Key Takeaways
- There are many European funds wearing the AI label, but their portfolios can vary dramatically.
- AI funds can be broadly divided into core, dynamic, and peripheral exposures.
- First Trust Bloomberg AI ETF scores highest according to Morningstar metrics.
For most investors, AI is a theme best played through a fund. Stock-specific risk is high in a rapidly evolving area where leadership can change within quarters, and a diversified basket gives investors a better chance of holding tomorrow’s winners without assessing each candidate.
There are dozens of products in the European market bearing the AI label. They look similar in the marketing material, but their portfolios—and their actual exposure to core AI—vary dramatically.
The Funds That Score Highest
First Trust Bloomberg AI ETF sits at the top of the European fund rankings. Its selection and weighting methodology both lean directly on AI-related revenue, producing the highest combined exposure score of any fund in our European coverage. The iShares AI Infrastructure UCITS ETF ranks second, reflecting a newer approach focused specifically on the picks-and-shovels segment of the value chain—and a strong fit with Morningstar analysts’ view that infrastructure is where both the exposure and the valuation argument are strongest.
The top 10 is split roughly evenly between ETFs and open-end funds, with BlackRock’s AI Innovation strategy appearing in both wrappers. The highest scorers are suitable for playing the undervalued core exposure story.
At the other end of the table, the lowest-scoring funds aren’t necessarily bad funds—they’re just doing something different.
Tangential strategies covering Big Data or quantum, and “AI adopter” funds targeting downstream beneficiaries like healthcare firms using AI, score poorly on core exposure by design. The Polar Capital AI Fund also ranks lower, because it often tilts into non-core names it deems to be the biggest downstream beneficiaries of AI, with holdings ranging from Delta Air Lines to Walmart. That looked unconventional last year. In February 2026, when software as a service or SaaS names sold off sharply, the fund’s early move out of software and into infrastructure produced a roughly 14-percentage-point outperformance over index-like peers in a single month.
Three Ways to Play AI Through Funds
That contrast points to the real choice facing investors. Broadly, AI funds fall into three buckets.
- Core AI funds—typified by First Trust Bloomberg AI—gives stable, beta-like exposure to the theme. They score highest on the metric and are well suited to investors who simply want the AI exposure in their portfolio to do what it says on the tin.
- Dynamic AI funds, such as CPR Invest Artificial Intelligence, keep exposure to the bellwethers but actively tilt as the theme evolves. They suit return-seeking investors comfortable with active risk and willing to accept tracking error versus the broader AI theme in exchange for the chance of alpha.
- Peripheral AI funds, such as the iShares AI Adopters & Applications UCITS ETF, target second-order beneficiaries—firms embedding AI to expand margins, or the power and data-centre providers enabling it. They score lower on core exposure by design but offer diversification away from crowded mega-cap tech.
The right choice depends on what investors are trying to do. But if the starting point is the valuation argument—that core AI is undervalued relative to the cash flows it is going to generate—then the funds at the top of our exposure table are the most direct way to express that view.
How the AI Fund Screen Was Built
To cut through the noise, Morningstar has built a combined thematic exposure score. This marries two factors: the equity research team’s bottom-up assessment of how much of each company’s revenue is genuinely tied to AI, weighted at 75%, and a manager research signal that measures how frequently a stock appears in specialist AI portfolios globally, weighted at 25%. Rolling those scores up at the fund level gives a single, comparable measure of how much core AI exposure each strategy is actually delivering.
This article is taken from the report “Opportunities in Artificial Intelligence,” published in May and written by Michael Field and Kenneth Lamont.
