Tech Giants' AI-Driven Debt Binge Drags on Global Bond Indices
0xBroomberg
Meta, Alphabet, Amazon and peers have sold over $300 billion in bonds since early 2025 to fund AI infrastructure. Those bonds have dropped an average of 3.3 points from issue price, spreading from the dollar market into global multi-currency indices and becoming the biggest drag on passive funds forced to hold them.
$300 billion in new debt — why are these bonds the worst performers?
Since early 2025, hyperscalers — tech firms that operate massive data centers, including Meta, Alphabet, Amazon and Oracle — have issued more than $300 billion in bonds to fund data centers and AI infrastructure.
Yet the bonds have fallen an average of 3.3 points from issue price, with a total return of negative 1.4% — ranking among the worst performers in every index they sit in.
This means → roughly 79% of these bonds have seen their spreads — the gap between a bond's yield and a benchmark rate, wider = more risk compensation demanded — widen since their first day of trading. Buyers are underwater.
The dollar market can't absorb it all — why go global?
The U.S. credit market is the world's largest bond market, but even it cannot digest supply at this scale alone.
The tech giants have turned to sterling, euro, yen and Swiss franc markets. About $80 billion of the debt was issued in non-dollar currencies.
In plain terms = one pool can't hold all the water, so they're filling several at once. Alphabet is now the largest non-financial issuer in the sterling investment-grade index; Amazon and Alphabet rank among the top four borrowers in the Swiss franc market.
Bryn Jones, head of fixed income at UBS Asset Management, warned: "They have to issue across all currency markets — some are building a presence from zero, and that creates enormous pressure."
Who is stuck holding the bag?
Many investors hold these bonds not by choice. Passive funds tracking corporate-bond indices, and active managers bound by benchmark weights, have been forced to absorb the new supply.
This means → the moment a tech giant issues, the bonds enter the index and funds must buy to stay on-benchmark — regardless of price.
JPMorgan strategist Nathaniel Rosenbaum noted the spread widening is "a byproduct of the investment-grade investor base trying to rationally price accelerating issuance."
What did Amazon's $25 billion deal set off?
Amazon's $25 billion bond sale this month was the direct trigger for the latest pressure wave.
Oversubscription on the deal was low by the standards of comparable AI-linked mega-issuances — and it landed in early summer, when corporate-bond supply typically slows. The market expected a breather; instead it got another giant deal.
Goldman Sachs strategists flagged "technical headwinds" from rising AI-related issuance as the reason credit is underperforming equities; Barclays analysts warned against chasing wider spreads given "rising concentration risk."
Could the risk spread beyond bonds?
Rufaro Chiriseri, head of European fixed income at RBC Wealth Management, said her team is underweight tech and warned: if AI capex disappoints in any form, "the reaction won't be confined to one market."
This reflects a widening anxiety beyond the hyperscalers themselves — data-center operator Prime Data Centers has already postponed a planned bond issue.
Jack Daley, portfolio manager at TwentyFour Asset Management, said his team avoided much of the new issuance: "These companies are obviously very strong, but the question is balancing that against supply and the market's ability to digest it. People will still say they're cheap relative to ratings, but if supply keeps coming, cheap can stay cheap for a long time."
What would it take for these bonds to stabilize?
JPMorgan strategists noted that insurers — one of the most stable buyer bases for corporate bonds — are still well below their concentration limits for the sector, and with yields at historic highs, they remain committed buyers.
In plain terms = the largest potential backstop hasn't maxed out its allocation yet. There is room.
But stabilization hinges on two variables: ① whether the pace of issuance slows and ② whether AI capex delivers returns — the first governs short-term digestion pressure, the second governs long-term credit quality.
Content is for reference only, not financial advice.