Illustration for: Wall Street's Biggest Bears Still Won't Short AI

Wall Street's Biggest Bears Still Won't Short AI

Even short sellers who call the AI supply chain fragile, including Steve Eisman, aren't willing to bet against AI stocks yet -- a signal that matters for anyone gauging how much longer the funding cycle has to run.

By the Numbers

6-12 month risk window
Amodei slowdown call
Declined, then held
Semiconductor reaction
None disclosed
Eisman's AI shorts
Sold Oracle stake
Eisman's move
TC
By the AI Desk
Edited by Trace Cohen · Early-stage VC & angel · Founder, New York Venture Partners
1 min read
ShareXLinkedInEmail

THE RUNDOWN

1

Steve Eisman, the investor who famously shorted mortgage bonds before 2008, calls the AI supply chain 'fragile' because it depends on two cash-burning labs -- yet he has not shorted a single AI name, only trimmed Oracle over its OpenAI exposure.

2

Wall Street analysts broadly still expect AI infrastructure spending to keep rising, arguing that even Anthropic CEO Dario Amodei's own call for a deliberate AI slowdown wasn't enough to change the investment thesis.

3

The absence of aggressive shorting is itself a data point: professional bears with strong track records calling bubbles are choosing to wait rather than bet against the trade, which historically front-runs a turn by months, not days.

4

For founders raising growth rounds against public AI comps, this suggests the public-market ceiling for AI valuations hasn't been tested yet by real short pressure -- comps remain support, not resistance.

TC

The VC Read · Trace's Take

Trace Cohen

The diligence signal isn't Eisman's words, it's his trade: selling Oracle over OpenAI exposure but not shorting OpenAI-adjacent names directly tells you where the perceived risk actually concentrates -- in companies levered to a single lab's continued spending, not in AI broadly. If you're on a cap table with heavy single-customer AI-lab revenue concentration, that's the exact exposure the smartest bear in the room just quietly de-risked around.

Analysis

Even Wall Street's most credentialed bears aren't willing to bet against AI stocks yet, according to The Information. Steve Eisman, the investor who famously shorted subprime mortgage bonds ahead of the 2008 crash, calls the AI supply chain 'fragile' because it leans on two cash-burning labs -- but he hasn't shorted a single AI name. His most aggressive move so far was selling his Oracle shares over its OpenAI exposure, not opening a short position against the trade itself.

The hesitation comes even after Anthropic CEO Dario Amodei called for a deliberate, global slowdown in AI development this month, citing safety risks including the possibility that AI-directed agent swarms could take over parts of the internet within six to twelve months. Semiconductor stocks dipped on the remarks during that trading session, according to CNBC, but Wall Street analysts broadly held their view that AI infrastructure spending stays on an upward path regardless.

“That's a meaningfully different signal than retail sentiment or analyst price targets, which can lag real conviction by months.”

What makes the non-shorting notable is precedent: professional short sellers with real track records calling bubbles -- Eisman among them -- have historically moved ahead of a turn, not after one, when they see it coming. Their continued absence from the AI short book is itself informative, suggesting the smart-money read is that valuations, however stretched, aren't yet detached enough from underlying revenue growth to justify the risk of shorting into a still-rising market. That's a meaningfully different signal than retail sentiment or analyst price targets, which can lag real conviction by months.

The risk this obscures: short sellers waiting for a cleaner entry point doesn't mean the trade is safe, only that the timing isn't obvious yet -- Eisman's own 'fragile' framing is a bear case sitting unexercised, not a bull case. A single disappointing earnings print from Nvidia, Microsoft or Alphabet could be the catalyst that moves bears from watching to acting, and that shift, when it comes, tends to happen fast rather than gradually.

ShareXLinkedInEmail

Key Sources

2 sources

THE WIRE in your inbox— Tech, startup & VC news with Trace's take. Free, no spam.