Unitree Robotics to Be Included in the World's Largest Humanoid Robot ETF
Nashnova编辑部
After listing on Shanghai's STAR Market, Unitree Robotics will be added to KOID — the first U.S. and world's largest humanoid-robot ETF — run by KraneShares. The fund holds 51–58 global stocks across the full supply chain, with notable Chinese exposure, yet commercial deployment timelines remain far longer than market hype suggests.
What is KOID, and why does Unitree's inclusion matter?
KraneShares will add Unitree to two ETFs: KSTR, which tracks the STAR Market and is dual-listed on the NYSE and LSE, and KOID, listed on Nasdaq.
KOID is the first U.S. and largest global humanoid-robot thematic ETF, tracking the MerQube Global Humanoid Robot & Physical AI Index. This means → Unitree is not just gaining a listing label; it is entering the core pricing vehicle for the global robotics sector.
CEO Jonathan Krane called the IPO a milestone, citing a Morgan Stanley forecast: over 1 billion robots deployed worldwide in the next 25 years, a market worth $5 trillion.
What does this ETF actually hold?
KOID does not bet on a single robot maker. It covers the full humanoid-robot supply chain, holding 51 to 58 global stocks.
By sector: tech (semiconductors and electronics) accounts for roughly 45%, industrial manufacturing about 33%, and consumer goods about 13%.
By region: North America and Asia each take about 41%, Europe about 15%. In plain terms = this is neither a "China robotics fund" nor a "U.S. AI fund" — it bundles the entire global supply chain into one allocation tool.
How large is China's weight inside the fund?
Chinese supply-chain presence is pronounced: Kewell (servo drives), Luxshare Precision (precision structural parts), and Keli Sensing are among the A-share names already held; Leaderdrive and SHG cover precision-gear components.
Krane explicitly stated he will set no cap on Chinese holdings — "Chinese companies are already world leaders in core components." He sees China's edge in actuators (parts that convert electrical signals into mechanical motion), servo motors, and batteries, while the U.S. leads in software and AI — a natural complement.
On potential FCC restrictions on optical modules and robots, Krane argues commercial logic will prevail. He points to the EV supply chain: finished cars struggle to enter North America, yet battery makers like Gotion High-Tech have already built U.S. factories. The robotics chain, he says, will follow the same pattern.
Where does the technology actually stand?
Hardware is relatively mature. The weak link is the "brain." Unitree's WVLA 2.0, released in May 2026, merges a world model with a VLA architecture — an AI framework that lets a robot perceive its environment and execute actions simultaneously — but industry insiders acknowledge clear shortcomings.
Key gaps: no standardized benchmark for continuous-task success rates; limited generalization — robots still falter with varying table heights, lighting, or clutter; weak memory and error-handling over long task horizons.
This reflects a core tension: the gap between eye-catching social-media demos and real-world deployment is still substantial.
How far away is commercial deployment?
Multiple industry participants note that most achievements remain at the demo stage, with limited commercial use cases. Unstructured settings — hospitals, hotels, homes — are far harder than warehouses and factories.
In plain terms = the industry must first prove itself in controlled environments before it can talk about entering the messy real world; unit prices also need to fall further for adoption at the enterprise and consumer levels.
Krane himself concedes: the timeline for this journey is longer than the market's current enthusiasm. Unitree's IPO is a starting point, not a finish line.
Content is for reference only, not financial advice.