Your Diversified Etf Isn't As Diversified As You Think — And Here's The $700 Billion Reason Why

Editor's note: This is the first article in a four-part series on AI concentration risk in growth portfolios. The next articles examine the supply chain behind that concentration, the risks facing that supply chain and the corporate adoption timeline that will ultimately determine which companies in it earn their valuations.

You own an ETF with hundreds of holdings. That number is not a measure of how diversified you actually are.

Diversification protects you when your money is spread across independent economic outcomes. It does nothing for you when it is spread across the same outcome wearing different tickers.

Look inside a growth ETF today and that is exactly what you will find: Dozens of companies whose fortunes trace back to a single assumption — that the largest technology companies in America will keep spending hundreds of billions of dollars building AI infrastructure.

What 'diversified' actually means

Take the Schwab U.S. Large-Cap Growth ETF (SCHG). It currently holds 197 stocks. On paper, that looks like broad exposure. In practice, its top 10 holdings account for around 51% of the fund, and technology alone makes up almost 50% of the portfolio.

At the time of writing, Nvidia (NVDA) is 10.83% of the fund by itself. Apple (APPL) is another 8.93%. Microsoft (MSFT) is 7.31%. Add Amazon (AMZN) and Alphabet (GOOGL) and five companies account for roughly 35% of everything you own in that single ETF.

Invesco's QQQ tells the same story with a different label. At the time of writing, its top 10 holdings run 45% of the fund, and technology alone makes up more than 50% of the total.

Nvidia is nearly 9% of the fund by itself. You did not buy the Nasdaq-100. You bought a concentrated bet on a handful of companies whose revenue increasingly depends on the same capital spending cycle.

The dependency investors don't see

Here is where it gets more concentrated than the ticker count suggests. Microsoft, Amazon, Alphabet and Meta (META) are projected to spend somewhere between $700 billion and $725 billion combined on capital expenditures in 2026, an increase of roughly 60% to 77% over 2025.

The overwhelming majority of that spending funds AI data centers, chips and networking equipment.

That single number sits underneath nearly every AI-adjacent stock in your portfolio. Nvidia's revenue depends on it. Micron's (MU) revenue depends on it. Vertiv (VRT) and Equinix (EQIX) depend on it. Arista Networks (ANET) depends on it.

You may hold 10, 20 or 30 companies spread across a growth ETF, an S&P 500 fund and a semiconductor fund, and a meaningful share of every one of them is ultimately underwritten by the same four checkbooks.

One company's revenue is another's expense

This is the mechanic most investors miss. Microsoft's cloud revenue depends partly on selling AI capacity it built using chips bought from Nvidia. Nvidia's revenue depends on Microsoft, Amazon, Alphabet and Meta continuing to buy those chips at the current pace.

Arista Networks and Micron sell into the same data centers. Vertiv and Equinix build and operate them.

None of these companies are competitors with each other. They are counterparties in the same transaction, repeated at scale. When you own several of them at once, you are not diversifying your exposure to AI. You are stacking your exposure to whether four companies keep spending at a pace none of them has ever sustained before.

Why this concentration keeps growing without you doing anything

Here is the part that catches most investors off guard. You do not need to buy more of these companies for your concentration to increase. It happens automatically, every day the market is open, as long as the stocks keep rising.

Most growth ETFs are weighted by market capitalization, which means the biggest companies automatically claim the biggest share of your money. When Nvidia's stock price doubles, it does not just hand you a stronger return. It also earns a larger slice of every new dollar that flows into the fund afterward, which pulls in even more of the next dollar behind it.

Your exposure compounds on top of itself, with no rebalancing decision required and no trade of your own.

This is not a flaw in how these funds are built. Cap weighting has tracked the real economy reasonably well for most of market history.

The problem is that the mechanism cannot tell a company earning a bigger share of the index because it is genuinely capturing more of the economy from a company earning a bigger share because enthusiasm has pushed its price ahead of its fundamentals.

The fund sees only price, and it keeps shifting more of your money toward the answer regardless of which one is true.

A fund that looked reasonably diversified a year ago can look considerably more concentrated today, without a single decision in between.

Compare your fund's current top 10 weighting against an old statement if you kept one. If that number climbed meaningfully, the climb happened to you, not because of anything you chose.

Why this concentration is easy to miss

A traditional concentration check looks at how much of your portfolio sits in any single stock. That check will tell you SCHG is fine, because no individual company crosses 11% of the fund.

It will not tell you that Nvidia, Microsoft, Amazon, Alphabet, Broadcom (AVGO) and Micron collectively represent one economic bet dressed up as six separate ones.

Fund overlap tools miss it for the same reason. They compare ticker lists across funds. They do not ask whether the companies on those lists share a common revenue driver.

A tool can tell you that your growth ETF and your S&P 500 index fund both hold Nvidia. It will not tell you that they also both depend, through different tickers, on the same $700 billion capital spending assumption.

What to do with this

Pull the fact sheet for every fund you own and look past the top 10 list. Ask which companies derive a meaningful share of revenue from AI infrastructure spending, either directly or as a supplier into it. Add that exposure across every fund you hold, not just the one marketed as a technology fund.

You do not need to sell anything to benefit from this exercise. You need an honest number for how much of your total portfolio depends on four companies continuing to spend roughly $700 billion a year on a bet that has not yet fully proven out.

Once you have that number, decide for yourself whether it matches the risk you believe you are taking.

Count the dependencies, not the tickers. That is where your real concentration risk is hiding, and it will not show up on a standard diversification report.

Related Content

• Managing a Concentrated Stock Position: Too Much of a Good Thing
• Worried About an AI Bubble? 5 Ways to Ensure Your Portfolio is Prepared — Whether It Bursts or Not
• There's a New AI Bubble No One Is Talking About: What Companies Can Do to Keep Up
• Using Google AI Tools Can Give Your Advisory Firm the Edge — If You Do These 5 Things First
• How Businesses (and Advisers) Can Budget for AI Use When the Bill Keeps Changing

This article was written by and presents the views of our contributing adviser, not the Kiplinger editorial staff. You can check adviser records with the SEC or with FINRA.