The leveraged loan market's liquidity paradox deepened through March. Selling pressure concentrated in the most liquid names—those backed by household-brand LBOs and traded in size—while illiquid, thinly-traded paper held or widened less severely. $47 billion left broadly syndicated loan funds in the first quarter, according to LSTA flow data. The liquid paper absorbed the outflow.

Fear that artificial intelligence will disrupt—or outright destroy—the leverage thesis underpinning current valuations has driven the exodus. Allocators are repricing debt tied to companies whose business models assume stable operating leverage and predictable cash conversion. Software services firms, customer-data platforms, and mid-market industrials with thin IT moats have seen their loan tranches reprice 8 to 12 points below par since January, even when documentation is strong and the underlying credit has not missed a payment. Meanwhile, covenant-lite paper in less-liquid situations—regional healthcare roll-ups, niche manufacturing—has traded sideways or down only 3 to 5 points. The inversion is clean: liquidity became liability.

The mechanism is straightforward. When a family office or credit fund needs to reduce exposure quickly, it sells what clears. Liquid names have quote depth and immediate bids. Illiquid paper requires calls, negotiation, and days of price discovery. In a broad selloff, liquidity is the first attribute punished. The result is a two-tier market: liquid LBO loans now yield SOFR + 525 to 575 basis points at 92 to 94 cents on the dollar, while illiquid equivalents with worse covenants sit at 96 to 98 cents yielding SOFR + 450. The spread between liquidity classes has never been wider in the post-GFC era.

This is the second-order effect allocators must price. If AI fear persists—and consensus now assumes at least 18 months of valuation compression across leverage-dependent sectors—then liquid loan exposure becomes a forced-sale risk, not a diversification hedge. The debt that was supposed to provide quick exit optionality in stress is now the debt that bleeds fastest. Portfolio construction for credit books tilts toward illiquidity tolerance, not liquidity preference. Family offices holding syndicated loan allocations above 12% of liquid alternatives are repricing their own redemption assumptions.

Operators and allocators should watch three follow-on developments. First, whether April redemptions from loan funds exceed $18 billion, which would confirm this is a structural repositioning, not a March quarter-end flush. Second, whether any large buyout shop announces a markdown or restructuring tied explicitly to "AI-driven revenue assumptions"—that language will accelerate the reprice. Third, whether liquid loan-to-bond spreads compress below 150 basis points, which would signal that bond buyers see relative value and are willing to step into credit risk the loan market is shedding. All three data points should clarify by mid-April.

The loan market has not seen credit events yet. It has seen a liquidity reclassification.