Blog

The Next Banking Crisis is Hiding in Plain Sight

Outside the Banks

Private Credit Remains Under Pressure from AI Infra Debt

For commercial borrowers, an alternative channel for funds, outside regulated traditional banking system, has always existed. Private credit or direct lending falls in that space.  It is lending by non-bank financial institutions including private equity firms and asset managers. The borrowers are usually small to medium sized businesses who are already highly leveraged and may often not be able to borrow in corporate bond markets.

Currently there is lot of attention on private credit because large borrowers, particularly high-tech mega corporations have dipped into this market. Not because they cannot borrow in the traditional financial institutions. It has more to do with the fact that credit markets have maxed out and the AI ecosystem is looking for funds anywhere and everywhere.

It is also well understood that the underwriting norms and interest rates in the private credit market are markedly different from traditional channels. Some may argue that it is even lax.

That brings us to the crux of the market anxiety we witness today. In my previous post (March 2026), I had highlighted a potential problem in this industry when we noticed liquidity issues at several big players including BlackRock, Blue Owl, Morgan Stanley, etc. that restricted customer withdrawals. Most of them were lending to AI infra players then and continue to do so now.

While no major collapse has occurred since, market worries on this space continue.

Fitch Ratings’ U.S. Private Credit Default Rate (PCDR) was at 6.2% for the trailing 12 months (TTM) ended October 2026, down from a record 6.3% in September 2026. It has remained at or above 6.0% since April 2026.

Some fresh data can help understand the magnitude of the investments.

The Federal Reserve Financial Stability Report, November 2026 reports that debt issued to finance advanced computing rose to more than $680 billion by October 2026 (roughly five times the full-year 2023 total), accounting for about 31% of nonfinancial investment-grade corporate issuance.

Other projections are on similar lines.

Atrium’s tracking of U.S. data-center development credit found that special purpose vehicles (SPVs) structures and pure private credit accounted for about $158 billion.

Bank for International Settlements (BIS), in a note had estimated private credit lending   to tech firms globally to be over US$ 1 trillion in 2025.This includes software / SaaS and not exclusively AI ecosystem investments.

It is important, however, to remember that the figures may not reflect an accurate picture because much of the financing is off balance sheet in SPVs, GPU-collateralized facilities, and asset-based finance structures that are not fully captured in standard private-credit aggregates.

The message is clear. The sheer volume has attracted the attention of various regulators and multilateral lending institutions including the IMF. IMF’s Global Financial Stability Report, October 2026, warns that while global financial markets appear resilient due to solid economic growth and strong corporate earnings, underlying financial stability risks remain elevated.

It warns of a potential AI market correction. Weaker-than-expected returns in AI ecosystem could “spark sharp asset repricing, amplified by leverage and concentration”.

The problem is not that private credit financing structures are inherently bad. They are   perfectly legitimate and have played and continue to play a constructive role in the economy.  

The different underwriting paradigm, the off-balance sheet portfolio profile and regulatory scrutiny may not be the only concerns on this industry. The real problem sitting in plain sight is the scale and rapid growth of the underlying assets. They are ultra expensive, rapidly depreciating and susceptible to rapid technological advancements in AI industry. Further these credit assets are with a small pool of hyperscalers whose aggressive investments are not guaranteeing matching cash inflows, raising concerns of potential default.

It, certainly, would be redundant to point out that given the size of the investments, the implication of any market correction or adjustments would be massive and permeate global financial markets, including the US.

But in same vein, it is also true that the private credit industry has done the heavy lifting in funding and spawning the AI boom, particularly the data center build out boom. It certainly commands a lion’s share of the credit for the unprecedented transformation and incredible changes that AI has induced.

Default or no default in private credit may not be able to alter the innovations and productivity windfalls that will accrue to human civilization. Worst case scenario, many big names in private credit will go underwater, but AI will continue its incredible march.

It is also true that despite the copious notes of caution from regulatory institutions, the market has not yet seen any major defaults in this space. Certainly, that does not mean that market regulators should dilute or wish away guardrails.

The IMF report referenced earlier has pointed out the need for authorities to strengthen systemwide “monitoring of risks across the entire financing chain”. It has advocated closing the data gaps on leverage, contingent liquidity commitments and cross-sectoral and cross border exposures. Of course, these are necessary steps to mitigate systemic risks before they materialize.

That said, these advocacies cannot remain in paper. Unless these are translated into firm regulatory requirements, the threat of massive market adjustments or corrections looms large over the horizon.

More By  :  Naagesh Padmanaban

  • Views: 42
  • Comments: 0


(will not be published)
1500 characters left
Can't read? Reload
Please fill the above code for verification.