"Big Short" Investor Sells Google Stock, Warns Market Is One Giant AI Bet

Steve Eisman, the investor famous for correctly calling the 2008 housing market crash and who inspired one of the characters in "The Big Short," told CNBC that he sold his long-held stake in Alphabet. The reason: he wanted to cut his exposure to artificial intelligence, arguing that nearly the entire financial market has become one giant bet on the same theme.

One giant trade

The key line Eisman repeated in his CNBC interview is that the market today is "one trade." He explained that even people who believe their 401(k) retirement portfolio is diversified, split in the classic 60 percent stocks to 40 percent bonds allocation, are in reality holding one concentrated bet on the success of artificial intelligence.

It's all one trade. It's literally one trade. - Steve Eisman, investor

Eisman explained that even buying shares of investment banks like Goldman Sachs doesn't amount to real diversification, since those banks are now largely financing the AI buildout themselves. In his view, most new corporate bond issuance is also tied, directly or indirectly, to spending on AI infrastructure.

Why he sold Google

Eisman stressed that the decision to sell Alphabet wasn't driven by a negative view of the company's fundamentals. Google posted 24 percent overall revenue growth in the second quarter, with cloud segment revenue up 82 percent to $24.8 billion.

I sold Google a few months ago... I wanted to reduce my AI exposure. - Steve Eisman, investor

Despite the strong operating results, the company raised its 2026 capital expenditure forecast to $205 billion, and its free cash flow dropped to negative $5.9 billion, the first negative reading on that metric since 2004. According to Eisman, that figure, combined with the scale of spending by other hyperscalers, is exactly what rattled the stock market over the past week.

Risk for hyperscalers

The investor warned that markets would "go straight down" if any of the major cloud and AI providers, meaning Google, Microsoft, Amazon or Meta, were to pull back on their AI infrastructure spending. He also pointed to another potential bottleneck: the rising cost of the language models offered by companies like Anthropic and OpenAI.

Anthropic and OpenAI have models that are much more expensive than the competition today. And I think that's a potential bottleneck. - Steve Eisman, investor

Asked what would happen to the market if artificial intelligence failed to deliver on the business expectations placed on it, Eisman said bluntly that he'd expect a major correction. He currently says he's holding a significant portion of his portfolio in cash while he waits for a clearer picture.

What it means for investors

Eisman's comments echo a broader unease that has emerged in recent weeks over Big Tech's enormous outlays on AI infrastructure. Eisman himself pointed to ServiceNow as an example, a company that reported 24 percent revenue growth yet still saw its stock drop 38 percent, as investors began to question the pace of returns on AI investment.

For Polish investors and pension funds, whose portfolios often track global indexes dominated by American tech companies, Eisman's warning carries practical weight. A large share of stock market exposure, even in products marketed as diversified, now effectively depends on continued AI spending growth at a handful of American corporations.

Eisman isn't claiming artificial intelligence is a bubble destined to burst. His message is narrower: concentration risk in portfolios today is greater than most investors realize, and one bad signal from a single hyperscaler could trigger a sharp sell-off across the entire segment at once.