A debate is splitting bond desks on Monday: whether record borrowing by AI hyperscalers — Alphabet, Amazon, Meta, Microsoft and Oracle — is competing with US Treasuries for liability-matching capital and lifting long-term yields, or whether the duration effect is smaller than the AI-debt narrative suggests.
Risk.net reports that market participants fall on both sides of the argument. The question is whether the huge volume of debt issued by so-called hyperscalers — technology giants borrowing to build artificial intelligence data centres — now competes directly with Treasuries for the same investor capital, and whether that competition is helping push long-term interest rates higher.
Here is why the answer matters beyond the trading floor. Insurers and pension funds hold 30-year liabilities and need long-duration assets to match them.
AI Debt Creates Duration Supply
- Risk.net reports participants are split on whether hyperscaler debt is competing with Treasuries for liability-matching capital and lifting long-term yields.
- Research by Hugo De Vere, Srini Ramaswamy and Seth Searls argues AI issuers create duration supply from both public and private issuance, unlike banks, whose fixed-to-floating swap activity is largely duration-neutral.
- Alphabet, Meta and Microsoft were added to the S&P Global CDX North America Investment Grade Index in March 2026, reflecting demand for protection on those issuers.
AI Issuers Prefer Fixed-Rate Debt
According to research by Hugo De Vere, Srini Ramaswamy and Seth Searls, many investment-grade issuers — banks in particular — prefer floating-rate debt and use swaps to convert their fixed-rate bonds to floating, making their issuance effectively duration-neutral.
AI issuers behave differently. Like utilities, they prefer fixed-rate debt and keep the fixed-rate exposure on their books. When they issue private debt, which is typically floating, they use swaps to convert it to fixed. The conclusion from Hugo De Vere, Srini Ramaswamy and Seth Searls: hyperscalers are creating public duration supply from both private and public issuance.
Under the headline “AI Is Driving Up Treasury Yields: ‘It Just Touches Everything,'” Bloomberg journalists Davide Barbuscia, Ye Xie and Michael MacKenzie documented on August 17, 2026 how the AI financing wave was being felt across the Treasury market.
“It just touches everything.”
That phrase, quoted in Bloomberg’s August 17 reporting, summed up how traders described AI debt’s reach into rate markets.
The Case Against a Yield Shock
Research published by the Dallas Federal Reserve pushes back on the strongest version of the story. Its piece, titled “How AI debt financing impacts duration supply and interest rates,” examines whether the competition for capital is as direct as the bears assume.
The theoretical starting point: investors should, in principle, allocate across all available debt in proportion to risk-adjusted yield. That means new supply of any debt can draw dollars away from other debt. But the research notes that many frictions limit investment across the credit and duration spectrum, so while some supply effect is always expected, its size depends on the magnitude of issuance.
For many investors, general investment-grade corporate debt does compete directly with Treasurys — anyone tracking an aggregate bond index holds both. That is where the supply effect bites hardest, per the analysis.
Five Names Now Anchor the Credit Market
The hyperscaler cohort is small but liquid. Bloomberg reports five current constituents of the S&P Global CDX North America Investment Grade Index: Alphabet, Amazon, Meta, Microsoft and Oracle.
| Issuer | CDX IG Constituent | Added to Index |
|---|---|---|
| Alphabet (NASDAQ: GOOGL) | Yes | March 2026 |
| Amazon (NASDAQ: AMZN) | Yes | Earlier revision |
| Meta (NASDAQ: META) | Yes | March 2026 |
| Microsoft (NASDAQ: MSFT) | Yes | March 2026 |
| Oracle (NYSE: ORCL) | Yes | Earlier revision |
Alphabet, Meta and Microsoft were added during the latest index revision in March 2026. Bloomberg attributes the additions to strong demand for protection on those issuers, as credit investors wanted hedging tools while borrowing increased. The five names, Bloomberg notes, tend to be the most liquid hyperscaler credits.
Bloomberg reported on May 23, 2026 that the hyperscaler debt flood had brought a derivatives bonanza, as swap and index markets absorbed the hedging flow generated by the new issuance.
What the Dallas Fed Adds to the Ledger
A bank issuing a fixed-rate bond and swapping it to floating adds no net duration to the market. An AI issuer issuing fixed — or swapping private floating debt into fixed — adds duration twice over.
Lucas Baynes published “The AI buildout comes to the bond market” on Vanguard.com on August 19, 2026, examining how the data-centre financing wave was reshaping credit supply.
Where the Debate Stays Open
Risk.net frames it as a live split among market participants: insurers with 30-year liabilities to match are the swing vote, and whether their Treasuries allocations are being displaced by hyperscaler paper is precisely what the two camps dispute.
Fixed-rate issuance from AI builders, swap-driven duration creation from private credit, and index-level demand for protection on the biggest issuers all point the same way — more long-duration corporate supply than the investment-grade market has been used to absorbing.
The unresolved variable is magnitude.
For anyone tracking this, the watch items are the next CDX index revision — Alphabet, Meta and Microsoft were added at the latest revision in March 2026, reflecting strong demand for protection on those names — and the pace of new data-centre bond deals out of Alphabet, Amazon, Meta, Microsoft and Oracle.