VOO Is 40% Ten Companies. The "500" Is the Misleading Part

I'm always keeping an eye out for the false narratives that take the market by storm, and the most consequential one right now isn't about AI, interest rates, or geopolitics. It's written into the name of the fund most American portfolios are built on. You bought an S&P 500 ETF, so you own 500 companies, and no single stock can really hurt you. That's the story the ticker tells. In my opinion, it isn't true — and the gap between the story and the portfolio is where the actual risk lives.

Do the arithmetic. Put $10,000 into VOO, Vanguard's S&P 500 fund — a 0.03% fee workhorse now trading around $700 — and roughly $750 of it lands in Nvidia. About $658 goes into Apple. The top five positions alone absorb a quarter of your money, and the ten largest swallow more than $4,000 of every $10,000. At the end of 2025 the ten largest companies made up 40.7% of the entire S&P 500. You didn't buy 500 slices of roughly equal size. You bought ten big slices and 490 crumbs, and the ten big slices all sit in roughly the same part of the house.

A risk that ratchets up without your permission

Here's the mechanical part, and it's what makes this "sneaky." The S&P 500 is weighted by market capitalization. When a stock goes up, its weight in the index goes up, and every index fund — including VOO — must buy more of it, regardless of what it thinks. And all that passive money, which is enormous, flows to the biggest names first. The result is a ratchet: concentration grows automatically, reinforced by the very flows it creates. No rebalancing button exists. You never voted for it.

Put that in perspective. The top ten were about 19% of the S&P 500 in 1990, and again about 19% in 2015. The dot-com peak of 2000 reached roughly 27%. At the end of 2025: 40.7%. By this measure, the index has never looked like this before. And the shift happened fast — Nvidia overtook Apple as the largest S&P 500 company in June 2024, on a gap of about $30 billion. Within a year the gap had widened to around $800 billion. The fund just followed the price.

Five hundred stocks, one trade

The structural difference from 2000 is the key thing to understand. Back then the top names were telecoms, PC makers, and internet gateways — related industries, but separate businesses with separate customers. Today's top of the index is one trade: the AI buildout and its customers. Chip suppliers, data-center software, cloud capacity, and the model companies. If the AI story wobbles, they wobble together, and there are few offsetting exposures inside the index to absorb the shock.

We already got a preview. On June 23, 2026, a wobble in chip stocks knocked the S&P 500 down 1.4% in a single day and the Nasdaq down 2.2%. An index of five hundred companies that moves like a single stock on a chip headline is telling you what its real risk profile is. The market recovered — the S&P 500 closed at 7,656.98 on September 11, sitting about 2% below its August high. But recovery is not a refutation. It means the story hasn't broken yet.

Why it isn't 2000 — and the fine print in "cheap"

Where the engineer in me has to step in and stop the alarm: the 2026 top ten are not Pets.com with a market cap. They generate cash. Nvidia's revenue grew 85% in its most recent fiscal quarter to $81.6 billion. These are high-earning, cash-rich companies, and I'll say it plainly — a concentrated position in genuinely excellent businesses is not the same animal as a concentrated position in speculation.

However, the gap between what they cost and what they produce is the number that matters. The top ten hold roughly 41% of the index's weight but are expected to generate only about 32% of its earnings. The market is paying those ten companies a premium for everything else, and that premium is what has been doing the work: since 2023, the cap-weighted S&P 500 has beaten the equal-weighted version by roughly 32% — one of the largest divergences on record. (For context, from 2003 through 2022, the equal-weighted index ran ahead by about 1.5% a year.)

Which brings us to the "VOO looks cheap" narrative that has been doing the rounds lately. The index's forward P/E slipped below 20 times in late July, a level some screeners show near 19 now — historically unexpensive. The fine print: that denominator assumes 2026 earnings of roughly $323 per share, implying about 34% earnings growth, a figure one prominent paper called implausibly high. The cheapness is a function of the same AI story the concentration is built on. You can't pocket the low multiple and treat the story as though it isn't inside the fund.

The bottom line for your portfolio

VOO remains a fine default core — nothing else gives you five hundred large U.S. companies at 0.03%. The error is believing the diversification inside it is broader than it is. If VOO is your entire stock sleeve, then your stock sleeve is about 40% ten companies on one correlated trade, and you should hold that position knowingly, not accidentally.

If you want to dial the bet down, the options are mechanical, not mystical. The equal-weight version of the index — RSP — spreads every stock evenly, and by the yardstick I trust most, cash returned to shareholders, it actually pays you more for the privilege: a roughly 1.5% dividend yield versus VOO's 1.05%. Or keep VOO as the core and add a sleeve that doesn't move with the AI story — value, international, gold, or plain cash.

What would change my view here: as long as the top names' cash flow keeps proving out, this is a concentration problem in excellent companies, and the index earns its premium. If free cash flow at the top decays, or the weight-versus-earnings gap — 41% of the price, 32% of the profit — keeps widening, the "quality leadership" defense fails and what you hold becomes a directional AI bet wearing a 500-company label. That is the moment the 500 stops being just misleading and starts being a problem. Until then, in my opinion, the fund is fine and the label is the risk — and the fix is knowing exactly what you're holding.