The AI investment boom is beginning to resemble previous periods of excessive capital spending, raising concerns that the artificial intelligence industry could eventually face a painful market correction, according to Bank of America (BofA).
In a research note released Friday, the bank warned that intensifying competition, declining AI service prices, and rapidly expanding infrastructure investments could undermine future profitability.
Investment strategist Sebastian Raedler highlighted similarities between today’s AI expansion and historical capital cycles, suggesting that technological breakthroughs do not always translate into sustainable investment returns.
AI Investment Boom Reaches Historic Spending Levels
Bank of America pointed to the extraordinary growth in capital expenditure expectations among major technology companies.
According to Raedler, consensus forecasts for hyperscaler capital spending over the next 12 months have surged from approximately $330 billion in early 2025 to an astonishing $1 trillion.
That figure represents roughly 3% of US gross domestic product (GDP), exceeding the relative investment levels recorded during several major historical infrastructure booms.
For comparison, capital spending during the railroad expansion of the 1870s reached approximately 2.2% of US GDP.
Meanwhile, the technology boom of the 1990s accounted for around 1.1% of economic output.
These comparisons highlight the unprecedented scale of investment currently flowing into artificial intelligence infrastructure.
However, BofA cautioned that enormous spending commitments do not necessarily guarantee attractive long-term returns for investors.
Could AI Follow the Boom-and-Bust Pattern of Past Capital Cycles?
Historical investment cycles frequently follow a similar pattern.
Initially, strong demand and attractive profit margins encourage companies to expand production capacity aggressively.
As additional competitors enter the market, investment accelerates, often creating more capacity than the industry can profitably absorb.
Eventually, increased competition places downward pressure on prices and profit margins, reducing returns on the investments made during the expansion.
Raedler emphasized that many historical capital booms ultimately disappointed investors, even when the technologies behind them delivered substantial economic benefits.
This distinction is particularly important for artificial intelligence.
AI could transform productivity and business operations while still generating disappointing financial returns for companies that invest too aggressively.
Why the Cloud Computing Boom Was Different
Not every major technology investment cycle ends with widespread financial disappointment.
Bank of America identified the cloud computing expansion of the 2010s as an important exception.
During that period, high barriers to entry helped established providers maintain strong competitive positions.
Building and operating large-scale cloud infrastructure required significant financial resources, specialized expertise, and extensive technology networks.
Consequently, a relatively small group of companies dominated the market, limiting competitive pressure and helping preserve profitability.
The artificial intelligence industry, however, appears to be developing under different conditions.
Competition is becoming increasingly intense, particularly among companies developing and commercializing AI models.
AI Token Prices Collapse as Competition Intensifies
One of BofA’s biggest concerns is the rapid decline in prices charged for artificial intelligence services.
Competition has intensified at the AI model level, particularly following the emergence of increasingly capable Chinese open-weight models.
These developments have challenged established providers and contributed to substantial reductions in AI usage costs.
According to the bank’s analysis, average AI token prices have fallen approximately 50% since the summer.
While cheaper AI services could encourage broader adoption, they also create challenges for companies attempting to recover their enormous infrastructure investments.
Lower prices may reduce revenue per unit of usage, potentially weakening profit margins unless overall demand expands sufficiently.
The crucial question is whether increasing AI consumption can compensate for the decline in pricing.
Can Cheaper AI Services Drive Enough Demand?
Supporters of the AI investment boom frequently point to Jevons Paradox as a potential explanation for why falling prices could ultimately benefit the industry.
This economic principle suggests that improvements in efficiency can sometimes increase overall consumption rather than reduce it.
Applied to artificial intelligence, lower token prices could encourage businesses and consumers to use AI services more frequently.
In theory, the resulting growth in demand could offset lower prices and sustain revenue growth across the industry.
However, Bank of America highlighted emerging evidence that challenges this optimistic scenario.
Data from Ramp AI indicate that US business spending on artificial intelligence has declined slightly since August, even as the volume of tokens processed has continued increasing rapidly.
This divergence suggests that growing usage does not necessarily translate into equally strong revenue growth.
Raedler warned that weaker monetization amid intensifying competition resembles the problems that have undermined previous capital investment cycles.
AI-Related Stocks Face Growing Selling Pressure
Concerns about the sustainability of AI spending have already contributed to weakness across several technology-heavy stock markets.
According to the figures cited in BofA’s report, South Korea’s KOSPI index had declined approximately 20% since early October.
Meanwhile, the Philadelphia Semiconductor Index was trading around 13% below its recent peak.
Both benchmarks have significant exposure to companies benefiting from the global artificial intelligence expansion.
The declines suggest that investors are becoming increasingly cautious about elevated technology valuations and the profitability of AI infrastructure investments.
Semiconductor companies are particularly sensitive to changes in capital spending expectations because demand for advanced chips depends heavily on continued investment in data centers and computing infrastructure.
If major technology companies begin slowing their spending, semiconductor manufacturers and related suppliers could face additional pressure.
Bank of America Maintains Bearish Outlook on European Stocks
Beyond artificial intelligence, Bank of America continues to maintain a cautious outlook for European equities.
The bank expects the Stoxx Europe 600 index to decline approximately 10%, reaching 580 points by the second quarter.
Its strategists also favor defensive sectors over economically sensitive industries.
Bank of America maintains overweight positions in consumer staples and pharmaceutical companies, which generally offer more resilient demand during periods of economic uncertainty.
Conversely, the bank remains underweight banks, capital goods manufacturers, and semiconductor companies.
These sectors could prove more vulnerable to slowing economic activity, tighter financial conditions, or reduced corporate investment.
The positioning reflects BofA’s preference for businesses with relatively stable earnings over industries dependent on continued economic expansion.
What Could Determine the Future of the AI Investment Boom?
The next phase of the artificial intelligence market will likely depend on whether revenue growth can justify the industry’s rapidly expanding capital expenditure.
Continued increases in AI usage could support long-term growth, particularly if businesses identify more profitable applications for the technology.
However, intensifying competition and falling service prices may make it harder for providers to achieve the returns investors currently expect.
Another important factor will be the pace of infrastructure investment.
If companies continue expanding data center capacity faster than demand grows, the industry could eventually face excess capacity and declining profitability.
Conversely, stronger AI monetization and sustained demand could help the industry avoid some of the difficulties experienced during previous investment booms.
For now, Bank of America warns that the AI investment boom is displaying several characteristics of historical capital cycles, including rapidly accelerating spending and increasing competitive pressure.
The bank’s analysis highlights a crucial distinction for investors: artificial intelligence may deliver enormous technological progress without guaranteeing equally impressive returns for every company financing its expansion.
