AI, Diversification, and How Your Plan Is Positioned
Given how much coverage artificial intelligence and its stock market impact have been getting, we wanted to set out what is actually happening, how it affects your plan, and where responsibility for managing it sits. The full picture is considerably more reassuring than the headlines suggest.
How concentrated the index has become
AI's influence on mainstream indices is unusual by historical standards:
• The ten largest companies in the S&P 500 now represent around 41% of the index's total market value, a level roughly 14 percentage points higher than the peak reached during the dot-com bubble in 2000.
• The “Magnificent Seven” (Apple, Microsoft, Alphabet, Amazon, Nvidia, Meta and Tesla) account for roughly a third of the S&P 500 on their own, and contributed more than half of the index's total gains from 2023 to 2025.
• Technology and technology-adjacent sectors as a whole now make up over 40% of US market capitalisation.
If you hold a global tracker fund, or your investment manager runs a core allocation benchmarked against the S&P 500 or MSCI World, this concentration flows into your portfolio whether or not AI was ever a deliberate choice. That is simply how market-capitalisation weighting works: the largest companies get the largest slice, and right now they are disproportionately tied to one investment narrative.
Why AI has driven, and may continue to drive, genuine growth
The reason AI-linked companies have grown to dominate the index is not simply speculation. There is a substantial body of earnings, spending and productivity evidence behind it.
• AI is already an earnings driver, not just a share-price story. The “Magnificent Seven” contributed roughly half of the S&P 500's earnings growth in 2025 and are forecast by Goldman Sachs to contribute around 46% in 2026. In the first quarter of 2026, the group's aggregate earnings grew more than 50% year on year, comfortably ahead of forecasts.
• AI infrastructure spending is now driving the wider economy. Combined capital spending by the largest hyperscalers (Microsoft, Amazon, Alphabet, Meta and Oracle) is on course to exceed $700 billion in 2026. In the first quarter, business investment overtook consumer spending as the main driver of US GDP growth, contributing around 1.5 percentage points to a 2.0% annualised growth rate.
• Demand for AI-related cloud services is growing at a rate rarely seen in businesses of this size. Google Cloud revenue grew 63% year on year in the most recent quarter, with its contracted order backlog nearly doubling to more than $460 billion. Amazon Web Services grew 24% to 28%, its fastest pace in more than four years.
• PwC's widely cited long-run modelling estimates that AI could add as much as $15.7 trillion to global GDP by 2030, a scale it compares to the historical impact of steam power, industrial robotics and the early spread of IT.
This is why we don't dismiss the AI theme as hype, and why your investment managers continue to hold meaningful exposure to it on your behalf. The spending is real and is already showing up in company earnings and economic growth. That said, strong demand is not a guarantee that today's valuations will be justified by tomorrow's earnings.
On valuations
• The S&P 500's forward price-to-earnings ratio has run in the low-to-mid 20s through much of 2025 and 2026, against a ten-year average of roughly 19.
• The cyclically adjusted Shiller P/E rose above 40 during 2025, a level exceeded only once in over a century of data, in the run-up to the dot-com peak in 2000, when it reached around 44.
• The Federal Reserve, the Bank of England and the International Monetary Fund have all, over the past year, described equity valuations, and AI-linked technology valuations specifically, as stretched or historically elevated.
• Markets have already tested this. A sharp sell-off in AI-related shares in June 2026 saw the Nasdaq fall around 2% in a single session and roughly 5% over the following days, with names such as Oracle and Micron losing considerably more. January and February 2026 saw similar bouts of volatility over whether AI capital spending will deliver the earnings growth currently priced in.
• In each case so far, markets have subsequently recovered and gone on to reach new highs.
We include that last point not to suggest volatility doesn't matter, but because elevated valuations, sharp corrections and subsequent recoveries have all been features of this cycle. What no one can tell you with any consistency is which correction will be a buying opportunity and which will mark a more prolonged downturn. Historically, returns following periods of high valuation have been more muted than average over the next one to two years. We factor that into return expectations, but it is not a reliable signal for when to reduce or exit exposure.
The past month in numbers
September showed both sides of this picture clearly: strong gains for chip manufacturers, but a narrow market underneath.
• Chip manufacturers led the market. The iShares Semiconductor ETF rose around 11% over the month, its strongest since June, with AMD and Intel each up roughly 29%. This follows a record second quarter in which the Philadelphia Semiconductor Index rose 81%, taking its gain for the first half of 2026 to around 94%.
• Those gains are being backed by results. Memory-chip maker Micron reported record annual revenue of $133 billion on 30 September, up 257% on the previous year, with fourth-quarter revenue almost five times the level of a year earlier.
• Wider AI-related performance was mixed. Within the “Magnificent Seven”, Meta rose around 26% and Nvidia around 3%, while Amazon fell 4%. Broadcom, another major AI chip designer, also slipped around 4%.
• Beneath the headlines, gains were narrow. The Nasdaq-100 rose 3.3% and large US technology companies 4.5%, yet the S&P 500 finished the month slightly lower (down 0.45%) and US smaller companies fell 5.3%.
For your plan, the takeaway is the same one this email makes throughout: returns are increasingly concentrated in a small group of companies, which is exactly why broad diversification matters.
Why the concentration itself is less concerning than it first appears
This is the part we most want you to take away. A large share of what gets labelled “AI exposure” in the index sits inside large, profitable, multi-line companies for which AI is one growth driver among several, not speculative AI ventures.
Amazon is a useful example. Only around 18% of its roughly $717 billion in revenue comes from AWS, the cloud division most directly tied to AI infrastructure demand. Online retail and third-party marketplace services still account for over 60% of revenue, and its advertising business alone now generates more revenue than PayPal and eBay combined.
Alphabet tells a similar story. In its most recent quarter, Google Search advertising generated $60.4 billion of the group's $109.9 billion in revenue, more than half the total, while Google Cloud contributed around $20 billion, roughly 18%, while growing 63% year on year. YouTube advertising and a growing subscriptions business add further diversified revenue.
So when your investment manager holds Amazon or Alphabet within a broad equity mandate, they are not making a pure bet on AI spending paying off. They hold diversified, cash-generative businesses that would continue to earn substantial profits from other divisions even if AI-related investment disappoints. That is a materially different risk profile from a single-product company whose valuation depends entirely on AI demand, and one reason broad equity mandates behave differently to concentrated thematic AI funds, even when both show meaningful AI exposure on paper.
What your investment managers are actually doing about this
As part of our ongoing due diligence, we discuss positioning like this directly with the managers we appoint on your behalf. In broad terms, they are:
• Rebalancing portfolios back within your agreed risk tolerances where drift occurs, rather than reacting to headlines.
• Continuing to hold significant weight in sectors with little direct AI exposure, such as healthcare, financials, industrials and defensives, which provide ballast if sentiment towards technology shifts.
We review this positioning with each manager regularly as part of the manager selection and monitoring work behind your plan.
Our role in all of this
Our part is to make sure this is happening in a way that's appropriate for your plan, your time horizon and your capacity for volatility, and to resist the temptation, on your behalf and ours, to time markets on this quarter's headlines. History gives little support to the idea that investors, professional or otherwise, can reliably call the top of a theme like this. It gives considerably more support to staying invested in a well-diversified, professionally managed portfolio through periods exactly like this one.
About Aled Phillips : Aled is a Fellow of the PFS, a Chartered Financial Planner, and a Chartered Fellow of the CISI.
This article is for general information purposes and reflects our understanding of markets at the time of writing. Market data is to 30 September 2026 unless otherwise stated. It does not constitute personal financial or investment advice, nor a recommendation to buy, sell or hold any individual security or fund. The value of investments can fall as well as rise, and you may get back less than you invested. Past performance is not a reliable indicator of future results.
