Particle.news
Download on the App Store

Investors Pull Back as Hyperscalers Flood Markets With AI Debt

Rising borrowing costs from weaker demand are driving more AI finance into private, off-balance-sheet structures that can hide who bears the risk.

Overview

  • Major tech firms have pushed hundreds of billions of dollars of bonds into markets to fund AI data centers, with Morgan Stanley projecting roughly $570 billion of AI-related issuance in 2026 and about $236 billion already priced by May 31.
  • Investor demand has cooled sharply, with cover ratios for hyperscaler deals falling from roughly 5 times in February to below 2 times by July, a shift that forces issuers to offer wider spreads to place deals.
  • Weaker primary demand is showing up in the secondary market as wider credit spreads and selective repricing, including some SpaceX debt trading near junk levels and higher CDS costs for several large tech issuers.
  • Banks and analysts report a growing migration of financing into private credit, securitizations and off-balance-sheet joint ventures—structures that can keep project debt off corporate balance sheets and reduce disclosure.
  • If AI returns or demand for data-center capacity slow, analysts warn the effects could ripple through utilities, counterparties and investors, making cover ratios, spreads and the size and terms of shadow financing key indicators to watch.