Analyst sends stark $120 billion AI warning

Everyone's been asking the aforesaid question astir artificial intelligence: who's going to win? Chipmakers person soared. Software companies person rallied. Investors person bought into the thought that we're astatine the commencement of a large productivity boom.

But Matthew Mish atUBS is asking thing other entirely. He wants to cognize who's going to lose, and much importantly, however overmuch harm they'll bash connected the mode down.

Mish runs recognition strategy astatine UBS, which means helium spends his days reasoning astir firm debt. And what he's seeing has him worried. It's not astir whether AI works. He thinks it astir apt will. The occupation is speed.

AI isn't rolling retired gradually anymore. It's moving from labs into existent businesses faster than galore expected. That's large if you're acceptable for it. If you're not, and you've got a heap of indebtedness connected your equilibrium sheet, things could get disfigured beauteous quickly.

Here's the happening astir recognition markets. When banks and investors lend money, they presume the borrower volition person clip to set if contention heats up oregon exertion changes. Maybe 3 years. Maybe five.

Related: SaaS-pocalypse stresses $3 trillion backstage recognition market

But if AI compresses that timeline to 18 months oregon less, galore of those assumptions break. Revenue tin driblet faster than anyone planned for. Cash travel projections that looked coagulated six months agone abruptly don't clasp up.

More AI Stocks:

• Morgan Stanley sets jaw-dropping Micron terms people aft event

• Bank of America updates Palantir banal forecast aft backstage meeting

• Morgan Stanley drops eye-popping Broadcom terms target

Mish told CNBC that markets person been "slow to respond due to the fact that they didn't truly deliberation it was going to hap this fast." He's talking astir what helium calls a "rapid and assertive disruption scenario," which is Wall Street talk for things happening excessively accelerated for radical to get retired of the way.

UBS has been moving the math. They're looking astatine astir $3.5 trillion successful leveraged loans and backstage recognition that could beryllium sitting successful the blast zone. These aren't the safest borrowers. They're companies with weaker equilibrium sheets, higher indebtedness loads, and little country for error.

In UBS's baseline scenario, defaults could deed $75 cardinal to $120 cardinal by the extremity of this year. Default rates could ascent to astir 2.5% for leveraged loans and arsenic precocious arsenic 4% for backstage credit.

Those are manageable if everything other stays calm. But if disruption picks up speed, Mish thinks defaults could spell twice arsenic high. At that point, you're dealing with a repricing of hazard crossed the full recognition market.