Global Market Strategist
Global Market Strategist
In Brief
- AI-related bond issuance is reshaping credit benchmarks as hyperscalers fund infrastructure capex with debt, increasing concentration risk for passive investors.
- Increased supply has been matched with ample demand given that issuers’ credit fundamentals remain strong. Earnings downside amid persistent supply may pressure spreads.
- Active issuer selection is increasingly important to capture income while managing duration, leverage, and AI-related correlation risks.
Artificial intelligence (AI) is not just transforming the way we live and work; it is also reshaping the credit market. As hyperscalers ramp up their infrastructure spending to support AI development, they are turning to the bond market in unprecedented volumes. This surge in AI-related bond issuance is altering credit benchmarks and raising exposure risk for passive investors.
Strong credit fundamentals by issuers and robust demand have meant that the rise in supply is being easily digested by markets. However, the risk of persistent supply, when combined with emerging doubts about future earnings growth and the return on investment, could put pressure on credit spreads.
In this environment, active issuer selection is becoming more important for those seeking to capture income while managing risks around duration, leverage, and AI-driven correlations.
Magnitude: Fast and furious
In 2025, the U.S. technology sector’s net issuance of investment grade (IG) bonds reached USD 131billion, more than double the average annual issuance of USD 61billion over the previous five years. This trend shows no sign of slowing, with year-to-date net issuance already at USD 192billion, accounting for 27% of all IG net bond issuance.
J.P. Morgan Securities projects net issuance could rise to USD 230billion in 2026, a 76% increase on 2025, with technology and AI-related funding the dominant force. Issuance of investment grade debt is only expected to grow and could potentially reach a cumulative USD 2.1trillion by 2030 to fund the expected USD 5.5trillion in AI-related capital expenditure.
Materiality: A new kind of concentration risk
The concentration risk that has become a hallmark of the equity market, where a handful of AI-exposed mega caps dominate, is a growing concern in the bond market. A key difference is that bond indices are weighted by the amount of outstanding debt, not by market capitalization. This means passive investors are increasingly exposed to companies with the largest debt loads, regardless of their underlying business strength.
For companies with robust earnings, issuing more bonds can be an efficient way to optimize their capital structure. For others, however, it may signal a genuine need for funding and carry higher risk. The market’s ability to differentiate between these scenarios is reflected in both credit spreads and credit default swaps (CDS). CDS spreads capture the cost of insuring against default or loss and are used as a gauge of the market-implied probability of default. CDS spreads have widened across several hyperscaler names as markets reprice the risks associated with rising leverage (Exhibit 2).
The risk is that a negative shock, such as disappointing revenue growth or a pullback in capital expenditure, could simultaneously impact the largest weights in both bond and equity indices. Because benchmark-constrained funds must track index weights, any spread widening risks triggering repositioning that amplifies the move.
Further down the credit spectrum, the high yield bond market has also seen a pick-up in issuance, particularly from companies involved in data center development, power infrastructure, fiber networks, and other AI-adjacent infrastructure.
The technology sector accounts for 18.7% of total high yield new issuance year-to-date but remains a relatively smaller share of the overall high yield index at 8.6% and below the current 10% weight of tech in the IG bond index.
Investment implications
The rise in AI-related issuance across investment-grade and high-yield indices is a structural consequence of the AI capex cycle and is likely to persist as infrastructure spending grows.
For credit investors, the key question is whether yield offers adequate compensation for the rising exposure within the credit indices.
For now, hyperscaler fundamentals remain strong and downgrade risk appears low. Credit metrics such as debt-to-EBITDA are undemanding for AI-related issuers. However, should this change, taking passive exposure means accepting these risks by default, with limited ability to distinguish between issuers’ balance sheet strength, leverage trajectory, or capital discipline.
Investors can diversify exposure by looking across other sectors in the IG market, as well as across shorter- and intermediate-duration bonds, to reduce spread and rate sensitivity. Securitized credit is another higher quality option where income is driven by different collateral pools rather than corporate AI spending.
Select high yield can also offer incremental income, though the recent pick-up in data center and infrastructure-linked issuance makes credit selection especially important.