The next phase of AI issuance

Artificial intelligence is now part of daily life. Beyond its technological impact, AI has also emerged as one of the decade’s defining investment themes, reshaping global credit markets in the process.

As technology companies compete to build the data centres, cloud infrastructure, and power generation required to support AI capabilities, capital spending has picked up considerably. With industry estimates projecting expenditures of US$3-5 trillion over the next several years, these companies will increasingly rely on debt financing to fund AI buildouts.

For investors, this wave of issuance presents both opportunity and risk.

The rapid shift to debt-funded growth

In 2025, hyperscale technology companies, including Amazon, Google, Microsoft, Meta and Oracle, issued approximately US$160 billion in public, private, and asset-backed debt to support AI infrastructure.

Beyond volume, what stood out to investors is pricing. Historically, these issuers have enjoyed some of the tightest spreads in the investment-grade market. Recently, however, multi-billion-dollar deals have included new issue concessions of 10-20 basis points.

“These are best-in-class companies, but the sheer scale of capital spending means they’re no longer optimizing for the lowest cost of debt,” says Ilias Lagopoulos, RPIA’s Head of Investment Grade. “They’re focused on securing capital quickly, and markets are demanding a higher risk premium to absorb that supply.”

Issuers clearly believe long-term AI returns justify higher near-term financing costs. For investors, this has created an opportunity to earn incremental yield from companies with strong business fundamentals.

A market-transforming event

As AI funding shifts from internally funded capex to market-wide cross-asset issuance, the compositions of public and private credit markets are beginning to change.

One effect is the structure shift within the benchmark. Funding even 20% of projected AI capital spending through public markets would elevate companies such as Amazon to eclipse the largest borrowers (currently US banks) within the US investment grade benchmark. This increases correlation risk, as index performance becomes more dependent on a small group of technology firms pursuing similar strategies.

At the same time, the growing scale of supply has also pushed risk premiums higher. Some AA-rated issuers now trade wider than the broader investment-grade index and, in some cases, wider than lower-rated BBB bonds. Meta is a notable example, with spreads that remain elevated despite its market position and balance sheet strength.

Despite these challenges, the AI buildout is creating opportunities for investors willing to conduct deeper analysis.

Where opportunities are emerging

One opportunity lies in structured data centre financing. Technology companies are partnering with infrastructure specialists and financial institutions to fund projects through joint ventures and special purpose vehicles. These bonds often trade at wider spreads than parent company debt due to concerns about asset obsolescence or incomplete guarantees. Where cash flows are supported by long-term contracts with large technology firms, careful analysis of structure and protections can uncover opportunities for spreads to narrow over time.

Even among established investment grade issuers, temporary dislocations can create value. A recent example is Oracle, where concerns about funding needs pushed bond spreads to wider levels than some high yield BB-rated bonds. Once the company clarified that future expansion would be financed through a mix of debt and equity, spreads tightened over 40 basis points, rewarding investors with a longer-term view.

Where to exercise caution

The scale of the AI build-out also introduces new risks. Supply constraints, including power availability, labour shortages, and construction delays, could pressure issuers with weaker balance sheets. In the US, data centre demand is expected to outpace power supply by 2028, potentially creating bottlenecks.

There is also the risk of overbuilding. Rapid improvements in chip efficiency could reduce future demand for certain types of infrastructure. At the same time, the emergence of lower-cost AI models has raised questions about whether today’s spending levels are sustainable, increasing the risk of stranded assets.

The key difference between the AI buildouts versus past supply booms, like the flood of issuance by Financials following the global financial crisis, is that this wave is largely leverage increasing. That makes credit selection and structure analysis far more important than simply owning the theme.

A longer cycle with lasting consequences

For institutional investors, the AI buildout should not be viewed as a short-term trade. It is a multi-year financing cycle that will influence index composition, correlations, and relative value across global credit markets.

Active strategies that emphasize balance sheet discipline, funding plans, and covenant protection are best positioned to navigate the shift. The opportunity is not simply in backing AI growth, but in understanding where markets are mispricing risk.

Derrick Jumper is a Principal and the Head of Credit Research at RPIA. Derrick has over 17 years of experience in credit markets. He graduated from the Wharton School at the University of Pennsylvania with a Bachelor of Science in Economics and a dual concentration in Finance and Accounting.

Important Information

Information presented by RP Investment Advisors LP (“RPIA”) is for informational purposes only, does not provide investment, or other advice and should not be relied upon in that regard without seeking the appropriate professional advice. Forward-looking statements are subject to change based on economic and market conditions.