Why Bond Investors Aren’t Totally Buying the AI Hype

Equity markets have ridden the AI wave to record highs, but we as bond investors are taking a more circumspect view. Major US tech companies are now turning to public debt markets to help fund their AI infrastructure rollout – a US$1.5 trillion gap according to Morgan Stanley that private credit alone cannot fill. Unlike equity investors, we're less concerned about making bets on AI's future or Big Tech's success and are focused on downside risks.

The levels of capital raising that has dominated equity and private debt markets this year has now found its way into the public credit sphere. October saw a flurry of bond issuance with Meta raising US$30 billion, the largest corporate bond issuance for over two years. The offering was oversubscribed, attracting approximately US$125 billion in orders, which we understand represented record-breaking demand for a US investment-grade corporate bond. Alphabet followed suit in November with US$25 billion worth of notes while Oracle also issued US$18 billion of paper towards the end of the third quarter.

AI capex roadblocks

To us the capex rollout plans are not as gold-plated as the shiny prospectuses might suggest. A lack of power, delayed construction, cybersecurity threats, rapidly ageing technology are just a few of the factors which we need to consider before investing, even if leases are backed by a tech giant.

While these risks exist, the critical question is whether we're being adequately compensated. In Meta's case, the answer appears to be yes. The recently issued long-dated AA-rated paper trades at a 40-basis point pick-up versus equivalent debt - a reasonable premium that reflects both the infrastructure uncertainties and Meta's financial strength. With a cash balance of US$47 billion, free cash flow-to-debt above 30%, and low leverage, the additional spread seems fair given the firm's solid fundamentals and global presence.

Figure 1. AA-rated hyperscaler debt trades wider to the broader market

Source: Man Group and ICE Data Indices as at 7 November 2025.

Problems loading this infographic? - Please click here

The arrival of AI to the bond sphere has not been exclusive to the investment-grade market, with high-yield firms also issuing publicly traded debt in recent weeks. These include former bitcoin miners who have raised significant capital for the construction of data centres, with backing in the form of lease guarantees from several large firms. These companies' infrastructure plans come with aggressive deadlines and a heavy reliance on the supporting contracts. From a creditor perspective, while the backing from major and well-capitalised players is welcome and provides a degree of credit enhancement, supply chain risks alongside a lack of track record in this space for some of the weaker firms means that we're not fully jumping onto the AI bandwagon just yet.

The funding gap

This rush to the bond markets reflects the scale of investment required. The hyperscalers have quadrupled their capex over the last few years, nearing US$400 billion annually and are forecasted to reach US$3 trillion over the next half decade.

While these companies are formidably well capitalised and exceptionally profitable, enabling a degree of self-funding, there remains a significant funding gap. This shortfall has been filled by private lenders such as Blue Owl, who in September helped fund a US$27 billion deal with Meta along with Pimco. Another operator, CoreWeave, has secured several lines of debt financing from the likes of Blackstone and private credit divisions of bulge bracket banks. But, as we mentioned earlier, there’s still US$1.5 trillion that needs to be found.

Figure 2. For the five hyperscalers, consensus forecasts call for capex jump

Note: Hyperscalers include the following companies: Amazon, Google, Meta, Microsoft, Oracle. Source: BofA Global Research, Bloomberg, Visible Alpha, as at 10 November 2025.

Problems loading this infographic? - Please click here

As bond investors, we welcome diversification but a glut of supply of lower quality names in the AI space might be too much for markets to stomach. The hyperscaler frenzy continues, but we remain watchful of future AI slop.

All data sourced from Bloomberg unless otherwise stated.

Author: Jon Lahraoui, Director, Discretionary Credit at Man Group and Hugo Richardson, Associate Director, Discretionary Investment Specialist, at Man Group.