
AI ETFs: Compare Holdings, Fees and Hidden Costs
- Oliver Narramore

- Aug 14
- 9 min read
AI ETFs focused on Artificial Intelligence can spread one trade across dozens of companies, but their names rarely reveal what you own. They can reduce reliance on one market favourite, yet several funds may hold the same large technology companies, such as Microsoft, creating concentration risk.
For a beginner, compare the index, top holdings, overlap, and total cost before buying. An AI fund is a thematic investment, but spreading money across companies does not remove technology-sector risk. Its label is the starting point, not the research.
AI ETFs: Know What You Actually Own
An exchange traded fund, or ETF, holds a basket of investments and trades on an exchange like a share. When you buy one, you own a small slice of that basket.
Most AI-themed ETFs own businesses that build Artificial Intelligence systems, chips, cloud infrastructure, software, data tools, robotics, or security products. Nvidia, Microsoft, AMD, and Alphabet can appear in different combinations. A fund may also hold smaller names that supply test equipment, networking gear, or specialist software.
That differs from a fund that uses AI to select investments. In that case, Artificial Intelligence tools may help a manager screen shares or run a trading model. It doesn't mean the fund owns the businesses shaping the technology. For example, the iShares AI Innovation Active UCITS ETF is actively managed, while many familiar thematic funds follow an index.
A UCITS ETF meets European fund rules and is common on UK platforms. Still, UCITS status does not automatically make a fund suitable, SIPP eligible, or low-risk. Its value can fall sharply when investors cool on technology stocks.
An ETF can hold 50 companies and still behave like a concentrated technology investment if its largest positions and sectors move together.
That is concentration risk, where many holdings can still move together as one technology trade.
Before you invest, read the fund's objective. A phrase such as "AI infrastructure" points toward chipmakers, networking, data centres, and power-related suppliers. "Automation and robotics" usually casts a wider net across industrial machinery, healthcare equipment, and factory software.
An Accumulating share class normally reinvests income, while a Distributing class generally pays it out. Check the factsheet label for Accumulating before investing. The Accumulating label affects how income is handled, so confirm the current factsheet.
Compare Holdings Before You Compare Fees
Compare AI ETFs by holdings, fees, fund size, and index rules. The index methodology matters more than the marketing headline. It sets eligibility, position limits, weighting, and rebalancing for Artificial Intelligence exposure.
For example, the L&G Artificial Intelligence UCITS ETF tracks the ROBO Global index. It uses physical replication, meaning it buys the underlying shares rather than relying on a derivative contract.
WisdomTree's WTAI tracks the NASDAQ CTA index, as set out on the WisdomTree fund page. The WisdomTree Artificial Intelligence fund uses rules-based selection. Its approach can favour mega-cap AI winners, including Nvidia, over an even spread across the theme.
The snapshot below uses provider information checked on 14 August 2026. Holdings, weightings, fees, and fund size can change, so verify every figure against the latest factsheet before dealing.
The LSE ticker column refers to the ticker commonly used on the London Stock Exchange.
ETF | LSE ticker commonly used | Main exposure | What to inspect |
|---|---|---|---|
L&G Artificial Intelligence UCITS ETF | AIAG | Broad AI supply chain | Semiconductors, software, and specialist technology weights |
WisdomTree Artificial Intelligence UCITS ETF | WTAI | Rules-based AI companies | Nvidia weight and the index's inclusion rules |
iShares AI Infrastructure UCITS ETF | AINF | Hardware and infrastructure behind AI | Heavy technology-sector exposure |
iShares Automation & Robotics UCITS ETF | RBTX | Automation and robotics | Broader industrial and robotics mix |
For a fair performance comparison, use the same period and share class for every fund. The table also shows how each fund's Artificial Intelligence exposure differs.
Check the share class before comparing returns
Share classes can affect the return you receive. Use the checklist below to confirm the relevant class and identifier in the latest documents.
Ticker | First check | Second check |
|---|---|---|
AIAG | Accumulating status | Accumulating ticker or ISIN |
WTAI | Accumulating status | Accumulating ticker or ISIN |
AINF | Accumulating status | Accumulating ticker or ISIN |
RBTX | Accumulating status | Accumulating ticker or ISIN |
A Distributing class may follow different cash-flow rules, so check the relevant document rather than assuming the labels match.
AINF shows why a name can mislead. The iShares AI Infrastructure UCITS ETF reported an information technology allocation of 87.60% in the latest available data. That may suit someone who wants infrastructure exposure. It is less suitable if your portfolio already leans heavily towards tech.
Look for overlap in your existing portfolio
A global index fund already owns substantial positions in Microsoft, Nvidia, Alphabet, Amazon, Broadcom, and Meta. Adding AI ETFs may increase your exposure to the same companies instead of creating genuine diversification.
Microsoft can therefore remain a large holding after the purchase. Repeated mega-cap holdings can create concentration risk and weaken genuine portfolio diversification.
Start with your top ten holdings across every account. Then check whether the thematic ETF repeats them. A portfolio holdings and allocation breakdown can also show how a technology-focused portfolio can differ from a broad global approach.
Look beyond the biggest names, too. Recent L&G portfolio data included AMD, Nebius, Astera Labs, MediaTek, Datadog, Infineon, Credo Technology, and IonQ. IonQ brings quantum computing exposure, which is exciting but far less proven than established chip sales or cloud revenue.
Company metrics can help you judge the risk sitting inside the fund:
Revenue growth shows whether a company is selling more. A 20% revenue CAGR means sales grew by about 20% a year over the measured period, not that they will keep growing at that pace.
Forward P/E compares a share price with forecast earnings over the next 12 months. A high figure means investors already expect strong profit growth.
PEG compares valuation with expected earnings growth. A lower PEG can look cheaper, but the number becomes unreliable if analysts cut their forecasts.
High expectations can unravel quickly. A chipmaker may report strong revenue while spending heavily on factories, research, or new capacity. Therefore, watch free cash flow and capital expenditure alongside headline sales growth.
Photo by Tiger Lily
Fees Matter, but They Are Not the Only Cost
When comparing AI ETFs, focus on the cost of Artificial Intelligence exposure, not just the headline fee.
The Ongoing Charges Figure (OCF) and Total Expense Ratio (TER) are annual fund costs deducted within the ETF. You won't normally see a separate bill, and performance is reported after that deduction. Neither measure captures dealing costs, bid-ask spreads, FX charges, or taxes.
Published charges sit in a fairly tight range. On a £10,000 holding, the difference may seem small in one year. Over a decade, those annual deductions compound alongside the returns you miss.
The following UCITS ETF comparison uses the same £10,000 illustration:
ETF label | OCF or TER | First-year cost on £10,000 |
|---|---|---|
AIAG | 0.49% | About £49 |
WTAI | 0.40% | About £40 |
AINF | 0.35% | About £35 |
RBTX | 0.40% | About £40 |
These figures assume the holding starts at £10,000. As the investment value rises, the cash amount paid through the fund also rises.
A 0.14 percentage-point gap between 0.49% and 0.35% is £14 in the first year per £10,000. That does not automatically make the cheaper ETF better. It means the more expensive fund needs a reason to earn its place, such as a methodology or holding mix you genuinely prefer.
A useful performance comparison should also consider tracking difference and after-cost returns, not fees alone.
Share-class cost matrix
Share-class income treatment can change cash flow, but it does not automatically change the annual charge. This matrix separates reinvested income from paid income:
Comparison point | Reinvested-income treatment | Paid-income treatment |
|---|---|---|
Income received | Accumulating: income remains invested | Distributing: income is paid out |
Reinvestment | Accumulating: reinvestment happens within the class | Reinvestment is an investor choice |
Cash withdrawals | Accumulating: sell units when cash is needed | Cash comes from payments |
Unit price | Accumulating: income is reflected in value | Value adjusts after payment |
Compounding | Accumulating: reinvested income can compound | Paid income leaves the class |
Cash management | Accumulating: no scheduled payment | Payment dates may matter |
Records | Accumulating: monitor total value | Track payments separately |
Comparability | Accumulating: compare total returns | Compare total returns after distributions |
Charges | Accumulating: not automatically cheaper | Distributing: not automatically more expensive |
Selection | Accumulating: suits a reinvestment preference | Choose based on an income preference |
The share class does not by itself determine which option is cheaper. Compare charges and returns on the same basis.
Also separate fund charges from these costs:
Broker commission is what the platform charges when you buy or sell.
Bid-ask spread is the gap between the price available to buyers and sellers. Thinly traded ETFs can have wider spreads.
FX charges may apply when your broker converts pounds into another currency.
Taxes can include capital gains tax, and treatment depends on your account type and personal circumstances.
The trading currency does not tell you where a fund's investments are located. However, buying a USD share class through a GBP account may trigger conversion charges. Read the factors behind ETF bid-ask spreads before placing a large order in a niche fund.
How to Invest in an AI ETF Without Rushing
Your decision on how to invest should begin with the job the ETF has in your portfolio. A broad global fund may form the core and already hold Microsoft. AI ETFs focused on Artificial Intelligence can form a smaller satellite position for portfolio diversification. No universal percentage suits every investor, and the role, wrapper, share class, platform, currency, and dealing method remain separate decisions.
The mechanics of how to buy stocks and ETFs are similar. Research the fund carefully on your investment platform.
Set an amount you could keep invested through a painful decline. AI ETFs can rise quickly, but they can also fall sharply when interest rates increase, earnings disappoint, or data-centre spending slows.
Search for the ISIN as well as the ticker. Tickers can differ across exchanges, including the London Stock Exchange, while an ISIN identifies the exact share class. Check whether the exact fund and desired class are supported by your platform.
Confirm the account wrapper before dealing. ETFs are common holdings in a stocks and shares ISA and SIPP, but your broker must support the exact fund. Check whether the exact UCITS ETF and chosen share class are available in a stocks and shares ISA. Confirm whether that class is SIPP eligible. HMRC's ISA investment guidance explains the qualifying-investment rules.
Compare platform costs and order types. Trading 212, for example, has its own Invest account terms. A limit order gives you control over the highest price you will pay, which can help when spreads are wider.
Share-class decision Accumulating classes reinvest fund income. Their factsheets may label them Accumulating, and Accumulating may suit investors seeking automatic reinvestment. Select Accumulating when that matches your preference. Platform search results may show Accumulating beside the fund name. Compare that Accumulating label with the factsheet, then confirm Accumulating in the order. Keep Accumulating if reinvestment remains your preference. Distributing classes pay income out. A Distributing label identifies that approach, and Distributing may suit investors who prefer cash payments. Before dealing, ensure the platform displays Accumulating, the factsheet identifies Accumulating, and the confirmation records Accumulating. The selected Accumulating class should match your preference.
If you have compared the risks, platform charges, and available share classes, Buy now on etoro is one possible next step. It is an optional commercial example, not a recommendation. Check that the exact UCITS ETF is offered in your country and account type before funding an order.
Availability and eligibility need separate checks. The exact fund may be offered in your country but unavailable through your account. Wrapper eligibility is distinct from investment suitability. A fund available in a stocks and shares ISA may still be unsuitable for your risk tolerance. Confirm that the exact share class is SIPP eligible with your platform.
Currency adds another layer. An ETF may own US companies, earn revenue worldwide, and trade in dollars, while you measure future spending in pounds or euros. Currency moves can lift or reduce your return, even when the underlying shares are flat.
Long holding periods can soften the urge to chase weekly stock market news. Use a compound interest calculator to see how regular contributions, returns, and annual fees affect long-term growth over time.
Frequently Asked Questions
What is an AI ETF?
An AI ETF is a fund that holds a basket of companies linked to Artificial Intelligence, including chipmakers, cloud providers, software businesses, robotics firms, and data specialists. Buying one trade gives you exposure to several companies, but it does not remove technology-sector or concentration risk.
Are AI ETFs diversified?
AI ETFs can spread your investment across dozens of companies, but their largest holdings may move together. Check the top positions, sector weights, and overlap with your existing global funds before assuming the ETF adds genuine diversification.
What costs should I compare when choosing an AI ETF?
Compare the Ongoing Charges Figure or Total Expense Ratio, tracking difference, bid-ask spread, broker commission, FX charges, and applicable taxes. A lower annual fee is useful, but the cheapest fund may not have the index or holding mix you prefer.
Should I choose an Accumulating or Distributing AI ETF?
An Accumulating share class normally reinvests income within the fund, while a Distributing class generally pays income out to you. Check the latest factsheet and confirm that the platform order shows the exact share class and ISIN you intend to buy.
Are AI ETFs suitable for a stocks and shares ISA or SIPP?
Many UCITS ETFs can be held in a stocks and shares ISA or SIPP, but eligibility depends on the exact fund, share class, and platform. Confirm availability and wrapper eligibility before investing, then consider whether the fund's risk suits your circumstances.
Final Thoughts
AI ETFs offer a simpler route into the Artificial Intelligence theme, but simplicity should not replace research. Compare the index, holdings, overlap, and costs, then check the factsheet to confirm whether your chosen share class is Accumulating.
Size the position for a bad year, not an exciting headline. Investing works better when you can hold through the moments that make everyone else want to sell.
This article is for education only and is not personal financial advice.













































Comments