Winnie Cisar, Global Head of Credit Strategy at CreditSights, highlights a growing mismatch between expectations and reality for some AI companies, leading to increased costs for protecting tech firms' debt against default. This rise in protection costs is driven by mounting concerns about whether the massive investments in AI, particularly by Big Tech companies, will ultimately pay off.

Investors are showing signs of fatigue after a $300 billion AI debt binge, leading to increased selectivity or 'choosiness' for AI-related debt. This sentiment is fueled by the poor performance of recent AI financing deals and concerns about the long-term viability of these uncertain investments in new technologies with unproven demand.

The bond market is beginning to question the AI boom, as evidenced by a lackluster reception for a Meta Platforms transaction that aimed to replicate a successful previous bond sale. Analysts anticipate that the cost of financing for longer-term deals will continue to rise as supply weighs on the market and investors demand more assurances about the long-term returns of AI infrastructure investments.

While an abrupt funding shutdown for AI infrastructure is unlikely due to low initial debt levels and strong recurring revenues of hyperscalers, early signals indicate wider relative spreads, larger new issue concessions, and weaker post-deal performance for AI-related debt. Winnie Cisar also emphasized that AI is influencing productivity, financial conditions, and investor sentiment, requiring investors to carefully consider evolving market risks and credit selection within the AI ecosystem. Moody's has also raised concerns about the lack of transparency from hyperscalers regarding off-balance-sheet liabilities and the growing 'circularity' of financing deals.