The AI Money Carousel

Everybody is talking about AI. Your friends mention it constantly at dinner. You go to the barber and it comes up. A few weeks ago I had an MRI and the technicians were discussing how AI is improving diagnostic systems. Politicians who have no idea what it means are regulating it in parliament.

I am a boring investor by conviction. Put your money in the S&P 500, keep adding, don't touch it. I've preached this for years and I believe it.

So I didn't want to think too hard about the AI bubble. Why should I? I'm invested in the top 500 companies in the world.

I stopped thinking that way when I looked more carefully at what I actually own.

Thirty-three percent of the S&P 500 sits in seven companies. All of them deep in AI. Which means the index I've been calling "diversified" for years is, in a meaningful sense, a concentrated AI bet. Not entirely. But more than most people realise — and certainly more than the word "diversified" implies.

I don't dispute the index. A hundred years of 8–10% average annual returns is not an accident. I'm not here to tell you to sell.

But I've sat across from enough clients to know what happens when a portfolio drops 40% in a year. I know the theory — keep investing, don't panic, time in the market beats timing the market. I believe that theory. And I've watched perfectly rational people abandon it the moment it costs them real money.

A drop on the scale of 2000–2002, or 2008, concentrated in a handful of interconnected companies — that's not just a number. That's a test most investors don't know they're about to fail.

This article is not an argument against the S&P 500. It's not an argument against AI. It's an attempt to explain what's underneath the surface of something millions of people own without fully understanding. If you know the risks and you're comfortable — fine. But if you don't know, that's a different conversation.

The most successful companies in the world all have one thing in common

Look at the top of the S&P 500 today. The Magnificent 7 — Microsoft, Apple, Nvidia, Alphabet, Amazon, Meta and Tesla — account for over 35% of the entire index. If you own an S&P 500 tracker, more than a third of your money is riding on seven companies. All of them are betting heavily on AI. All of them are doing extraordinarily well — for now.

But look closer. Look at what these companies are actually doing with their money, and a strange picture begins to emerge.

Microsoft invests billions into OpenAI. OpenAI runs on Microsoft's Azure cloud — so Microsoft profits from its own investment. Nvidia sells the chips that power OpenAI's models. OpenAI has now taken a stake in AMD, a Nvidia competitor — while Nvidia itself has pledged $100 billion back into OpenAI. Oracle's entire future revenue growth is tied to a $300 billion cloud contract with OpenAI. Amazon, Google and Meta are all building data centres at a scale never seen before, buying chips from the same small group of suppliers who depend on these same giants as customers.

The money is moving. But it is largely moving in a circle.

What is circular investing — and why does it matter?

Think of it this way. Imagine five friends who agree to buy dinner for each other every night. On paper, they each spend and earn the same amount. They look busy, they look prosperous — but no new wealth is actually being created. The moment one of them loses their job and can no longer pay, the whole arrangement collapses.

Where is the real customer?

The revenue gap nobody talks about

The numbers tell a sobering story. AI companies will need $2 trillion in annual revenue by 2030 to justify the infrastructure they are currently building — and at today's projections, they are short by $800 billion. OpenAI, the company at the centre of all of this, is expected to lose $5 billion this year alone — despite commitments worth many times its projected revenue.

The combined free cash flow of Amazon, Google, Meta and Microsoft is projected to shrink by 43% between 2024 and early 2026 — not because the businesses are failing, but because capital expenditure on AI infrastructure is consuming it faster than it can be replenished. Goldman Sachs expects Big Tech to spend over $1 trillion on chips and data centres over the next five years. Where does the revenue to justify this come from?

Have we been here before?

The comparison to the dot-com bubble is imperfect — but instructive. And I want to be transparent about why I keep returning to history. I studied it formally — international relations and history — and it shaped how I see almost everything, including markets. The patterns of human behaviour around money are remarkably consistent across centuries. The euphoria, the denial, the rationalisation, the panic. When I look at the AI boom today I am not just running numbers. I am recognising a feeling. I have felt this feeling before. Most people currently invested in AI have not.

The bull case is real and I do not dismiss it. AI may well be the most significant technological revolution since the industrial revolution. But history does not ask whether the technology is real. It asks whether the price is right. And whether the money flowing through the system represents genuine new value — or the same dollars moving between the same hands, getting counted again and again.

What this means for your portfolio

I am not predicting a crash. Nobody can. But I am suggesting that the moment to think about this is before the music stops — not after.

I know what courage in markets actually feels like — not the bravado of riding a bull run, but the quiet conviction of buying when everyone else is selling. In the 2022 correction, when sentiment was at its worst and valuations had collapsed, I invested in Amazon, Nvidia and the S&P 500. Nvidia alone went up 800% from my entry point. Not because I am exceptional — but because lower valuations create better odds. That principle works in every market cycle, without exception. The inverse is equally true. High valuations driven by circular momentum rather than real earnings create poor odds. That is where we are today.

I speak to friends regularly who are buying every small dip — treating a 1% or 2% correction as a buying opportunity, adding more of the same concentrated positions. I understand the instinct completely. It has worked, consistently, for three years. But I find myself asking a quiet question: do they understand what they actually own? Do they know that more than a third of their S&P 500 tracker is riding on seven companies whose balance sheets are increasingly tied to each other? Are they investing with conviction — or following a trend that has not yet ended? There is a difference. And that difference matters enormously when the trend eventually does.

The AI revolution is real. The bubble may also be real. Both things can be true at the same time.

The question worth asking yourself today is a simple one: if the circle breaks, where are you standing?