The $5 Trillion CapEx Wall
The narrative surrounding Artificial Intelligence is shifting from software capabilities to physical realities. New data suggests the global build-out of AI infrastructure—primarily data centers and energy grids—will carry a price tag exceeding $5 trillion over the next five years. While equity markets have previously celebrated this spending, credit markets are beginning to show signs of indigestion.
According to recent reports, AI data center and project financing deals have surged to $125 billion year-to-date, a staggering increase from just $15 billion during the same period in 2024. Major technology firms, including Amazon, Alphabet, and Meta, have flooded the market with nearly $90 billion in investment-grade bonds since September to fund this expansion. However, this massive supply is testing investor appetite. Oracle Corporation recently saw its credit default swaps (CDS)—the cost to insure against default—spike to their highest levels since 2009, signaling that public markets are becoming wary of the leveraging required to sustain this arms race.
Risk Transfer and the “Shadow” Build-out
A structural shift is occurring in how this infrastructure is financed. To protect their own balance sheets, hyperscalers (the massive tech giants) are increasingly using complex leasing structures to offload risk onto smaller developers and intermediate operators. While this keeps headline debt levels manageable for the tech titans, it concentrates risk in the private sector.
Regulators and institutional investors are taking note. The Bank of England has issued warnings regarding financial stability risks tied to AI debt, and distressed debt veterans like Oaktree’s Howard Marks have flagged the potential for a supply glut. The concern is simple: if the monetization of AI software lags behind the physical construction of data centers, the borrowers holding the debt for the steel and concrete may face a liquidity crunch.
Despite these “bubble” fears, the demand for capital remains absolute. Traditional lenders, wary of the “irrational exuberance,” are tightening terms or demanding exorbitant coverage ratios—some seeking up to 150% of construction costs as collateral.
What This Means for Borrowers
The divergence between the “AI Dream” and the “Infrastructure Reality” creates a specific environment for borrowers. Public markets are becoming volatile and expensive for all but the highest-rated issuers.
For infrastructure developers and mid-market operators, the implication is clear: Reliance on broad capital markets is becoming risky as volatility increases. While banks struggle to underwrite the repayment timelines of unproven technologies, Bond Capital focuses on the tangible. We understand that regardless of software adoption rates, the physical assets—power generation, cooling systems, and real estate—hold intrinsic value. In a market where public liquidity is fearful, private credit offers the certainty required to pour the concrete.
