The humanoid robotics sector has captured investors’ imagination as a potential multi‑year growth story, promising machines that can walk, talk and perform tasks once reserved for humans. Two exchange‑traded funds that chase this narrative have taken markedly different routes, and the result in 2026 is a striking performance divergence of about thirty percentage points. One fund loads its basket with the most recognizable brand names that are building finished robots, while the other spreads its bets across the underlying components and subsystems that any robot—no matter the logo—must incorporate. This split reflects a deeper philosophical disagreement about where value will accrue: will the winners be the visible OEMs that brand the machines, or the quieter suppliers that provide the motors, sensors and control silicon? Understanding this fault line is essential for anyone trying to decide whether to place a concentrated wager on a few headline acts or to diversify across the broader enablement ecosystem. The following analysis walks through the construction of each fund, explains why the component‑focused approach has outperformed so far this year, and offers concrete guidance on how investors might position themselves given their own outlook on the timing and scale of humanoid adoption.
The Roundhill Humanoid Robotics ETF (ticker HUMN) was designed for investors who want pure play exposure to the companies that are actually assembling humanoid robots. Its portfolio is therefore heavily weighted toward a handful of high‑profile names: Tesla carries close to nine percent, UBTech Robotics accounts for just over six percent, and XPeng, Shenzhen Dobot and Rainbow Robotics each contribute roughly four to five percent. By construction, the fund avoids diluting the theme with generic industrial automation firms, instead concentrating on the entities that are most likely to benefit if consumer‑grade robots begin to ship in volume. This concentration gives the fund a high beta to the fortunes of those marquee names; when the market perceives progress in Optimus, Walker or similar programs, HUMN tends to rally sharply. Conversely, any setback—whether a missed milestone, a supply‑chain hiccup or a shift in investor sentiment—can translate into an outsized drag on the fund’s net asset value because there are few other holdings large enough to cushion the blow.
In the first half of 2026, exactly that scenario played out for HUMN. Tesla’s stock retreated amid broader concerns about valuation and execution timelines for its Optimus prototype, while UBTech faced pricing pressure in the Chinese service‑robot market. Together, the top two positions represent roughly fifteen percent of the fund’s assets, so when both slid in July the ETF lost about a quarter of its value in a single month. The remaining holdings—names such as NVIDIA, Harmonic Drive Systems and Teradyne—are meaningful but individually too small to offset the decline of the leaders. Consequently, HUMN finished the year‑to‑date period up only eleven percent, a modest gain that pales beside the double‑digit returns posted by many broader technology indices. The episode illustrates a classic risk of concentrated thematic bets: high upside potential paired with vulnerability to idiosyncratic shocks that affect the few names driving the majority of performance.
Contrast that with the KraneShares Global Humanoid Robotics and Physical AI Index ETF (ticker KOID), which takes a deliberately different tack. Rather than loading up on finished‑robot makers, KOID tracks an equal‑weighted index of companies that supply the sensors, actuators, motion‑control chips, harmonic reducers, ball screws and inference silicon that every humanoid platform must purchase. Because the index is equally weighted, the ten largest holdings together constitute less than twenty‑four percent of the fund’s total assets, meaning that no single company can dominate the portfolio’s movement. This structure intentionally captures the “physical AI” layer—the hardware and embedded intelligence that enable perception, locomotion and decision‑making—regardless of which OEM eventually wins the consumer race. In effect, KOID is a bet on the indispensable build‑out of the robotics supply chain, a thesis that holds even if the timeline for mass‑market humanoid slips or if several OEMs share the market.
The equal‑weighting approach proved its worth when the same July turbulence that hammered HUMN struck the market. While Tesla and UBTech dipped, the breadth of KOID’s holdings meant that the impact of any one stock’s decline was diffused across dozens of other positions. Companies such as key semiconductor firms, precision‑mechanics makers and industrial sensor producers either held steady or even rose on expectations of continued capital spending on automation. As a result, KOID absorbed the shock with far less volatility and still posted a year‑to‑date gain of roughly forty‑two percent through August, outperforming HUMN by more than thirty points. The fund’s one‑year gain through the same date stood at about twenty‑three percent, demonstrating that the equal‑weighted, component‑focused strategy not only limited downside but also captured upside from the broader automation rebound that accompanied the humanoid narrative.
Cost and liquidity considerations further differentiate the two vehicles, although they are unlikely to be the primary driver of the performance gap. KOID’s most recent prospectus shows a gross expense ratio of seventy‑nine basis points and a net ratio of sixty‑nine basis points as of July 2026. HUMN does not publish an expense ratio in its latest filing, so investors should consult the current prospectus before making any switch; however, even a modest difference in fees would struggle to explain a thirty‑point return disparity. On the liquidity front, HUMN holds just under one billion dollars in assets, translating into tighter bid‑ask spreads and a lower likelihood of significant premiums or discounts to net asset value. KOID, being a newer and smaller fund, may exhibit wider spreads and occasional NAV deviations, which could add a modest transaction cost for active traders. For long‑term, buy‑and‑hold investors, these liquidity nuances are typically secondary to the fundamental portfolio construction that drives returns.
Tax efficiency also plays a role when contemplating a move between the two funds. In a tax‑advantaged account such as an IRA or 401(k), selling HUMN triggers no immediate tax consequence, making a reallocation frictionless. In a taxable account, the decision hinges on the investor’s cost basis. Those who purchased HUMN near its fifty‑two‑week low of around twenty‑nine dollars and now sit close to the current price of thirty‑one dollars have only modest embedded gains and can shift without generating a large tax bill. Conversely, buyers who entered near the fifty‑two‑week high of thirty‑eight dollars are sitting on a loss that could be harvested to offset other capital gains. A pragmatic tactic some investors employ is a partial swap: maintaining a core position in HUMN to retain direct leverage to Tesla’s Optimus or UBTech’s Walker, while adding a satellite allocation to KOID for broader supply‑chain exposure. This blend lets investors capture upside from a potential OEM breakthrough while cushioning the portfolio against the volatility that can accompany any single brand’s stumble.
To appreciate why the component‑focused approach has prevailed thus far, it helps to view the humanoid robotics market through the lens of its supply chain. Before a robot can walk off a factory floor, it needs a suite of precision parts: high‑torque actuators, low‑backlash gearboxes, high‑resolution encoders, force‑torque sensors, powerful inference chips and robust communication buses. These components are sourced from a diverse set of suppliers that often serve multiple end markets—industrial automation, automotive, aerospace and consumer electronics—thereby providing a degree of natural diversification. When headlines focus on a glitzy demo or a missed milestone, the underlying demand for these building blocks frequently remains intact because OEMs continue to prototype, test and iterate. Moreover, many of the component manufacturers are benefiting from secular trends such as the rise of edge AI, the proliferation of collaborative robots in logistics and the ongoing push for factory digitization. Hence, even if the timeline for mass‑market humanoid adoption slips, the physical‑AI layer can still experience steady growth driven by adjacent automation needs.
Looking forward, the relative performance of the two ETFs will hinge on how quickly humanoid robots transition from pilot programs to volume production. If OEMs such as Tesla, UBTech and XPeng succeed in scaling shipments over the next eighteen to twenty‑four months, the concentrated bets in HUMN could begin to pay off handsomely, narrowing or even reversing the current gap. In that scenario, the fund’s high exposure to the winners would amplify returns, while the equal‑weighted KOID might lag as its diversified holdings capture only a fraction of the OEMs’ upside. Conversely, if the path to volume remains fraught with technical hurdles, regulatory delays or soft consumer demand, the component suppliers may continue to enjoy steady orders from OEMs that are still investing in research and development, keeping KOID’s performance robust while HUMN stagnates or declines. Investors should therefore calibrate their allocation according to their conviction about the timing of commercialization: a bullish outlook on near‑term volume justifies a larger tilt toward HUMN, whereas a more cautious or long‑term view favors a core position in KOID with a smaller, tactical exposure to the OEM‑focused fund.
From a practical standpoint, investors can implement several approaches to align their portfolios with their outlook. First, define the investment horizon: if you are prepared to hold for three to five years or longer, the supplier‑centric KOID may serve as a stable core that captures the secular growth of automation and edge AI. Second, assess risk tolerance: those uncomfortable with the potential for double‑digit monthly swings might limit HUMN to a modest satellite allocation—perhaps ten to fifteen percent of the thematic exposure—while placing the remainder in KOID. Third, consider using a barbell strategy: pair a small, high‑conviction position in the OEM‑focused fund with a larger, diversified base in the physical‑AI fund, rebalancing semi‑annually or when the weight drifts beyond predefined bands. Fourth, keep an eye on valuation metrics; if HUMN’s price‑to‑sales or price‑to‑earnings multiples become stretched relative to its growth prospects, it may be prudent to take profits and redeploy into KOID. Finally, remember that thematic investing works best when paired with a disciplined process: set clear entry and exit rules, monitor key fundamentals, and avoid letting short‑term noise dictate long‑term strategy.
To inform those decisions, monitor a set of leading indicators that can signal shifts in the humanoid robotics ecosystem. Watch for announcements of volume production ramp‑ups from major OEMs, especially any disclosure of unit shipment targets for the next fiscal year. Track capital expenditure plans of automotive and electronics manufacturers that are investing in collaborative‑robot cells, as these often precede larger humanoid orders. Follow the sales trends of component makers: rising revenues for harmonic reducers, precision ball screws, high‑bandwidth motor controllers and AI inference chips can hint at increased robot building activity. Pay attention to regulatory developments—safety standards for collaborative robots, liability frameworks for autonomous machines and any government subsidies for advanced manufacturing can accelerate adoption. Lastly, consider macro‑economic cues such as interest‑rate trends and corporate confidence indices, which influence willingness to fund costly automation projects. By synthesizing these data points, investors can form a more nuanced view of whether the market is pricing in an imminent OEM breakthrough or a longer, steadier build‑out of the enabling infrastructure.
In summary, the thirty‑point performance chasm between the two humanoid‑robot ETFs in 2026 is less a clash of opposing views on Tesla or UBTech and more a reflection of how portfolio construction shapes outcomes when a theme is still evolving. The OEM‑focused HUMN offers direct leverage to the success of a few headline names but carries the attendant concentration risk. The supplier‑oriented KOID spreads bets across the essential hardware and software layers, delivering smoother returns and capturing growth from the broader automation wave even if humanoid shipments lag. For investors seeking exposure to this exciting theme, a sensible starting point is to allocate a core position to KOID to capture the inevitable build‑out of physical‑AI infrastructure, then layer a smaller, tactical position in HUMN if you believe the OEMs are poised for near‑term volume. Keep the allocation aligned with your time horizon and risk appetite, use tax‑efficient vehicles when possible, and rebalance periodically to maintain the intended risk‑return profile. Staying informed on both OED progress and component‑sector health will allow you to adjust the mix as the story unfolds, positioning you to benefit whether the winners turn out to be the robot brands or the parts that make them move.