By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
The panel agrees that 'AI ETFs' like AIS, CHAT, and BOTZ are not equivalent and expose investors to distinct sector risks. They caution against relying solely on 2026 YTD performance, highlighting cyclical capex, liquidity, and demand risks.
Risk: Capex cyclicality and potential demand collapse, leading to forced selling and multiple compression in AIS due to its liquidity and concentration risks.
Opportunity: AIS's exposure to the physical infrastructure for data centers, which may have a 'floor' due to utility capacity constraints.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
If you're searching for the best AI ETF in 2026, three funds dominate the conversation: CHAT (Roundhill Generative AI & Technology ETF), BOTZ (Global X Robotics & Artificial Intelligence ETF), and AIS (VistaShares Artificial Intelligence Supercycle ETF). All three carry the AI label but their 2026 returns could not be more different. One is up over 100%, one is up 73%, and one is down 7%. The gap tells you everything you need to know about how to pick an AI ETF.
CHAT, BOTZ, and AIS all claim the artificial intelligence label. But beneath the shared branding, they hold almost entirely different companies, target different parts of the AI economy, and have produced returns that bear no resemblance to each other. Before putting money into any AI ETF, understanding what it actually owns is more important than what the name implies.
Active (aims to replicate BITA VistaShares Artificial Intelligence Supercycle Index)
Index (Indxx Global Robotics & Artificial Intelligence Thematic Index)
H1 2026 return
+66%
+122%
+1%
AI angle
Generative AI applications & platforms
AI infrastructure & semiconductors
Robotics & physical automation
All data as of 06/24/26.
CHAT vs BOTZ vs AIS: Holdings Comparison
The performance gap starts to make sense the moment you look at the portfolios.
CHAT is a concentrated active fund built around the generative AI ecosystem — the companies building, running, and commercializing large language models. Its top five holdings are NVIDIA (6.39%), SK hynix Inc (5.10%), Alphabet (5.04%), Knowledge Atlas Technology JSC (4.68%), and Broadcom (4.06%). It's a recognizable roll call of the companies the market most directly associates with the ChatGPT era: hyperscalers, chip designers, and AI application developers. If you want a portfolio that reads like "generative AI," CHAT delivers it.
AIS takes a different bet entirely. Rather than owning the AI platforms themselves, it owns the infrastructure that AI runs on: semiconductors (SK hynix, Micron, AMD), power infrastructure (Vertiv Holdings, GE Vernova, Legrand), and data center supply chain. The fund's thesis — the "AI Supercycle" — is that the physical buildout required for AI compute represents a multi-year spending wave that benefits companies well outside Silicon Valley. Its industrial products sector alone represents nearly 14% of the portfolio. AIS doesn't own the AI. It owns the grid, the memory, and the cooling systems that AI can't run without.
BOTZ is the oldest of the three by nearly seven years, launched in 2016 when AI meant something very different. Its index tracks robotics and automation companies — physical systems that move, weld, operate, and perform surgery. The top five holdings are ABB (9.81%), Keyence (9.57%), Fanuc (9.00%), NVIDIA (8.51%), and Intuitive Surgical (6.35%). NVIDIA makes an appearance here too, but the fund is otherwise populated with industrial automation names: Japanese precision engineering firms, Swiss industrial conglomerates, surgical robot companies. This is not a software AI portfolio. It's a factory-floor AI portfolio.
AI ETF Performance in 2026: Why Returns Have Diverged
H1 2026 has been a defining test of which part of the AI economy the market is rewarding — and the verdict so far is unambiguous: infrastructure and software win, physical automation loses.
AIS has been the biggest beneficiary. AI infrastructure spending has accelerated dramatically in 2026, driven by hyperscaler capital expenditure programs and the build-out of inference capacity to support widespread AI deployment. The companies that supply memory, power conditioning equipment, and data center infrastructure have seen earnings revisions move sharply upward. AIS's +122% is not speculative momentum — it's tracking actual earnings growth among its core holdings.
CHAT's +66% reflects the continued strength of the large-cap generative AI ecosystem. The active management gives CHAT the ability to tilt toward winners and away from underperformers in a market moving quickly. It has used that flexibility effectively.
BOTZ's +1% is the more complicated story. Robotics and industrial automation are not broken businesses — they are long-cycle capital investments tied to manufacturing capex, which has been under pressure in 2026 from tariff uncertainty and slower global industrial output. Fanuc and Keyence, BOTZ's two largest non-NVIDIA holdings, are exposed to Japanese and Chinese manufacturing investment cycles that have been sluggish. Intuitive Surgical remains a strong business but trades at a valuation that already reflects years of growth. BOTZ is not a bad fund — it's a fund that tracks a part of the economy that is moving slowly in an environment where AI software and infrastructure are moving fast.
Expense Ratios: CHAT vs BOTZ vs AIS Fee Comparison
All three funds charge similar fees. CHAT and AIS both carry a 0.75% expense ratio — identical. BOTZ charges 0.68%, a marginal discount. On a $50,000 position, the annual difference between BOTZ and CHAT is $35. That's genuinely noise relative to the performance divergence on display in 2026.
Unlike comparing QQQ vs QQQM — where the two funds are literally identical underlying products — the cost difference here is irrelevant compared to portfolio construction. These three funds are making fundamentally different investment bets. Choose based on the bet you want to make, not the 7 basis point fee gap.
Which AI ETF Should You Buy? The Case for CHAT, AIS, and BOTZ
CHAT: for investors who want pure generative AI exposure. CHAT's portfolio maps almost directly to the companies that media coverage, earnings calls, and analyst reports discuss when they talk about AI. It's actively managed, which provides a margin of flexibility that index-based AI funds lack. The expense ratio is high for what it offers — the active portfolio is not radically different from what a tech index would give you — but for investors who want a focused generative AI tilt, CHAT delivers it cleanly.
AIS: for investors who want to own the AI infrastructure buildout. AIS is the most unconventional choice of the three — and in 2026, the most rewarding. Its bet is that AI cannot run without power, memory, and cooling, and that the companies supplying those inputs will see sustained demand for years regardless of which AI platform "wins" at the application layer. This is the AI equivalent of owning picks and shovels during a gold rush. The fund is also the newest (launched December 2024) and the smallest by AUM ($914.4M), which means less liquidity history and higher concentration risk in a young portfolio. Investors comfortable with those caveats have been handsomely rewarded.
BOTZ: for investors with a longer horizon on physical AI and robotics. BOTZ is the contrarian pick. Robotics and industrial automation represent genuine long-term AI applications — warehouse automation, surgical robotics, manufacturing precision, autonomous systems — and the current underperformance reflects a cycle, not a structural collapse. The fund has been around since 2016, has the largest AUM of the three ($3.42B), and is the most diversified geographically with significant exposure to Japanese and European industrial firms. For investors who believe physical automation will catch up to software AI in the next market cycle, BOTZ offers the most established vehicle. It is not a 2026 trade. It may be a 2028 trade.
CHAT vs BOTZ vs AIS: Our Verdict for 2026
For investors who want AI exposure now, with proven 2026 performance: AIS or CHAT. AIS targets the infrastructure layer that is driving actual earnings growth. CHAT targets the application layer where market attention is concentrated. Both have beaten the market significantly in H1 2026, and both make sense for investors with 3–5 year horizons who believe AI spending will continue to accelerate. AIS is the higher-conviction, higher-concentration bet. CHAT is the more diversified generative AI play with familiar names.
For investors who want to diversify across the AI economy: Consider holding both CHAT and AIS rather than choosing one. Their portfolios have almost no overlap — CHAT owns software and platform companies, AIS owns semiconductors and infrastructure. Together, they provide broader AI theme coverage than either does alone.
For investors focused on physical automation and robotics: BOTZ remains the most established dedicated vehicle for that theme. Just go in with clear eyes: this is a different bet than owning AI software or infrastructure. Its return profile will diverge significantly from CHAT and AIS in both directions over time.
Bottom Line: Picking the Right AI ETF for 2026
The lesson from CHAT vs BOTZ vs AIS is not that some AI ETFs are good and some are bad — it's that "AI ETF" is a label broad enough to cover almost anything. CHAT owns ChatGPT's closest relatives. AIS owns the power and memory that ChatGPT runs on. BOTZ owns the robots that may eventually run the factories building all of it. The right fund depends on which part of that chain you want to own — and how soon you need it to deliver. In 2026, the market has been very clear about which one it's rewarding first.
Four leading AI models discuss this article
"The 2026 winners reflect cyclical AI capex, not a durable alpha, so a diversified, multi-theme exposure is prudent."
Today's piece crystallizes a familiar risk: labels like 'AI ETF' can be dangerously misleading. AIS, CHAT, and BOTZ track very different parts of the AI economy, and 1H 2026 performance largely reflects cyclical capex and sentiment on infrastructure and software, not a single sustainable theme. AIS's double‑digit rally relies on hyperscale data-center demand that could decelerate; CHAT's outperformance hinges on active bets, which can underperform in drawdowns; BOTZ remains exposed to manufacturing cycles. The missing context includes full-year results, liquidity, and macro/regulatory risks. Relying on 2026 year-to-date as a guide is irresponsible.
Even if AIS and CHAT have outperformed so far, the gains may prove temporary; a downturn in data-center capex or AI hardware pricing could erase much of AIS's earnings gains. Also, active management risk in CHAT and concentration risk in AIS could undercut performance in adverse markets.
"The market is currently pricing the physical infrastructure of AI as a defensive, high-growth utility, which provides a more sustainable foundation for returns than the speculative software-layer premiums found in broad AI funds."
The 2026 performance divergence between AIS, CHAT, and BOTZ highlights a critical transition: the market is shifting from 'AI hype' to 'AI utility.' AIS’s 122% return is a classic industrial-supercycle play, capitalizing on the massive, non-discretionary CapEx required for data center power and cooling. While CHAT offers exposure to the software layer, it carries significant valuation risk if hyperscalers pull back on software spending to prioritize infrastructure. BOTZ is currently a value trap, suffering from cyclical manufacturing headwinds, but it remains the only play on the eventual physical integration of AI into the global supply chain. Investors must understand that these are not 'AI' funds; they are distinct sector bets on different phases of the compute stack.
The massive outperformance of AIS could signal an over-extended infrastructure bubble, where current prices already discount years of future utility-grade earnings growth, leaving it vulnerable to a sharp correction if data center build-outs hit regulatory or energy-grid bottlenecks.
"AIS's +122% return reflects a cyclical capex surge that the article mischaracterizes as structural, creating significant mean-reversion and concentration risk for new entrants."
The article conflates correlation with causation. Yes, AIS (+122%) crushed BOTZ (+1%) in H1 2026, but this reflects a cyclical capex surge, not structural superiority. The article treats infrastructure spending as durable, yet hyperscaler capex is notoriously volatile—it spikes during deployment waves, then contracts. CHAT's +66% is presented as 'proven' but sits between the two, suggesting mean reversion risk. The real issue: all three funds are levered to AI hype. When sentiment shifts (regulatory pressure, margin compression, capex pullback), the performance ranking could invert. AIS's youth ($914M AUM, launched Dec 2024) and concentration in cyclical infrastructure suppliers create downside tail risk the article minimizes.
If hyperscaler capex sustains at 2026 levels through 2027-28 (plausible given inference scaling), AIS's infrastructure thesis compounds and BOTZ's 'cycle' argument becomes a multi-year headwind, not a temporary dip.
"AIS's infrastructure bet carries higher concentration and liquidity risks than the article acknowledges despite its 2026 returns."
The article correctly highlights how CHAT, AIS, and BOTZ deliver divergent exposures—generative AI platforms, power/memory infrastructure, and robotics—explaining the 2026 performance gap. AIS's +122% return tracks real hyperscaler capex tailwinds in semiconductors and data-center equipment, while BOTZ lags on slower industrial automation cycles. Yet the piece underplays AIS's December 2024 launch, $914M AUM, and 0.75% fee, which embed liquidity and concentration risks that could amplify drawdowns if AI buildout slows. Investors should weigh whether infrastructure earnings revisions can sustain multiples already pricing in multi-year growth.
AIS's infrastructure thesis could unravel faster than expected if hyperscaler capex peaks in 2027 or memory prices collapse, turning its recent outperformance into a classic momentum trap rather than a durable edge.
"AIS liquidity and concentration risk could trigger non-linear downside if capex slows, not captured by the current cyclicality narrative."
Claude’s take rightly flags capex cyclicality, but he underweights AIS-specific liquidity and concentration risks. AIS launched Dec 2024 with ~$914M AUM; in a drawdown, redemptions could force forced selling and compress liquidity, widening spreads and depressing multiples even if macro capex remains intact. That non-linear dynamic isn’t captured by a 'cyclical tailwinds' thesis and could invert performance quickly if hyperscaler spend cools or memory pricing slips.
"The physical energy-grid bottleneck provides a structural floor for infrastructure-heavy ETFs like AIS that software-centric funds lack."
ChatGPT is correct about liquidity, but everyone is missing the energy-grid bottleneck. AIS isn't just a capex play; it is a bet on power availability. Even if hyperscaler software spending stalls, the physical infrastructure for data centers is constrained by local utility capacity, not just corporate budget cycles. This creates a 'floor' for AIS that software-heavy funds like CHAT lack. The risk isn't just volatility; it is a regulatory wall hitting the entire sector.
"Energy constraints redirect capex flows but don't prevent demand destruction—the real floor is inference economics, not grid availability."
Gemini's energy-grid constraint is real, but it's a *ceiling*, not a floor. Utility capacity limits *where* data centers can expand, not *whether* capex continues—it just redirects to grid-constrained regions or forces capex into power infrastructure itself (benefiting AIS). The actual floor risk is demand: if AI inference economics deteriorate or hyperscaler ROI collapses, capex stops regardless of available grid. That's the underexplored tail.
"Grid bottlenecks raise power costs that could accelerate AI capex cuts beyond pure demand weakness."
Claude rightly flags demand collapse over grid limits as the binding floor, but this misses the cost channel: utility constraints will raise power prices, eroding hyperscaler AI inference margins and hastening capex cuts. Those cuts would land first on AIS's concentrated suppliers, turning ChatGPT's liquidity risk into rapid multiple compression if earnings revisions follow.
The panel agrees that 'AI ETFs' like AIS, CHAT, and BOTZ are not equivalent and expose investors to distinct sector risks. They caution against relying solely on 2026 YTD performance, highlighting cyclical capex, liquidity, and demand risks.
AIS's exposure to the physical infrastructure for data centers, which may have a 'floor' due to utility capacity constraints.
Capex cyclicality and potential demand collapse, leading to forced selling and multiple compression in AIS due to its liquidity and concentration risks.