▲ AI data center (file photo)
An analysis has emerged that the artificial intelligence (AI) boom has been propping up economic growth while simultaneously acting as a factor driving up U.S. Treasury yields.
The New York Times (NYT) reported on the 20th (local time) that five major hyperscalers—Alphabet, Amazon, Meta Platforms, Microsoft, and Oracle—are increasing bond issuances instead of using cash, fueling the rise in U.S. Treasury yields.
According to Refinitiv data, debt issuance by these five companies did not exceed an annual average of 30 billion dollars (approx. 42 trillion won) between 2020 and 2024, but surpassed about 140 trillion won last year, and has already broken through approximately 280 trillion won this year.
Microsoft is the only one among them that has not raised funds in the bond market over the past year.
Asset management firm Vanguard projected that the total volume of AI-related bond issuances, extending beyond these five companies, will exceed approximately 1,400 trillion won annually from 2027 to 2030.
"For most of the past decade, major tech companies funded their AI investments with operating cash flow, but that era is coming to an end," said Lucas Baines, a senior investment strategist at Vanguard.
The borrowing burden is also being reflected in financing costs.
The yield spreads over U.S. Treasuries at the time of issuance rose for Alphabet from 0.63 percentage points in April to 0.85 percentage points this month, and for Amazon from 0.55 percentage points in November last year to 0.8 percentage points in July.
For Oracle, which has the lowest credit rating among them, the spread climbed from 1.05 percentage points in September last year to 1.45 percentage points in February this year.
The average spread on investment-grade corporate bonds of similar maturity is a little over 1 percentage point.
The expansion of AI investment fueled by debt financing has raised growth forecasts for the economy as a whole. Generally, stronger economic growth increases the incentive for the Federal Reserve (Fed) to maintain high interest rates to curb inflation.
Some analysts suggest that the recent rise in U.S. long-term Treasury yields is partially a reflection of expectations that "AI-driven growth expectations will prolong the Fed's tightening stance."
Matt King, founder of Satori Insights, said, "Now that hyperscalers have to borrow money, additional capital expenditures are no longer free," adding, "The rise in real interest rates is driving up costs for the rest of the economy as well."
The 30-year U.S. Treasury yield surged to its highest level since 2007 this week before retreating due to the Treasury Department's expanded buyback measures, and then turned upward again on the 20th, with the surge in AI-driven corporate bond issuances pointed to as one of the backdrops for this trend.
(Photo: AP, Yonhap News)
※
Copying, redistribution, and unauthorized use in AI training are strictly prohibited.