The market for highly rated corporate credit has diverged, with bonds from AI-related firms, particularly hyperscalers, facing caution from investors. In contrast, traditional issuers like financial and industrial firms are seeing strong demand. While investors aren't concerned about AI companies defaulting, the massive and unpredictable borrowing required for data centers, chips, and AI infrastructure is leading bond buyers to demand significant concessions and reconsider portfolio concentration limits. Gross debt issuance from hyperscalers is projected to reach a record $420 billion next year, a 60% increase from 2026 estimates, according to Goldman Sachs data.
This heightened selectivity is evident in recent bond sales. Google parent Alphabet, for instance, had to offer a substantial concession for its August debt sale. Conversely, insurance broker Aon's $13.5 billion acquisition financing this month attracted $65 billion in orders, with pricing on its 30-year tranche tightening by 35 basis points. Spreads on AI-related issuers have widened to approximately 115 basis points, compared to 78 basis points for the broader investment-grade market, reflecting the continuous stream of supply that investors are struggling to absorb.
AI issuers are being compelled to offer concessions, with some double-A credits pricing closer to triple-B spread levels. Concentration risk is also a growing concern, as some institutional investors are nearing single-name exposure limits when aggregating debt from related structures. While the widening spreads in AI-related debt reflect supply-and-demand dynamics rather than credit quality concerns, and offer attractive compensation for investors, the clear message from the bond market is a growing selectivity. Many investors are now looking for opportunities outside of hyperscaler debt.
