Bonds issued by tech titans that are ‘hyperscaling’ their artificial intelligence (AI) infrastructure have taken hits in recent weeks.
Investors are now selling off the debt accumulated by these companies, highlighting growing concern over tech groups turning to debt markets to finance their AI ambitions, effectively asking public investors to bankroll their expansion.
This wave of debt-fuelled investment has raised red flags and intensified anxiety among fixed-income investors, as risks mount from the billions being spent on AI.
The Debt-Fuelled Growth in AI Spending
Despite holding vast reserves of cash, big tech companies are issuing debt faster than ever to fund their AI expansion. The Financial Times reports that Google, Amazon, Microsoft and Meta are projected to spend more than £303.7 billion ($400 billion) on data centres alone in 2026.
All four companies are likely to use some liquid cash to fund construction.
JP Morgan revealed that the four collectively hold around £265.7 billion ($350 billion) in liquid cash and investments, expected to generate roughly £550.4 billion ($725 billion) in operating cash flow in 2026.
The firm also noted that building AI infrastructure could cost more than £3.7 trillion ($5 trillion), a scale that will ‘likely require participation from every public capital market as well as private credit, alternative capital providers and even government involvement’.
Meanwhile, Bank of America reported that tech giants borrowed about £56.9 billion ($75 billion) in bonds and loans during September and October alone to finance data-centre construction, signalling a structural shift in which AI capital expenditures (capex) now exceed internal cash-flow limits.
Long-Term Debt
The spread, or premium in yield, demanded by investors for tech firms’ bonds climbed to 0.78 per cent, the highest level since US President Donald Trump introduced tariff plans that rattled markets earlier this year.
The company that suffered most within the group is Oracle. Data obtained by the Financial Times shows that Oracle’s bonds have fallen nearly 5 per cent since mid-September.
According to Bloomberg, the company has about £72.8 billion ($96 billion) in long-term debt, balances that have risen sharply as part of a series of leasing deals to provide computing power for OpenAI’s ChatGPT.
OpenAI reportedly told Oracle that its AI software could generate £227.7 billion ($300 billion) in revenue over the next five years.
Analysts remain divided over the sector’s reliance on large financial commitments from a small group of AI firms to sustain growth. Credit-rating agency Moody’s has warned of risks tied to this concentration.
Meanwhile, George Pearkes of Bespoke Investment Group said the AI debt cycle is still in its ‘early stages’, noting that what worries him most is the potential increase in supply rather than a full ‘sell-off’.
Debt Financing Amplifies Risks
Historically, these AI companies have funded innovation through strong internal cash flows. However, the scale and speed of current development are straining their own resources.
The growing reliance on debt financing could amplify risks if AI-driven returns fail to meet market expectations or if tighter economic conditions limit credit availability.
Bank of America warned that the sector’s heavy borrowing spree has placed their financial profiles under increasing pressure, especially given the pace of spending on AI.
Originally published on IBTimes UK
