AI debt binge pushes $250B into bond market, lifting Treasury yields

Technology companies have raised $250 billion in global debt markets this year to fund AI infrastructure, creating a new source of supply that is pushing long-term US Treasury yields higher.

The artificial-intelligence buildout has added a new force to the Treasury market: technology companies raised $250 billion in debt globally this year, contributing to the May rout that pushed 30-year yields to their highest since 2007, according to Morgan Stanley.

"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," Thomas Urano of Sage Advisory in Austin said. "It's kind of like thinking about a federal stimulus package."

AI-related issuance accounts for roughly 15 percent of the duration supplied by total Treasury issuance, according to Srini Ramaswamy, a senior financial economist at the Federal Reserve Bank of Dallas. Oracle's five-year credit default swap surged to 150 basis points from about 30 basis points in the past year, reflecting the company's sharply increased debt load. Swap activity converting shorter-dated debt into long-term fixed-rate exposure added about $50 billion in 10-year-equivalent supply in the fourth quarter alone, Ramaswamy estimated.

The scale of AI-driven borrowing means the recent Treasury selloff reflects more than just inflation and Fed policy. Real yields have risen while inflation expectations stayed contained — a pattern consistent with an investment boom that pushes up capital demand now but could improve productivity over time, said Jonathan Hill, head of US inflation market strategy at Barclays. The question for investors is whether the supply pressure eases as AI infrastructure matures or accelerates as the buildout intensifies.

Hyperscalers Reshape the Bond Market

Meta Platforms, Oracle and other hyperscale technology companies have become unlikely stars of the investment-grade bond market. Oracle, once a minor issuer of long-term debt, is now one of the largest suppliers of duration risk — a measure of interest-rate exposure watched closely by insurers and pension funds — in the investment-grade universe, Ramaswamy said.

The borrowing reflects the unique capital structure of AI infrastructure. Data centers combine assets with very different lives: AI chips may need replacement every few years, but buildings, land and power connections can last 20 to 30 years. Companies building long-lived assets have a strong incentive to lock in long-term fixed-rate funding, pushing them to issue bonds with maturities that compete directly with government debt.

Alphabet took a different route, announcing an $80 billion stock sale on June 1 to fund its AI spending, showing the firms have multiple fundraising options. But the debt route has dominated: Morgan Stanley's $250 billion figure captures issuance across the technology sector globally this year.

A Structural Shift or a Cyclical Surge?

The AI buildout's impact on Treasury yields may persist for years. The Treasury sold $540 billion in 10-year notes over the past year — a figure that looks less dominant when set against the $50 billion in swap-driven 10-year-equivalent supply Ramaswamy estimated for just one quarter, a number he said is probably higher now.

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 absorbing unusually heavy government borrowing.

"The issuance numbers on this are dramatic," Hill said. "It's running in the background and it truly is the big story."

For bond investors, the key question is whether the supply pressure is structural or cyclical. If hyperscaler capex continues to grow toward $1 trillion annually as Urano projects, the competition between corporate and government debt for investor dollars will only intensify, keeping upward pressure on long-end yields even if the Fed begins cutting rates.

This article is for informational purposes only and does not constitute investment advice.