High-grade AI debt attracts junk bond investors as data centre funding costs rise

Investment-grade bonds associated with AI infrastructure projects are increasingly attracting investors typically found in the high-yield market, according to Bloomberg. This trend was evident when Blackstone-backed QTS Realty Trust raised $3.9 billion through bonds to finance a Georgia data center linked to Microsoft. The bonds, despite receiving investment-grade ratings, offered a yield of approximately 7.23%, above levels typically seen in the BB-rated market.

A similar pattern emerged in July when BlackRock raised $12.5 billion for a Texas data center project, with some debt carrying a yield of around 7.53%. These bonds were marketed to investors across both investment-grade and high-yield markets. For private capital managers, this development highlights how the substantial financing needs of AI infrastructure are causing traditional credit strategy boundaries to blur.

Companies have borrowed over $410 billion this year to fund data centers and other AI-related investments, according to Bloomberg data. High-yield investors have previously shifted into higher-quality corporate debt when pricing becomes attractive, as seen during the Covid-19 pandemic. However, the scale of the current AI investment cycle is creating a different context.

Major technology companies often rely on equity and operating cash flow for capital expenditure, but the enormous upfront costs associated with AI infrastructure are now prompting them to increasingly turn to debt markets. Some technology debt is already trading in the secondary market at yields more commonly associated with speculative-grade securities.

The financing costs could also impact the economics of AI investments, as rising borrowing costs increase companies' weighted average cost of capital, potentially putting pressure on projects with uncertain or distant returns. The financing requirement is expected to persist, with Vanguard estimating that hyperscalers could spend nearly $800 billion on AI this year, followed by more than $1 trillion annually from 2027 through 2030.

Debt markets are anticipated to provide a significant portion of this funding. Investors are demanding higher compensation due to uncertainty around additional borrowing following individual transactions. High-yield and distressed investors are showing growing interest in investment-grade AI-related debt where yields have risen sufficiently to meet their return targets.

However, the US junk bond market is relatively small, limiting the potential for high-yield investors to rotate into higher-quality AI debt. The surge in AI-related capital expenditure is driving record corporate bond issuance, with TD Securities raising its forecast for 2026 US investment-grade issuance by $100 billion to $2.2 trillion.

European issuance has also accelerated following a summer slowdown. Companies like Alphabet and ByteDance have recently tapped the bond market with record-high yielding debt, while private credit providers are increasingly financing AI projects, such as Eagle Point Credit Management's $1.3 billion loan for an Anthropic AI data center in Texas.

Moody's has upgraded numerous US and European collateralized loan obligation tranches, while Jefferies Credit Partners is seeking €1 billion for a private credit continuation vehicle.