Best AI ETF 2026: AIAI vs WTAI vs RBOT vs XAIX for EU Investors
AI ETF comparison 2026 for EU investors: AIAI, RBOT, WTAI, XAIX UCITS funds. TER, AUM, holdings overlap, performance, and US ETF alternatives explained.
- €7.8B
- XAIX fund size — largest AI UCITS ETF (0.35% TER)
- IE00BK5BCD43
- correct AIAI ISIN (IE00BK5BCH80 is L&G Clean Energy)
- +44%
- AIAI 2026 YTD to late July (WTAI +42%, RBOT +34%, XAIX +34%)
- 4.7%
- NVIDIA weight in XAIX — not a top-10 holding in AIAI or RBOT
Fact-checked against primary sources on · figures re-verified on regulator, issuer or SEC filings — not copied from other sites
Best AI ETF 2026: AIAI vs WTAI vs RBOT vs XAIX for EU Investors
TL;DR
EU investors have four primary UCITS ETF options for thematic AI exposure: L&G Artificial Intelligence (AIAI), WisdomTree AI (WTAI), iShares Automation & Robotics (RBOT), and Xtrackers AI & Big Data (XAIX). TERs range from 0.35% (XAIX) to 0.49% (AIAI). Fund sizes (justETF, late July 2026) range from €1.2B (WTAI) to €7.8B (XAIX) — XAIX, not RBOT, is now the category leader. The popular US AI ETFs — BOTZ, AIQ, IRBO, ROBO — are not available to EU retail investors under MiFID II. Overlap is lower than you would think — NVIDIA is not even a top-10 holding in AIAI or RBOT (2.5% in WTAI, 4.7% in XAIX) — and methodology differs significantly: AIAI and WTAI are near-equal-weight, XAIX is mega-cap-tilted and adds big data, RBOT skews to industrial automation. Many investors consider XAIX the best blend of cost, AUM, and AI exposure.
Why AI ETFs Matter in 2026
Generative AI has been the dominant equity theme since November 2022 and the cluster of beneficiaries has expanded far beyond Nvidia. Building meaningful exposure as a retail investor through individual stock picking requires monitoring chip cycles, hyperscaler capex commentary, foundation model competitive dynamics, and antitrust risk simultaneously. A diversified AI ETF outsources that complexity for 30–50 basis points per year.
The AI ETF category has grown from roughly $8 billion AUM globally in 2022 to over $45 billion in 2026, with European UCITS products absorbing the bulk of EU retail flows. Performance has been outstanding — the average AI ETF has approximately doubled from January 2023 to April 2026, though performance dispersion across products is wider than most investors expect.
This guide focuses on the four UCITS-domiciled funds available to EU retail investors via standard European brokers (XTB, Trading 212, Trade Republic, Saxo, Interactive Brokers, BOSSA). US-listed AI ETFs are referenced for context only — they cannot be purchased by EU retail clients due to MiFID II PRIIPs requirements.
Investment Thesis
The AI ETF thesis rests on four legs:
Leg 1: Capex cycle. Hyperscaler 2026 capex of approximately $330 billion flows directly to GPU vendors (Nvidia, AMD), networking (Broadcom, Arista), memory (Micron, SK Hynix), and equipment (ASML, Applied Materials). Most AI ETFs concentrate in these names.
Leg 2: Software monetization. Microsoft Copilot, Salesforce Agentforce, ServiceNow Now Assist, and Adobe Firefly are early-stage but materially repricing software per-seat economics. AI ETFs that include software exposure participate in this monetization curve.
Leg 3: Productivity diffusion. Beyond pure-play AI, productivity gains accrue to consumer (Meta), enterprise software, and selected industrials. Broader AI ETFs (XAIX, AIAI) capture this diffusion; narrower robotics ETFs (RBOT) miss it.
Leg 4: Diversification vs single-stock risk. Holding NVDA at 2–5% of an ETF is materially less binary than holding NVDA at 25% of a portfolio. ETF wrappers reduce idiosyncratic event risk meaningfully.
The bear case: thematic ETFs systematically underperform broad-market alternatives over long horizons because they buy at peak hype (high valuations) and sell at troughs (forced rebalancing). This is documented in academic literature on thematic fund flows. The AI thesis carries elevated valuation risk and many investors consider thematic exposure as a satellite holding rather than a core position.
Top Picks Breakdown
L&G Artificial Intelligence UCITS ETF (AIAI / IE00BK5BCD43)
| Metric | Value |
|---|---|
| ISIN | IE00BK5BCD43 (tickers AIAI / XMLD / AIAG — note: IE00BK5BCH80, sometimes quoted, is L&G's Clean Energy ETF) |
| TER | 0.49% |
| AUM | €1.8 billion (justETF, July 2026) |
| Holdings | 53 stocks |
| Domicile | Ireland |
| Distribution | Accumulating |
| Index | ROBO Global Artificial Intelligence Index |
| Top 5 holdings | Palo Alto Networks, Datadog, Ambarella, Cloudflare, Samsara (each ~2.4–2.6%; near-equal-weight) |
| US weight | ~73% |
| Asia weight | ~10% (Taiwan 5.3%, China 4.6%) |
AIAI tracks the ROBO Global AI Index, which screens for revenue derived directly from AI products and services. The index methodology produces the purest-play AI exposure of the four UCITS options — only ~3% of holdings have minimal AI revenue contribution. Trade-off: highest TER at 0.49%, and the near-equal weighting means it behaves like a mid-cap software/semis basket rather than a mega-cap fund. At €1.8B it is now comfortably liquid.
Best for: investors who want concentrated, equal-weighted AI exposure without mega-cap dominance — it was the best performer of the four in 2026 (+44% YTD to late July).
WisdomTree Artificial Intelligence UCITS ETF (WTAI / IE00BDVPNG13)
| Metric | Value |
|---|---|
| ISIN | IE00BDVPNG13 |
| TER | 0.40% |
| AUM | €1.2 billion (justETF, July 2026) |
| Holdings | 66 stocks |
| Domicile | Ireland |
| Distribution | Accumulating |
| Index | NASDAQ CTA Artificial Intelligence NTR |
| Top 5 holdings | AMD 4.9%, Marvell 3.7%, Micron 3.3%, Astera Labs 2.9%, Alphabet 2.5% (NVIDIA 2.5%) |
| US weight | ~75% |
| Asia weight | ~7% (Korea 4.7%) |
WTAI tracks a Nasdaq-curated AI index with an equal-weight tilt that caps mega-caps at lower weights than market cap might suggest. This produces low NVDA concentration (~2.5% vs ~4.7% in XAIX) but higher exposure to the second tier of AI beneficiaries — AMD, Marvell, Micron, Astera Labs, SK hynix, Nebius, Credo.
Best for: investors who want AI exposure without overconcentration in a few mega-caps. Decent compromise between purity and diversification.
iShares Automation & Robotics UCITS ETF (RBOT / IE00BYZK4552)
| Metric | Value |
|---|---|
| ISIN | IE00BYZK4552 |
| TER | 0.40% |
| AUM | €4.5 billion (justETF, July 2026) — second-largest after XAIX |
| Holdings | 157 stocks |
| Domicile | Ireland |
| Distribution | Accumulating |
| Index | iSTOXX FactSet Automation & Robotics |
| Top 5 holdings | Snowflake 3.7%, Advantest 3.2%, Keyence 3.1%, Siemens 3.1%, Emerson 3.1% |
| US weight | ~56% |
| Japan weight | ~14% |
| Europe weight | ~20% (Germany 7.5%, Switzerland 6.2%) |
RBOT is the most diversified of the four (157 names, near-equal-weighted). It's not a pure AI ETF — it's an automation and robotics ETF mixing industrial automation (Keyence, Siemens, Emerson, Rockwell), test equipment (Advantest) and AI software (Snowflake, ServiceNow, SAP). The ~14% Japan and ~20% Europe weights differentiate it from US-tilted competitors.
Best for: investors who want broader automation thematic exposure with meaningful international diversification. Lower correlation to the Magnificent 7 than other AI ETFs.
Xtrackers AI & Big Data UCITS ETF (XAIX / IE00BGV5VN51)
| Metric | Value |
|---|---|
| ISIN | IE00BGV5VN51 |
| TER | 0.35% (lowest in category) |
| AUM | €7.8 billion (justETF, July 2026) — largest in category |
| Holdings | 100 stocks |
| Domicile | Ireland |
| Distribution | Accumulating |
| Index | Nasdaq Global Artificial Intelligence and Big Data |
| Top 5 holdings | Apple 5.6%, Microsoft 5.1%, Meta 5.1%, Walmart 4.8%, NVIDIA 4.7% |
| US weight | ~83% |
| Asia weight | ~12% (Korea 6.6%, China 4.9%) |
XAIX combines AI and big data exposure under a single thematic index. The 0.35% TER undercuts all competitors and the €7.8B fund size makes it the category's liquidity benchmark. Holdings skew most heavily to mega-cap tech (plus data-heavy non-tech names such as Walmart and Bank of America) — the top 10 represent ~46% of the fund versus ~25% for near-equal-weight AIAI.
Best for: cost-conscious investors who want core US mega-cap AI exposure with some big-data diversification. Often the default recommendation for first-time AI ETF buyers — though it was the laggard of the four in 2026 (+34% YTD to late July vs +44% for AIAI).
Savings in one bank, investments in another? See it all in one place — and how many months it could carry you.
See your Freedom Runway — freeValuation Analysis
All four ETFs trade at meaningful premiums to the broad US equity market on every standard valuation metric:
| Metric | XAIX | WTAI | AIAI | RBOT | S&P 500 |
|---|---|---|---|---|---|
| Forward P/E | ~38x | ~36x | ~42x | ~28x | ~22x |
| Price/Sales | ~7.5x | ~6.8x | ~8.2x | ~4.5x | ~2.7x |
| EPS growth (3y forward) | ~22% | ~24% | ~28% | ~16% | ~11% |
| PEG | 1.7 | 1.5 | 1.5 | 1.8 | 2.0 |
On absolute multiples, AIAI is most expensive (purest AI = highest growth, highest multiple), while RBOT is cheapest because it includes industrial automation names trading at 18–20x. On growth-adjusted PEG, AIAI and WTAI screen most attractive.
Compared to the Magnificent 7 cohort blended forward P/E of ~31x, all AI ETFs trade at a small premium because they include faster-growing pure-plays (Palantir, ARM, AMD) at the expense of slower compounders (Apple, Tesla).
The historical context: thematic tech ETFs averaged forward P/E of ~22x in 2019, peaked near 45x in late 2021, fell to ~25x in 2022, and have expanded back to ~38x. Current valuations are above pre-COVID norms but below 2021 peaks.
EU Investor Access
All four UCITS ETFs are available via:
- XTB — commission-free up to €100k/month, Polish broker
- Trading 212 — commission-free, available across most EU countries
- Trade Republic — €1 flat commission, German-based
- Saxo Bank — institutional pricing, advanced features
- Interactive Brokers — best for >€50k accounts, lowest FX cost
- BOSSA / mBank / ING — Polish bank brokers, higher commissions but native PLN account
Listed currencies: EUR, USD, GBP across multiple European exchanges (Xetra, LSE, Euronext, Borsa Italiana). For Polish investors, EUR-listed shares minimize FX cost since most Polish brokers charge 0.1–0.5% PLN/EUR conversion.
Tax efficiency: All four are accumulating UCITS ETFs domiciled in Ireland. This means dividends are reinvested internally without triggering withholding tax events for the holder, and Ireland's tax treaty network minimizes underlying US dividend withholding to 15%. For Polish investors, holding in IKE/IKZE shelters capital gains from the 19% Belka tax — strongly recommended for thematic positions held >5 years.
US ETFs not available: AIQ ($10.2B, 0.68% TER), BOTZ ($3.4B, 0.68% TER), ARTY (formerly IRBO — iShares Future AI & Tech, $4.0B, 0.47% TER) and ROBO ($2.1B, 0.95% TER) are popular US-listed alternatives but not available to EU retail investors due to MiFID II PRIIPs requirements. Some EU investors gain access via professional client classification (€500k+ portfolio + experience tests) at Interactive Brokers, but most retail clients cannot.
Real-World Example Portfolio
A €100,000 portfolio with thematic AI tilt for an EU investor with 5+ year horizon might look like:
| Position | Allocation | Amount | Rationale |
|---|---|---|---|
| iShares Core MSCI World (IWDA) | 45% | €45,000 | Global core |
| Xtrackers AI & Big Data (XAIX) | 15% | €15,000 | Primary AI thematic, lowest cost |
| L&G AI UCITS (AIAI) | 8% | €8,000 | Pure-play AI satellite |
| iShares Nasdaq 100 (CNX1) | 10% | €10,000 | US tech beta |
| Vanguard FTSE All-World (VWCE) | 10% | €10,000 | Diversification |
| iShares Core EM (EIMI) | 5% | €5,000 | EM exposure |
| Cash / short bonds | 7% | €7,000 | Dry powder |
The combined 23% in AI thematic (XAIX + AIAI) is meaningful but not dominant. Splitting between XAIX (broader) and AIAI (purer) reduces single-index methodology risk while keeping blended TER low.
For investors who already hold individual Mag 7 names directly, the AI ETF allocation should be sized lower (5–10%) to avoid double-counting NVDA, MSFT, GOOGL exposure.
Risk Factors
Concentration risk. XAIX has ~25% in its top 5 holdings; the other three are near-equal-weighted (top 5 ≈ 12–16%). NVIDIA is the largest position in none of them (4.7% in XAIX, 2.5% in WTAI, outside the top 10 in AIAI and RBOT), so a 30% NVDA drawdown costs an ETF holder roughly 1–1.5%.
Methodology drift. Thematic indices rebalance based on revenue exposure and analyst classification. As more companies tag themselves "AI" for marketing purposes, index providers face challenges distinguishing pure-play from rebadged exposure. AIAI's stricter methodology mitigates this; XAIX is more permissive.
Currency risk. EUR-denominated ETF shares track underlying USD assets without hedging. A 10% EUR/USD move translates ~1:1 to ETF NAV. Many investors consider unhedged exposure appropriate for long-horizon equity holdings, but it adds volatility.
Capex cycle reversal. If hyperscalers materially cut AI capex in 2026 or 2027, the entire AI ETF complex de-rates simultaneously. This is the systemic risk the diversification cannot eliminate.
Liquidity risk in stress. All four funds now exceed €1B (WTAI €1.2B is the smallest; XAIX €7.8B the largest), so spreads are tight in normal markets; they can still widen in stress (March 2020 levels). For €5,000+ trades, check spreads carefully.
China geopolitical risk. TSMC concentration (Taiwan-based, ~20% of NVDA's COGS) and ASML China export restrictions both flow into AI ETF holdings. A Taiwan escalation would disrupt the entire AI hardware supply chain.
Time Horizon Considerations
Short-term (0–12 months): AI ETFs are highly correlated with tech sector beta. Don't time them. If you must trade short-term, individual stocks offer better risk/reward than diversified ETFs.
Medium-term (1–3 years): This is the AI capex cycle window. Most ETF investors entering 2025–2026 are positioning for this horizon. Performance dispersion across the four UCITS funds is likely to be low (within 5–10%) because top holdings overlap heavily.
Long-term (3–10 years): The harder question. Thematic ETFs have historically underperformed broad market alternatives over decade-plus horizons because they include companies near peak relative growth. The 2010s "internet" thematic funds delivered ~9% CAGR while QQQ delivered ~17%. Whether AI breaks this pattern depends on whether AI compounds productivity gains across the broad economy or remains concentrated in a few infrastructure providers.
The most defensible AI ETF holding period is 3–7 years with active rebalancing back to target weights when allocations drift more than 5 percentage points from target.
FAQ
Q: Which AI ETF should I buy as a first-time EU investor?
A: For most investors, XAIX (Xtrackers AI & Big Data) offers the best blend of low cost (0.35% TER), the largest fund size (€7.8B) and mega-cap-anchored exposure. Investors who want less mega-cap concentration pick AIAI (equal-weight, 0.49%) or WTAI (0.40%).
Q: Can I buy BOTZ or AIQ as an EU retail investor?
A: No. Both are US-listed and not registered for EU retail distribution under MiFID II. Use UCITS alternatives (AIAI, WTAI, RBOT, XAIX) instead.
Q: Is RBOT really an AI ETF?
A: It's primarily an automation and robotics ETF with significant AI overlap. About 50–60% of holdings have meaningful AI revenue exposure, but the fund includes industrial automation names (Fanuc, KEYENCE) that are not pure AI plays. Choose RBOT if you want broader automation; choose AIAI/WTAI/XAIX if you specifically want AI.
Q: How much overlap is there between AI ETFs and Nasdaq-100?
A: It depends on the fund. XAIX overlaps heavily with the Nasdaq-100 (Apple, Microsoft, Meta, NVIDIA, Amazon, Alphabet in its top 10); AIAI, WTAI and RBOT are dominated by mid-caps that are not Nasdaq-100 heavyweights. Holding XAIX plus a Nasdaq-100 ETF double-counts the mega-caps — size accordingly. If a broad-based holding fits better than a narrow AI theme, our best technology ETF in Europe guide covers the diversified Nasdaq-100 and tech-sector options.
Q: How do I track AI ETF positions across multiple brokers and currencies?
A: For consolidated tracking of UCITS ETFs across XTB, Trading 212, Trade Republic, and Interactive Brokers, Freenance supports multi-broker import with cost basis tracking and FX-adjusted unrealized P&L — useful when holding the same ETF on multiple exchanges.