Strip out the AI darlings from the S&P 500, and the index barely moved over the past two years. With them, it more than doubled.
From May 2024 to June 2026, the S&P 500 posted a 142% gain. Remove the AI stocks, and that number collapses to 16%. The gap between those two figures tells you everything about where the market’s real engine lives, and how narrow that engine has become.
A 45% concentration that should worry you
AI stocks now account for 45% of the S&P 500’s total market capitalization. That’s an all-time high for any single thematic cluster within the index. Nearly half of the most widely tracked equity benchmark on the planet is riding on a single technological thesis.
The usual suspects are driving this. The “Mag 7,” which includes Apple, Microsoft, and Nvidia, have been the gravitational center of this rally. Their collective weight has pulled the entire index upward, masking what is otherwise a pretty underwhelming performance from the other 493 companies in the basket.
The bull case: scarce assets and big forecasts
Capital Economics forecasts the S&P 500 will reach 7,250 by the end of 2026, banking on continued momentum from the AI rally and supportive economic policies.
Tom Lee from Fundstrat has been vocal about what he calls “scarce assets” as key drivers for 2026. His list includes AI hardware companies like Nvidia, AMD, Intel, and Micron, alongside energy infrastructure plays such as GE Vernova and Caterpillar. The logic is straightforward: building out AI infrastructure requires chips, power, and physical equipment. Companies that supply those things occupy a bottleneck position.
ETF performance data supports this framing. Funds built around AI-related themes, including Fundstrat’s own Granny Shots ETF, have posted notable weekly gains against both the S&P 500 and the Russell 2500.
The $1.4 trillion question
Behind the rally sits a growing mountain of debt. AI-linked borrowing has reached $1.4 trillion, a figure that encompasses everything from corporate bonds issued by hyperscalers to leveraged positions in AI-adjacent equities.
The sustainability question isn’t about whether AI is real. The question is whether the current pricing already reflects several years of future revenue growth, and what happens to a leveraged market if those revenue projections slip even slightly.
What this means for investors
For anyone holding a vanilla S&P 500 index fund, you are running a concentrated AI bet whether you intended to or not. Nearly half your portfolio’s value is tied to one theme.
The 16% figure for the ex-AI S&P 500 reveals something important about diversification. Investors who thought they were diversified by owning “the whole market” through an index fund were actually making a massive sector bet.
If credit conditions tighten or if any major AI company reports disappointing revenue growth, the $1.4 trillion in AI-linked borrowing could unwind quickly.
Capital Economics may well be right that the index hits 7,250 by year-end. But the path to that number runs through one of the most concentrated markets in modern history, with 45% of the S&P 500 in a single thematic cluster and $1.4 trillion in AI-linked debt underpinning the rally.