36Kr Publishes Article on Unitree’s Ambition to Lead the Humanoid Robot Market

On January 23, 2026, 36Kr published an article about Unitree’s ambition to lead the humanoid robot market.

The available publication information shows only that 36Kr published an article about Unitree and the humanoid robot market. The excerpt provided does not confirm whether “the lead” in the headline refers to shipments, revenue, operational performance or technical capability, nor does it identify the comparison group or assessment period. The article therefore cannot be treated as proof of Unitree’s market ranking or competitiveness; at this stage, it should be read as raising issues around the market. For the article’s full argument, see the original 36Kr report. One way to consider evaluation criteria is through a view of physical AI as a field made up of multiple layers, discussed in Business+IT Publishes Experts’ Discussion on Physical AI “World Models”, published by this newspaper. The excerpt provided, however, does not specify the world model’s technical approach or the discussion’s conclusions. Unitree should likewise be assessed using verifiable metrics for each layer, including which tasks its robots can perform and under what conditions, and how their operation is measured after deployment, rather than by hardware performance alone. The excerpt contains no figures that would allow such comparisons. Another way to assess commercial viability is to examine how robots are offered and the data from their actual use. leaderobot.com Publishes Article on Optimus’s Offering Model and Robot Rentals focuses on rentals, but its summary does not confirm results or profitability. The availability of rentals alone is not enough to judge adoption or revenue potential; information such as utilization rates and maintenance costs is needed. Similarly, BitRobot Releases a 100-Hour 3D Task Dataset gives the dataset’s size, but the summary does not establish how it relates to the training or task performance of any specific robot. The existence of data should be evaluated separately from the ability to reproduce tasks on a physical robot. Accordingly, connecting the ambition to lead with a specific competitive position will require examining comparison groups and metrics alongside task conditions, reproducibility on physical robots, operating status and offering models. These are not confirmed for Unitree in the excerpt, and examples involving other companies or datasets cannot be attributed to the company as evidence of its performance.