Investors are offloading the debt of major U.S. technology companies as concerns mount over Silicon Valley’s aggressive spending on artificial intelligence (AI). This trend is now rippling through the bond market.
Bonds issued by leading “hyperscalers”—tech giants such as Alphabet Inc., Meta Platforms Inc., Microsoft Corp., and Oracle Corp., which are heavily investing in massive data centre projects—have declined in recent weeks.
Such bonds are yielding 0.78 percentage points more than U.S. Treasuries, according to Bank of America data — the most since April, when volatility erupted in markets following President Donald Trump’s tariff announcements. In September, the spread was only 0.5 percentage points — a notable shift in investor attitudes.
The widening spread reflects growing investor unease over the way tech giants are turning to debt markets to finance their massive investments in AI infrastructure.
“The important thing the market woke up to in the past two weeks is that it’s the public markets that are going to need to finance this AI boom,” said Brij Khurana, fixed income portfolio manager at Wellington Management.
SoftBank capitalizes on AI surge
Meanwhile, Japan’s SoftBank Group has made a bold move to capitalize on the surging AI wave. The conglomerate announced Tuesday that it has sold its entire stake in U.S. chipmaker Nvidia for $5.83 billion as it doubles down on its “all in” bet on ChatGPT creator OpenAI.
The firm said in its earnings statement that it sold 32.1 million Nvidia shares in October. It also disclosed that it sold part of its T-Mobile stake for $9.17 billion.
“We want to provide a lot of investment opportunities for investors, while we can still maintain financial strength,” said SoftBank’s chief financial officer, Yoshimitsu Goto, during an investor presentation.
“So through those options and tools we make sure that we are ready for funding in a very safe manner,” he said in comments translated by the company, adding that the stake sales were part of the firm’s strategy for “asset monetization.”
Trillion-dollar AI buildout raises debt and energy concerns
JPMorgan estimates that building AI infrastructure will cost more than $5 trillion, requiring financing from every corner of public and private markets—ranging from traditional bonds to private credit and even potential government participation.
The enormous scale of AI investment has raised concerns over potential overcapacity, long-term profitability, and energy consumption. Tech titans Google, Amazon, Microsoft, and Meta are expected to spend more than $400 billion on data centres in 2026, on top of over $350 billion this year.
Large tech firms, including mainstream tech companies, are increasing their borrowing for AI-related infrastructure, despite holding around $350 billion in liquid assets. Analysts estimate that the companies may produce $725 billion in operating cash flow by 2026. Yet, despite this strong position of liquidity, a flood of debt from such high-rated issuers is chasing new money into credit markets, indicative of a wider push towards more leverage to fuel the AI capital expenditure boom.
In recent weeks, Meta, Alphabet, and Oracle have each hit the bond market with massive offerings—some extending up to 40 years in maturity.
Meta last month secured a $27 billion private debt deal with investors, including Pimco and Blue Owl Capital, to finance its “Hyperion” data centre in Louisiana. The company followed that up with a $30 billion bond sale at the end of October—the largest corporate bond issuance since 2023.
Alphabet issued $25 billion in bonds in early November, including $17.5 billion in the U.S. and $7.5 billion in Europe. Meanwhile, Oracle sold $18 billion in bonds in September to fund infrastructure leases such as OpenAI’s “Stargate” data centre in Abilene, Texas.
Oracle’s bonds under pressure
Analysts say Oracle’s debt has been particularly hard hit. A Financial Times index tracking Oracle’s bonds issued before its latest sale has dropped nearly 5% since mid-September, compared to a 1% decline for the broader Ice Data Services index of U.S. high-grade tech debt.
Oracle now has approximately $96 billion in long-term debt, accumulated through deals to lease computing power to the makers of ChatGPT. This OpenAI collaboration is expected to generate $300 billion in revenue over five years.
However, Moody’s has cautioned that there are risks associated with Oracle’s dependence on a handful of AI firms to aid growth.
Some analysts view the recent sell-off as a healthy sign of price discovery in a new credit cycle. George Pearkes, macro strategist at Bespoke Investment Group, noted that as long as they are still pricing incremental risk, it’s a good sign.
Don’t just read crypto news. Understand it. Subscribe to our newsletter. It's free.
