AI's Debt Reckoning Looms as Hyperscalers' Hidden Borrowing Hits $1.65T

Summary

Tech giants are in the midst of an unprecedented borrowing binge to fund AI infrastructure, with Amazon among those raising capital expenditure guidance despite already massive spending plans. S&P Global reports that AI hyperscalers and related entities have issued $225 billion in bonds through mid-2026, putting them on pace for a record $400 billion in annual issuance—a 973.7% jump from the prior year.

However, market participants are showing signs of strain under the weight of this debt flood. Bond spreads are widening as investors demand higher risk premiums, and economists warn that demand for both corporate and government debt may not endure indefinitely. This comes as the federal government competes for investor capital amid a projected $2 trillion budget deficit, while the Federal Reserve no longer provides purchasing support for Treasury securities.

The true scope of AI infrastructure financing extends far beyond visible bond markets. Nikkei's analysis reveals that so-called hidden debt at U.S. tech giants has exploded to $1.65 trillion over just four years—exceeding the $1.35 trillion in official debt on their books. These off-balance-sheet commitments include long-term GPU purchase agreements, server leases, and data center arrangements that are technically disclosed but not reflected as formal debt.

Moody's similarly flagged $1.2 trillion in off-balance-sheet liabilities, with over $820 billion attributed to data centers still under construction. While these arrangements comply with accounting rules, they represent massive future obligations that will eventually crystallize into official debt. Credit rating agencies and economists caution that the convergence of record corporate borrowing and historic government deficits could soon strain investor appetite.

• Combined corporate and government debt issuance at historic levels outside pandemic era, straining capital markets
• Sustainability of AI funding model depends on continued investor appetite amid Federal Reserve withdrawal from Treasury markets

Editorial Opinion

The AI infrastructure buildout is being financed through historically unprecedented borrowing that already exceeds pandemic-era levels. While investors remain eager for the high yields offered by tech giants today, the sheer scale of issuance—coupled with nearly $1.65 trillion in hidden obligations—suggests a financial reckoning awaits if lending sentiment cools. The simultaneous record government deficits and Federal Reserve exit from Treasury purchasing eliminate traditional shock absorbers, leaving this AI funding model vulnerable to credit market disruption. Whether this debt-fueled expansion represents necessary infrastructure investment or a bubble-building exercise will become clear only in hindsight.