History as a mirror: when will the capital spending boom turn into a bubble burst?
TechFlow Selected TechFlow Selected
History as a mirror: when will the capital spending boom turn into a bubble burst?
The report warns of an AI bubble burst within 6 to 12 months, advising investors to adopt a neutral stance in the short term and underweight stocks in the medium term, while paying attention to forward-looking indicators such as analyst expectations and GPU costs.
Navigating Web3 tides with focused insightsAuthor: Dong Jing
Source: Wall Street Insights
From 19th-century railways to 21st-century artificial intelligence, every major technological breakthrough in history has triggered a capital expenditure boom—but such frenzies often end in bubble bursts.
BCA Research’s November report titled *When Capex Booms Turn Into Busts: Lessons From History* reviews four typical capital expenditure booms, revealing the core logic behind their collapse from prosperity to bust, and issues a warning about the current AI frenzy.
The report identifies five common patterns: investors ignore the S-curve of technology adoption, revenue forecasts underestimate price declines, debt becomes the primary financing source, asset prices peak before investment declines begin, and capex collapses exacerbate economic recessions. These patterns are already emerging in today’s AI sector—stagnating adoption rates, token prices plunging over 99%, surging corporate debt, and falling GPU rental costs.
Based on historical comparisons, BCA Research concludes that the AI boom is following a classic bubble trajectory and is expected to end within the next 6 to 12 months. The report advises investors to maintain a neutral stock allocation in the short term, moderately underweight equities in the medium term, and closely monitor forward-looking indicators such as analyst forecast revisions, GPU rental costs, and corporate free cash flow.
The report highlights additional concerns in the current economic environment: U.S. job openings have fallen to a five-year low. If the AI boom fades without a new bubble to offset its impact, the ensuing economic downturn could be more severe than the 2001 internet bubble burst.
Historical Precedents: The Collapse Trajectories of Four Capital Frenzies
BCA states that capex booms stem from collective investor optimism about the commercial potential of new technologies. However, history repeatedly shows this optimism often脱离s the objective realities of technology adoption, ultimately collapsing due to supply-demand imbalances, debt accumulation, and inflated valuations.
The 19th-century U.S. and U.K. railway booms demonstrated the destructive power of overcapacity.
The report notes that the success of the Liverpool–Manchester Railway in 1830 ignited a speculative frenzy in Britain, with railway stock prices nearly doubling between 1843 and 1845.
By 1847, railway construction spending surged to a record 7% of British GDP. Tightening liquidity triggered the October 1847 financial crisis, causing the railway index to plummet 65% from its peak.
The U.S. railway boom culminated in the Panic of 1873, forcing the New York Stock Exchange to close for ten days. Between 1873 and 1875, corporate bond defaults reached 36% of face value.
After U.S. rail mileage peaked at over 13,000 miles in 1887, oversupply led to a collapse in freight rates, with approximately 20% of U.S. rail lines entering bankruptcy by 1894.
The 1920s electrification boom exposed the fragility of pyramid-like capital structures.
The report notes that household electrification rose from 8% in 1907 to 68% in 1930, though progress was largely urban.
Wall Street played a deep role in this boom, promoting utility stocks and bonds as safe investments "suitable for widows and orphans." By 1929, holding companies controlled over 80% of U.S. electricity generation.
After the 1929 stock market crash, the largest utility conglomerate, Insull, collapsed in 1932, reportedly wiping out the life savings of 600,000 small investors. U.S. electric utility construction spending peaked at $919 million in 1930, then plunged to $129 million by 1933.
The late-1990s internet boom confirmed that innovation does not guarantee profitability.
BCA notes that U.S. nonfarm business productivity grew at an annualized rate of 3.1% between 1995 and 2004, far exceeding later periods.
However, tech-related capex as a share of GDP soared from 2.9% in 1992 to 4.5% in 2000, placing immense strain on corporate balance sheets.
The report points out that telecom sector free cash flow peaked at the end of 1997 and declined continuously, crashing in 2000. After rising sixfold between 1995 and 2000, the Nasdaq Composite Index fell 78% over the following two and a half years.
Multiple oil booms perfectly illustrate the recurring cycle of supply-demand imbalance.
BCA states that after vast oil reserves were discovered in eastern Texas in 1930, daily production exceeded 300,000 barrels within 12 months—but the Great Depression caused oil prices to crash to just 10 cents per barrel.
In 1985, Saudi Arabia abandoned production limits, driving oil prices down to $10 per barrel.
Between 2008 and 2015, the U.S. shale oil boom pushed crude output from 5 million to 9.4 million barrels per day. In 2014, OPEC's refusal to cut production sent oil prices tumbling from $115 per barrel at mid-year to $57 by year-end.
Five Common Patterns: The Inevitable Path from Boom to Bust
Reviewing the rise and fall of four典型 booms, BCA Research identifies five recurring patterns that serve as key benchmarks for assessing the current AI boom. Specifically:
Pattern one: Investors ignore the S-curve of technology adoption.
Technology adoption is never linear but follows an S-shaped curve: early adopters → mass adoption → laggards. Stock prices typically rise during the first phase and peak around the midpoint of the second phase—when adoption growth turns negative.
The current AI landscape exhibits this pattern: while most companies express intent to increase AI use, actual adoption has stalled, with some metrics even declining in recent months. This divergence between intention and action is a classic sign of late-stage mass adoption.
Pattern two: Revenue forecasts underestimate price declines.
New technologies initially command pricing power due to scarcity, but prices inevitably fall sharply as adoption spreads and competition intensifies. Internet traffic grew at a 67% annualized rate between 1998 and 2015, yet unit data transmission prices dropped sharply. Solar panel prices have fallen steadily since inception, dropping 95% from 2007 to today.
The AI industry is repeating this pattern: since 2023, faster chips and better algorithms have driven token prices down over 99%. Although new applications like video generation are emerging, user willingness to pay remains unclear.
Pattern three: Debt becomes the core financing source.
In early stages, companies may fund capex through retained earnings, but as investment scales up, debt increasingly becomes the primary funding channel.
In October 2025, Meta announced a $27 billion off-balance-sheet data center financing deal via special purpose vehicles; Oracle, after securing a $38 billion loan, raised another $18 billion in bond markets, bringing its total debt near $96 billion.
More alarming are “new cloud providers” like CoreWeave: by October 2025, CoreWeave’s credit default swap rate had risen from 359 basis points at the start of the month to 532 basis points.
Pattern four: Asset prices peak before investment declines.
Historically, asset prices such as stocks tend to peak before actual investment spending begins to decline. Even as investment falls from highs, absolute levels may remain elevated, further worsening overcapacity. This means investors who wait for clear signs of declining investment often miss the optimal exit window.
Pattern five: Capex collapse and economic recession reinforce each other.
Technology bubble bursts typically unfold in two phases:
First, the hype fades and overcapacity becomes evident; second, collapsing capex drags down the broader economy, leading to deteriorating corporate profits—a vicious cycle.
The report notes that the 2001 U.S. recession was not triggered by weakening fundamentals but by a collapse in capex following the internet bubble burst. The emergence of the housing bubble in 2002 temporarily offset the fallout, but it remains uncertain whether a new bubble will emerge to cushion the impact of an AI bust.
Risk Signals in the AI Boom: A Turning Point Within 6–12 Months
Based on historical parallels, BCA Research believes the AI boom is tracking a classic bubble path and is likely to end within the next 6 to 12 months. This assessment rests on multiple risk signals already visible in the AI sector.
In terms of adoption, real-world AI deployment is lagging behind capital market enthusiasm, with enterprise adoption stalling and consumer willingness to pay for AI applications still unproven.
On pricing trends, the sharp drop in token prices reflects deflationary pressure, while the commercial viability of new applications like video generation remains questionable.
Regarding debt risks, AI-related firms are increasingly reliant on debt financing, and credit risks at some companies are beginning to surface.
The report recommends monitoring four key leading indicators:
First, analyst revisions to future capex expectations—if rising forecasts plateau, it could signal danger;
Second, GPU rental costs, which began declining after May 2025;
Third, free cash flow among hyperscalers—still high in absolute terms but showing signs of deterioration;
Fourth, the arrival of a “metaverse moment,” when a major AI company announces a significant project but its stock price falls—a clear sign of shifting market sentiment.
For investors, BCA Research recommends a “moderately defensive” stance. Maintain neutral equity exposure in the short term (3 months), moderately underweight equities in the medium term (12 months), and increase defensiveness in the coming months.
Specifically, closely track the four leading indicators above to avoid reactive adjustments only after capex declines become evident. Additionally, consider allocating to defensive sectors and high-quality bonds to hedge against potential volatility in AI-related assets.
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
