The Era of “Hyper-Scale” Borrowing
The artificial intelligence revolution has transitioned from a software narrative to a hard asset reality, and the bill is coming due. As of late 2025, the financing landscape for AI infrastructure has shifted dramatically from corporate cash reserves to the debt markets. According to recent data, AI data center and project financing deals have surged to $125 billion this year, a nearly tenfold increase from $15 billion during the same period in 2024.
While equity valuations grab headlines, debt is quietly financing the steel, concrete, and gigawatts required to power the next generation of compute. JPMorgan estimates that the total build-out cost could reach between $5 trillion and $10 trillion, creating an unprecedented demand for capital that public markets alone cannot satisfy. This marks a structural shift: the largest technology companies are no longer just cash-rich fortresses; they are becoming aggressive borrowers, utilizing complex instruments like synthetic leases and asset-backed securities (ABS) to keep liabilities off their balance sheets.
Private Credit Steps Into the Gap
With traditional banks constrained by regulatory caps and public markets wary of overexposure, private credit has emerged as a critical liquidity provider. Morgan Stanley estimates that private credit markets could supply over half of the $1.5 trillion specifically needed for data center construction through 2028.
This influx of private capital is not merely filling a void; it is reshaping risk. Major players like Oracle and Meta have recently tapped debt markets for billions to fund campuses, yet smaller operators and new entrants are increasingly relying on non-bank lenders. The flexibility of private credit allows for the bespoke structuring required by these complex, capital-intensive projects—projects where traditional bank covenants often fail to account for the unique cash flow dynamics of high-performance computing assets.
The “Glut” Risk and Market Discipline
However, this flood of liquidity brings volatility. Warnings from the Bank of England and Oaktree Capital regarding a potential “glut” of data center capacity have introduced a note of caution. We are seeing signs of stress in specific pockets; Oracle’s credit default swaps (CDS)—insurance against default—recently spiked to their highest levels since 2009, reflecting investor anxiety over the sheer scale of capital expenditure relative to immediate returns.
The market is currently bifurcated: “Risk-On” exuberance drives valuations for the winners, while a “Risk-Off” reality sets in for those over-leveraged on unproven technology. The emergence of “circular deals,” where tech giants finance their own customers to buy their chips, adds a layer of opacity that sophisticated lenders must pierce.
What This Means for Borrowers
For borrowers in the infrastructure and technology space, the message is clear: capital is available, but the terms are tightening for those without tangible asset backing. The ease of 2024’s financing environment has been replaced by a more discerning market in late 2025.
Bond Capital’s View: The physical build-out of AI requires massive capital. While equity grabs headlines, debt finances the steel and concrete. We understand complex CapEx needs. However, as the market froths with complex financial engineering and “synthetic” structures, Bond Capital maintains strict credit discipline. We look past the AI hype to underwrite the fundamental value of the asset—the land, the power rights, and the hardware. In a market debating a bubble, we remain focused on the bedrock of tangible security.
