The $1 Trillion AI Debt Binge: Moody's and Jefferies Warn Big Tech's Spending Spree Risks a Credit Crisis

Moody's central warning is blunt: AI infrastructure spending is running at a "trillion-dollar annual clip," forcing companies that were historically asset-light into an unprecedentedly asset-heavy model .

The $662 Billion Phantom Liability: Off-Balance-Sheet Leases

Perhaps the most concerning risk identified by Moody's is a massive hidden liability embedded in accounting rules. As of end-2025, the five major US hyperscalers had accumulated approximately $662 billion in future data center lease commitments that have not yet commenced, remaining entirely off their balance sheets under US GAAP (ASC 842) .

Combined with other commitments—including GPU supply contracts and servers—the total undiscounted future lease exposure reaches approximately $969 billion . A separate Nikkei analysis of financial statements found that Alphabet, Microsoft, Amazon, Meta, and Oracle carried $1.65 trillion in total off-balance-sheet commitments, mostly tied to data-center leases, GPU contracts, and servers .

Moody's accounting analysts David Gonzales and Alastair Drake calculated that the unrecorded $662 billion is equivalent to 113% of the combined adjusted debt of the five hyperscalers . As these leases commence, more than half a trillion dollars in data-center assets will begin appearing on corporate balance sheets, pressuring traditional credit metrics.

Moody's explicitly warned that companies using off-balance-sheet leasing to finance AI infrastructure risk a "material" deterioration in their credit profiles, putting the ratings agency at odds with its rival S&P over what has become one of the most popular financing methods for data center capacity .

The Circular Economy of AI: Self-Referential Spending

Both Moody's and Jefferies have flagged the interconnected, circular nature of AI financing as a growing fault line. Bloomberg documented a recurring "playbook": cloud computing companies and chipmakers—led by Nvidia—help fund leading AI developers, which then become some of their largest customers, creating a self-referential cycle of spending .

Microsoft has invested more than $13 billion in OpenAI, with the largest tranche of $10 billion from a deal announced in early 2023. Amazon and Alphabet's Google agreed to invest as much as $4 billion and $2 billion, respectively, in similar arrangements .

UBS has compared these interconnected deals to the telecom bubble at the turn of the millennium, noting concerns that companies invest in each other and rely on mutual spending . J.P. Morgan acknowledged the "scale and circularity" of large partnership announcements among AI model developers, hyperscalers, and chip companies, noting the capital commitments span multiple years with contingent execution targets .

Jefferies delivered a particularly stark warning: the biggest risk to the AI-driven tech cycle is not a chip supply disruption but a "sudden realisation by investors that the hyperscalers and the likes of OpenAI and Anthropic will not be able to make a return on their investment" . The brokerage argued that funding concerns could "trigger a sudden unwillingness to fund these investments which will then be aggravated by the circular arrangements between the main players" .

Record Bond Issuance: $108.7 Billion in a Single Quarter

Big Tech companies have flooded credit markets with an unprecedented wave of debt. Tech companies issued a record $108.7 billion in corporate bonds in the last three months of 2025, roughly double the previous quarterly record, with heavy issuance continuing into 2026 . Oracle, Meta, Alphabet, and others borrowed heavily to finance AI data centers and associated energy systems .

Bloomberg reported that these mega-bond offerings risk overwhelming buyers and weakening credit markets on both sides of the Atlantic, creating the largest wave of corporate bond supply in history . Investors warned that the unprecedented volume could disrupt passive credit funds, which are the primary buyers of this new supply . Some projections suggest tech firms could rely on debt to the tune of $1.5 trillion by 2028 for AI infrastructure expansion .

Eroding Free Cash Flow and Leverage Deterioration

Moody's is blunt about the erosion of credit quality. The race to build AI infrastructure at a trillion-dollar annual clip is "eroding free cash flow" at the hyperscalers, forcing even the world's most cash-rich companies to increase balance-sheet risk .

The agency wrote: "The transition from asset-light to asset-heavy models requires unprecedented levels of investment," draining free cash flow and pushing balance-sheet risk higher at six major technology companies: Microsoft, Amazon, Alphabet, Meta, Oracle, and CoreWeave .

Moody's has flagged Oracle and CoreWeave as AI's weakest credit links . Jefferies noted that hyperscalers are increasingly relying on debt rather than internal cash flows to finance AI investments, raising the risk of leverage deterioration if AI revenue fails to materialize on schedule .

Summary of Key Figures

The Bottom Line: A Credit Risk Regime Shift

The consensus across Moody's, Jefferies, and other analysts is clear: the AI infrastructure buildout has created a credit risk regime shift. Massive upfront spending, enormous hidden lease liabilities, circular deal structures, and declining free cash flow are all testing the credit profiles of the world's largest technology companies in ways never seen before. Moody's Analytics Chief Economist Mark Zandi has sounded the alarm, saying the leverage problem could inflict broader economic damage than previous market corrections, including the dot-com era .

The key question remains unanswered: will the revenue from AI applications ever justify the trillions being poured into its infrastructure?