Big Tech’s AI borrowing boom tests US bond market
Big Tech is on track to become the dominant force in the US corporate bond market, as the race to build AI infrastructure drives an unprecedented surge in borrowing that is beginning to worry investors.
By the end of the decade, around half of the 10 largest borrowers in the US bond market will be so-called hyperscalers, such as Alphabet, Amazon, Meta, Microsoft and Oracle, according to forecasts from Apollo Global Management.
This would mark a huge shift for a market long dominated by banks and telecoms groups, whose steady cash flows traditionally insulated bond investors from the volatility of tech-heavy equity markets.
Instead, credit portfolios are becoming increasingly exposed to AI.
The change is being driven by the sheer scale of capital expenditure required to build data centres, power infrastructure and custom chips to support generative AI.
Morgan Stanley estimates hyperscalers and adjacent sectors will raise about $400bn (£291.17bn) from the US investment-grade bond market in 2026 alone, up from $170bn last year and just $44bn in 2024.
Meanwhile, banks are expected to borrow less as regulatory changes allow them to hold less capital on their balance sheets, accelerating Big Tech’s rise up the borrowing league tables.
Credit markets feel the strain
The growing concentration is already visible in benchmark indices, with JPMorgan saying AI and data-centre-related issuers now account for 14.5 per cent of its US Liquid Index.
The firm expects that figure to exceed 20 per cent by 2030.
That worries some investors, who fear that credit markets are becoming more correlated with equities just as questions mount over whether AI spending will generate sufficient returns.
“What appears diversified across issuers and sectors increasingly represents a single macro trade on AI,” Apollo analysts warned in their latest credit outlook.
But the concern is not only the hyperscalers themselves. The AI investment boom has also boosted borrowing across utilities, industrials and energy firms supplying power and equipment to data centres.
This has extended AI exposure far deeper into credit portfolios than issuer lists suggest.
Those concerns have only resurfaced as tech stocks have wobbled in recent months and investment-grade credit spreads, the premium companies pay over US treasuries, sit close to historic lows, leaving little margin for error if sentiment turns.
Investors managing trillions of dollars have begun to trim exposure amid concerns of a so-called AI bubble, and many see echoes of earlier cycles in which heavy borrowing preceded a reassessment of risk.
“There’s fear in markets, and everyone’s looking for the next shoe to drop,” said Brian Kloss, a portfolio manager at Brandywine Global, who has been taking profits in top-rated credit.
The effects are already visible at the issuer level. For example, after Oracle sold $18bn of bonds last autumn, its credit spread jumped by more than 0.75 percentage points, according to S&P Global data.
Barclays strategist Dominique Toublan asked how markets would cope if hyperscalers were forced to tap bond investors quarter after quarter to fund AI expansion.
“If they have to borrow $10bn every quarter, how is the market going to react?” he said.
Still safe
Others argue that these risks are being overstated.
The largest hyperscalers remain among the most cash-rich firms in the world, giving them significant capacity to absorb higher debt without threatening their ratings.
Nathaniel Rosenbaum, a strategist at JPMorgan, said the surge in issuance was “a net positive” for credit quality overall, while John Lloyd of Janus Henderson said that giants such as Alphabet and Meta could increase leverage substantially and still retain strong balance sheets.
“If AI blows up, it will be bad for their equity,” Lloyd added. “But their credit would likely still be very solid.”
Even so, the scale of borrowing is hard to ignore, and Apollo’s chief economist, Torsten Slok, has warned that competition between corporate bond issuance and US Treasury issuance could push yields higher across the economy.
What’s more, rising long-term yields influence mortgage rates, corporate investment decisions and government finances, giving bond investors outsized power over the economic outlook.