The artificial-intelligence boom already has fueled a record stock-market surge. Now it’s driving up long-term Treasury yields too.
Meta Platforms, Oracle and other technology companies have raised $250 billion in debt markets globally this year, according to Morgan Stanley, borrowing at a scale that would have been hard to imagine only a few years ago.
The recent surge in AI-related infrastructure investment is partly behind the May rout that pushed 30-year Treasury yields to their highest level since 2007, analysts say, alongside inflation fears and shifting expectations for Federal Reserve policy. Treasury yields have come down from their May highs but remain above the levels where they began the year, reflecting in part the wave of bond issuance.
“We’re talking $750 billion, $850 billion of capex spending on an annualized basis, and it’s expected to ramp up to close to a trillion next year,” said Thomas Urano of Sage Advisory in Austin. “It’s kind of like thinking about a federal stimulus package or some kind of infrastructure spend at a federal level.”
UNLIKELY BOND-MARKET STAR
Corporate borrowing to pay for AI investment could help shape Treasury valuations for years to come, investors say.
Hyperscalers are raising vast sums to finance data centers, power systems and computing capacity, and long-term financing offers them several advantages - though Alphabet’s GOOGL.O $80 billion stock-sale announcement shows the firms have multiple fundraising options.
Data centers combine assets with very different lives: AI chips may need to be replaced every few years, but buildings, land, power connections and other infrastructure can last 20 to 30 years. Companies building long-lived assets have a strong incentive to lock in long-term fixed-rate funding.
Oracle, once a relatively minor issuer of long-term debt, has become one of the largest suppliers of duration risk - a measure of interest-rate exposure watched closely in the bond market and sought after by long-term investors such as insurance companies - in the investment-grade universe, said Srini Ramaswamy, a senior financial economist at the Federal Reserve Bank of Dallas.
Ramaswamy estimates that AI-related issuance accounts for roughly 15 per cent of the duration supplied by total Treasury issuance. “That’s a huge number,” he said.
The cost of Oracle’s five-year credit default swap ORCL5YUSAX=MG - used by investors to speculate on credit quality or hedge debt holdings - has surged to 150 basis points from around 30 basis points in the last year, reflecting concerns about the company’s sharply increased debt load.
BORROWING LONG AND SHORT
Though the reported long-term borrowing by the large technology firms has soared, institutional incentives mean that the real numbers may even be higher. Many institutional investors face internal limits on how much debt they can hold from any one company, and on duration, Ramaswamy adds. Private lenders also typically prefer floating rate loans.
As a result, technology borrowers may issue shorter-dated or floating-rate debt and then use interest-rate swaps to convert that exposure into longer-term fixed-rate funding.
Ramaswamy estimates that this swap activity alone accounted for about $50 billion in 10-year-equivalent supply in the fourth quarter, and that the figure is probably higher now.
The Treasury, by comparison, sold $540 billion in 10-year notes in the past year.
That makes the recent Treasury selloff more complex than a simple story about the Fed and inflation. And it can be seen in real yields, which strip out inflation expectations.
In a typical inflation scare, long-term inflation expectations rise sharply, said Jonathan Hill, head of U.S. inflation market strategy at Barclays.
This time, real yields have risen while inflation expectations have stayed relatively contained — a pattern Hill says is consistent with an AI-driven investment boom that is pushing up capital demand now, but could improve productivity and ease inflation over time.
The scale of U.S. government borrowing has also been driving the move. Higher yields raise debt-service costs; higher debt-service costs require more issuance; and more issuance can push yields higher still, Hill said.
None of this means AI borrowing is the dominant force in Treasuries. Fed policy, inflation data, fiscal deficits and global demand for safe assets still matter more. But the AI buildout is creating a new source of supply at a moment when the market is already being asked to absorb unusually heavy government borrowing.
“The issuance numbers on this are dramatic,” said Hill. “It’s running in the background and it truly is the big story.”
(Reporting by Karen Brettell; Additional reporting by Chuck Mikolajczak; Editing by Colin Barr and Andrea Ricci )
