Artificial intelligence has gone from tailwind to threat for parts of the software sector, and the shock is no longer confined to equities. A rapid repricing of listed software names, driven by fears that new AI tools will erode traditional business models, is now colliding with a vast pile of corporate borrowing that depends on those same cash flows. At the center of the worry is a $1.5 trillion slice of the U.S. credit market that Morgan Stanley says is more exposed to AI disruption than investors have been willing to price in.
The concern is not just that software stocks have sold off, but that lenders and bondholders have barely adjusted. With a majority of software loans sitting in the riskier tiers of the ratings spectrum, and a meaningful share tied to companies most vulnerable to AI competition, the gap between equity anxiety and credit complacency is starting to look like a systemic vulnerability rather than a passing market squall.
How an AI shock in equities became a credit story
The starting point for the current stress is a sharp slide in software valuations as investors reassess how quickly AI could cannibalize established products. In research highlighted by By Siddarth, analysts describe how Concerns about artificial intelligence disrupting large parts of the software industry have been building for months, with each new model release prompting another round of earnings downgrades. The latest wave followed new generative AI tools that promise to automate coding, customer support and even complex workflow design, undermining the pricing power of legacy platforms that once looked unassailable.
What turns this into a credit problem is the sheer volume of debt that sits behind those business models. According to Concerns reported in Feb by Reuters, Morgan Stanley has mapped the potential fallout from an AI-driven repricing of software revenues onto a $1.5 trillion universe of U.S. leveraged loans and high yield bonds. The bank argues that if equity markets are right about the speed and depth of disruption, then credit spreads on many software borrowers are still far too tight for the risks they actually face.
Why software lending looks especially fragile
Under the surface, the structure of software borrowing makes the sector more vulnerable than headline numbers suggest. A majority of the software sector’s exposure is tied to lower credit ratings, with 50% of the loans holding a “B- or lower” credit rating. That leaves lenders with little cushion if earnings fall, since many borrowers already sit just a notch or two above distress and have limited flexibility to absorb revenue shocks without breaching covenants or needing fresh capital.
Specialist credit analysts have been flagging the same mismatch between market pricing and underlying risk. In a piece titled Where the Risk in Software Lending, By Bill Alpert notes that Software stocks are down sharply while loan pricing has barely moved, leaving Morgan Stanley cautious that credit markets are underestimating how long the pressure could last. The bank’s analysts point out that in past downturns, such as the early 2000s tech bust, credit spreads eventually caught up with equity volatility, but only after lenders had already extended generous terms that proved hard to renegotiate.
The AI catalyst and the Anthropics of disruption
What makes this cycle different is the speed at which AI tools are being deployed into the very workflows that many software borrowers sell. In a widely discussed market commentary, a video titled Software Stocks Selloff: What’s Behind the Rout? describes how the proximate cause of the latest leg down was Anthropic releasing a tool that could effectively automate tasks that mid-market software vendors currently charge subscription fees to handle. When a single AI release can wipe out a chunk of a company’s addressable market, traditional models of gradual technological disruption start to look outdated.
Credit analysts at Morgan Stanley have tried to quantify that shift by looking at which borrowers have the highest exposure to AI-driven competition. One summary of their work notes that the bank has warned that the rise of artificial intelligence could disrupt the software industry in ways that directly affect loan performance, with Morgan Stanley highlighting that even modest revenue erosion could push highly leveraged issuers into restructuring. While the bank still judges a wave of large, immediate defaults as unlikely, it stresses that the distribution of outcomes has shifted in a way that credit markets have not fully internalized.
Inside the $1.5 trillion exposure
The headline figure that has grabbed attention is the $1.5 trillion in U.S. credit that Morgan Stanley links to software and adjacent technology borrowers. According to a detailed breakdown cited by WHTC in Holland, Morgan Stanley said 20% of that market has direct exposure to AI-driven disruption, either because the borrowers sell software that can be replicated by generative models or because their customers are likely to cut spending as they adopt cheaper AI alternatives. That concentration means a relatively small set of technological breakthroughs could have an outsized impact on a large pool of loans and bonds.
The quality of that exposure is another red flag. A separate summary of the same research notes that a majority of the software sector’s exposure is tied to lower credit ratings, with 50% of the loans rated “B- or lower,” which historically have shown much higher default rates in downturns. If AI adoption accelerates faster than expected, that tail of weaker credits could become a focal point for losses across collateralized loan obligations and private credit funds that have loaded up on software paper.
What lenders and investors are missing
Despite these warning signs, pricing in the loan market has remained relatively calm compared with the equity rout. Story by Bill Alpert notes that while Software names have been hit hard, spreads on many loans have barely budged, even as high-yield bonds in other sectors have roughly doubled in cost during past stress episodes. That suggests lenders may be relying too heavily on the sector’s historical resilience, when recurring revenue and low churn insulated software from typical cyclical downturns, and are underestimating how a structural technology shock could change that pattern.
Part of the complacency may stem from the view that large-scale defaults remain unlikely, a point Morgan Stanley itself acknowledges. In a detailed recap of the bank’s stance, Morgan Stanley is quoted as saying that the risk of large, immediate losses is unlikely, even as it warns that software faces a more challenging backdrop if Concerns about AI were to materialize quickly. That nuance may be getting lost in translation, with some investors hearing reassurance rather than a call to reprice risk before the market forces them to.
