When the index isn't what you think it is
12th June 2026
There is a version of investing that feels safe because it sounds simple. Buy the index, track the market, diversify across hundreds of companies. For decades, that logic held reasonably well. The S&P 500 was a genuine cross-section of the American economy: energy, manufacturing, finance, retail, healthcare, technology. A rough reflection of how the world actually worked.
That is no longer quite true.
In May 2026, the S&P 500 hit a series of all-time highs. Headlines were positive. Markets were up. For anyone glancing at their portfolio statement, things looked fine. But underneath the headline number, something worth paying attention to was happening. According to data cited by Bank of America strategist Michael Hartnett, as the index reached those record levels, only around 5% of individual S&P 500 stocks were simultaneously hitting new 52-week highs. The last times that pattern appeared at a market peak were 1929, 1973, and 2000.
That tells you something about the difference between what the index says and what the index is actually doing.
That is no longer quite true.
In May 2026, the S&P 500 hit a series of all-time highs. Headlines were positive. Markets were up. For anyone glancing at their portfolio statement, things looked fine. But underneath the headline number, something worth paying attention to was happening. According to data cited by Bank of America strategist Michael Hartnett, as the index reached those record levels, only around 5% of individual S&P 500 stocks were simultaneously hitting new 52-week highs. The last times that pattern appeared at a market peak were 1929, 1973, and 2000.
That tells you something about the difference between what the index says and what the index is actually doing.
The index and the rally are not the same thing
The S&P 500's gains in 2026 have been driven by a very small number of companies, almost all of them connected to artificial intelligence infrastructure spending. The Philadelphia Semiconductor Index (SOX) rose roughly 95% in just 65 days between late March and early June 2026, according to Bank of America research cited in the Argonaut Capital Partners May 2026 commentary. At its peak, the SOX was trading approximately 76% above its 200-day moving average. To put that in context, the equivalent deviation at the peak of the dot-com bubble in 2000 was 55%. At the height of the Mississippi Company bubble in 1720, it was 73%.
Information technology now represents over 31% of the S&P 500 by market capitalisation. Energy, the sector the world currently needs to actually power those AI data centres, sits at roughly 3% to 4%. That gap between what markets are pricing and what the physical economy requires is one of the more striking disconnects in recent financial history.
Goldman Sachs' market breadth measure, updated in May 2026, was the weakest it had been since late 2023, even as the index itself was setting records.
None of this means the market is about to collapse. It means the market is narrower than it looks.
Why this matters for long-term investors
Most people investing through a global index fund assume they are broadly diversified. They are, to a degree. But if a significant portion of that index is concentrated in a handful of technology and semiconductor companies at historically elevated valuations, then the risk profile of the portfolio is not what the label suggests.
This is not a new observation. We wrote about S&P 500 concentration recently, noting that the top ten stocks in the index control around 40% of its total weight. The 2026 AI rally has taken that concentration further and faster than most analysts expected.
The Argonaut Capital Partners commentary from May 2026 draws a direct comparison to the final phase of the dot-com boom, when capital was being pulled out of the broader market to fund a concentrated surge in a single theme. Narrow market breadth of this kind has, historically, been prone to sharp reversals. That does not mean the underlying technology is without merit. It means the prices paid for that technology may already reflect, and in some cases exceed, the most optimistic possible outcomes.
For a long-term investor, particularly one managing a portfolio across currencies, jurisdictions, and decades, the question is not whether AI will be transformative. It almost certainly will be. The question is whether that transformation is already fully priced at current valuations, and what happens to your portfolio if the answer turns out to be yes.
What a genuinely diversified portfolio looks like
Diversification is frequently misunderstood. Holding 500 stocks is not diversification if 40% of the value sits in ten of them. Real diversification means spreading risk across sectors, geographies, asset classes, and time horizons in a way that does not depend on any single theme continuing to perform.
That means global equities, not just US equities. It means some exposure to sectors currently unloved by the market, which historically is often where future returns are found. It means being thoughtful about currency risk, particularly relevant for British expats holding sterling-denominated assets while living and spending in another currency. And it means having a clear plan that does not require you to predict which sector wins the next decade.
The question worth asking of any portfolio is simple: if the AI trade corrects sharply, what happens? If the honest answer is "a lot," that is worth addressing before it becomes relevant.
The S&P 500's gains in 2026 have been driven by a very small number of companies, almost all of them connected to artificial intelligence infrastructure spending. The Philadelphia Semiconductor Index (SOX) rose roughly 95% in just 65 days between late March and early June 2026, according to Bank of America research cited in the Argonaut Capital Partners May 2026 commentary. At its peak, the SOX was trading approximately 76% above its 200-day moving average. To put that in context, the equivalent deviation at the peak of the dot-com bubble in 2000 was 55%. At the height of the Mississippi Company bubble in 1720, it was 73%.
Information technology now represents over 31% of the S&P 500 by market capitalisation. Energy, the sector the world currently needs to actually power those AI data centres, sits at roughly 3% to 4%. That gap between what markets are pricing and what the physical economy requires is one of the more striking disconnects in recent financial history.
Goldman Sachs' market breadth measure, updated in May 2026, was the weakest it had been since late 2023, even as the index itself was setting records.
None of this means the market is about to collapse. It means the market is narrower than it looks.
Why this matters for long-term investors
Most people investing through a global index fund assume they are broadly diversified. They are, to a degree. But if a significant portion of that index is concentrated in a handful of technology and semiconductor companies at historically elevated valuations, then the risk profile of the portfolio is not what the label suggests.
This is not a new observation. We wrote about S&P 500 concentration recently, noting that the top ten stocks in the index control around 40% of its total weight. The 2026 AI rally has taken that concentration further and faster than most analysts expected.
The Argonaut Capital Partners commentary from May 2026 draws a direct comparison to the final phase of the dot-com boom, when capital was being pulled out of the broader market to fund a concentrated surge in a single theme. Narrow market breadth of this kind has, historically, been prone to sharp reversals. That does not mean the underlying technology is without merit. It means the prices paid for that technology may already reflect, and in some cases exceed, the most optimistic possible outcomes.
For a long-term investor, particularly one managing a portfolio across currencies, jurisdictions, and decades, the question is not whether AI will be transformative. It almost certainly will be. The question is whether that transformation is already fully priced at current valuations, and what happens to your portfolio if the answer turns out to be yes.
What a genuinely diversified portfolio looks like
Diversification is frequently misunderstood. Holding 500 stocks is not diversification if 40% of the value sits in ten of them. Real diversification means spreading risk across sectors, geographies, asset classes, and time horizons in a way that does not depend on any single theme continuing to perform.
That means global equities, not just US equities. It means some exposure to sectors currently unloved by the market, which historically is often where future returns are found. It means being thoughtful about currency risk, particularly relevant for British expats holding sterling-denominated assets while living and spending in another currency. And it means having a clear plan that does not require you to predict which sector wins the next decade.
The question worth asking of any portfolio is simple: if the AI trade corrects sharply, what happens? If the honest answer is "a lot," that is worth addressing before it becomes relevant.
A note on what we are not saying
We are not predicting a crash. We are not recommending selling technology stocks. We are pointing at data that professional fund managers are paying close attention to right now, and suggesting it is worth understanding.
Markets can remain irrational for longer than feels comfortable. The Argonaut commentary puts it well: as the final phase of the dot-com boom demonstrated, knowledge of an overextension does not tell you when it ends. What it does tell you is that concentration risk is real, that breadth matters, and that a portfolio built on a single narrative is more fragile than one built on a plan.
If your investments are heavily weighted toward US large-cap technology through index funds, it is worth at least knowing that. Whether to act on it depends on your own goals, timeline, and risk tolerance, which is exactly what a proper financial plan is designed to clarify.
If you want to talk through what your portfolio actually looks like beneath the surface, we offer a free, no obligation initial chat via the shiny button below..
We are not predicting a crash. We are not recommending selling technology stocks. We are pointing at data that professional fund managers are paying close attention to right now, and suggesting it is worth understanding.
Markets can remain irrational for longer than feels comfortable. The Argonaut commentary puts it well: as the final phase of the dot-com boom demonstrated, knowledge of an overextension does not tell you when it ends. What it does tell you is that concentration risk is real, that breadth matters, and that a portfolio built on a single narrative is more fragile than one built on a plan.
If your investments are heavily weighted toward US large-cap technology through index funds, it is worth at least knowing that. Whether to act on it depends on your own goals, timeline, and risk tolerance, which is exactly what a proper financial plan is designed to clarify.
If you want to talk through what your portfolio actually looks like beneath the surface, we offer a free, no obligation initial chat via the shiny button below..
