Robotics ETFs and How to Actually Invest in the Physical AI Theme

What is a robotics ETF, and is it actually the best way in?

A robotics ETF is a basket fund that buys a range of companies tied to automation, industrial robotics, and — increasingly — physical AI and humanoid robots, so you get exposure to the theme without picking a single winner. For most retail investors this beats trying to hand-pick individual humanoid robot stocks, because the category is still young, few of the most talked-about companies (Figure AI, for instance) are even publicly traded, and the ones that are can swing 50%+ in a year on a single earnings call or a Musk tweet. The three funds that dominate search results and assets under management — BOTZ, ROBO, and IRBO — take noticeably different approaches to constructing that basket, and as of early July 2026 their year-to-date returns diverge so sharply that picking the “wrong” one has mattered more than getting the theme right in the first place.

That divergence is the first thing to understand before putting money to work: this is not a single trade. It’s three different bets wearing the same “robotics” label, plus a much smaller pool of pure-play stocks for investors who want direct exposure rather than a diversified wrapper.

The three flagship robotics ETFs, compared

Fund	Ticker	Expense ratio	~1yr return (as of early July 2026)	What it tracks
Global X Robotics & Artificial Intelligence ETF	BOTZ	0.68%	~14% (roughly 2% YTD)	Indxx Global Robotics & AI Thematic Index
ROBO Global Robotics & Automation Index ETF	ROBO	0.95%	strong 2026 performer; YTD cited as ~18–20% depending on source	ROBO Global Robotics & Automation Index
iShares Robotics and Artificial Intelligence Multisector ETF	IRBO	0.47%	~82% (roughly 54% YTD)	NYSE FactSet Global Robotics and Artificial Intelligence Index

Two things jump out. First, cost varies a lot for what is nominally the same theme — IRBO’s 0.47% expense ratio is barely half of ROBO’s 0.95%, which compounds meaningfully over a multi-year hold. Second, and more strikingly, IRBO was the standout 2026 performer among the three, up roughly 54% year-to-date and about 82% over the trailing year, while BOTZ lagged both its peers and the broader market, gaining only around 2% YTD versus the S&P 500’s roughly 9%. ROBO landed in between, though its exact year-to-date figure isn’t fully settled — one widely cited estimate puts it near 20% YTD (about 40% over twelve months), while data aggregators calculate it closer to 18.22% YTD. Either way, ROBO clearly outran BOTZ in 2026, even though the two funds charge comparable “thematic ETF” fees.

The lesson isn’t that IRBO is simply “the good one” going forward — index construction differences, not stock-picking skill, drive most of the gap. IRBO’s multisector index and lower cost happened to line up well with 2026’s rally in AI-adjacent industrials and semiconductor names; that alignment can just as easily reverse. What matters for due diligence is that “a robotics ETF” is not a single number, and the fund’s underlying index methodology explains far more of the return gap than its ticker or marketing name does.

What’s actually inside these funds?

All three funds hold a mix of industrial automation suppliers (think motion-control and sensor makers), semiconductor and AI-compute names, and a smaller allocation to pure-play robotics and humanoid companies. That mix is precisely why none of them is a clean bet on humanoid robots specifically — you’re buying a broad automation-and-AI-infrastructure basket that happens to include some robotics names, alongside chipmakers and software firms whose fortunes are also tied to data-center AI demand, cloud spending, and the broader tech cycle. This overlap with mainstream AI/semiconductor exposure is worth checking against whatever tech or AI funds you already own before assuming a robotics ETF meaningfully diversifies your portfolio; for readers new to the category, our explainer on humanoid robots is a useful primer on which companies are actually building the hardware these funds are betting on.

Humanoid robot stocks: the direct-exposure alternative

If a diversified ETF feels too diluted, the direct-exposure alternative is buying individual humanoid robot stocks — but the field of investable pure-plays is genuinely narrow. UBTECH Robotics listed on the Main Board of the Hong Kong Stock Exchange on 29 December 2023 under ticker 9880.HK, becoming the first humanoid-robot company listed on the HKEX main board, with its Walker S humanoid making a stage appearance at the listing ceremony. Walker S is UBTECH’s industrial humanoid, built with 41 servo joints and force feedback for smart-manufacturing tasks like logistics, assembly, and inspection — a concrete, shipping product rather than a demo reel, which is more than can be said for most names in this space.

Tesla is the other publicly traded name investors reach for, though buying Tesla stock is a bet on the whole company, not a pure Optimus play. Tesla ended production of the Model S and Model X in early May 2026 specifically to convert its Fremont assembly line to Optimus humanoid-robot manufacturing, with output expected to begin around late July or August 2026, initially at low volume. A second, higher-volume line targeting the Gen 4 Optimus variant is planned at Giga Texas for around summer 2027. That’s a real production timeline, not just a promise — but it also means Optimus revenue is still mostly ahead of the company rather than behind it, and Tesla’s share price remains driven primarily by its vehicle and energy businesses in the meantime.

Notably absent from any public market: Figure AI, arguably the most closely watched humanoid startup, which closed a Series C round on 16 September 2025 that exceeded $1 billion in committed capital at a $39 billion post-money valuation, led by Parkway Venture Capital with participation from Brookfield Asset Management, Nvidia, Macquarie Capital, and Intel Capital, among others. Figure has raised roughly $1.9 billion in total since its 2022 founding — real institutional conviction — but retail investors currently have no way to buy shares directly. For a broader map of who’s raising what across the sector, see our roundup of recent robotics funding rounds.

How big could this market actually get?

The bull case for staying in this theme long-term rests less on any single stock’s chart and more on the size of the market these companies are chasing. A Barclays Research report titled “The Future of Work: AI Gets Physical,” released 14 January 2026, forecasts the humanoid-robotics market growing from roughly $2–3 billion today to as much as $200 billion by 2035 under an optimistic scenario. The report attributes that growth path to a roughly 30-fold reduction in production cost over the past decade, driven by advances in AI reasoning, actuators, and batteries. That’s a genuinely large total-addressable-market claim from a mainstream sell-side research desk, not a startup pitch deck — but it’s also explicitly a forecast, and the same report frames it as an “optimistic” scenario rather than a base case, which is worth keeping in mind when translating a 2035 headline number into a 2026 valuation.

Per-unit cost projections for Optimus and similar platforms are even less settled: public estimates for near-term unit costs range from roughly $50,000 to $100,000, while long-run targets once production scales toward a million units a year are cited anywhere from $20,000 to $30,000. These are forward-looking figures built mostly on public statements rather than audited numbers, so treat them as directional rather than a spreadsheet input.

So how should you actually approach investing in this theme?

A reasonable framework looks less like picking a single “best” ticker and more like separating the decision into two layers. The diversified layer — a low-cost robotics ETF such as IRBO — gives broad exposure to the automation and physical-AI supply chain without requiring you to guess which humanoid platform wins, and its 0.47% expense ratio is a meaningfully lower cost drag than ROBO’s 0.95% for a multi-year hold. Recognize, though, that all three flagship funds also carry heavy overlap with semiconductor and AI-infrastructure names, so check for double-counting against existing tech holdings before treating “robotics ETF” as a diversifier on its own.

The direct-exposure layer — individual names like UBTECH or Tesla — is a smaller, higher-conviction satellite position rather than a core holding, given how few pure-play, liquid, publicly traded humanoid-robot companies actually exist today. UBTECH’s Hong Kong listing also means US investors are taking on considerations tied to a China listing (currency, disclosure standards, geopolitical risk) that a domestic ETF doesn’t carry. Tesla’s Optimus timeline is real but still pre-revenue at scale, with the higher-volume Giga Texas line not expected until around 2027, so the stock’s near-term price action will continue to be driven by its automotive and energy segments rather than robotics headlines alone.

Whichever layer you weight more heavily, size the position to reflect that this remains an early-stage, volatility-prone theme — the roughly 50-point performance gap between IRBO and BOTZ over a single year is itself evidence that thematic ETFs in this space are not interchangeable, and individual humanoid stocks can move even more sharply on production-timeline news or a single funding round. For the broader macro and policy backdrop shaping this market, our Physical AI market outlook and the Market, Investment & Regulation hub track how funding, regulation, and adoption are evolving together, and the glossary is a quick reference if any of the technical or fund terminology above is unfamiliar.

Robotics ETFs and humanoid robot stocks are not two versions of the same bet — they’re a diversified index play and a handful of concentrated, still-early single-company wagers, and conflating them is the most common mistake investors make when they first look at this theme. Decide which layer you actually want exposure to, check the fee and overlap details before assuming any two “robotics” funds are similar, and size individual stock positions bearing in mind that production timelines for Optimus, Walker S, and their eventual competitors are still measured in quarters and years, not weeks.