Zhejiang Humanoid received an order for 2,000 humanoid robots for the clothing industry

English translation from Unibot. Source published 2026-05-18.
The Zhejiang Humanoid Robotics Innovation Center announced on May 12 that it had signed a strategic partnership with Jack Technology and received an order for 2,000 humanoid robots adapted for sewing production. According to Gasgoo, the company described the deal as the first mass supply of humanoids in the global apparel industry. The announcement is significant because garment processing combines flexible materials, tight tolerances and repetitive manufacturing steps that have been difficult to automate with general-purpose humanoid systems.
Humanoid sewing robots face tough test
The author of the original article describes the clothing industry as a demanding testing ground for embodied AI systems. Fabrics vary in material and shape, and they can wrinkle, shift, and become deformed when handled.
The Zhejiang Innovation Center noted that alignment deviations for cut parts such as collars and pockets should be kept within ±2mm, while cutting and sewing operations require movement precision of 0.3-0.5mm.
These limitations help explain why garment manufacturing remains a challenging target for advanced automation. Unlike rigid parts in structured industrial cells, fabric requires continuous sensing and adjustments during manipulation. That is why the announced order is remarkable not only in scale, but also in the nature of the work performed.
NAVIAI development for clothing production
Zhejiang Innovation Center said it began specialized research and development of its NAVIAI humanoid robot in September 2024, focusing on the key processes of pattern machines in garment production lines. By July 2025, the company said the template machine's operating design had passed proof of concept.
According to Gasgoo, this included the full work cycle: opening and closing templates, fabric manipulation, stacking and unloading.
The company attributes its results to its proprietary hardware and software platform, as well as a proprietary sensor and control system for embodied intelligence. According to Zhejiang Innovation Center, the system achieves precise sensing and fine manipulation in fabric handling, and the flexible fabric separation success rate exceeds the industry average. This statement comes from the company itself; the source does not contain comparative benchmarks or independent validation.
Telecontrol and training data collection strategy
One of the important details of the report is the use of remote telecontrol for data collection. Zhejiang Innovation Center said that this has created a complete path for long-term data accumulation in garment manufacturing scenarios. This is especially important for humanoid designers because deformable materials are difficult to model in advance, and learning from repeated real-world operations is often the key to improving the quality of manipulations.
The company also noted that targeted training methods allow NAVIAI to work with fabrics of different sizes and compositions - from polo to denim. Based on the results presented, the robot performs single-step operations in less than 10 seconds and achieves alignment accuracy with placements of less than 2 mm on multiple layers of fabric. These metrics indicate an attempt to move from individual demonstrations to repeatable cycles and tolerances measured on the production line.
What does an order of 2,000 units mean?
The reported deal stands out because it ties the deployment of humanoids to a specific industrial workflow rather than to a general agenda of future labor automation. If deployed at the scale stated, it will place humanoids in a production environment where variability in materials is the norm and errors in accuracy quickly impact the quality of subsequent operations. This is a more demanding inspection than simple transport or container handling tasks.
However, several operational details remain unclear in the original article: the timing of the rollout, the number of production sites involved, and the proportion of the process that will be performed autonomously versus supervised work. Economic aspects are also not disclosed. These questions will be key in assessing whether this order will become a repeatable model for the apparel industry or remain a highly customized project.
Even with these gaps, the announcement shows that humanoid suppliers are targeting sectors outside of the automotive and warehouse industries. The clothing industry serves as a useful guide because success in it depends on the controlled handling of soft materials at industrial speeds and with high precision. If Zhejiang Humanoid can translate its development efforts into sustainable field operations, the entire industry will have a clearer reference point for understanding the role of humanoids in precision manufacturing.