Moody’s Ratings warned on Tuesday, October 27, 2026, that heavy borrowing to fund artificial intelligence infrastructure is expanding private credit beyond traditional corporate lending, and that a wave of loans coming due between 2027 and 2029 will test weaker borrowers’ ability to refinance at higher interest rates. The warning lands as the five largest United States hyperscalers and CoreWeave are expected to spend nearly $1 trillion on capital projects in 2027, financed through cash flows, debt, joint ventures and special-purpose vehicles, according to Moody’s. It is not a prediction that the AI buildout will stop, but a flag that the financing behind it now carries risks of its own.
What Moody’s Actually Flagged
The first point is how broad the spending has become. Moody’s said AI financing needs now extend beyond data-centre construction to land, power, semiconductors, graphics processing units and tensor processing units. In other words, the bill is not just for buildings full of servers. It covers the ground they stand on, the electricity that feeds them, and the chips inside them, each with its own suppliers, lead times and lenders.
The second point is who is doing the lending. Private credit, lending by non-bank firms outside the traditional bank-loan market, is growing as a channel for this spending.
A telling detail in the Moody’s assessment is that about 17 percent of the 607 borrowers it tracked used payment-in-kind interest, which lets a borrower defer cash interest payments while adding the unpaid interest to its debt burden. The tool buys time, but it also means the debt keeps growing even when the borrower is not paying it down.
The 2027–2029 Refinancing Wall
The timing is what turns a financing trend into a risk. Moody’s said a wave of loans coming due between 2027 and 2029 will test weaker borrowers’ ability to refinance at higher interest rates. A Reuters analysis earlier this year found that loan maturities at United States business development companies would peak in 2028 and 2029, putting the heaviest refinancing load right in the middle of that window.
Recent stress in CoreWeave debt has already highlighted investors’ growing concerns about highly leveraged AI borrowers, according to Reuters. CoreWeave is one of the highest-profile names in the neocloud wave, and any wobble in its debt is read as a signal for the wider group of companies funding AI capacity with borrowed money.
The scale of the financing structures now in use shows why the warning carries weight. SpaceX closed a $40 billion Apollo deal to lease AI chips, and banks lined up $63 billion to finance Broadcom chips for Anthropic. These are not ordinary equipment purchases. They are large, structured commitments whose value depends on AI demand staying strong long enough to pay them back.
The “Circular” Financing Worry
Moody’s also flagged “circular” financing risks, as technology companies, chipmakers and AI developers increasingly invest in, lend to and buy from one another. The concern is correlation: if the same group of companies sits on both sides of many deals, a downturn does not hit one borrower in isolation. Exposures that looked diversified can weaken together, because they ultimately depend on the same AI spending cycle.
That is the deeper version of the refinancing worry. A single company missing a payment is a credit event. A network of companies that owe each other, supply each other and invest in each other is a system, and systems can seize up faster than any one balance sheet suggests.
Why It Matters
Most debate about the AI boom focuses on chips, power and demand: whether there are enough graphics processors, enough electricity, and enough customers to justify the spending. Moody’s warning adds the other half of the equation, which is time. Even if demand holds, the debt used to build capacity must be rolled over, and the terms of that rollover will depend on interest rates and investor confidence in 2027, 2028 and 2029, not on today’s optimism.
For weaker borrowers, that creates a squeeze with two jaws. Higher rates raise the cost of refinancing, while any cooling in AI demand would cut the revenue meant to service the debt.
The same caution has been sounded from other directions this year. The Bank of Korea warned that an AI slowdown could cut chip exports by $100 billion, a reminder that the financing story and the hardware story are two sides of one bet on continued AI growth.
Background
Moody’s Ratings is one of the major credit ratings agencies, and its assessments feed into how lenders price risk across corporate debt markets. Its warning, reported by Reuters on October 27, 2026, comes in a year of record AI infrastructure announcements, from data-centre builds to multi-billion-dollar chip financing deals, and as investors have begun asking more pointed questions about how that spending is funded and when it must be repaid.
0 Comments