Systematic Insights: Is High Yield an AI Sweet Spot? | Insight Investment

Data center-related issuance has ramped up in US high yield markets, but overall, AI exposure is still relatively modest. This may create opportunities for selective access to AI-linked growth, while also offering diversification if broader AI-related markets see a pullback.

AI exposure remains modest in high yield

Since Q4 2025, many of the “hyperscalers” (Amazon, Alphabet, Meta, Microsoft and Oracle¹) have aggressively issued investment grade bonds to help finance data center construction and other capex needs (Figure 1).

High yield AI-driven issuance was more modest by comparison, until April 2026, when issuers raised a record $41bn (Figure 1), accounting for 90% of the month’s gross high yield issuance. Most deals were AI-focused data center-related deals.

Figure 1: High yield AI-driven issuance spiked this quarter³

Although AI high yield issuance has since cooled off in May, it has been enough to raise the high yield market’s AI exposure.

Nonetheless, high yield remains modestly exposed to the AI boom at ~6% by market cap, when including legacy issuers pivoting their businesses to service the AI industry. This is significantly less than the investment grade market’s exposure and both are dwarfed by the S&P 500’s market cap exposure at over 60% (Figure 2).

Figure 2: High yield’s AI exposure has increased but remains low²

AI capex shows no sign of slowing down. Current estimates indicate $4trn to $8trn over the next five years³. Investment grade, high yield, loans, private credit, bank financing and other debt structures will likely play roles in financing it, across multiple currencies.

The high yield market may offer a balance of lower overall AI exposure with potential AI-related security selection opportunities.

does high yield offer a sweet spot of AI spread opportunities but low exposure?

In high yield, AI sectors currently offer significantly higher spreads than non-AI sectors (Figure 3). We expect AI-related issuance to drive security selection opportunities through new sources of dispersion, without significantly disrupting the overall sector mix.

Figure 3: AI-related high yield sectors may offer some compelling credit spread opportunities⁴

This may allow for opportunities to exploit potential mispricing within AI-focused sectors like hardware, or within other areas of the AI value chain like power supplies. In some cases, investors can potentially access close to double-digit yields, a significant premium over the investment grade hyperscalers.

Non-AI high yield sectors include businesses tied to more traditional cyclical, industrial, energy and service-oriented cashflow drivers. This may help insulate the market from wider AI-market pullbacks.

Pinpointing potential value in AI-related high yield deals

We believe bottom-up investors should focus exposure on potential higher quality issuers, those subject to potentially pessimistic pricing and those that offer healthy coupon income that may offer sufficient compensation for issuer-level risk.

Our systematic models have pinpointed a potential overweight candidate in the “neocloud” sector (which provides high-performance computing clusters) on value grounds. It has also highlighted potential value in a legacy business (on value and quality grounds) whose fastest-growing segment is data centers. Our models have also flagged potential underweights across the AI value chain, such as in certain hardware supplier and a power generation companies, on potentially inadequate credit spread compensation for risk.