TL;DR — A third of every dollar in the S&P 500 now sits in just six companies. An index fund marketed as diversified carries the concentration risk of a handful of mega-cap tech bets.
Buy the most boring, most diversified investment product retail investors have — an S&P 500 index fund — and roughly a third of every dollar lands on six companies. Not five hundred companies in roughly equal measure, which is what “diversified” implies to most people who buy the pitch. Six. Nvidia, Apple, Microsoft, Amazon, Alphabet, and Broadcom currently account for about 32.7% of the index’s total market capitalization, based on current SPY holdings. Add Meta and Tesla and the top ten names climb past 38%. Nobody voted for that allocation. It happened automatically, and it keeps happening every time a paycheck lands in a 401(k).
The Math Behind the “Diversified” Label
S&P 500 index funds don’t hold five hundred companies in equal amounts. They’re cap-weighted, which means each company’s share of the fund is proportional to its share of the index’s total market value. That sounds neutral — the market decides, not a fund manager — but it creates a mechanical feedback loop. When a stock goes up, its weight in the index goes up too, which means every new dollar flowing into the index fund sends a larger share to that stock than it did the day before. Winners get bigger allocations automatically, which pushes their prices up further, which increases their weight again. Nothing about this requires anyone to believe a stock is worth more. It just requires new money to keep arriving, and new money arrives constantly, on autopilot, every time someone’s retirement contribution clears.
Six Names, One Third of the Index
Run the current numbers and the concentration is stark. Nvidia alone is 7.70% of the S&P 500. Apple is 6.49%. Microsoft is 5.27%. Amazon is 4.20%. Alphabet, counting both its share classes as the single company it is, comes to roughly 6.15%. Broadcom adds 2.89%. Combined, that’s about 32.7% of the index sitting in six businesses, four of which — Microsoft, Amazon, Alphabet, and Nvidia’s largest customers — we’ve covered directly in the past two weeks of earnings reporting. Widen the lens to the top ten, adding Meta, Tesla, Berkshire Hathaway, and Eli Lilly, and the share of the index crosses 38%. An index fund holder isn’t spreading risk across the American economy in any meaningful sense anymore. They’re running a concentrated bet on a handful of companies, wrapped in a product that’s marketed and sold as the opposite.
The Feedback Loop Nobody Votes On
What makes this different from an active fund manager making a concentrated bet on purpose is that nobody is making a decision here at all. Passive index funds now hold a substantial and growing share of total U.S. equity market capitalization, and that money flows in through retirement contributions, target-date funds, and robo-advisor allocations largely without anyone examining what they’re actually buying. The fund doesn’t ask whether Nvidia at 7.70% of the index is a reasonable weighting. It buys 7.70% of every incoming dollar’s worth of Nvidia because that’s the rule, and the rule doesn’t have an opinion. The concentration isn’t a bet anyone placed. It’s what happens when enough capital runs on a formula for long enough.
Why This Isn’t Quite 1999 — and Why That’s Not Entirely Reassuring
The obvious comparison is the dot-com top, when technology stocks reached a similar share of the index before the crash that followed. The comparison has a real limit: today’s largest holdings are, for the most part, businesses that generate tens of billions of dollars in actual operating cash flow every quarter, not pre-revenue concepts trading on narrative alone. Microsoft’s backlog, Amazon’s operating income, and TSMC’s order book are real, auditable numbers, not projections. That’s a legitimate difference, and it’s worth taking seriously rather than waving away as denial.
What the profitability argument doesn’t address is correlation. Real earnings don’t make six stocks move independently of each other. Right now, they largely move together, on the same macro inputs — AI capex expectations, interest rate paths, a handful of shared customers and suppliers. A portfolio that’s 33% concentrated in businesses whose stock prices react to the same news on the same day isn’t diversified just because the businesses themselves are profitable. Profitability changes how a drawdown might resolve. It doesn’t change how correlated the ride there is.
What This Actually Means for a Passive Portfolio
Anyone holding “the market” through a standard index fund is holding a bet on the AI infrastructure cycle whether they intended to or not, at a size most of them never chose deliberately. That’s not a reason to abandon index investing, which still beats most attempts to pick individual winners over time. It’s a reason to know what’s actually inside the box before assuming the label on it — “diversified,” “low-risk,” “broad market” — still describes what’s in there. The S&P 500 of today concentrates risk in a way the S&P 500 of a decade ago didn’t, and the fund’s name hasn’t changed to reflect it.
We’ve written about several of the businesses now carrying disproportionate weight in that index directly — Microsoft’s backlog, Alphabet’s capital allocation shift, and Nvidia’s position at the top of TSMC’s allocation queue. Understanding those companies individually is no longer optional for anyone who thinks they’re just holding “the index.” Increasingly, the index is them.
Sources: S&P 500 index weightings via State Street SPDR S&P 500 ETF Trust (SPY) holdings data, accessed August 2026. Third Pole Markets holds no position in the securities named as of publication. This is not investment advice — see our About page for our full disclosure policy.
Index concentration data is published by S&P Dow Jones Indices, the official source for S&P 500 methodology and weights.
