QTS's $3.25 Billion Loan: AI Data Centers Tap Into the Leveraged Loan Market
Taylor Wilson
Blackstone-backed data-center operator QTS more than tripled its loan to $3.25 billion and scrapped a planned $1 billion bond sale; the deal marks the first large-scale spill of AI infrastructure financing into the leveraged-loan market.
Why did this loan balloon to three times its original size?
QTS initially planned a far smaller loan, but investor appetite was so strong that JPMorgan, the lead arranger, upsized it more than threefold to $3.25 billion — while dropping a separate $1 billion bond offering entirely.
This means → a single channel — loans — swallowed the volume that was supposed to need two legs, loans plus bonds.
Proceeds go to refinancing construction debt and other existing obligations, with the remainder covering operating costs. This is not growth capital — it is swapping out older, costlier debt on assets already built.
Why are investors lining up?
The loan pays the benchmark rate plus 225 basis points and carries a Baa3 rating from Moody's — the lowest rung of investment grade, the threshold that lets CLO holders buy in.
In plain terms = comparable investment-grade bonds currently trade at an average spread of roughly 94 basis points. QTS's loan offers materially more yield for the same rating bracket — a magnet for return-hungry capital.
The collateral matters most: more than ten operational data centers with high-profile tenants like Microsoft already generating steady cash flow. Oaktree's David Rosenberg put it directly: "You want a high probability of getting your principal back — either strong parent support or contracted cash flows."
Why is AI financing spilling into the leveraged-loan market now?
Until this deal, AI infrastructure funding was almost entirely a bond-market story. Amazon, SpaceX, and others have raised over $335 billion globally through bonds this year alone.
This means → the bond market is crowded. Issuers need to diversify, and leveraged loans — typically higher-rate instruments bought mainly by CLOs — are the natural overflow channel.
MetLife's co-head of leveraged finance John Yovanovic was blunt: "We're still in the early innings of AI buildout. The investment-grade market has absorbed a lot this year; there's room on our side."
Is this a one-off or the start of a trend?
QTS is not alone. CoreWeave closed a $3.1 billion leveraged loan in May; bitcoin miner TeraWulf is expected to issue its first leveraged loan as well.
Barrow Hanley portfolio manager Nick Losey said: "I would be shocked if we don't see more of these."
This reflects a hunger on the market's own side — leveraged-loan desks have been starved of new M&A-driven supply for years. The AI buildout provides a fresh pipeline. Rosenberg noted: "When CLOs are digesting a mountain of refinancings, genuinely new deals carry value in their own right."
Can the QTS template be replicated?
QTS's core selling point is "operational assets + marquee tenants + stable contracted cash flow." That combination earned an investment-grade rating and made the loan eligible for CLO portfolios.
In plain terms = not every AI data center can take this path. Projects still under construction, or facilities without signed tenants, carry a fundamentally different risk profile and may not replicate this rating or pricing.
Whether QTS becomes a template — pulling more AI data-center assets into the leveraged-loan market — is the key signal for whether this financing trend has legs.
Content is for reference only, not financial advice.