AI ETFs: What They Own and What They Cost

AI ETFs Are Not One Thing

The single most useful thing to understand before buying an "AI ETF" is that the label covers at least three completely different products. Two funds can both put "artificial intelligence" in their marketing and share almost no holdings. One might be 60% mega-cap US software. Another might be half Japanese and European robotics manufacturers. A third might be nothing but chipmakers.

So the question "what is the best AI ETF?" has no answer until you decide which kind of AI exposure you actually want. The rest of this guide sorts the field into three buckets, tells you what each owns, and shows why the most expensive option is often the one you need least.

Three Ways an ETF Gives You AI Exposure

Every fund marketed around AI falls into one of three groups. They differ in what they hold, what they cost, and how concentrated the bet is.

Broad tech is the cheapest and least concentrated, and it already owns the names that have driven the AI story. Thematic funds are the purest expression of the theme but the most expensive and narrow. Semiconductors are the infrastructure layer, covered in depth in our semiconductor ETF guide. Most investors overpay by reaching for the thematic bucket when the broad bucket already covered them.

The Dedicated AI Funds: AIQ, BOTZ, ROBO, QTUM

When people picture an "AI ETF," they usually mean the thematic funds. These are actively curated baskets built around a specific slice of the theme. They are worth knowing precisely because they are not interchangeable.

• AIQ (Global X Artificial Intelligence & Technology, 0.68%) holds about 50 companies that build or heavily use AI and big-data software. It tilts toward large US technology and semiconductor names, making it the closest thing to a pure AI-software bet.
• BOTZ (Global X Robotics & AI, 0.68%) holds about 45 robotics and industrial-automation companies, with meaningful Japanese and European exposure. It is AI applied to physical machines rather than AI software.
• ROBO (Robo Global Robotics & Automation, 0.95%) is older and broader, spreading weight more evenly across many smaller robotics and automation names, at the highest fee of the group.
• QTUM (Defiance Quantum, 0.40%) blends quantum computing and AI-adjacent hardware. Despite the name, most of its return has come from AI and cloud companies rather than quantum revenue, which barely exists yet.

You Probably Already Own the Biggest AI Winners

Here is the part the fund marketing does not lead with. The companies that have captured the most value from AI so far, Nvidia, Microsoft, Alphabet, Meta, Amazon, and Broadcom, are already the largest holdings in the broad funds most investors own.

VGT (Vanguard Information Technology) holds Nvidia, Microsoft, Broadcom, and Apple as top positions, at a 0.10% fee. QQQ (the Nasdaq-100) holds the same mega-caps plus Alphabet, Meta, and Amazon. Even a total US market fund like VTI holds all of them, just at smaller weights. In many cases these broad funds hold a higher combined weight in the true AI leaders than a thematic AI fund does, because thematic funds deliberately spread into smaller, purer names to justify the theme.

What the AI Label Costs You

The fee gap is not a rounding error. A broad tech fund charges around 0.10%. A thematic AI fund charges around 0.68%. That is a difference of 0.58 percentage points every year, and it compounds.

On a $25,000 position held for 20 years, the roughly 0.58% annual fee difference quietly costs thousands of dollars, money that leaves your account regardless of whether the AI theme delivers. You are paying that premium for the curation and the specific tilt, not for better access to Nvidia or Microsoft, which the cheap fund already gives you. Whether the tilt is worth 7x the fee is the real question, and our fee calculator puts an exact dollar figure on it. For the mechanics of why small percentages matter so much over time, see the expense ratios guide.

The Concentration Trap With Thematic AI Funds

Narrow funds cut both ways. The same concentration that produces eye-catching returns in a good year produces brutal drawdowns in a bad one. Three specific risks apply to thematic AI funds.

Top-heavy holdings. Many AI funds put a large share of assets in a handful of mega-cap names. When those few stocks stumble, the whole fund drops with them, and the "diversification" of holding 50 stocks is an illusion if the top 10 drive most of the movement.

Priced for perfection. AI leaders already trade at valuations that assume years of rapid growth. That growth may well arrive, but it is already in the price, so the return depends on companies meeting expectations that are unusually high. A theme this popular has little margin for disappointment.

Hidden overlap. The most common mistake is owning an AI fund on top of a broad tech or total-market fund and believing you have added diversification. In reality you have doubled down on the same names. This is the exact problem covered in our guide on thematic ETFs, where narrow bets quietly stack on top of exposure you already hold.

Where AI ETFs Fit, If at All

For most long-term investors, the honest answer is that a broad, low-cost fund already delivers the AI exposure that matters, and a dedicated AI ETF is optional. That is not a knock on the theme. It is a statement about how much of it you already own.

A dedicated AI or robotics fund can make sense as a deliberate satellite position: a small, intentional overweight for someone who specifically wants more of the theme than a broad fund provides, understands the concentration, and is comfortable with the fee. The key word is small. Used that way, AIQ, BOTZ, or a semiconductor fund adds a tilt without letting a single narrow bet dominate the portfolio.

What does not work is treating an AI ETF as a core holding or as a substitute for broad diversification. The theme is real, the fees are high, and the overlap with what you already own is larger than the marketing suggests. Decide which of the three buckets you actually want, check the overlap, weigh the fee, and size it accordingly.