Back to Flash News

Eisman: About 65% of the S&P 500’s weight has become an AI trade

Investor Steve Eisman, who became famous for shorting subprime mortgages before the 2008 financial crisis, said that the weight of AI-related companies in the S&P 500 index could be as high as 65%, and that the entire stock market has in effect become a trade centered on artificial intelligence.

He pointed out that the information technology sector accounts for about 40% of the index’s weight; if companies such as Google and Amazon, which are actually tech companies but are not classified in that sector, are included, the proportion is close to 50%. Adding utilities that supply power to data centers, large banks and alternative asset managers that provide financing, industrial companies involved in construction such as Eaton, and power infrastructure companies such as Quanta Services and GE Vernova, the weight can easily reach 65%. Nvidia, under Jensen Huang’s leadership, has invested in or loaned money to many companies through circular financing, including about $20 billion to Anthropic, about $30 billion to OpenAI, about $20 billion to SpaceX, about $2.5 billion to CoreWeave, and about $2 billion to Nebius; these funds then flow back to buy Nvidia chips, forming what he called a “virtual vertical holding company.”

Eisman believes that Huang’s resignation means the actual dissolution of this structure. Companies that were previously seen as a whole will have to develop independently, and investors will have to assess separately their financing capabilities, operating conditions, management teams, cash flow, balance sheets and earnings prospects. AI is no longer a single trade that can be made as a whole; stock picking and valuation become important again.

ABAB AI Insight

As index weight extends from a single technology sector into utilities, finance and industrials, it shows that the capital path has expanded from chip purchases to the entire chain of power, construction and financing, creating a highly correlated single risk exposure. Circular financing turns supplier capital into customer purchasing power, accelerating demand and capital concentration in the short term, but also binding upstream and downstream parts of the industry chain to the same core node. When a personnel change occurs at that node, the binding relationship is broken, and market pricing shifts from an “overall AI beta” to individual company fundamentals and independent financing capability, with capital once again flowing in a differentiated way according to company quality.

Cognitive law
Binding creates a false whole: when a single node simultaneously serves as supplier, funder and demand anchor, the market prices dispersed entities as the same trade; once the node loosens, the whole immediately breaks down into independent credit.