AI Debt Boom Sparks Warnings

Oaktree's Danielle Poli warns that AI-driven debt issuance could create future cracks in credit markets, particularly in private credit and leveraged loans. She also said higher interest rates may stay elevated longer than markets expect, while investors are still finding 8-10% yields. The piece is primarily cautionary commentary rather than a direct market-moving event.

Analysis

The bigger issue is not AI capex itself, but the liability chain being created to fund it. When a capital-intensive growth theme is financed with increasingly loose private-credit and leveraged-loan structures, the first-order beneficiary is the asset-light platform layer, while the second-order losers are lenders and refinancers sitting one turn down the stack. That setup typically looks benign until the market needs to roll debt at a meaningfully higher coupon; then the stress shows up first in covenant-lite paper, then in secondary discounts, and only later in headline default rates.

The rate message is more important for credit than for equities. If policy stays restrictive longer, the duration of “good” yields extends, but so does the window in which floating-rate borrowers refinance into worse terms or burn through liquidity. That is constructive for cash-rich lenders and plain-vanilla bond buyers, but negative for highly levered private-market sponsors that relied on an easier terminal rate assumption. In practice, the most vulnerable names are those underwriting near-term EBITDA growth to justify aggressive leverage on AI-adjacent projects.

The contrarian read is that the market is underpricing dispersion rather than systemic stress. A broad credit blowout is unlikely near term because cash is still available and default cycles lag by quarters, not weeks; the better short is the weakest marginal underwriter or sector with concentrated exposure to sponsor-backed AI infrastructure. Over the next 6–18 months, the catalyst is not an AI demand miss, but a refinancing calendar meeting a higher-for-longer rate regime and forcing repricing in private credit marks before public markets fully acknowledge it.

Key Decisions for Investors

• Short a basket of lower-quality leveraged loan / CLO-sensitive credit proxies via JNK or HYG put spreads over 3–6 months; risk/reward favors downside if refinancing spreads widen even modestly.
• Long high-quality short-duration credit such as SGOV or BIL as a parking place for dry powder while earning policy-supported yield; this is a defensive carry trade if rates stay elevated longer than expected.
• Pair trade: short private-credit beneficiaries with heavy sponsor exposure against long cash-rich public lenders / asset managers (for example, short BKNG? no) — better expressed as short a regional bank ETF proxy with high CRE/sponsor exposure versus long large-cap balance-sheet lenders; monitor for spread widening over 1–2 quarters.
• Avoid chasing AI infrastructure debt until issuance clears and secondary concessions appear; if spreads on new AI-related paper cheapen 75–100 bps from current levels, that is the first tell that demand is becoming rate-sensitive.
• For event-driven exposure, buy downside protection on leveraged-loan ETFs into the next refinancing wave; the clean catalyst is 6–12 months, not days, so structure longer-dated puts rather than tactical short stock.