Anyone who bought Roundhill Humanoid Robotics ETF (CBOE:HUMN) sought pure exposure to the humanoid robot buildout without diluting the theme with generic industrial automation names. HUMN concentrates on marquee humanoid brands, headed by an 8.93% weight in Tesla and a 6.42% position in UBTech Robotics. A second humanoid ETF, launched a year ago, has taken a different route through the same theme and pulled roughly 30 points ahead of HUMN in 2026. That fund is KraneShares Global Humanoid Robotics and Physical AI Index ETF (NASDAQ:KOID), and the gap is material.
Why Investors Bought HUMN in the First Place
The humanoid fund is the more familiar way to play this theme. Its portfolio reads like a highlight reel of the biggest names in the space, with Tesla at 8.93%, UBTech at 6.42%, XPeng at 4.74%, Shenzhen Dobot at 4.15%, and Rainbow Robotics at 4.14%. If a humanoid robot actually ships in volume this decade, most of the companies behind it are sitting right inside this fund. For investors who want a bet on named humanoid programs rather than the broader factory-automation ecosystem, that construction lines up pretty well with the stated objective.
Where the Concentration Backfired in 2026
The same construction put HUMN behind this year. Its top two positions alone approach 15% of assets, and both moved sharply lower in mid-year. HUMN gave back roughly 26% in July and finished year-to-date up only 11.0%, versus a 41.8% gain for KOID over the same stretch. The Tesla and UBTech weights that define HUMN’s identity dictate its short-term outcome. When the aspirational names are correct, little in the portfolio offsets them, because enabling-technology positions such as NVIDIA at 4.84%, Harmonic Drive Systems at 3.24%, and Teradyne at 3.58% are smaller.
What KOID Does Differently
The physical AI fund tracks an equal-weighted index of global humanoid robotics and embedded intelligence companies. Physical AI covers the sensors, actuators, motion-control chips, harmonic reducers, ball screws, and inference silicon that every humanoid program has to buy, no matter which brand ends up winning the consumer race. Because the fund is equal-weighted, the top 10 positions represent only 23.44% of assets, so a single stumble in a marquee name has a muted impact on the whole portfolio. That structure is exactly why this fund absorbed the same July that took 26% out of the humanoid fund and still holds a 23.15% one-year gain through August 25.
Costs favor KOID as well. Its most recent prospectus lists a 0.79% gross and 0.69% net expense ratio as of July 31, 2026. HUMN does not report an expense ratio in its most recent filing, so verify cost comparisons against the current prospectus before switching, but the fee gap is unlikely to be the deciding factor. The performance driver is portfolio construction.
Real Tradeoffs to Weigh
The physical AI fund is a diluted humanoid bet by design. Its equal weighting pulls in factory-automation and physical-AI names that would benefit even if humanoid programs stall out, which is a feature for most investors but a limitation for anyone wanting direct leverage to Tesla’s Optimus or UBTech’s Walker programs. The humanoid fund provides that leverage more directly. The physical AI fund also carries new-fund risk per its prospectus, which means wider bid-ask spreads and greater NAV premium or discount risk than a larger, more established vehicle. The humanoid fund, with $994.03 million in assets, is the more liquid of the two.
How to Think About the Swap
The switch is cleanest in a tax-advantaged account, where selling HUMN creates no taxable event. In a taxable account, holders who bought HUMN near its 52-week low of $29.17 and sit close to the current $30.63 price have modest embedded gains and can move without friction; anyone who bought near the $38.17 high is sitting on a loss that could be harvested. A partial swap, keeping HUMN for the pure humanoid tilt and adding KOID for supplier diversification, is one construction some investors have used.
Reading the Gap Going Forward
The 30-point gap between KOID and HUMN this year reflects equal weighting and component exposure rather than directional views on Tesla or UBTech. If humanoid unit shipments accelerate and marquee names recover, HUMN’s concentration works in reverse, and the gap narrows. For investors whose thesis is the buildout of the physical AI supply chain rather than a specific brand winning, KOID currently expresses that view more directly, and the year-to-date numbers reflect the difference.
Contact [email protected] for any questions or corrections.
Top Gaining Stocks
Top Losing Stocks