In a recent televised interview, billionaire investor Ray Dalio warned that the United States is sprinting toward a debt crisis that could materialise before the close of the decade. The national debt has already breached the $40 trillion threshold, an inflation-adjusted increase of more than $10 trillion since 2019. Servicing that liability now costs over $1 trillion a year – roughly one-fifth of the entire federal budget – leaving shrinking room for Social Security, health care, defence and other priorities.
Dalio argues that the problem is not merely the size of the debt pile but the speed at which it is growing. When the rate of new borrowing outpaces repayment, the share of cash flow devoted to interest payments rises, squeezing discretionary spending. He also highlighted that geopolitical frictions have dampened foreign appetite for Treasury securities, while the rapid expansion of the artificial intelligence sector is being financed almost entirely with debt, tightening credit conditions in riskier markets such as auto loans.
AI-driven corporate borrowing strains the investment-grade bond market
Tech giants that are racing to build AI infrastructure have begun to feel the pressure in the bond market. Oracle, the most aggressive borrower in this space, saw its five-year credit default swap (CDS) spread climb to about 203 basis points – the widest level in 18 years – and its rating was pushed to BBB-, just above junk status. Similar widening has been observed for Amazon, Microsoft, Meta and Alphabet, whose spreads over Treasuries rose toward 80 basis points in September, up from roughly 50 basis points earlier in the year.
Rising spreads on hyperscaler issuances
Demand for these AI-related bonds is also cooling. The cover ratio, which measures how many dollars of orders are received for each dollar of bonds sold, fell from a robust 5-to-1 in February to roughly 2-to-1 by July. Meta’s $25 billion bond sale, for example, attracted $96 billion in peak orders, a noticeable drop from the $125 billion that chased a comparable issue a year earlier. Morgan Stanley estimates that AI-linked global debt issuance will top $570 billion by the end of 2026 – more than double the pace of the previous year – and that tech now represents about 18 % of all U.S. investment-grade bond supply, the highest share on record.
Off-balance-sheet data-center financing
Complicating the picture, a substantial portion of AI-related borrowing is hidden in special-purpose vehicles that sit off the borrowers’ balance sheets. BlackRock, for instance, is arranging more than $12 billion of bonds for Meta’s 1-gigawatt data centre in El Paso, Texas, through a vehicle dubbed Project Sopaipilla Holdings – one of the largest single-infrastructure financings ever recorded. This layer of leverage resides a step removed from the companies whose names appear on the data centres, making risk assessment more opaque for traditional investors.
International bond markets react: France, the United States and safe-haven flows
A broader sell-off in sovereign bonds has amplified these dynamics. When French government securities surged higher amid fiscal and political worries, investors fled to perceived safe havens – notably U.S. Treasuries and German bunds. The 10-year Treasury yield hovered around 5.3 % while France’s 10-year OAT settled near 4.9 %, both at levels not seen since 2002. The rush to safety has tempered the upward pressure on U.S. yields, but they remain elevated as the AI boom and heavy corporate borrowing compete with sovereign debt for investor capital.
Analysts caution that the current environment could tighten further if bond spreads continue to widen and cover ratios keep shrinking. While the five biggest hyperscalers still generate massive free-cash-flow, projections indicate that cash generation could dip to zero or turn negative this fiscal year as AI-related capital expenditures surge – the sector’s total capex is expected to exceed $690 billion, an 80 % year-over-year rise. The combination of a mounting national debt, aggressive corporate borrowing and shifting global investor sentiment creates a scenario that Dalio likens to a “perfect storm” threatening a fiscal “heart attack.”
