Federal Reserve officials are beginning to question whether the frenzied investment driving the buildout of the artificial intelligence sector is getting out of hand. As tech companies pour hundreds of billions of dollars into data centers, chips, and power infrastructure, some central bankers are mulling whether this rapid expansion is creating risks for the broader financial sector. For now, some of the officials who have tackled the subject call for vigilance, dashed with a sense that a financial crisis mirroring what happened 20 years ago with housing, or the dot-com shakeout before that, is probably not in the offing. Still, the immense scale of investment, the uncertain returns for an unproven technology, the rise of complex financing structures, and the increased use of debt have moved artificial intelligence finance onto central bankers’ radar.
The debate exposes a significant division within the central bank. Some policymakers view the massive capital spending cycle as a normal, albeit highly enthusiastic, response to a major technological shift. They argue that because highly profitable companies with fortress balance sheets are leading this spending, the leverage does not pose an immediate threat to financial stability. Other officials are far less sanguine. They warn that the circular web of contractual commitments connecting data centers, energy providers, hardware makers, and private lenders could create a systemic vulnerability, raising the question of whether the artificial intelligence industry is becoming another structural risk.
With the financial system increasingly concentrated around a handful of dominant technology companies, the debate is moving beyond simple interest rate paths. Central bank supervisors are now forced to analyze the macro-level linkages of the digital infrastructure boom. If the expected revenue from these massive artificial intelligence investments fails to materialize, the sudden unwinding of this capital cycle could send shockwaves through the financial system, putting the Federal Reserve’s regulatory capabilities to a severe test.
The Too Big to Fail Debate: Jeff Schmid’s Systemic Warning
The sharpest warning from inside the Federal Reserve came recently from Kansas City Fed President Jeff Schmid. Speaking at an economic conference hosted by his bank, Schmid urged policymakers to look past the immediate hype of the technology and analyze the underlying financial structures supporting the buildout. His remarks marked a significant shift in the central bank’s vocabulary, bringing back terms that dominated the 2008 financial crisis.
Schmid’s Warning on Structural Leverage
Schmid did not mince words when addressing the scale of the artificial intelligence infrastructure boom. He argued that the sheer volume of capital flowing into the sector deserves close scrutiny from a financial stability perspective. He asked whether the financial situation involved in building out the artificial intelligence sector bears watching, noting that policymakers must correlate what is happening in artificial intelligence, from a pure scale standpoint, to some of the other experiences the country has had that could create a systemic problem.
Schmid went further, asking a question that caught the immediate attention of Wall Street: “Is this industry becoming another too big to fail?” This phrase, historically reserved for systemically important banks whose collapse would ruin the broader economy, represents the most direct warning yet from a Federal Reserve official regarding the risks of the technology boom.
Schmid is particularly concerned about how the industry is financing itself and how the interconnected linkages of these agreements could act as a vector where a problem starts at one stage and then rapidly propagates. He questioned whether the circular motion of a contractual commitment, such as a multi-decade contract connecting a data center, an energy provider, and a local community, is becoming overleveraged. If a spark starts a flame, Schmid warned, the highly leveraged circular commitments could make it difficult to contain the damage, creating a cascading default risk across the corporate credit markets.
The Jackson Hole Influence and Vocabulary Shift
As the president of the Kansas City Fed, Schmid hosts the annual Jackson Hole Economic Policy Symposium, which brings together international central bankers, academics, and policymakers to discuss the most pressing issues facing the global economy. By raising these systemic concerns ahead of the landmark conference, Schmid is signaling that the financial stability risks of the artificial intelligence buildout are officially on the regulatory agenda.
This vocabulary shift represents a deeper evolution in how the Federal Reserve views its role. While the public and the financial markets typically focus on the Fed as a rate-setting body, the central bank’s supervisory and financial stability divisions require no formal vote to act. If regulatory teams begin to view the artificial intelligence sector as a potential source of systemic risk, they can implement stricter lending standards, increase capital requirements for banks exposed to tech debt, and step up oversight of the private credit markets that have funded much of this expansion. This regulatory tightening could cool the capital cycle long before any formal interest rate cuts or hikes occur.
The Sanguine Counter-Perspective: John Williams and Market Volatility
Not everyone inside the Federal Reserve shares Schmid’s alarmist perspective. Other senior officials view the current capital cycle through a more traditional economic lens, arguing that high levels of investment are a natural feature of major technological transitions.
Williams on Tech Cash Flows and Manageable Debt
Federal Reserve Bank of New York President John Williams represents a more relaxed, market-oriented view. In a recent interview, Williams dismissed the idea that the artificial intelligence sector is in a classic, dangerous bubble. Instead, he characterized the current environment as a period of very high excitement and enthusiasm around a transformational new technology.
Williams pointed out a critical difference between the current tech boom and previous financial crises. During the subprime mortgage crisis of 2008, the financial system was highly exposed to low-quality, leveraged loans held by fragile financial institutions.
Today, however, the massive capital expenditures driving the artificial intelligence buildout are being managed and funded by some of the most profitable, cash-rich corporations in human history. Tech giants like Microsoft, Alphabet, Meta, and Amazon generate hundreds of billions of dollars in annual operating cash flows. While these firms are borrowing money to support their investments, Williams noted that their high earnings easily cover their debt service obligations, meaning he is not particularly worried about financial stability from corporate leverage right now.
Volatility as an Intractable Valuation Exercise
Williams also addressed the stock market volatility that has occasionally rattled tech investors. He explained that investors are attempting to solve an almost intractable problem in real time: calculating exactly how large the eventual economic benefits of artificial intelligence will prove to be.
Because nobody can predict the future utility or profitability of these models with absolute certainty, any changes in corporate earnings or capital expenditure forecasts will naturally lead to sharp shifts in market valuations.
From Williams’s perspective, this volatility is a healthy, natural part of the price-discovery process. When a company announces that its data center spending is rising faster than expected, stock investors may sell off the shares to adjust their near-term profit expectations. However, this market correction does not represent a systemic threat to the banking system. As long as the corporations borrowing the capital remain highly profitable and capable of servicing their debts, the wider financial system remains secure, even if individual stock prices suffer temporary declines.
Mary Daly’s Balanced Assessment of Corporate Risk
San Francisco Fed President Mary Daly has taken a middle-of-the-road position, acknowledging both the scale of the risk and the unique financial strength of the companies involved. Daly, whose district covers the Silicon Valley technology hub, has a front-row seat to the rapid deployment of artificial intelligence capital.
Daly admitted that if an analyst looked simply at the growth rate and the sheer amount of capital currently flowing into the artificial intelligence space, they could easily conclude that the situation is very worrisome. The rapid escalation of data center construction and chip orders has created an unprecedented demand for capital and electrical power.
However, Daly also noted that several unique factors offset this anxiety. She emphasized that many of the largest financial commitments are being made by companies with massive cash cushions and exceptionally strong balance sheets.
Furthermore, these companies have demonstrated a high degree of flexibility in how they deploy their projects. If a specific artificial intelligence application fails to generate immediate revenue, these firms can easily pause, delay, or adjust their future capital spending plans without facing a threat of corporate bankruptcy. This built-in financial flexibility acts as a vital shock absorber, protecting the wider economy from a sudden, chaotic collapse of the technology capital cycle.
The Velocity of the Buildout vs. Historical Booms
While Federal Reserve officials debate the systemic risks from a qualitative perspective, independent economic research provides a striking quantitative look at the sheer scale and speed of the ongoing artificial intelligence capital expenditure boom.
Measuring the Velocity of the AI Buildout
Analysis compiled by Apollo Global Management Chief Economist Torsten Slok puts the current technology cycle into a historic perspective. According to current consensus estimates, US tech giants and hyperscalers could spend about $916 billion on capital expenditures over the next 12 months, with that figure projected to climb to nearly $1.2 trillion the following year.
Slok’s research highlights that the speed of the AI buildout may matter far more than its overall size relative to the economy. The data-center capex boom is projected to rise by 2.5 percentage points of GDP, climbing from just 0.6% of GDP in 2023 to a projected 3.1% in 2027.
To put this acceleration into perspective, Slok compared it to previous major investment booms:
- The late-1990s telecom and fiber buildout peaked at just 1.2% of GDP in 2000, rising by only 0.4 percentage points before collapsing and tipping the economy into a mild recession.
- The mid-2000s housing boom added a smaller 2.2 percentage points of GDP from its mid-1990s start through its peak at 6.6% of GDP in 2005.
- On this measure, the current data-center buildout is growing twice as fast as the housing boom at its fastest pace, and almost three times faster than the dot-com telecom cycle.
While the absolute level of the housing boom was larger, the rapid acceleration of the artificial intelligence cycle means that any sudden slowdown would have an outsized impact on national GDP. Slok warned that because the buildout has grown so fast, the capital cycle could eventually unwind at a similar pace, creating a significant drag on economic growth if tech giants decide to scale back their infrastructure investments.
A Multi-Trillion Dollar Squeeze in Private Credit Markets
This rapid capital deployment is also creating what industry insiders call a capital suction across both public equity and private credit markets. The demand for financing to construct massive, gigawatt-scale data centers is so large that it is drawing capital away from other sectors of the economy.
According to market disclosures, major technology firms are exploring partnerships to secure massive debt guarantees for individual data center projects. For example, prominent developers have entered into talks to secure up to $250 billion in debt guarantees to fund single, massive multi-gigawatt data center installations.
This level of borrowing is beginning to squeeze the private credit and corporate bond markets, driving up borrowing costs for smaller, non-tech companies that must compete for the same pool of institutional capital. This capital concentration supports Jeff Schmid’s concern that the financial system is becoming dangerously exposed to the success or failure of a single technology buildout.
Concentration Risks and the Dot-Com Redux
The concentration of capital is also visible in the public equity markets, where a tiny group of technology giants now dominates the performance of the entire stock market. This extreme concentration has drawn comparisons to the height of the dot-com bubble in March 2000.
An analysis by Bank of America Global Research highlighted that the AI Big 10—a group of ten dominant technology stocks including Nvidia, Microsoft, Alphabet, Amazon, Meta, Apple, Tesla, Broadcom, Micron, and Advanced Micro Devices—now accounts for a massive 41 percent of the S&P 500’s total market value.
This 41 percent concentration is not just an arbitrary statistic; it matches the exact share that technology and telecom stocks held at the peak of the dot-com bubble before the market crashed. It also aligns closely with other historic concentration peaks, such as the Nifty Fifty group of blue-chip stocks, which held a 40 percent market share in 1972, and Japanese equities, which commanded 44 percent of the global index in 1989.
This high level of concentration means that the performance of the broader United States economy has become highly dependent on the continued profitability and capital spending of just ten corporate boards. If a major tech provider or cloud platform suffers an unexpected operational failure or a sharp decline in earnings, the resulting market correction would instantly wipe out trillions of dollars in paper wealth, impacting consumer confidence and corporate investment across every sector of the economy.
The Microeconomic Reality: Soaring Capex and the Nvidia Dependency
This macro-level debate is supported by real-world corporate actions, where companies are spending historic sums to secure their positions in the artificial intelligence race.
For instance, aerospace pioneer SpaceX reported that its second-quarter capital expenditures hit an astonishing $18.4 billion, with the vast majority of that cash outflow going directly toward purchasing advanced processing chips and building out dedicated AI data centers. Elon Musk publicly committed that the company’s future computing infrastructure will be built exclusively on specialized processors, illustrating the deep, systemic dependency that modern corporations have developed on a single hardware manufacturer.
This deep hardware dependency creates a fragile supply chain. If the dominant chipmakers or their manufacturing partners experience any production bottlenecks, or if geopolitical tensions restrict the trade of advanced semiconductors, the global data-center buildout would instantly grind to a halt.
With billions of dollars in pre-committed debt and contractual obligations tied to these projects, a hardware supply disruption could quickly trigger a wave of corporate defaults, validating Jeff Schmid’s warning that the artificial intelligence supply chain has become a major source of systemic vulnerability for the global financial system.
Resolving the Regulatory Challenge
The intense debate playing out inside the Federal Reserve proves that the artificial intelligence revolution has officially transitioned from a technological novelty into a major macroeconomic force. Whether policymakers view the current capital expenditure cycle as a healthy response to technological innovation or a dangerous, overleveraged bubble, they can no longer ignore the sheer scale and speed of the buildout.
As the central bank continues to monitor the financial stability risks of this multi-billion-dollar expansion, the outcome of this debate will shape the future of corporate regulation. If the Federal Reserve decides that the artificial intelligence industry has indeed become too big to fail, it will implement a new era of stricter oversight, permanently changing how Wall Street banks, private credit funds, and technology giants finance the digital infrastructure of the modern age.
