The stock market has recently been swift in reacting to the perceived risks faced by software companies amid the artificial intelligence (AI) boom. However, UBS analyst Matthew Mish suggests that credit markets could be the next to reflect the disruption risks associated with AI.
Mish, in a research note released Wednesday, indicates that tens of billions of dollars in corporate loans, particularly those held by software and data services companies owned by private equity, are expected to default in the next year due to AI-related challenges.
"We’re pricing in part of what we call a rapid, aggressive disruption scenario," Mish, who serves as the head of credit strategy at UBS, explained in an interview with CNBC.
The analyst, together with his team, has hastily revised forecasts for this year and beyond as anticipations on AI disruption have been accelerated by the latest models from companies like Anthropic and OpenAI.
"The market has been slow to react because they didn’t really think it was going to happen this fast," Mish observed. "People are having to recalibrate the whole way that they look at evaluating credit for this disruption risk because it’s not a ‘27 or ‘28 issue.”
This month, investor concerns around AI reached a peak, shifting the narrative from considering the technology as a uniform growth opportunity for tech firms to a competitive scenario where firms such as Anthropic and OpenAI challenge existing businesses. While software firms felt the first impacts heavily, the repercussions have rolled across various sectors, including finance, real estate, and logistics.
According to Mish and other UBS analysts, a baseline scenario anticipates between $75 billion to $120 billion in new defaults for borrowers in leveraged loans and private credit by the year's end.
CNBC derived these figures based on Mish’s estimates, indicating a potential increase in default rates by up to 2.5% for leveraged loans and up to 4% for private credit by late 2026, within markets estimated at $1.5 trillion and $2 trillion respectively.