The Great Risk Transfer
The narrative of the Artificial Intelligence boom has fundamentally shifted. For the past two years, the story was written in equity markets and valuations; today, it is being written in the credit markets. As the race to build the physical backbone of AI accelerates, the world’s largest technology companies are aggressively maneuvering to keep the colossal costs off their balance sheets.
Recent data indicates that tech giants, including Microsoft and Meta, have executed tens of billions in lease agreements and off-balance-sheet financing to construct data centers. By utilizing private credit partners to hold the assets, these corporations protect their own credit ratings while fueling a construction boom that rivals the greatest infrastructure projects of the last century. However, this financial engineering has a cost: it pushes risk away from blue-chip balance sheets and into the darker corners of the private lending market.
The Numbers: A Lending Frenzy
The scale of the capital requirement is staggering. UBS reports that AI-related project finance deals surged to $125 billion year-to-date, up from just $15 billion in the same period last year. JPMorgan and McKinsey now estimate the total infrastructure tab—covering data centers and power grid upgrades—could reach between $5 trillion and $7 trillion by 2030.
This insatiable demand for capital has birthed a dangerous exuberant froth. Reports from private credit desks describe borrowers requesting loans covering 150% of construction costs, justifying the over-leverage based on theoretical future valuation uplifts once the servers go live. When borrowers begin pricing perfection into their debt structures, the market has officially entered a “Risk-On” state of potential irrationality.
Regulators are noticing. The Bank of England recently issued warnings regarding the concentration of debt in AI infrastructure, and credit default swaps (CDS) for major players like Oracle have spiked, signaling that institutional investors are quietly hedging against the possibility that the AI payoff may not arrive on schedule.
Bond Capital: Financing Steel, Not Speculation
At Bond Capital, we analyze these trends through the lens of asset integrity. The physical build-out of AI—the steel, concrete, HVAC systems, and power substations—requires massive, flexible capital. This is a tangible manufacturing revolution disguised as a tech trend.
However, we distinguish between financing the asset and financing the hype. While the broader market chases synthetic leases and aggressive loan-to-value ratios based on projected software revenues, the soundest investments remain rooted in physical collateral. The current lending frenzy offers a cautionary tale: liquidity is abundant, but discipline is scarce. For borrowers in the infrastructure space, the goal should be securing partners who understand the complexity of heavy capex cycles, rather than predatory lenders merely looking to capitalize on the bubble.
As the market floods with capital willing to take equity-like risks for debt-like returns, Bond Capital maintains its directive: we fund business fundamentals, ensuring that when the hype cycle cools, the foundation remains solid.
