You know that scene in The Dark Knight when the Joker burns the mountain of cash and says "it's not about the money, it's about sending a message"?
Well. Artificial intelligence is burning mountains of value in the credit market — and most investors are still debating whether the fire is even real.
The warning nobody wanted to hear
Matthew Mish, head of credit strategy at UBS, dropped a report this week that should be keeping a lot of private equity and private credit managers up at night. The thesis is simple and brutal: AI disruption, which has already slaughtered software stocks in recent weeks, is about to spill over into leveraged loan and private credit markets.
We're talking about a combined market of $3.5 trillion. That's not chump change.
In the base case — I repeat, the base case, not the apocalyptic one — Mish projects between $75 billion and $120 billion in new defaults by the end of 2026. Fresh defaults. Companies that will simply stop paying their debts because AI made their businesses obsolete or drastically less valuable.
And the guy made a point of warning: there's a tail-risk scenario (that fancy "tail risk" term Wall Street loves tossing around to sound smart at cocktail parties) where those defaults double. At that point, it's no longer a localized crisis. It's a credit crunch. A systemic shock. Dominoes falling.
"We didn't think it would happen this fast"
The most revealing part of Mish's CNBC interview was the admission that he and his team had to scramble to update their projections. The latest models from Anthropic and OpenAI accelerated the disruption timeline in ways nobody anticipated.
"The market was slow to react because nobody thought this would happen so quickly," Mish said. "People are having to recalibrate the entire way they assess credit for this disruption risk, because it's not a 2027 or 2028 problem."
Read that again: it's not a 2027 or 2028 problem. It's right now.
This reminds me of something Taleb would repeat until he was blue in the face: the risk that breaks you isn't the one you're monitoring on your pretty dashboard — it's the one you filed under "unlikely" and shoved in a drawer.
The three AI buckets — and where the danger lives
Mish sorted companies into three categories in the context of this revolution:
Bucket 1: The foundational model builders — Anthropic, OpenAI. Startups today, potential public giants tomorrow. The kings of the game.
Bucket 2: Investment-grade companies like Salesforce and Adobe. Strong balance sheets, can deploy AI to defend themselves. They'll take hits, but they survive.
Bucket 3: Software and data services companies controlled by private equity, loaded with debt and squeezed margins. The perfect target for disruption.
Guess which bucket the winners of this transformation are coming from?
"If this really becomes rapid and severe disruption — as we're increasingly believing — the winners will very unlikely come from that third group," Mish said.
Damn, if the UBS guy is saying this in a public report, imagine what's being said behind closed doors.
Is tech private equity dead?
Not coincidentally, that same week, a Lightspeed partner declared on national television: "Tech private equity, in its current form, is dead."
And it makes sense. The classic PE playbook in tech was: buy a software company with recurring revenue, lever it to the gills, slash costs, flip it in 3-5 years. It worked because software had reasonable competitive moats and predictable revenue.
AI threw a grenade at that playbook. When a language model can replace entire functions that these companies sell as a service, what's left? Debt. Debt with no cash flow to pay it off.
So now what?
The stock market has already started pricing in this risk — we've seen sell-offs in sectors as diverse as finance, real estate, and trucking. But the credit market, historically slower and less liquid, is still crawling through this reassessment.
When credit wakes up, it doesn't happen gradually. It happens all at once. It's a shock.
Mish said "we're not calling the tail-risk scenario yet, but we're moving in that direction."
Translation from econ-speak: "The shit hasn't hit the fan yet, but you can already smell something burning."
The question you should be asking yourself: if the guys who make a living lending money to these companies are getting nervous, why the hell would you be relaxed?
