Tech borrowers account for 39% of $139.8 billion in distressed US leveraged loans

Take technology out of the distressed pool and $85.4 billion, or 61% of it, is still there, owed by borrowers in other sectors ^[18]. Tech's 39% is the largest share, and no other sector comes close in the JPMorgan data ^[6][7]. Software is one part of the tech figure. Its own distressed volume hit a record $25 billion to $40 billion during the AI-related selloffs, according to Crypto Briefing's account of the report ^[26].

The concentration is clearer by borrower, or rather by the size of the borrowers now falling below 80 cents. About 141 issuers have loans under that line, against roughly 106 a year ago, a rise of a third ^[10][24]. Distressed dollar volume rose nearly 90% over the same year ^[4]. On those figures the average distressed issuer is about 40% larger than last year's, with roughly $990 million of loans below 80 cents ^[20][19]. Software providers CDK Global, Qlik Technologies and Quest Software are among the biggest contributors, according to Bloomberg, as reported by Quartz ^[11].

A loan bought at 60 cents returns about 67% if the borrower repays at par ^[25]. Nearly half the distressed pool, $65 billion of $139.8 billion, trades below that price ^[21]. Payment defaults have not yet tested the market's doubt. The trailing 12-month payment default rate on the loan index was 0.93% in July ^[12]. Part of the gap, according to Crypto Briefing, comes from liability management exercises and distressed exchanges ^[14]. In these, borrowers push out maturities or swap loans for new ones on terms worse for lenders, and avoid a formal default (the borrower keeps paying, just on different paper) ^[14]. A lender can take a loss in one of those swaps that never appears in the payment default rate ^[14].

Defaults could rise to meet the prices. JPMorgan projects leveraged loan defaults at 4.50% next year, up from 2.25% in 2026 and well above its 2.75% projection for high-yield bonds ^[9]. The 4.50% is about 4.8 times the July payment default reading, though it may be a different measure ^[22]. Exchanges could instead keep the default rate low while lenders absorb losses through worse terms ^[14]. Or software borrowers could refinance what comes due, and their loans could climb back toward the near-par prices stronger borrowers already get ^[13].

I think holders of tech-heavy loan books take most of their losses along the second path, and the default statistics record those losses late, if at all ^[14]. JPMorgan's own projection argues for the first path, and the two can run together ^[9]. This view is wrong if software loans recover toward par while payment defaults stay under 1% ^[12][13]. In that case the 60-cent marks were a price that misjudged the value, and buyers of the paper collected the 67% ^[25].

The sources do not break out how much of the $54.4 billion sits in CLOs or loan funds. Quartz adds that loans originated before 2024 largely did not account for AI as a meaningful business risk ^[27]. If that holds, the year a fund's software loans were written tells an allocator more than the fund's sector label.