S&P 500 concentration in AI stocks: how much of the index is the AI bubble?

What does S&P 500 concentration mean for the AI bubble?

Concentration describes how unevenly an index is spread across its members. The S&P 500 holds about 500 US companies, yet the largest few can carry a weight far above their share of the count. When the largest few are AI leaders, the index becomes a vehicle for the AI theme, whatever the name of the fund says.

This matters for the AI bubble because a bubble inside a concentrated index reaches every holder. A saver who owns the index through a pension or an exchange traded fund holds the AI theme without choosing it. Measuring concentration shows how much of the theme you hold and how much the other companies in the index offset it.

How does index weighting raise AI concentration in the S&P 500?

The S&P 500 weights its members by market value, the share price multiplied by the number of shares. A company whose price rises 2x sees its weight double, and the index buys no extra shares to do so: the weight simply grows. Rising prices therefore push the largest companies higher in the ranking, and concentration climbs when a theme leads.

Equal weighting offers a useful comparison. An equal weight version of the index gives each company the same share, so the AI leaders carry a small weight. Setting the 2 versions side by side shows how much of the return comes from the leaders. A wide difference between them signals a narrow market.

Which four measures show S&P 500 concentration in AI stocks?

The first measure is the weight of the top 10 companies. The second is their share of total index profit, which tells you whether the weight is earned. The third is the effective number of stocks, a figure that converts the weights into the count of equal companies that would give the same spread. The fourth is the valuation gap between the leaders and the rest.

Used together, the 4 measures give a fuller picture than any single one. A high weight with a matching share of profit points to concentration that earnings support. A high weight with a lower share of profit points to concentration that rests on expectations. Record the 4 figures with a date, and compare them quarter after quarter.

How do you count AI companies to measure S&P 500 concentration?

The sector labels of an index do not isolate AI, so you build the list yourself. Use 3 groups: companies that sell chips and equipment, companies that sell cloud capacity and models, and companies that sell software or services with AI features. Add a fourth group of power and data centre suppliers if you want the physical build-out. Write the rule down so that the count stays comparable.

A consistent rule matters more than a perfect one. Two readers with different lists will reach different weights, yet each can track a trend if the list stays fixed. Share your list with a date when you share a number. That habit lets another reader recompute your result and trace any gap to a specific company.

Does S&P 500 concentration in AI stocks match their share of profits?

Compare the weight of your AI list with its share of index profit. If the profit share is close to the weight, investors pay a fair price for earnings that exist. If the profit share is far lower, investors pay for earnings they expect. The EQOS Asia AI study records about $200B a year invested in AI while 95% of projects fail in production (EQOS Asia AI study, April 2026).

Spending by AI leaders becomes revenue for other AI leaders, so profit in the list can be partly circular. Check whether the revenue of chip and cloud companies comes from end customers or from other companies in the same list. The share that comes from end customers is the part that outlasts a pause in spending.

What do past index peaks say about S&P 500 concentration and bubbles?

Around the dot-com peak of 2000, technology companies held a large share of US indexes, and concentration fell after prices dropped. High concentration has not always ended in a crash, since leaders can grow into their weight. The mother guide, Is the stock market in a bubble because of AI stocks, places the AI case beside that period.

The lesson is to read concentration with a second signal. Concentration plus a widening valuation gap plus rising debt is a stronger warning than any one of them. Concentration with falling valuation and rising profit reads as leaders growing into their weight. Pair the weights with the funding data to avoid drawing a conclusion from one number.

How can EQOS Asia help you read S&P 500 concentration and the AI bubble?

The EQOS Asia AI study scores the enterprise AI market at φ(S) = 0.337, in the fragile zone where a shock can tip the system (EQOS Asia AI study, April 2026). The score adds the view of the buyers of AI, which index weights cannot show. Use the Λ simulator to estimate a score for your own organisation in 10 minutes.

For a portfolio with a large US index position in Singapore, ASEAN, or Hong Kong, the EQOS Asia pricing page lists the diagnostic that applies the scoring to your case. Your fund list goes in, and a score and a ranked set of options come out. That output gives your investment committee a shared number to discuss.

Frequently asked questions

About the author

Stéphane Jambu

President, EQOS Forecast Asia

Stéphane Jambu leads EQOS Forecast Asia and works as a consultant in e-commerce, search visibility, and AI. LinkedIn profile