AI Debt Attracts High-Yield Investors: A New Trend

Companies developing artificial intelligence (AI) infrastructure are increasingly drawing in junk bond investors, even for debt classified as investment-grade. This trend is driven by the attractive yields offered, which are pulling high-yield investors, dubbed ‘tourists,’ into technology debt that traditionally carries lower risk.

The intense competition for capital underlines this shift, with firms investing hundreds of billions into AI infrastructure. Over $410 billion has already been borrowed this year for data centers and other AI-related investments, showcasing the significant upfront costs involved.

Key Debt Examples

• QTS Realty Trust Inc. sold $3.9 billion in bonds for a Microsoft-tied facility, achieving high-grade ratings but yielding approximately 7.23 percent. This yield level is typically associated with middle-tier junk bonds.
• BlackRock Inc. is paying a 7.53 percent yield on blue-chip securities for a Texas data center project, further illustrating the elevated compensation for investment-grade ventures.

In secondary markets, high-grade notes from major companies like Oracle Corp. and SpaceX are also trading at yields comparable to those of speculative-grade debt. This suggests that these entities would need to offer similar compensation if they were to seek debt investors again.

Market Implications

While higher compensation can broaden the buyer base for AI debt, the US junk bond market has inherent limitations due to its smaller size and lower liquidity compared to its high-grade counterpart. Experts note that this dynamic could eventually constrain debt issuance for lower-rated tech businesses.

Firms with smaller financial cushions would find wider spreads a more significant obstacle to their funding costs. This could lead to increased borrowing costs for tech companies across the spectrum.