Credit spreads for tech companies driving the AI buildout are widening and are expected to expand even more later this year as debt levels increase. That could put stress on the heavily indebted tech infrastructure builders known as neoclouds, as well as the larger cloud computing hyperscalers.
While all the circular investment among cloud computing companies, artificial intelligence software firms, and chipmakers could help the neoclouds skirt some of this expense, it also means that pricier corporate debt could ripple through the sector.
A spread refers to the difference in yield between two bonds that have the same maturity but different credit qualities. Typically corporate bonds are quoted by the difference above a Treasury security of similar maturity. Widening spreads suggest that investors see greater default risk and thus want to be paid more for it.
Hyperscalers are continuing to boost their investments, with Google raising its capex projections for both 2026 and 2027, and the question of credit quality is becoming more of a concern for investors.
Analysts for Mizuho warned last weekend about wider spreads specifically in the context of the smaller, more highly indebted neoclouds.
Neocloud CoreWeave has total debt of about 739 times its amount of equity, while Nebius has a debt-to-equity ratio of 131.
By comparison, the big cloud companies have much smaller relative debt loads.
Widening spreads and a changing risk backdrop Corporate credit spreads are expected to widen out later this year and into 2027.
A lot of tech financing – especially at the frontier – happens outside of standard bond issuance channels and some of it is kept off the books entirely.
The Bank for International Settlements warned this month that ample use of debt in the context of circular financing could lead to a bust.
