The enormous spending spree by the technology giants building the infrastructure behind the artificial intelligence boom is beginning to attract greater scrutiny from credit markets.
Apollo Global Management chief economist Torsten Slok has reportedly warned that rising credit-default swap (CDS) spreads on hyperscaler debt suggest investors are becoming increasingly concerned about the financial foundations of the AI investment cycle.
CDS contracts
CDS contracts provide protection against a company’s debt default. Reportedly, according to Apollo, the gap between CDS spreads for major hyperscalers and those of large banks has widened to around 60 basis points, having been broadly negligible in October 2025.
It is argued that this is significant because bank CDS spreads have remained relatively stable.
The implication is that investors may not simply be reacting to the huge volume of new bonds being issued. Instead, they could be reassessing the credit fundamentals of companies such as Amazon, Microsoft, Alphabet and Oracle as they borrow heavily to finance data centres, chips and other AI infrastructure.
Concern
Apollo points to three particular concerns: rising leverage, negative free cash flow and uncertainty over whether the enormous investment will generate sufficient returns before the underlying technology and equipment depreciate.
That does not necessarily mean the AI boom is about to collapse. The major hyperscalers remain large, established businesses with substantial revenues and access to capital. Indeed, Apollo itself is reportedly notes that the bond market continues to absorb enormous amounts of issuance.
Credit markets
Nevertheless, the changing behaviour of the credit markets provides another indication that investors are beginning to ask harder questions about the economics of AI.
For years, the central question was how quickly artificial intelligence would transform business. Increasingly, another question is emerging: how much debt can the AI revolution carry before investors demand a greater return for the risk?
If borrowing costs continue rising while AI revenues fail to keep pace with infrastructure spending, the industry’s extraordinary investment cycle could face a very different financial environment.
