Steve Eisman warns that more than 65% of the S&P 500 is now tied to artificial intelligence, leaving investors far less diversified than they might think.

Yet the investor famous for The Big Short says he owns Nvidia Corp. (NASDAQ:NVDA) and GE Vernova Inc. (NYSE:GEV), two major beneficiaries of the AI infrastructure boom.

“I own Nvidia, for example, and I watch and I worry,” he said in an interview with In the Money’s Amber Kanwar.

Suppliers Are Collecting the AI Cash

Eisman favors companies receiving AI infrastructure spending over those financing the buildout.

“I would be looking at the companies that are getting that cash flow,” he said in an interview on Oct. 8.

Nvidia supplies the chips powering data centers. Its revenue rose 106% from a year earlier to $96.2 billion in the quarter ended July 26.

GE Vernova sells gas turbines and electrical equipment needed to power AI data centers. Its companywide backlog reached $176 billion at the end of June, while data-center orders in its Electrification business topped $5 billion in the first half of 2026, more than double its full-year 2025 total.

But Eisman is taking precautions. He has partially hedged four unnamed AI-related holdings while remaining mostly long.

He worries that OpenAI and Anthropic could struggle to finance their enormous commitments, potentially slowing spending throughout the supply chain.

The S&P 500’s Hidden AI Exposure

Eisman has long recommended buying the S&P 500 and leaving it alone. He now warns the index, tracked by the SPDR S&P 500 ETF Trust (NYSE:SPY), provides limited protection against an AI downturn.

Information technology represents roughly 40% of the index by his count. Add Alphabet, Amazon and other technology-related companies, and the figure exceeds 50%.

Include utilities and industrial firms supplying AI infrastructure, and Eisman estimates exposure rises above 65%.

“It’s mostly all one trade. And to be perfectly frank, that makes me nervous,” he said in his podcast on Oct. 9.

The warning comes amid broader market pain. More than 40% of S&P 500 stocks have fallen in 2026, despite the index gaining roughly 14%, Reuters reported Monday.

Even bonds may offer limited shelter. Eisman estimates more than half of this year’s corporate bond issuance is AI-related, exposing lenders to the same investment cycle.

For investors seeking to avoid AI exposure, his preference is clear.

“I would just buy short-term treasuries,” he said.

On Polymarket, traders assign approximately 6% odds to a qualifying AI industry downturn by year-end. The contract requires at least three specified stress events within a 90-day period, rather than an ordinary market correction.

Eisman isn’t predicting a collapse, and he’s still long Nvidia and GE Vernova.

Image: Shutterstock

Read Also: Google Could ‘Eat OpenAI’s Lunch,’ Warns Cal Newport