Debt financing for AI-related data centers faces weakening market demand

Huo Xing Finance reports that on August 7, tech companies are heavily borrowing to finance AI investments, yet even in smaller markets, investors are growing increasingly cautious about such debt. According to informed sources, two of the last three commercial mortgage-backed securities (CMBS) issued to finance data centers were forced to widen their pricing spreads to attract sufficient demand—among them, CMBS offerings by CyrusOne, backed by KKR, and QTS Realty Trust, backed by Blackstone. Meanwhile, risk premiums on CMBS linked to data centers have risen across the board over the past 12 months. In mid-July, Pure Data Centres (Pure DC), a UK data center operator backed by Oaktree Capital, abandoned its planned record €1 billion bond issuance in favor of bank financing. At the time, the company was marketing an unsecured bond, but signs of market saturation in AI-related data center debt had emerged, compounded by sharp declines in CoreWeave’s stock and bond prices following news that Meta was building its own cloud infrastructure. Investors adopted a cautious stance on terms, leading the company to conclude that its concurrent bank loan offers provided more favorable conditions, prompting it to cancel the bond offering. This wave of resistance has been dubbed the “Luddite trade” by markets—investors are now viewing AI data centers as risk assets akin to traditional office and retail properties, concerned about overbuilding, tenant concentration, and obsolescence leading to asset depreciation. In contrast to past enthusiasm for AI-themed debt, investors are now tightening pricing standards for data center mortgages, demanding higher risk compensation. This trend creates a subtle tension with the flood of AI capital expenditures—while tech giants continue ramping up investments, pricing signals in the debt market are beginning to issue cautious warnings.