Pimco Sees AI Data Center Bond Spreads Tightening
Pimco expects spreads on bonds backing AI data centers may narrow as construction slows, with hyperscaler-backed debt possibly outperforming tech senior unsecured debt.
Updated on Nov. 17, 2026 in Data Centers

Live Poll
Do you expect borrowing costs to rise over the next year?
As of Nov. 10, investment-grade bonds backed by hyperscaler-leased data centers traded about 146 basis points wider than comparable tenant debt, down from near 180 basis points in late September. Pimco expects these spreads may tighten as data-center construction slows.
Why it matters
A flood of AI-linked debt has put pressure on credit markets, and Pimco's view suggests slower build-out could ease that pressure. Permitting constraints and higher costs are delaying construction and curbing new bond supply.
Spread data covers investment-grade bonds tied to hyperscaler-leased facilities, measured against comparable tenant debt. Consensus forecasts put combined capital spending by five hyperscalers above US$1.2 trillion a year, though whether that pace holds is not yet settled.
The players
Pacific Investment Management Co (Pimco)
Pimco is a large global bond investment manager whose credit research shapes how institutional investors price debt tied to infrastructure and technology borrowers.
Microsoft Corp
Microsoft is one of the largest cloud computing providers and a major hyperscaler whose data-center leasing and spending influence demand for AI infrastructure debt.
Amazon.com Inc
Amazon operates Amazon Web Services, one of the largest cloud platforms, and is among the hyperscalers whose data-center spending supports this bond market.
The details
Pimco's analysis points to permitting constraints and higher costs as the forces slowing data-center construction, which both curbs new bond supply and keeps computing capacity scarce, supporting utilization and pricing at existing facilities. The firm is more cautious on leveraged infrastructure providers whose interest expenses, leases and chip-purchase commitments persist.
Timeline
Late September 2026: data-center bond spread near 180 basis points.
Nov. 10, 2026: spread about 146 basis points wider than comparable tenant debt.
Next two years: hyperscaler AI spending forecast above US$1.2 trillion annually.
The Tech Race
The current AI data-center build-out invites comparison with the telecom fiber overbuild of 2000-2001, when debt-financed capacity outran demand and later weighed on lenders. Pimco's expectation that construction will slow suggests lenders may avoid a repeat of that oversupply.
Changes in data-center bond spreads can shift borrowing costs for the companies building AI infrastructure, which may influence the pace of new cloud and AI services. Slower construction could also affect how quickly new computing capacity becomes available to users.
The takeaway
Watch how data-center bond spreads move as construction slows, since they offer an early read on whether AI infrastructure spending is cooling. Shifts in that debt market can signal broader changes in how much capacity the industry actually builds.
Further reading
Explore more on Data Centers to see how AI infrastructure financing is shaping the wider market.
Live Poll
Do you expect borrowing costs to rise over the next year?



