities of the pyramid-shaped capital structure.

The report indicates that the proportion of households with electricity rose from 8% in 1907 to 68% in 1930, but this process was primarily concentrated in urban areas.

Wall Street is deeply involved in this craze, with utility company stocks and bonds being promoted as “safe assets that even widows and orphans can invest in.” By 1929, holding companies controlled over 80% of the electricity generation in the United States.

The report states that after the stock market crash of 1929, the largest utility group Insull went bankrupt in 1932, allegedly causing the life savings of 600,000 small investors to vanish. Construction spending by American electric utilities peaked at about $919 million in 1930 and plummeted to $129 million in 1933.

The internet boom in the late 1990s proved that innovation does not equate to profit.

BCA stated that from 1995 to 2004, the annualized growth rate of productivity in non-farm U.S. businesses reached 3.1%, far exceeding the subsequent period.

However, the share of technology-related capital expenditure in GDP surged from 2.9% in 1992 to 4.5% in 2000, leading to immense pressure on corporate balance sheets due to over-investment.

The report indicates that the free cash flow in the telecommunications industry peaked at the end of 1997 and has been declining continuously, with a significant drop in 2000. The Nasdaq Composite Index rose sixfold between 1995 and 2000, followed by a 78% crash in the next two and a half years.

Multiple oil booms perfectly illustrate the cyclical nature of supply and demand imbalances.

BCA stated that after the discovery of huge oil reserves in East Texas in 1930, the daily production exceeded 300,000 barrels within 12 months, but the worsening Great Depression caused oil prices to plummet to 10 cents per barrel.

In 1985, Saudi Arabia abandoned production limits, causing oil prices to fall to as low as $10 per barrel.

From 2008 to 2015, the U.S. shale oil boom drove crude oil production up from 5 million barrels per day to 9.4 million barrels per day, while in 2014, OPEC's refusal to cut production caused oil prices to drop from $115 per barrel in mid-year to $57 per barrel by the end of the year.

Five Common Laws: The Inevitable Path from Prosperity to Collapse

Reviewing the rise and fall of four typical booms, BCA Research has summarized five common patterns that provide key benchmarks for assessing the current direction of the AI boom. Specifically:

The first rule is that investors overlook the S-curve of technology adoption.

The adoption of technology has never been a linear progression, but rather follows an S-shaped curve of “early adopters acceptance - widespread adoption - laggards follow.” Stock prices typically rise in the first stage, peaking in the middle of the second stage when the growth rate of adoption turns negative.

The current AI field is showing this characteristic: most enterprises express a willingness to increase AI usage, but the actual adoption rate has shown signs of stagnation, with some indicators even declining in recent months. This divergence between “willingness and action” is a typical signal of the later stage of technology adoption entering its second phase.

The second rule is that revenue forecasts underestimate the extent of price declines.

In the early stages of new technology, it has pricing power due to its scarcity, but as the technology becomes widespread and competition intensifies, prices are bound to drop significantly. From 1998 to 2015, the annual growth rate of internet traffic reached as high as 67%, but the unit price of information transmission simultaneously fell sharply. Since the introduction of solar panels, their prices have continued to decline, dropping by 95% from 2007 to the present.

The AI industry is repeating its past mistakes: since 2023, the introduction of faster chips and better algorithms has led to a decline in Token prices of over 99%. Although new applications such as video generation have emerged, users' willingness to pay for such applications remains unclear.

The third rule is that debt becomes the core reliance for financing.

In the early stages of a boom, companies can usually meet their capital expenditure needs through retained earnings, but as the scale of investment expands, debt will gradually become the main source of financing.

In October 2025, Meta announced a $27 billion financing agreement for data centers through an off-balance-sheet special purpose entity; Oracle, after securing a $38 billion loan, also raised $18 billion in the bond market, bringing its total debt close to $96 billion.

What is even more alarming is the “new cloud vendors” like CoreWeave. As of October 2025, CoreWeave's credit default swap rate has risen from 359 basis points at the beginning of the month to 532 basis points.

Rule four is that asset price peaks occur before a decline in investment.

Historically, asset prices such as stocks during a capital expenditure boom often peak before actual investment spending begins to decline. Even if investment spending falls from its peak, its absolute value may remain high, further exacerbating overcapacity. This means that if investors wait for a clear signal of “investment decline” before taking action, they often miss the best opportunity.

Rule five is that the collapse of capital expenditures and economic recession mutually exacerbate each other.

The burst of a technological bubble is usually divided into two stages:

The first stage is the retreat of technological speculation, revealing overcapacity; the second stage is the collapse of capital expenditure dragging down the overall economy, which in turn leads to the deterioration of corporate profits, forming a vicious cycle.

The report indicates that the 2001 economic recession in the United States was not triggered by a deterioration in economic fundamentals, but rather stemmed from the collapse of capital expenditures following the burst of the internet bubble. The rise of the real estate bubble in 2002 temporarily alleviated the impact of the internet bubble's burst, but it remains uncertain whether a new bubble will emerge to offset the effects of the AI hype collapse.

Risk Signals of the AI Hype: Turning Points in 6 to 12 Months

Based on a comparative analysis of historical patterns, BCA Research believes that the AI boom is following the historical bubble trajectory and is expected to end within the next 6 to 12 months. This judgment is based on the multiple risk signals that have already emerged in the current AI sector.

From a technological adoption perspective, the actual implementation speed of AI has not kept pace with the fervent expectations of capital. The adoption rate on the enterprise side has stagnated, and consumers' willingness to pay for AI applications has not yet been fully validated.

From the price trend, the significant decline in Token prices has shown deflationary pressure, while the commercial value of new applications such as video generation remains questionable.

From the perspective of debt risk, the financing structure of AI-related companies increasingly relies on debt, and the credit risk of some companies has begun to emerge.

The report suggests focusing on four key forward-looking indicators:

First, the analysts' expectations for future capital expenditures have been revised. If the expectations that have been rising continuously start to level off, it could be a warning sign.

Secondly, the cost of GPU leasing has started to decline after May 2025.

Thirdly, the free cash flow situation of large-scale enterprises is still at an absolute high level recently, but it has shown a deteriorating trend.

The fourth is the emergence of the “metaverse moment,” which occurs when a certain AI company announces a major project, yet its stock price falls instead. This will be a clear sign of a shift in market sentiment.

For investors, BCA Research recommends adopting a “moderate defensive” strategy at present. In the short term, that is, for the next 3 months, maintain a neutral allocation to equities; in the medium term, that is, for the next 12 months, moderately underweight equities, and in the coming months, further enhance defensiveness.

Specifically, it is important to closely monitor the aforementioned four leading indicators and avoid making passive adjustments only when investment spending clearly declines; at the same time, attention can be given to defensive sectors and high-quality bonds to hedge against potential significant fluctuations in AI-related assets.