By Adam Whiteley, Head of Global Credit, Insight Investment, a BNY Investments company.
AI investment is creating an unusual challenge for credit investors. The heaviest spenders on AI infrastructure are generally high-quality companies with strong fundamentals and among the world’s largest, highest-rated borrowers. However, financial strength does not protect their bonds from a substantial rise in debt issuance. Relatively insensitive to borrowing spreads, they are likely to continue issuing regardless.
According to S&P Global estimates, AI-related borrowing could reach $400bn in 2026, potentially making this a record year for investment grade (IG) issuance. Markets have absorbed the additional supply so far, but repeated waves of borrowing could outpace investor demand and push credit spreads higher.
The companies could slow investment to protect their credit ratings and should remain able to service their debts even if AI investment failed to deliver. As we see it, the main risk for bondholders is not default, but greater price volatility as companies issue debt across different markets.
The implications extend beyond individual technology companies. As issuance from hyperscalers (large cloud-computing and technology businesses) grows, their bonds could become a larger part of global credit indices. The technology weighting of the Bloomberg Global Aggregate Corporate Index has already increased during 2026. Much of the new debt also has relatively long maturities, reflecting the expected lifespan of the infrastructure being financed. This could extend the index’s duration and increase its sensitivity to changing bond yields.
IG investors may therefore acquire greater exposure to the AI investment cycle, which is already an important influence on equity markets. This makes diversification within fixed income increasingly valuable.
High yield as an unlikely source of diversification
Counterintuitively, high-yield bonds could offer one route away from this concentration. AI-related borrowing is considerably less prevalent in high yield than in IG. In May 2026, issuers in data centres, cloud computing, semiconductors or digital infrastructure represented 6.2% of US high-yield issuance, compared with almost 18% of IG issuance. Our analysis suggests high yield has one-third of IG’s exposure to AI-related capital raising.
The two markets also have different sector compositions. Software companies, which many investors view as potentially vulnerable to AI-driven disruption, are primarily represented in IG. High yield has greater exposure to areas such as basic industry, energy and capital goods, where the disruptive effects of AI may be more limited.
This does not make high yield “safe” in the conventional sense. Lower-rated companies present greater default risk, while high yield bonds can experience substantial losses during periods of economic weakness or market stress. The potential attraction is that the source of risk is different.
By accepting more traditional credit risk, investors may gain diversification from the financing pressures and technological disruption affecting parts of the IG market.
Could emerging market corporates offer another escape route?
Emerging market (EM) corporate bonds could extend this argument further. Although EM countries supply many of the raw materials needed to build AI infrastructure, the accompanying surge in bond issuance remains primarily concentrated among US companies.
EM corporate debt can provide exposure to a broader range of borrowers, rather than concentrating credit exposure among large developed market issuers.
For global investors concerned about increasing dependence on the US technology sector and AI investment cycle, EM corporates could provide a second source of diversification alongside developed market high yield.
Again, these bonds should not be regarded as a conventional safe haven. EM securities can be less liquid and riskier than their developed-market counterparts. Accounting standards, trading arrangements, settlement and custody may also present additional challenges. Country-specific, commodity, currency and company-level risks mean that careful security selection remains essential.
Rethinking where the risk lies
The AI boom is unlikely to turn IG technology companies into weak borrowers overnight. Many remain highly profitable, retain strong ratings and generate substantial cashflows. The concern is that the volume of prospective issuance could keep their credit spreads under pressure even if their underlying businesses remain sound.
Credit risk is therefore not simply a matter of selecting the highest-rated borrowers. Investors must also consider where their portfolios are concentrated, which risks they are being paid to accept and whether apparently different investments ultimately depend on the same economic theme.
High yield and EM corporate bonds may deserve renewed attention. Neither is a traditional safe haven, but their lower direct exposure to AI-related borrowing could make them valuable sources of diversification. In an AI-driven credit cycle, avoiding concentration may prove just as important as avoiding credit risk.
