By Anirban Sen NEW YORK, March 17 (Reuters) - Investors are offloading software loans in debt vehicles at a discount, in the latest sign of pain in the software industry, which is being upended by AI.
Debt Investors Reduce Software Exposure Amid AI-Induced Credit Fears
By Anirban Sen
AI Disruption Drives Shifts in Software Debt Markets
NEW YORK, March 17 (Reuters) - Investors are offloading software loans in debt vehicles at a discount, in the latest sign of pain in the software industry, which is being upended by AI.
In recent weeks, several managers of collateralized loan obligations (CLOs) have started exploring ways to reduce their exposure to software, as they grapple with the prospect of a wave of rating downgrades on junk bonds and potential defaults down the line, according to three CLO managers and several credit industry analysts.
The push to reduce exposure shows how the pain in private credit and software is still working through the system after the software rout in January and February that was largely triggered by the release of Anthropic's latest AI tools, which raised fears of widespread disruption across the technology and professional services industries.
CLO Managers Respond to Market Turbulence
"Software is a sector where there is more selling coming from CLO managers than there is buying right now," said Jim Egan, co-head of securitized products research at Morgan Stanley, adding that there was elevated exposure to software within broadly syndicated loans (BSLs), which are corporate loans that are arranged by investment banks and sold to a wide group of credit investors like CLOs. Egan added CLOs currently have lower exposure to riskier companies, which are "CCC" rated, compared to a year ago.
CLOs, which buy up small chunks of numerous individual leveraged loans, in recent years capitalized on the credit boom and bought up loans that backed hundreds of software buyouts in the height of a dealmaking boom during and after the pandemic. During the same period, CLOs also hoovered up their holdings in other non-software sectors that are now faced with the existential threat emerging from AI. According to initial estimates from JPMorgan analysts, around $40 billion to $150 billion of U.S. CLO holdings fall within sectors that are most associated with AI risk.
Software Sector Exposure in CLOs
The software and services sector accounts for about 15% of the collateral in currently outstanding syndicated CLO deals in the U.S., according to a Feb. 20 estimate from Morgan Stanley, which added that software alone makes up roughly 12% of CLO holdings, making it the single-largest subsector by concentration. Software exposure in direct lending is estimated to be about 19% based on private-credit focused CLOs, Morgan Stanley said in a March 17 note.
Widening Spreads and Market Impacts
Spread Trends and Investor Sentiment
Spreads on CLOs, which are the risk premium that companies pay on the bonds over Treasuries, have widened over the past few weeks as fears of a meltdown in the $1.8 trillion private credit industry have spooked investors.
“We’re seeing some CLO managers reduce exposure to software — particularly where positions were overweight or ahead of refinancing activity," said Al Remeza, associate managing director at Moody’s Ratings. "At the same time, many view the current environment as a buying opportunity, especially for companies they believe are least vulnerable to AI disruption.”
Recent Loan and Bond Sales
A mix of investment-grade notes - which are senior unsecured corporate bonds - and high-yield leveraged loans of some software makers, including Intuit, Dayforce, and Citrix, were sold between a range of 89 cents and 98 cents on the dollar in late February and earlier in March, according to data compiled by the Trade Reporting and Compliance Engine (TRACE), which was developed by the Financial Industry Regulatory Authority to track over-the-counter fixed-income transactions.
A few months ago, those same bonds and loans were trading at a premium, the data shows. Reuters could not determine which specific software companies CLOs sold. Dayforce and Citrix did not respond to requests for comment.
Intuit's Performance Amid AI Fears
To be sure, while spreads across the software industry have widened, the spread on Intuit's investment-grade bond that matures in 2033 is largely in line with the level at which it was issued in 2023, while its credit rating was upgraded to 'A' from 'A-' by S&P Global in October last year. Intuit's shares are down about 32% so far this year, as AI disruption fears have weighed broadly on the enterprise software industry.
"We bet the entire company on data and AI nearly ten years ago, when we declared our strategy to be an AI-driven expert platform to deliver done-for-you experiences. Our strategy is working; in the first half of our fiscal year 2026, we delivered 18 percent revenue growth while expanding margins," an Intuit spokeswoman said in an email to Reuters.
Assessing AI Risk in Debt Portfolios
Challenges for Distressed Debt Investors
While the current bout of selling could present a unique buying opportunity for distressed debt investors, several credit industry analysts cautioned that the buyer base for large swathes of these loans is thin, adding that most large private credit firms and direct lenders are unlikely to participate in large software loan deals in the near term as they grapple with investor scrutiny amid rising redemption requests at their flagship funds.
Frameworks for Evaluating AI Threats
"The majority of the CLO community is really taking its time to think about how to come up with a framework to assess AI risk, more on the single-name level, to really scrub their book to identify which are the names that are more prone to AI risk," said Joyce Jiang, head of U.S. CLO Research at Morgan Stanley. "They're still in the middle of doing that, so in the near term we think it's not likely that there's going to be dip buying from the CLO community at a full scale."
Market Rotation and Future Outlook
This is likely to be exacerbated by the fact that CLO managers, who have relatively less exposure to software, are not yet seeing a strong enough reason to buy up loans that are coming to market, said Gavin Zhu, head of U.S. CLO Research at Barclays.
"It's a bit more difficult to suddenly and opportunistically rotate back into software without a true catalyst. And I think that mi
