Blackstone's data center arm, QTS Realty Trust, has completed an investment-grade bond issuance at yields nearing junk-bond levels to finance a data center for Microsoft, underscoring the escalating cost of capital for AI infrastructure buildouts.
According to sources familiar with the matter, the five-year, $3.9 billion bond sale drew demand of approximately $23 billion, representing an oversubscription ratio of about six times. This level far exceeds the average 3.8 times oversubscription seen in blue-chip bond offerings this year, highlighting strong market appetite for the deal.
However, robust demand comes at a steep price. The final issuance yield was tightened by just 40 basis points from initial guidance, settling at 7.228%—a level more akin to high-yield debt, reflecting investor caution over the risks associated with new-generation data center assets. For companies planning to deploy trillions of dollars in AI infrastructure over the coming years, this serves as a warning signal.
Strong Subscription, But Limited Pricing Power
The offering, internally code-named "Project Odyssey," was expanded by roughly $1 billion from the preliminary terms shared with investors last week. Citigroup, Goldman Sachs, JPMorgan, and Morgan Stanley had previously hosted a roadshow call with investors to discuss deal details. Proceeds will fund the development of a data center in Georgia tied to Microsoft.
Despite the demand multiple significantly exceeding comparable bond averages, the final yield compression remained relatively modest. The yield was set at 7.228%, just 40 basis points below the initial guidance. For an investment-grade rated bond, this yield level is exceptionally high, bordering on the pricing range of speculative-grade debt. This partly signals investors' tightening risk appetite for such asset classes.
As previously reported, AI infrastructure investment is driving up Treasury yields, with effects rippling across capital markets. For companies seeking to finance large-scale AI projects through bond markets, the current pricing environment implies higher capital costs, which could materially influence investment decisions.
Comments