Corporate bond buyers get picky with flood of AI debt

A split has emerged in the highly rated corporate bond market, with bonds issued by AI-related firms receiving cautious attention while traditional issuers such as financial and industrial firms draw strong bidding. Portfolio managers are not worried about default risks from hyperscalers and other AI-linked companies, but rather, the excessive borrowing required to finance data centers, chips and AI infrastructure is prompting bond buyers to demand significant concessions and to limit portfolio concentration.

It is projected that gross debt issuance from hyperscalers will reach a record $420 billion next year, a 60% increase from 2026 estimates. Goldman Sachs data indicates that overall US corporate issuance through August rose by 30% to $1.9 trillion. Fixed income professionals are being particularly selective in investing within hyperscaler debt, with Colby Stilson, head of fixed income at Brown Advisory, stating that their investment conviction needs to be very high due to concerns about supply and return on invested capital.

In contrast, corporate bond spreads remain near historically tight levels, and new deals are often heavily oversubscribed. Investors have cash to deploy, but many are increasingly favoring opportunities outside the AI investment boom. The divergence in demand is evident in recent bond sales, with Google parent Alphabet having to offer large concessions in its August debt sale, while Aon's $13.5 billion acquisition financing drew significant orders.

The heightened demand for scarcer bonds is reflected in the widening spreads for AI-related issuers, which have remained persistently wider at around 115 basis points compared to 78 basis points for the broader investment grade market.