Big Tech's AI Debt Will Crash the Economy

Major technology companies are taking on unprecedented levels of debt to fund artificial intelligence (AI) infrastructure, sparking warnings from analysts and investors about systemic risks to financial markets. Firms like Alphabet, Meta, Amazon, Microsoft, and Oracle have collectively raised over $100 billion in 2025 alone—more than three times the historical average—to finance data centers and AI development. This marks a strategic shift from prior years when these companies relied primarily on their substantial cash reserves for capital expenditures.

Oracle Under Scrutiny Amid Aggressive AI Spending

Oracle has emerged as a focal point of concern due to its aggressive capital expenditure and debt financing. The company raised $18 billion in investment-grade bonds in 2025 to accelerate its AI cloud infrastructure, but its stock has since plunged nearly 40% from its September peak. Investors are questioning the sustainability of its spending, with forecasts showing negative free cash flow reaching $24.3 billion by fiscal 2028.

S&P Global Ratings downgraded Oracle’s outlook to negative in early November 2025, citing a strained credit profile from anticipated capex and debt issuance. Five-year credit default swaps (CDS) on Oracle have surged to their highest level in three years, reflecting growing leverage risk.

Despite strong earnings from AI leaders like Nvidia, investors are entering a “show me the money” phase, demanding clearer proof of returns on these massive investments.

Credit Markets Face Structural Shifts and Widening Spreads

The flood of tech debt is reshaping the corporate bond market. TD Securities predicts investment-grade credit spreads could widen to 100–110 basis points in 2026, up from 75–85 basis points in 2025, driven in part by a projected record $2.1 trillion in U.S. investment-grade issuance. Goldman Sachs warns that a continued shift toward debt financing by hyperscalers would “increase the macro risks associated with the AI build-out.”

JPMorgan strategist Matthew Bailey notes that “a flood of data center financing could cause supply indigestion,” particularly in dollar-denominated markets. While demand remains strong—Meta’s $30 billion bond offering attracted a record $125 billion in orders—there are growing fears about lower-quality AI-related issuers entering the market.

Systemic Risks and the Threat of an AI Bubble

Analysts are drawing parallels between today’s AI spending boom and the dot-com bubble of the early 2000s, when excessive debt funded fiber-optic infrastructure that outpaced demand. A Morgan Stanley report estimates Big Tech will spend $3 trillion on AI through 2028, with only half covered by internal cash flows. If AI revenue growth slows, the result could be massive overcapacity and stranded assets.

A Massachusetts Institute of Technology study cited in recent reports found that 95% of organizations see zero return from generative AI projects, raising doubts about the revenue assumptions underpinning current investments. Hedge fund Man Group has warned of a potential “glut of lower-quality names” and “AI slop” entering the market.

Mixed Investor Sentiment and Future Outlook

Despite mounting concerns, many investors remain bullish on megacap tech stocks due to their strong earnings, competitive advantages, and durable cash flows. UBS estimates that 80–90% of planned AI capex is still funded by internal cash, suggesting debt remains a supplementary tool.

However, the increasing use of leverage, especially among firms with weaker balance sheets, and the rise of interconnected, circular financing deals—such as banks issuing debt to fund data centers tied to tech firms—introduce new systemic vulnerabilities.

As Bob Savage of BNY notes, “We’re in a ‘show me the money’ phase,” where the burden of proof is shifting to companies to demonstrate that AI investments will generate returns justifying their debt loads.

Analysts are drawing strong parallels between the current AI investment surge and the dot-com bubble due to massive infrastructure spending, significant debt financing, and the potential for overcapacity and stranded assets if AI demand does not meet lofty expectations.

Key Parallels and Concerns

Infrastructure Overbuild: In the late 90s, telecom companies laid excessive fiber-optic cable in anticipation of demand that never fully materialized, resulting in “dark fiber” and catastrophic overcapacity. Today, major tech companies are investing hundreds of billions in AI data centers, leading to concerns that this physical infrastructure could similarly be overbuilt relative to actual near-term demand.
Financing Gap and Debt: A Morgan Stanley report estimates that while Big Tech may spend around $3 trillion on AI infrastructure through 2028, only about half can be covered by projected internal cash flows, necessitating massive borrowing. The reliance on debt, including off-balance-sheet special-purpose vehicles, increases risk if the market experiences a correction.
Valuations vs. Revenue: During the dot-com era, many companies had little to no revenue but massive valuations. While today’s leading AI companies (like Microsoft, Google, etc.) are highly profitable, a significant disconnect remains between the scale of investment and current AI-specific revenue, leading some to question whether valuations are justified.
Speculative Spending: The influx of capital into the sector, including practices like “circular spending” where one company funds another to buy its products, may artificially inflate demand signals and create an over-hyped environment.

Key Differences

Despite the concerns, analysts also point out key differences that may prevent a full dot-com style crash:

Stronger Fundamentals: Leading AI companies are generally very profitable with strong balance sheets, unlike many of the unprofitable startups that dominated the dot-com bubble.
Real-World Utility: AI technology is already widely adopted and providing tangible economic value and productivity gains across various industries, giving it a more solid foundation than the largely unproven business models of the early internet.
Demand Visibility: There is strong underlying demand for current data center capacity, and the buildout is considered a strategic national priority in many countries, suggesting a potentially more durable investment cycle.

Overall, the risk is less about a complete collapse of the technology and more about the potential for sharp market corrections in overvalued stocks and a period of lower returns on investment if demand growth slows down.

Note: We do use YouTube Video’s under the “Fair Use” Act under the Copyright Law:

“Fair use is a doctrine in the United States copyright law codified in Section 107 of the Copyright Act of 1976.¹ It provides for the legal, non-licensed citation or incorporation of copyrighted material in another author’s work without requiring permission from the rights holders, such as for commentary, criticism, news reporting, research, teaching or scholarship.⁰¹ The U.S. Copyright Office Fair Use Index should prove helpful in understanding what courts have to date considered to be fair or not fair but it is not a substitute for legal advice.²“