Hyperscalers Alphabet, Amazon, Meta Platforms and Oracle have issued nearly $223 billion in bonds this year to fund AI buildouts, more than double all of 2025. The wave of supply is helping push Treasury yields higher, but it is also handing income investors yields as high as 8% on largely investment-grade debt.
Alphabet, Amazon, Meta Platforms and Oracle have issued nearly $223 billion in bonds this year, as of August 20, according to LSEG data cited by CNBC. That is more than double the amount seen in all of 2025. The companies are borrowing to fund their artificial intelligence buildout, and Wall Street is competing for the debt.
Hyperscaler supply pressures Treasury yields
Krishna Guha, head of economics and central bank strategy at Evercore ISI, said high-quality hyperscaler debt competes directly with government debt because relative-value investors use it to isolate issuer risk while hedging out broader market risk. That dynamic, combined with rising issuance across currencies, helps explain why sovereign yields are moving higher globally, he said.
The 30-year Treasury yield hit a 19-year high on Tuesday when it topped 5.33%, then eased on Wednesday after the Treasury Department said it would step up its bond buybacks. However, the yield moved higher again on Thursday. Wall Street has also pointed to the ballooning deficit and rising inflation fears, and total U.S. government debt passed the $40 trillion mark as of Tuesday, more than doubling in a decade.
Where the yields stand today
Investors can currently grab yields anywhere from 4.75% to 8% on hyperscaler bonds, depending on the issuer and maturity, said Dominic Pappalardo, chief multi-asset strategist for Morningstar Wealth. He noted the bonds are largely investment grade and longer dated, and while rising debt levels make the issuers riskier, these are large, profitable companies with cash on hand and growing businesses.
Leslie Falconio, head of taxable fixed income strategy at UBS Americas, said demand has come from insurance companies, pension funds and foundations even though the supply surprised the market. According to CNBC: "This is a great time to lock in these yields." She added that investors will see pockets of credit-spread widening as new supply comes to market.
Vishal Khanduja, head of broad markets fixed income at Morgan Stanley Investment Management, estimates investors can earn about 6.5% for long-end bonds issued by investment-grade AI leaders. He cautioned that credit quality and structural quality vary widely across the cohort, so results years down the line could differ significantly between issuers.
What investors are watching
Khanduja's team screens each bond for the off-taker, or the customer buying the underlying project, and whether the deal includes a tie-in requiring the off-taker to pay if it doesn't take the lease. He also weighs the location of the project, its power agreements, municipal approvals, timeline and the contractor's track record.
Falconio said investors already holding a core bond fund have some exposure to hyperscalers, since about 8% of the investment-grade index sits in such assets. Pappalardo warned that some of the AI investment behind the debt is still speculative, since the companies are banking on revenue and profits that haven't yet materialized. He added that continued issuance is likely, which could keep flooding the market with supply.
Source: CNBC
Trading involves risk.
Risk Disclaimer
