PGIM caps AI-related debt exposure at 15% in CLOs, signaling institutional caution on tech lending

PGIM, one of the largest asset managers in the world, anchored a recent collateralized loan obligation with a feature that would have seemed unnecessary two years ago: a hard cap limiting AI-exposed debt to 15% of the deal’s portfolio.

The numbers behind the nerves

PGIM’s own analysts estimate that roughly 11% of US CLO portfolios currently carry exposure to sectors facing near-term AI disruption. European CLOs are somewhat less exposed at around 7%.

Software and technology-adjacent loans typically make up 12% to 15% of US CLO collateral pools. JPMorgan has put a dollar figure on the problem: up to $150B in US CLO-held leveraged loans sit in sectors that face meaningful AI disruption risks. The entire US CLO market is roughly a $1 trillion ecosystem.

Price action has already started to reflect the anxiety. Software names within CLO portfolios have underperformed recently, with AI-related volatility cited as a contributing factor.

Why a CLO cap matters more than it sounds

For anyone unfamiliar with the machinery of structured credit, a CLO bundles hundreds of leveraged loans into a single vehicle, then slices it into tranches with different risk profiles. Senior tranches get paid first and carry the least risk. Equity tranches absorb losses first and earn higher yields for doing so.

Capping exposure to a technology trend rather than a traditional sector classification is a novel structural mechanism. It acknowledges that AI disruption cuts across conventional industry lines, affecting everything from customer service software to data analytics to cybersecurity.

Edwin Wilches, co-head of PGIM’s securitized products group, has been vocal about the need for active management and structural protections in CLO senior tranches. The firm has published multiple analyses on how AI-driven volatility affects CLO tranche performance stability.

A broader shift in credit risk philosophy

CLO managers across the leveraged loan landscape have been quietly adjusting portfolios in response to AI-related risks, reducing exposure to companies whose business models look most vulnerable to automation or substitution.

The $150B figure from JPMorgan raises systemic questions. Leveraged loans are a critical funding mechanism for mid-market and growth-stage companies, many of which operate in precisely the sectors most exposed to AI substitution. If CLO managers collectively pull back from these credits, borrowing costs for tech-adjacent companies will rise, potentially accelerating the very distress the managers are trying to avoid.