Morgan Stanley projects global AI-related debt issuance will hit $570 billion in 2026. By the end of May, roughly $236 billion had already been issued, putting the market on pace to grow at approximately four times the rate of the prior year.

The money machine behind AI infrastructure

Hyperscaler investment-grade bond issuance, meaning debt from the massive cloud and AI platform operators, surpassed $100 billion in 2025. Morgan Stanley expects net supply in 2026 to climb another 30-50%, landing somewhere between $130 billion and $150 billion.

The bank has been actively structuring major deals in this space. One notable example is a $3 billion debt facility arranged for TeraWulf, the AI and data center infrastructure company, with backing from Google. That deal was announced in September 2025. Core Scientific’s expanded credit facilities represent another piece of the puzzle.

Utilities are getting pulled into the orbit

Data centers are voracious consumers of electricity. US investment-grade utility issuance hit approximately $135 billion in 2025, and Morgan Stanley expects that figure to climb to around $145 billion in 2026, driven largely by the need to power new data center capacity.

The risk nobody wants to talk about (but 34% already are)

A Bank of America survey from May 2026 found that 34% of fund managers now identify data center debt as a potential systemic risk. AI-linked issuers are taking on significant leverage to build infrastructure whose revenue streams depend on continued exponential growth in AI demand.

TeraWulf itself originally emerged from the crypto mining world before pivoting toward AI workloads. When $236 billion gets issued in five months and the full-year target is $570 billion, the underwriting standards for deals issued in month eleven might not match the rigor of deals issued in month two.