How Your Index Fund Got So Tech-Heavy Without Changing a Thing

Open the fact sheet for a plain S&P 500 index fund and you will not find a line item called “technology bet.” You will find a sentence saying the fund seeks to track a broad benchmark of large U.S. companies, a single-digit expense ratio, and a list of top ten holdings. Nothing in that document announces a decision. And yet the largest positions in nearly every broad U.S. equity fund sold today are the same handful of mega-cap technology and platform companies — Apple, Microsoft, Nvidia, Alphabet, Amazon, Meta Platforms, and Tesla, the group markets have taken to calling the Magnificent Seven.

The useful thing to understand about index fund tech concentration is that no one chose it. It is an output of three mechanical rules stacked on top of each other: how index weights are calculated, how sectors are labelled, and how retirement portfolios are assembled from overlapping funds. Each rule is defensible on its own. Together they produce a portfolio whose risk profile is considerably narrower than the word “diversified” implies. This piece takes those three rules apart in order.

1️⃣ The first rule: weight follows market value, not conviction

A capitalisation-weighted index assigns each constituent a weight proportional to its float-adjusted market value — share price multiplied by shares available to public investors, divided by the index total. That formula is published in the index provider’s own methodology documents for the S&P 500, and close cousins of it govern the major global and total-market benchmarks. There is no committee deciding that one company deserves twenty times the weight of another. The multiplication does it.

That design has a property investors rarely state out loud: the index becomes more concentrated precisely when the largest companies perform best. If a stock doubles while the rest of the market is flat, its weight in the index rises, which means the fund that tracks the index now has more of your money exposed to it than it did before — and it got there without a single trade decision by anyone you could call and question. Concentration is not a failure of the mechanism. It is the mechanism working.

The scale of today’s version is what makes it a portfolio question rather than a trivia question. A MarketWatch report frames the combined market value of the seven as exceeding the entire economic output of almost every individual country in the world. Treat that as a secondary characterisation rather than a filing-derived figure — it is a magnitude comparison from a news summary, not an audited number, and market values move daily. The interpretive point survives the imprecision: a group this large relative to the index cannot be a rounding error in a cap-weighted fund, because weight is value.

The same report pushes the argument one step further than most concentration commentary does, arguing the group is now large enough to move the economy itself — and flagging that as a problem rather than a milestone. That reframing matters for a fund holder, because it breaks the usual mental separation between “my portfolio risk” and “the macro backdrop.” If seven companies’ capital spending and hiring are big enough to show up in national aggregates, then the diversifying effect you were counting on — your equities doing badly while the broader economy holds up, or vice versa — gets thinner. Note the authority limit again: this is a journalistic characterisation of scale, not a measured elasticity from an official statistical agency.

This is also why “I own the whole market, so I’m diversified” and “I have meaningful single-name risk” are not contradictory statements. Owning all 500 names tells you how many positions you hold. It tells you nothing about how your dollars are distributed across them.

Sector labels and why the technology column understates the problem

The second rule is classification. Investors commonly sanity-check concentration by reading the sector breakdown on a fund page and comparing the Information Technology line against a comfort threshold. That check is weaker than it looks, because the Global Industry Classification Standard — the taxonomy most fund sector tables are built on — does not sort companies by whether their business is technological. It sorts them by the nature of their revenue.

Under that taxonomy, the Magnificent Seven are distributed across three different sectors:

The split is not arbitrary. Alphabet and Meta were moved into Communication Services in the 2018 GICS restructuring because the overwhelming majority of their revenue is advertising, and an advertising business genuinely does behave differently from an enterprise software business. Amazon’s classification follows its retail sales. Tesla’s follows the fact that it manufactures and sells cars.

But a sector table answers “what does this company sell?” when the question a concentration-conscious investor is actually asking is “what happens to these positions at the same time?” Those are different questions, and in the past several years they have produced different answers. When a shift in expectations about artificial-intelligence capital spending or long-term interest rates moves mega-cap growth equities as a bloc, the move does not respect sector boundaries. The consequence is concrete: an investor who reads a moderate Information Technology weight and concludes the portfolio is balanced has measured the smaller part of the exposure. Two of the seven sit in a column most people never associate with technology risk at all, and two more sit in a column most people associate with shopping and cars.

No public classification system is designed to measure correlated drawdown risk, and it would be unfair to blame a revenue taxonomy for failing at a job it never claimed. The error is on the reading side — using a revenue taxonomy as a risk taxonomy.

Overlap: the same seven names, counted once per fund

The third rule is the one that does the most damage in ordinary retirement accounts, and it has nothing to do with index construction. It is additive.

A typical workplace plan menu offers a large-cap blend fund, a large-cap growth fund, a total stock market fund, sometimes an explicit technology or innovation sleeve, and a target-date fund that itself holds several of the others. An investor diversifying across that menu — three funds instead of one, which feels like risk reduction — is often buying the same mega-cap positions three times, because each of those products is capitalisation-weighted over an overlapping universe. Large-cap growth benchmarks tilt toward exactly the names whose weights have grown most. Total-market funds add thousands of small companies, but those companies enter at weights so small that they barely dilute the top of the portfolio.

Fund-level disclosure makes this checkable, though not instantly. Here is what each document actually reports and how current it is:

The practical implication of that table is narrow but real: a fund’s marketing page sector pie is not the primary record. The holdings file and the SEC filings are. If an investor wants to know their true aggregate exposure to any single company across several funds, the arithmetic requires pulling each fund’s weight for that name from the issuer’s own disclosure and multiplying by the dollars held in each fund. Tedious, unglamorous, and the only version of the number that is actually theirs.

⚠️ Where this explanation stops holding

Three honest limits, because a mechanism argument that admits none is a sales pitch.

• Concentration is not a forecast. Nothing in the arithmetic above says mega-cap technology companies will underperform. The mechanism describes the shape of the risk, not its direction. These are among the most profitable enterprises in corporate history by their own reported results, and a concentrated portfolio of exceptional businesses can do well for a long time.
• The alternatives carry their own concentration. Equal-weighted indices reduce single-name dominance but mechanically increase exposure to mid-sized companies and require regular rebalancing, which has turnover and cost consequences. Capping schemes introduce rules about when a weight gets trimmed — a different judgement, not the absence of one. International diversification runs into the fact that the same U.S. mega-caps carry substantial weight in global developed-market benchmarks too.
• Exact weights are a moving target. Any specific percentage quoted in an article like this one is stale by the time it is read. That is a reason to go to the issuer’s current holdings file rather than to trust a number in prose — including the numbers in prose anywhere on a finance site.

What would change the answer

The mechanism stops being decisive at the point where correlation breaks. Cap-weighting concentrates dollars in the largest companies; the risk that matters is whether those companies fall together. If the seven begin to diverge meaningfully from one another — if an advertising-revenue shock, an AI capital-spending reassessment, or a hardware demand cycle starts hitting some of them and not others — then their combined index weight becomes much less alarming, because the bloc stops behaving like a single position. The group’s collective momentum, not its size, is what converts arithmetic into fragility.

So the observable question is not how heavy the top of the index is. It is whether these companies’ earnings and share prices are still moving in step. That is answerable from each company’s quarterly SEC filings and from each fund’s own holdings disclosure, which is a less exciting place to look than a market commentary — and the only one where the numbers are the issuer’s rather than someone’s characterisation of them.