U.S. Private Credit: New Cracks in the Post-Lehman Era, a Precursor to AI Bubble Transmission?

Ask AI · How the AI boom is reassessing the debt-paying ability and credit risk of software companies

Reporter Ouyang Xiaohong

When a private equity credit fund tells investors that its assets are still “valued at net asset value” but can only be redeemed at 5%, the market may be losing trust, and not just liquidity.

In spring 2026, the issue with U.S. private credit is that “its true value only becomes visible when someone wants to exit.” Blue Owl (Blue Owl Private Credit Investment Company) disclosed in early April that two retail-oriented private credit funds had redemption requests reaching 40.7% and 21.9% in the first quarter, both ultimately processed at the quarterly 5% limit.

On Monday (April 6), Blue Owl’s stock price fell to $8.45, hitting a new all-time closing low. Due to Blue Owl’s high credit exposure to software borrowers, it is more susceptible to shocks from AI risk revaluation narratives. The public market is viewing Blue Owl as a representative target betting on continued pressure in U.S. private credit.

Meanwhile, a private credit fund under the global asset management firm Barings (Baring) also triggered redemption limits due to surging withdrawals. Redemption limits are originally part of private credit product terms, but when they appear repeatedly at leading institutions, market perceptions of liquidity commitments in this asset class change.

Does this invisible “bank run” hint that U.S. private credit might rewrite the next round of global risk pricing? At the same time, it is worth noting that the ultimate risk bearers have quietly shifted from banks’ on-balance-sheet to funds, insurance, and wealth management products.

“Post-Lehman” New Cracks

Signs of cracks have already appeared. On March 30, Fed Chair Powell said the Federal Reserve is closely monitoring the U.S. private credit sector but has not yet seen signs capable of dragging down the entire financial system. On April 6, JPMorgan CEO Jamie Dimon stated that private credit may not pose systemic risk, but if the credit cycle weakens, losses could exceed expectations, and issues with transparency and valuation could amplify selling impulses.

The most common mistake in the market is trying to identify this cycle’s risks using the contours of the last crisis. So, could the risk in U.S. private credit evolve into “the next Lehman crisis”?

The typical risk structure of 2008 involved high leverage on bank balance sheets, nested securitization chains, and pressure on payment and clearing systems; today’s private credit pressures resemble risk moving from banks “off the books.” Fed research in 2025 shows that by Q2 2024, U.S. private credit assets totaled about $1.34 trillion, nearly $2 trillion globally; by Q1 2024, large U.S. banks’ commitments to BDCs (Business Development Companies, which pool private credit resources for SMEs) were about $56 billion. Banks remain part of the chain but more as connectors and leverage providers rather than the sole final risk repository.

This is why “seeming stability of net asset value” is not enough to reassure markets. Many private loans are not traded continuously like public bonds; they rely more on models, quarterly assessments, and non-public comparable assumptions for pricing. Usually, this mechanism smooths out portfolio volatility, but when investors truly want to exit, the question becomes: at what price, when, and to whom?

Why Now?

Private credit is under dual pressure from macroeconomic and microfundamental factors.

On the macro side, Middle East conflicts have pushed oil prices higher, and inflation remains sticky. In early April, Jamie Dimon warned that Middle East conflicts could impact oil and commodities, making inflation more persistent and interest rates higher than previously expected. Meanwhile, institutions like Morgan Stanley, Goldman Sachs, and Barclays have generally delayed expectations for the next Fed rate cut to around September 2026. For private credit, “high interest rates lingering” is the real danger, as it means borrowers relying on floating-rate financing will endure longer periods of high cash interest payments; those dependent on refinancing for survival face higher costs, greater difficulty, or even the risk of being unable to refinance.

But the current pressure on U.S. private credit is not just macro interest rates or fund liquidity issues. A deeper background is that long-term capital pursuit of high returns continues to drive capital into high-valuation, high-volatility sectors like software and AI. In recent years, in a low-interest, highly liquid environment, demand for “high-yield, low-volatility, publicly substitutable” products has risen steadily, with private credit expanding and gradually taking on more risk outside the banking system. In a sense, this echoes the 2000 dot-com bubble: capital first bets on narratives, then uses narratives to inflate valuations and financing capacity.

Goldman Sachs research suggests that the current AI boom has similarities to the 2000 internet bubble, but is not a simple replay—there are localized overheating, aggressive valuation, and capital expenditure issues, but leading AI companies have stronger profitability and cash flow foundations, making it more like “a localized bubble or bubble signs” rather than a full-scale internet bubble replay.

More specifically, the focus is on the most sensitive area—software sector borrowers.

In recent years, software companies, with high gross margins, subscription revenues, and “light assets with high cash flow,” have become favored borrowers in private credit; but now, AI is changing this narrative. There is growing concern that AI may erode some software companies’ profitability models, pricing power, and debt repayment ability, and software is precisely the core risk exposure in private credit portfolios. In other words, the market’s worry is no longer just about “whether the economy will slow,” but whether the business models of borrowers are being revalued.

Interestingly, just one day before the redemption pressure drew market attention, Blue Owl announced that its “Asset Special Opportunities Fund IX” had raised about $2.9 billion by March 31, 2026, exceeding its initial $2.5 billion target. Unlike products with quarterly liquidity arrangements, this type of fund targets institutional investors, has longer lock-up periods, and emphasizes asset-backed, opportunistic allocations and downside protection. For the market, this indicates that the current problem with U.S. private credit is not about overall capital outflows but about where capital is leaving and where it is flowing to.

The coexistence of Blue Owl’s redemption limits and oversubscription reflects a re-tiering of capital within private credit: from semi-liquid, retail-oriented, corporate credit-focused products to more closed, institutionalized strategies emphasizing asset support and downside protection.

This suggests that Blue Owl’s redemption limits, falling stock prices, and the re-pricing of AI risks in software borrowers are collectively pushing up industry risk premiums.

Market Leading Signals

True mature market observation never just looks at statements but focuses more on prices, terms, and behaviors. Revaluation often occurs first in financing conditions, not in managers’ monthly net asset values.

Redemption limits seem to be evolving into an industry phenomenon. Blue Owl’s two private credit funds faced redemption requests, ultimately paid out at the 5% quarterly cap; Baring has also followed suit with similar redemption limits. Previously, several large institutions had also experienced redemption limits or approached redemption caps. Redemption limits do not automatically mean loss of control, but once they appear simultaneously at multiple leading institutions, they are no longer just “normal product terms” but a warning that liquidity commitments across the entire asset class are being re-priced.

Moreover, the public market has expressed unease. Listed Business Development Companies and alternative asset managers’ stocks have come under pressure first. If the public market continues to discount BDCs, it essentially signals: book net asset value may look good, but it does not reflect real liquidity conditions and realizable prices. The secondary market is discounting, challenging the valuation narratives of the primary market.

Bank financing is also shifting. Major U.S. banks have begun tightening support for private credit funds. Banks are pragmatic—they consider discounts, advance rates, and financing spreads. When banks believe assets are harder to dispose of, less certain in value, and refinancing environments more fragile, they will respond with higher prices, shorter tenors, and stricter collateral requirements. For professional investors, these signals are often more informative than a reassuring letter to LPs.

The Bank for International Settlements’ March study pointed out that private credit loans to SaaS (Software as a Service) companies have risen to over $500 billion, about 19% of direct lending; fears that AI could disrupt traditional SaaS models have led to declines in software stocks and deeper discounts on BDCs.

Defaults have not disappeared—they are just delayed. In late March, reports indicated that private credit institutions are allowing some borrowers to delay cash payments and accept looser terms to avoid immediate defaults. Because of this, such risks do not clear instantly like in public markets but first manifest as: bad news not fully priced in yet, but financing conditions have already changed.

Who Is Taking on the Risk?

In this structure, regulators and investors need to correct not “whether risks exist” but “whether they are still being evaluated using the last crisis’s coordinates.”

Powell’s statements and Dimon’s comments suggest that, for now, related risks have not reached the level of bank capital failure.

If the 2000 internet bubble’s risks mainly manifested in equity markets, today’s vulnerabilities are embedded in the non-bank credit chain, including private credit, BDCs, insurance, and wealth products. This cycle’s risks may not explode first in the public equity markets like in 2000 but could gradually propagate within the non-bank system along the chain of “valuation lag—tightening financing—rising redemptions—rising defaults.”

If private credit risks are not just a symptom of the AI bubble, then the overhyped AI narrative is accelerating the exposure of cracks in this non-bank credit system.

This explains why the current risk structure differs from 2000 or 2008. The 2000 dot-com burst primarily hit equity valuations; the 2008 financial crisis first overwhelmed bank capital and securitization chains; but this time, the more concerning aspect is the expansion of high-valuation financing, high-yield allocations, and non-bank credit driven by AI narratives, forming a more covert and less quickly resolvable risk transmission mechanism.

Some analysts believe this is not a 2008-style banking capital crisis. Instead, it resembles an opaque, non-bank, valuation-lagged credit market stress test under higher interest rates and AI revaluation environments. It could develop into a significant credit risk event, but more likely in a “multi-point outbreak” from late 2026 to 2027: rising defaults among software and service borrowers, continued redemption limits on retail private credit products, and then potential amplification through banks, insurance, and mortgage-backed securities. Morgan Stanley estimates that private credit default rates could reach 8% in the second half of 2026 to the first half of 2027, indicating substantial credit deterioration.

Data from Crunchbase shows that in Q1 2026, global startup financing reached about $297 billion, with AI-related funding about $240 billion, accounting for roughly 80%. This indicates that global AI investment and financing are not cooling but accelerating, with funds increasingly concentrated. Further, AI funding remains heavily focused on leading U.S. companies and a few mega-deals, making the impact of the AI boom’s turn more likely to transmit through valuation revaluation, credit tightening, and risk appetite decline into the U.S. private credit market.

Thus, the cracks in U.S. private credit, after the AI capital boom enters its latter stages, are early signs of overheating in equity markets, software credit revaluation, and vulnerabilities in non-bank finance overlapping. Private credit risks are not the root problem but may be the first front where the AI bubble begins to influence risk pricing.

The world spent over a decade repairing banks, but now the market is only beginning to realize that new cracks may be outside banks.