Is This an AI Bubble? Nobody Knows. Here Is the One Number You Can Check

Updated: Oct 26
Nobody can tell you whether this is a bubble. Not me, not the analyst on television, not the fund manager with the confident voice. Every bubble is obvious afterwards and invisible while you stand inside it. What you can know, today, in ten minutes, is how much of your own money is riding on the answer. For most index investors in September 2026, the figure was larger than they thought: about 41% of the S&P 500 sat in just ten companies.
That is the whole page. The rest is why prediction is not available, and why concentration is.
Why did everyone start asking in 2026?
Because the biggest companies of the AI era started going public. On 12 June 2026 SpaceX listed on the Nasdaq: the largest IPO in history, around $75 billion raised, a valuation near $2 trillion, and a first full trading day about 19% above the offer price. Roughly 30% of the shares went to retail investors, where a large IPO normally reserves single digits. Anthropic filed confidentially with the SEC on 1 June; OpenAI filed too, on a timetable that keeps moving.
The question arrived in my inbox in fifteen different outfits: yes, but is this an AI bubble?
Can anyone predict a bubble?
No. Hear that as a position, not a dodge. In 1999 the cleverest people in the room were not saying "this is a bubble." They were explaining why the old rules no longer applied.
And here is the part the commentary always skips. Every crash arrived from a direction almost nobody was watching. In 1987 the market fell more than 20% in one day with no news event to explain it. In 2008 the damage came not from the technology everyone was arguing about, but from mortgages, which most people considered boring and safe. In 2020, from a virus.
So when someone tells you confidently that this looks like a previous bubble, notice what they are doing. They are pattern-matching against the last crash, which is the one thing history reliably does not repeat. Nobody rang a bell in advance last time. Nobody will ring one this time either.
What is actually knowable?
Three numbers, as of September 2026, including one that argues against my own caution.
Concentration was at a record. The ten largest US stocks made up about 41% of the entire S&P 500. Not 41% of the technology sector. Of the whole index.
Prices were high by any long measure. The Shiller CAPE ratio, which compares prices to ten years of earnings, sat around 40.7, the second-highest reading in the entire record.
The companies make real money. Today's largest technology companies generated roughly $350 billion of free cash flow between them in their last financial year. Not slogans with a website. Some of the most profitable businesses ever built.
Expensive and profitable is a different animal from expensive and empty. It is still expensive.
How much of my index fund sits in ten companies?

You own the index. You have been told, correctly, that this makes you diversified. It does. It also means that of every €10,000 in an S&P 500 fund, roughly €4,100 rode on ten names in September 2026, most of them selling the same story about the same technology.
The Nasdaq-100 is narrower still: the hundred largest non-financial companies on the Nasdaq, technology-heavy by construction, its top ten around half the fund. Put the two top tens side by side (Nasdaq-100, S&P 500) and most of the names are the same, at different weights. A woman holding both funds has not spread her money twice. She has bought the same companies twice, at two different intensities.
And in 2026 the rules themselves moved. In May, Nasdaq introduced a fast entry rule: a company large enough to rank in the top forty now joins the Nasdaq-100 after about fifteen trading days, instead of three months. SpaceX joined on 7 July, three and a half weeks after listing. Every Nasdaq-100 fund had to buy it. Not because anyone judged it a good investment, but because the rule said so.
If you hold one of those funds, you bought shares in a rocket company that July. Nobody asked you. That is not a scandal, it is how index investing works. It is why the composition is worth knowing.

Nobody can tell you whether this is a bubble. Anybody can tell you how much of their money is riding on the answer. Prediction is not available. Concentration is.
What is an IPO, in plain words?
A company sells shares to the public for the first time, and the early owners finally have a market to sell into. So when a company debuts near a trillion, "getting in early" is not what is happening. The early part happened in a private room. You are offered the late seat, at a price that room found acceptable.
A note: this is education, not financial advice. I don't know your personal situation, so I can't advise you on it. Nothing here is a recommendation to buy, sell or hold any security, fund or IPO. For advice on your own situation, talk to an independent, licensed financial adviser.
What did I learn inside finance?
Twenty years inside finance, and I have sat through several of these moments. Every time, two things were true at once. Everybody knew somebody making easy money on the story of the day. And nobody wanted to be the careful one, because careful looked like slow.
I have been the careful one and I have been the tempted one. What I have never been, not once, is the one who knew.
Women do not have an earning problem. They have an owning problem. And "I will wait until the AI thing calms down" is that problem wearing a very reasonable coat.
What should you do this month?
Look at what you own. Open your investment app. Write down every fund and every share you hold, and how much money sits in each. Not what you think you own. What is actually there.
Then one question at your Money Date: if the biggest line fell by half, what actually changes in my life?
That question was never yours to solve at the level of the market. Yours is sitting in your app right now.
Q. Is the stock market in an AI bubble? A. Nobody knows, and anyone who says they do is guessing. As of September 2026 the facts were: record concentration (ten stocks about 41% of the S&P 500), a CAPE ratio near its all-time high, and very large real profits at the biggest companies. Expensive, and profitable.
Q. If I own an S&P 500 index fund, am I diversified? A. Across five hundred companies, yes. But about 41% of the money sat in ten of them in September 2026, and a Nasdaq-100 fund holds largely the same ten at higher weights. Owning both is not twice the spread. See also In Gold We Trust on where capital goes in a crash.
Q. Should I sell before the bubble bursts? A. That is timing, and timing does not work: see Should I Invest at an All-Time High?. The useful question is not "when will it fall" but "how much of what I own would fall together, and could I live with that."
About the author
Céline Pujdak is the Founder of Cello Wealth. She spent two decades inside one of Europe's largest banks, managing billions. Cello Wealth turns that experience into financial education for women who earn well but do not yet own enough.
Originally shared with Cello Notes readers on Sunday 13 September 2026. Get the Sunday letter: Cello Notes on Substack.
Read next: Should I Invest at an All-Time High?, In Gold We Trust, Do You Know What You Own Together?.
About Cello Wealth
Cello Notes is published by Cello Wealth, founded by Céline Pujdak. Cello Wealth helps women understand their money, build investment portfolios and take control of their long-term wealth.
Educational purposes only. Not financial advice. I don't know your personal situation and cannot advise on it. Nothing here is a recommendation to buy, sell or hold any security, fund or initial public offering, including any company named; companies and IPOs are discussed as public news and as illustrations of how markets work. Market figures are stated as of September 2026 and are not updated. The value of investments can fall as well as rise; past performance is not a reliable indicator of future results.
Sources:
CNBC: SpaceX's historic IPO, size, valuation and about 30% retail allocation, 13 June 2026
CNBC: Anthropic confidentially files IPO prospectus with the SEC, 1 June 2026
Benzinga: ten largest US stocks a record share of the S&P 500, 2026
GuruFocus: S&P 500 Shiller CAPE ratio, around 40.7, September 2026
Nasdaq-100 holdings and weights (QQQ), top ten concentration



Comments