You might think you have a diversified portfolio because you own several ETFs. But if those funds all hold the same AI-linked companies, you could be taking a much bigger bet on artificial intelligence than you realise.
That’s increasingly relevant as AI has become a significant part of global sharemarkets. A broad global shares ETF can give you exposure to the companies driving the AI boom without you ever specifically choosing an AI investment. Add an S&P 500 fund, Nasdaq ETF or semiconductor fund, and the overlap can quickly grow.
The result is a portfolio that can look diversified by fund while being much less diversified underneath. That’s not necessarily a problem, but if you own several ETFs, it’s worth checking how much of your portfolio ultimately rests on the same companies and the same AI story.
Here’s how to work it out.
Why concentration matters – and why AI?
Concentration isn’t necessarily a bad thing. If a company, sector or theme performs strongly, having more exposure to it can lift your returns.
AI is a particularly useful example right now because it has become such a significant driver of the sharemarket. Many of the companies at the centre of the AI boom are among the largest companies in major indexes, so investors can pick up sizeable exposure simply by owning broad-market funds.
That can work in your favour if the companies continue to perform strongly. But it also means a shift in expectations around AI could affect several parts of your portfolio at once. If the same companies appear across your global, US, technology and semiconductor ETFs, the impact can be greater than the number of funds you own might suggest.
There’s another way concentration can creep up: a company that performs strongly becomes a larger part of an index, while multiple ETFs continue to hold it.
That doesn’t mean you need to avoid AI exposure or eliminate every bit of overlap. It just means it’s worth knowing how much exposure you have, where it comes from and whether you’re comfortable with it.
Start with what you mean by “AI exposure”
There’s no single definition of an AI company. Some businesses build AI models and software. Others, like Nvidia, make the chips, provide cloud infrastructure or operate data centres. Then there are businesses using AI to improve existing products or cut costs.
Depending on how broadly you define it, AI exposure could include:
- AI models and software
- Semiconductor designers and manufacturers
- Cloud and data-centre providers
- Networking, memory and power infrastructure
- Cybersecurity and data businesses
- Companies incorporating AI into existing products
- Businesses expected to benefit from AI-driven productivity
Your answer will depend on where you draw the line. If every company using machine learning counts, it may be almost impossible to build a portfolio with little AI exposure. If you only count businesses generating meaningful revenue from AI products, the universe is much smaller. So decide what you’re measuring before you start calculating.
Four ETFs can still mean a lot of the same companies
Consider an investor who owns:
- A broad global-market ETF
- An S&P 500 ETF
- A Nasdaq-100 ETF
- A semiconductor ETF
Four funds sounds diversified. But the number of ETFs you own isn’t the same as the number of different risks you’re taking. Those funds can contain many of the same companies, increasing your exposure to particular businesses or themes without you realising it.
And sector breakdowns won’t necessarily reveal the overlap. A semiconductor manufacturer, data-centre operator and electricity provider might all benefit from increased AI infrastructure spending while sitting in different sectors.
The same goes for a broad global shares ETF. Because many broad-market indexes are weighted by market capitalisation, the world’s largest technology-related companies can make up a substantial part of the fund.
A thematic ETF can therefore increase concentration rather than diversification. That might be intentional, but it’s worth knowing when it happens.
How to work out your actual exposure
The basic calculation is: Portfolio exposure to a company = your portfolio weight in the ETF × the company’s weight within that ETF
For example:
- 50% of your portfolio is in Fund A
- Company X makes up 8% of Fund A
- 30% is in Fund B
- Company X makes up 10% of Fund B.
Your exposure to Company X would be:
- Fund A: 50% × 8% = 4%
- Fund B: 30% × 10% = 3%
- Total exposure: 7%
Simply adding the 8% and 10% weights would give you the wrong answer because it ignores how much of your portfolio is actually invested in each fund.
Repeat the calculation for companies appearing across multiple funds, then group them according to the AI definition you chose.
A simple way to audit your portfolio
You can do the exercise in a spreadsheet.
1. List everything you own
Include ETFs, managed funds and individual shares. If your superannuation forms part of the same long-term investment plan, consider whether it belongs in the audit too.
2. Work out your current portfolio weights
Use current market values rather than what you originally invested. A holding that has performed particularly well can become a much larger part of your portfolio than you intended.
3. Get the latest fund holdings
Check each ETF provider’s website and use its most recent holdings data. An old comparison article may no longer reflect what the fund owns.
4. Clean up company names
The same business can appear under different names or share classes. Consolidate these before calculating your exposure.
5. Calculate your look-through exposure
Multiply your portfolio allocation to each fund by the weight of each underlying company, then add contributions from different funds together.
Pearler investors can start with the holdings and current market values in their account, then use each ETF provider’s latest holdings file to do the rest. If you use Pearler Automate, it’s also worth checking whether your existing investment instructions are continuing to send new money towards an allocation that has already become larger than intended.
6. Categorise the exposure
One useful framework is:
| Category | What it could include |
| Direct AI | AI models, software or products are central to the business |
| AI infrastructure | Chips, cloud, networking, data centres and power |
| AI adopters | Existing businesses using AI to improve operations or products |
| AI-adjacent | Businesses whose demand or valuation is partly linked to AI spending |
There’s inevitably some personal judgement required. The aim isn’t to produce a perfectly precise “AI percentage”, but to make your assumptions visible.
7. Compare it with your investment plan
Ask:
- Would a major fall in AI-related valuations materially affect my portfolio?
- Is one company having an outsized influence on my returns?
- Does each ETF have a clear role?
- Would I still want these holdings if the AI narrative cooled?
- Has my exposure increased deliberately or simply because some investments performed well?
Look beyond the biggest holdings
Checking the top 10 holdings of each ETF is useful, but it won’t reveal every concentration risk.
Several relatively small holdings can still be exposed to the same underlying theme. Semiconductor companies, for example, might operate in different markets while responding to the same changes in AI infrastructure spending.
Likewise, data-centre companies and utilities might look diversified by company and sector while sharing exposure to electricity prices, interest rates and infrastructure investment.
When auditing your portfolio, look for common drivers, not just duplicate company names.
What can you do if you find more overlap?
Finding concentration doesn’t automatically mean you need to sell. Depending on your investment plan, you might:
- Keep the exposure
- Direct future contributions elsewhere
- Stop adding to an overweight allocation
- Gradually rebalance your portfolio towards your targets
- Simplify overlapping funds
- Change your target allocation
- Sell holdings after considering tax and transaction costs
Selling can trigger a capital gain or loss, while redirecting future contributions may take longer to change the portfolio but avoids an immediate sale.
If your concern is specifically about the social or environmental impact of AI, that’s a separate question. Knowing how much AI exposure you have is one thing; deciding what level fits your values is another.
The point isn’t to avoid AI but to know what you own
You can own several ETFs and still have a portfolio concentrated in the same companies, themes or economic drivers. That’s not necessarily a bad thing. But it’s much easier to make deliberate investment decisions when you can see what’s happening underneath the fund labels.
A look-through audit can show whether AI is a relatively small part of your diversified portfolio, an intentional satellite allocation or a much larger exposure that has accumulated across several funds.
Once you can see it, you can decide whether it still fits the plan you actually have, rather than reacting to whichever AI headline lands in your feed that day.
—
This article contains general information only and does not constitute personal financial advice. It does not recommend acquiring, holding or selling any investment. ETF holdings and classifications change over time; review current issuer documents and consider obtaining advice from an appropriately licensed financial adviser before changing a portfolio.
