(Bloomberg) — Anywhere from $40 billion to $150 billion of leveraged loans packaged into US collateralized loan obligations could be disrupted by the artificial intelligence boom, according to JPMorgan Chase & Co.
That’s because those loans fall within sectors most associated with AI risk, according to the Wall Street lender, which released its estimate as a recap of the SFVegas 2026 conference where software’s impact on corporate CLOs was “the topic du jour.”
Most Read from Bloomberg
CLOs offer investors exposure to floating-rate debt, rather than fixed-rate corporate bonds. They do this by bundling leveraged loans into bond-like products with differing levels of risk and reward, which are then sold. Lately, CLO managers have been sifting through their portfolios to determine which loans are most exposed to AI’s impact, following a sharp selloff in software loans sparked by Anthropic PBC’s release of its powerful Claude chatbot.
“AI-pocalypse? That seems over the top,” JPMorgan strategists led by Rishad Ahluwalia wrote in their Thursday note. “While the focus on software is valid, we have told investors we feel it is more important (if more nebulous at this stage) to consider the impact of broader AI disruption on CLO credit risk.”
To come up with their $40 billion-$150 billion estimate, the strategists looked at a simplified screen for CLO AI credit risk using market price and ratings information. Still, they acknowledge that the approach needs more nuance and refining, citing the healthcare sector as an example where proprietary data issues and a complex regulatory environment make it difficult to offer more clarity to investors.
The strategists also highlighted concerns around loan refinancing risks, noting that around $51 billion of software debt rated B- or lower is set to mature in 2028, and a further $50 billion in 2029.
“In addition, the large software exposure in private credit suggests the limited ability of private markets to refinance syndicated assets, unlike in the past when public-to-private takeouts were common,” they wrote.
At the conference, attendees were also concerned about price risk should the labor market weaken or anxiety around AI prompt a broader selloff, the strategists said.
“To be fair, our economists expect the diffusion process of AI into the economy to be more gradual,” they wrote. “However, the leveraging of financial markets to AI also carries risks of an unpleasant reset of expectations, and our cautious 2026 outlook for CLOs also plays to this theme.”