For thirty years, the index fund was sold as the common-sense choice: no bet, no expert to pay, just the whole market in a single product. The promise held as long as the index was a basket. It wavers the day the basket comes down to a few pieces of fruit. Today, the saver who believes they are diversified is in fact heavily exposed to a handful of companies, all hanging on the same story: artificial intelligence.
1 A cushion that no longer absorbs
The product reputed to be the most prudent has become, without warning, one of the most concentrated.
The original promise
Owning "the whole market" in one fund
The idea behind the index fund is simple and sound: rather than betting on a few stocks, you buy the entire index, pay tiny fees and capture the average return of the markets. For decades, it has been the safest route to long-term saving, the path of retirement plans. Its principle, market-capitalisation weighting, looks harmless: each company weighs in the index in proportion to its stock-market value.
Top 7 stocks
≈ 34%
of the S&P 500 rests on the shoulders of the "Magnificent Seven" in early June 2026.
Top 10 stocks
≈ 41%
a record level: ten stocks out of five hundred now decide where the index goes.
The silent reversal
Nothing has changed in the fund's machinery; everything has changed in its content. The same product that once spread risk across hundreds of companies now concentrates it on a few. Every dollar invested in a fund tracking the S&P 500 sends close to a third of its value to just seven companies. The advertised prudence hides a bet that has become very tight.
2 A century-old record
To gauge the anomaly, you have to look far back: before the euro, before television, before the war.
Beyond 1932 and 2000
The highest concentration in nearly a century
According to CRSP data, the top ten stocks reached 37.7% of the US market at the end of October 2025, surpassing the previous peak of May 1932, in the depths of the Great Depression, when they topped out at 37.3%. The bar has since been pushed higher still. The other great precedent, the internet bubble of 2000, is also behind us. Two dates that evoke nothing reassuring.
Previous peak
May 1932
The only moment, in nearly a century, when concentration came close to today's.
Shiller CAPE
41.5
In May 2026: a valuation level never exceeded, except in March 2000 (44.2), just before the crash.
The cautious lesson
A record is not a prophecy. Concentration had already passed its 2000 levels by the end of 2020; anyone who fled then would have missed several remarkable years of gains. A historic peak does not say when the top will be reached: it only says the safety net is thinner than it looks.
3 The autopilot
To understand concentration, look not at the companies but at the rule that buys them.
A blind rule
The index fund has no opinion
A fund that tracks the index buys without judging: it allocates each dollar in proportion to each company's market value. The more a stock rises, the more it weighs, so the more the fund buys of it on the next inflow, which pushes it up further. It is a self-reinforcing loop. The rise manufactures the rise, regardless of earnings, by the sheer logic of the flow.
How the loop closes
①
Money follows size. Capital entering an index fund goes first to the largest caps: they carry the most weight in the formula.
②
Size attracts money. The more passive capital a stock attracts, the bigger it grows; the bigger it grows, the more capital it attracts. Cause and effect blur into one.
③
Judgement disappears. None of these decisions rest on the company's real value. The autopilot buys what is already expensive, expensively, by construction.
The illusion of diversification
Owning five hundred lines gives a feeling of protection. But if a third of the fund depends on the same seven stories, and those seven stories are in reality the same bet on AI, the diversification is apparent, not real. You do not hold five hundred risks: you hold a few, repeated five hundred times at unequal weights.
4 "Too big to fail", stock-market edition
Concentration threatens more than portfolios: it also reshapes the real economy.
The banking analogy
Giants turned into infrastructure
After 2008, some banks were called "too big to fail": their collapse would have dragged down the whole system. Today's seven giants resemble that, but on the scale of the entire economy. Cloud, online advertising, payments, chips and AI models all rest on a handful of players. These are no longer just share prices: they are mandatory waypoints of digital life.
Weight of a single stock
≈ 22%
Nvidia alone accounts for nearly a fifth of the seven giants' combined market cap.
Market cap reached
$5tn
The level touched by the largest of them in late 2025: a value greater than the GDP of most countries.
Stock-market concentration thus compounds an economic concentration. When a company becomes both the heaviest in the index and an unavoidable supplier to its own rivals, the risk ceases to be merely financial. An incident on one of these infrastructures, technical, regulatory or strategic, does not just lower a share price: it shakes a dependency shared by the whole economy.
5 The domino effect
The very mechanism that amplifies the rise can, reversed, amplify the fall.
When the loop runs backwards
The autopilot sells without judging too
The day the seven giants stumble, the small saver's supposedly prudent holding falls along with everything else, since it is made up largely of them. If the drop pushes some to pull their money out, the funds must sell, and sell the heaviest stocks first, that is, those same seven names. Selling drives prices down, the fall prompts more selling: the loop of the rise runs in reverse.
The honesty of the analogy
①
What brings it close to a bank. Like a systemic institution, the mechanism amplifies shocks procyclically: it buys when prices rise, it sells when they fall.
②
What sets it apart. An index fund has neither debt nor a "run on the counter" like a bank: the risk is not insolvency, but amplification and the erosion of price formation.
③
The real danger. When a growing share of the market is held passively, no one debates the fair price any more: they follow. The market gradually loses its function of judgement.
The scale of the flows
These movements are not theoretical. During index rebalancing, the entry or exit of a major stock can force tens of billions of dollars of mechanical buying or selling in the mega-caps, spread over months. The autopilot moves mountains: as long as it climbs, we cheer; the day it descends, we discover it has no brake.
6 The imperfect fix
Faced with the problem, one solution dominates the 2026 conversation: equal weighting. It too has a flip side.
Giving everyone the same weight
The equal-weight index, a reversed discipline
Rather than weighting each company by its size, the equal-weight index gives them all the same weight. At regular intervals, it trims what has risen and tops up what has fallen: a "sell high, buy low" discipline that the classic index, by construction, cannot replicate. That asset managers openly push it for 2026 speaks volumes: the idea of a protective index is now being questioned from within.
But there is no free exit. Equalising the weights means betting structurally against the largest companies: in the years when the giants lead the dance, like most of the recent ones, equal weighting lags behind. It reduces concentration risk at the cost of another bet, against the prevailing current. It is not insurance, it is a choice: trading a risk you know for another, which you accept.
The real message
The point is not to crown the "right" index, but to grasp that none is neutral. The cap-weighted one bets on the leaders' continuity; the equal-weighted one bets on their exhaustion. Choosing an index is already taking a position. The only mistake would be to believe you are taking none.
7 The social epilogue
One last loop remains, slower, reaching beyond the stock market.
The chase for margins
Satisfying a shareholder gone passive
When a growing share of capital is held by funds that do not judge, but demand returns, the pressure on margins becomes permanent and anonymous. To sustain it, the giants automate, replace, compress costs. In the short term, profits swell and the index with them. But these profits ultimately rest on revenues, hence on jobs, hence on the consumption that feeds them.
Here the snake bites its own tail: automating to boost margins can, at scale, dry up the very revenues that keep demand turning. We devoted an entire analysis to this contradiction lodged at the heart of algorithmic capitalism. Stock-market concentration is its financial mirror: the same companies that dominate the index are the ones leading that transformation.
Read alongside
The demand paradox in the age of AI →
"The robots produce, but who will buy?" Automating work boosts margins in the short term but dries up the income that keeps consumption turning. The social epilogue to the phenomenon analysed here.
The compass
For the saver, a useful truth: "passive" does not mean "risk-free". The index fund remains an excellent tool, provided you know what it really holds and do not mistake the number of lines for the diversity of risks. For the citizen, a broader question: an economy where a few players concentrate at once the stock-market value, the digital infrastructure and the power to automate has never existed at this scale. Learning to watch it is already a way of guarding against it.
Sources: Morningstar · The Motley Fool · MacroMicro · Pensions & Investments · Tema ETFs
One last loop remains, slower, reaching beyond the stock market.
Here the snake bites its own tail: automating to boost margins can, at scale, dry up the very revenues that keep demand turning. We devoted an entire analysis to this contradiction lodged at the heart of algorithmic capitalism. Stock-market concentration is its financial mirror: the same companies that dominate the index are the ones leading that transformation.
"The robots produce, but who will buy?" Automating work boosts margins in the short term but dries up the income that keeps consumption turning. The social epilogue to the phenomenon analysed here.
The same fragility seen from the financing side: when AI-linked credit risk migrates toward long-term savings. A single story underpins both the index and the chain of financing.