Global X Robotics & Artificial Intelligence ETF
Global X (Mirae Asset)
Tracks the Indxx Global Robotics & Artificial Intelligence Thematic Index — industrial automation, non-industrial/surgical robots, and autonomous vehicles. Concentrated (~60-65 holdings, top 10 ~60%) and Japan/Europe-tilted rather than US-megacap.
Top holdings
Swiss industrial automation
Japanese factory-automation sensors
Japanese industrial-robot leader
Main US AI-compute exposure, capped near ~8% at rebalance
Surgical robotics
Pneumatics/automation
Performance
July 2026: 7.2% price: $37.94 (Jun 30) -> $35.22 (Jul 31), financialcontent daily closes — 5yr NAV +1.82% ann. and since-inception +10.46% ann. are still Jun-30-2026 issuer vintages; no Jul-31 standardized table yet. July was an AI-hardware rout (SOX -20.6%) in which NVDA held up (+0.3%), yet BOTZ's industrial core fell ~7.2% — the non-AI tilt gave no shelter. A Jul 30-31 bounce ($33.68 -> $35.22) did not recover the month.
Recent moves
The March annual rebalance trimmed NVIDIA back toward the ~8% single-name cap; no structural change.
Our take
The purest 'physical robotics' play of the AI group — its industrial-automation core (ABB, Keyence, Fanuc) lagged the 2026 AI-compute trade and the fund has roughly round-tripped its 2021 peak, so it is a bet on a robotics capex cycle that hasn't yet re-accelerated. Iconic and liquid, but a flat 5-year record is the honest headline.
The structure explains the record better than the theme does.
A revenue-purity screen selects companies that *sell* robots — Japanese and Swiss capital-goods manufacturers, plus a surgical-robotics sleeve — and largely excludes the companies capturing AI economics through software and compute. So the fund is levered to machine-tool order books and factory capex, not to model training. Two mechanical consequences the marketing omits. First, currency: with Keyence, Fanuc, SMC, Daifuku and ABB unhedged, a meaningful share of any given year's return is a JPY/CHF translation effect independent of robotics fundamentals. Second, concentration without conviction: the top 10 is roughly 59-60% of assets, yet the largest positions appear capped near 9.6%, so BOTZ neither diversifies like a broad basket nor lets its best idea run. It is not leveraged and has no daily reset, so compounding decay is not a mechanism here — the drag is thematic mis-specification plus a 0.68% fee on a top-heavy set of large, liquid single names.
- 2026-07Issuer holdings at 31 Jul 2026 put Keyence at 10.59% — above the ~9.6% level inferred as a cap, and ahead of ABB 9.33% and NVIDIA 8.85%. Aggregators still carry Jun-30 weights, so providers disagree.
- 2026-07July 2026 inverted the attribution: in an AI-hardware rout (SOX -20.6%) NVIDIA held at +0.3% and Broadcom at +3.1%, yet BOTZ fell 7.2%. The industrial core, not the capped AI sleeve, was the drag.
Thesis
BOTZ is marketed as an AI fund but built like an industrial one: 45.8% Industrials vs 35.2% Info Tech (Global X, Jun 30 2026). Its dominant exposure is the global factory-automation capex cycle — ABB 9.64%, Keyence 9.41%, Fanuc 8.64%, SMC 4.60% — priced mostly in CHF and JPY, plus a ~9% surgical-robotics sleeve led by Intuitive Surgical at 5.97%. The AI label rests on a capped NVIDIA position (9.31%) and Alphabet (2.45%). Read as an AI-compute vehicle, the label misdescribes the holdings; read as a machine-tool and medical-robotics cyclical, the structure is internally consistent.
Tracks the Indxx Global Robotics & AI Thematic Index: modified market-cap weighting across 62 holdings (issuer, Jul 2026), screened on revenue from industrial automation, non-industrial and surgical robots, and autonomous vehicles. Observed weights top out near 9.6%, consistent with a single-name cap that trims a winner like NVIDIA rather than letting it compound — though the cap level and reconstitution schedule are not published on the issuer page.
Assessment
- Genuine non-US exposure — Japanese and Swiss automation leaders most US-domiciled AI funds hold at zero weight.
- Nearly 10 years live (inception Sep 12 2016) with 10.46% annualized NAV since inception through Jun 30 2026.
- ~$3.19B AUM and tight liquidity make it a usable, low-friction expression of the automation theme.
- Rules-based and transparent: a published revenue-purity screen with observable weight caps, not discretionary stock-picking.
- Sector reality contradicts the ticker: 45.8% Industrials vs 35.2% Info Tech means this is a capital-goods fund wearing an AI label.
- Unhedged JPY/CHF exposure injects a currency return the thesis never mentions and the investor is not compensated for.
- Top 10 is roughly 59-60% of assets, but caps near 9.6% prevent winners compounding — concentration risk without concentration upside.
- Index reconstitution cadence is not disclosed on the issuer page, so the lag between a new robotics entrant and its inclusion is unknowable from public materials.
- 0.68% buys index construction, capping and rebalancing across a top-heavy set of large, individually liquid holdings.
Record
Issuer NAV returns to Jun 30 2026: 1yr +16.24%, 3yr +9.96%, 5yr +1.82% ann., since-inception +10.46% ann. The 5-year number is the honest headline — roughly flat real return across the most explosive stretch in AI history, because the industrial core round-tripped the 2021 thematic peak while the AI move accrued to compute and software the index screens out. Attribution is lopsided: the NVIDIA sleeve is doing disproportionate work in the 1- and 3-year figures, and it is capped. Window sensitivity is severe — stockanalysis.com shows a trailing 1-year of 3.79% to Jul 20 2026 versus the issuer's standardized 16.24% to Jun 30. That ~12-point gap reflects a different measurement window (rolling trailing vs month-end standardized period), moving both the start and end points, not three weeks of price action. Treat any single trailing number here as noise.
- 2026-07Trailing one-year total return is +5.01% to 31 Jul 2026, against the issuer's +16.24% NAV figure standardized to 30 Jun — roughly 11 points lower one month on. The Jun-30 table is a prior vintage.
- 2026-07That gap between the trailing and standardized figures was not mostly window mechanics: BOTZ fell about 7.2% on price during July 2026 ($37.94 to $35.22). Real price action drove much of the spread.
Risks & fit
- A stalled industrial capex cycle (Japanese machine-tool orders, China factory automation) caps the fund regardless of AI progress.
- Yen or franc weakness silently erodes USD returns even if the underlying businesses execute.
- Chinese exposure (Shenzhen Inovance, 4.07%) carries policy, listing-access and localization-substitution risk.
- The ~9% surgical-robotics sleeve adds hospital-capex and reimbursement risk unrelated to either AI or factory automation.
- Thematic-flow reflexivity: a 2016-vintage ticker draws momentum money at peaks and redemptions at troughs, worsening entry timing.
The mis-specification read fails if the industrial core starts leading rather than lagging: Japanese machine-tool orders and Fanuc/Keyence order books inflect to sustained year-over-year growth, and forward returns broaden so the industrial sleeve — not the capped NVIDIA position — drives the gain. If the 5-year annualized figure climbs back toward the 10.46% since-inception rate on breadth rather than one holding, the critique is wrong. Conversely, if physical robotics commercializes and the value still accrues to compute and software, the revenue-purity screen is structurally on the wrong side of its own theme.
Suits an allocator explicitly wanting ex-US industrial-automation and capital-goods exposure, accepting unhedged JPY/CHF as part of the return stream, with a horizon long enough to sit through a capex cycle. Poorly matched to anyone seeking AI-compute or AI-software exposure — the index screens most of that out. Also a mismatch for portfolios already holding NVIDIA at size, since the capped 9.31% sleeve duplicates existing exposure while diluting the industrial thesis.
0.68% expense ratio (issuer, Jul 2026). Peer set: ARTY (formerly IRBO) 0.47%, ROBT 0.65%, ROBO 0.95%. BOTZ sits mid-pack — the fee buys index construction, capping and rebalancing over a portfolio whose top names are large, liquid single stocks.