The orders for humanoid robots are "a tiny bit" numerous, while the deliveries are "a little bit" scarce.
During this year's "Double 11", intelligent robots became the "top stars" on major e - commerce platforms.
JD.com data shows that in the first week after the start of "Double 11", the transaction volume of intelligent robots increased by 5 times year - on - year. On the Tmall platform, intelligent robot brands such as Weilan, Yun Shenchu, and Ubtech all achieved double - digit growth. Not only on e - commerce platforms, since October, procurement amounts such as 126 million yuan, 260 million yuan, and 630 million yuan have appeared in the announcements of various humanoid robot companies time and time again. The industry is using real orders to prove the speed of its commercial implementation.
"Difficulty in Delivery" Challenges Mass - Production Capability
On the consumer side, the sales of intelligent robots are very popular. A reporter from "IT Times" noticed that on the JD.com platform, the sales page of the newly launched Xiaobumi robot by Songyan Power shows that the first batch of 500 units has been sold out. For the new R1Air robot from Unitree, more than 3,000 people made pre - orders before its official sale. At the same time, many humanoid robot sales pages are marked with estimated delivery times such as "Expected to be in stock in May 2026" and "The first batch will be delivered starting from March 2026".
There are also more and larger orders on the B - side. Statistics show that this year, there have been 17 orders worth over 10 million yuan for embodied intelligence and humanoid robots, and 6 companies including Zhipu, Ubtech, and Songyan Power have received orders worth over 100 million yuan.
When orders come in like a tide, the next competition is about who can complete key links such as mass production, quality control, and after - sales maintenance first and win with delivery capabilities. Among them, Zhipu Robot has performed quite well, mass - producing more than 2,000 general humanoid robots in 2025, which is at the leading level in the industry.
Putting Zhipu aside, the reality is that the production capacity of most humanoid robot companies may not match such a large order volume. Take Ubtech, a leading humanoid robot company, as an example. This year, Ubtech has won large orders multiple times. Just for the Walker series of humanoid robots, it has received orders worth over 630 million yuan throughout the year. The set annual delivery target is several hundred humanoid robots, but the specific delivery data has not been announced yet.
Looking back at the actual delivery volume in 2024, Ubtech only delivered 10 humanoid robots and suffered a huge loss of 1.16 billion yuan.
Globally, Elon Musk also planned to reach a mass - production target of one million units by the end of 2026. However, as of the third quarter of 2025, the trial - production scale was still less than 1,000 units. It can be seen that the mass - production and delivery of humanoid robots still face many difficulties.
"What matters is not the size of the order amount, but the actual closed - loop delivery ability." Chen Chen (a pseudonym), an industry insider from Digital China, told a reporter from "IT Times". Currently, the mass - production and delivery of humanoid robots face two major bottlenecks: technology and cost.
No Highly Mature Scenarios Have Emerged
From the perspective of orders, currently, the large - scale implementation of humanoid robots is mostly concentrated in industrial manufacturing scenarios. On October 16, Zhipu Robot released its new - generation industrial - grade interactive embodied operation robot, the Elf G2, and received orders worth hundreds of millions of yuan. The implementation scenarios are in automobile parts production lines and consumer electronics manufacturing. Recently, Ubtech has also reached cooperation agreements with several automobile companies and plans to deploy humanoid robots in automobile manufacturing factories.
Yao Maoqing, the president of the Embodied Business Department of Zhipu Robot, told a reporter from "IT Times" at the Elf G2 press conference that, first of all, in industrial manufacturing scenarios, there are pain points such as poor labor stability and high personnel mobility, which affect industrial production capacity and quality. Secondly, the industrial environment is relatively controllable, and the operation instructions are clearer, making it an ideal testing and application place for humanoid robots.
Even though the industrial scenario has many advantages, it alone cannot support the development of the entire humanoid robot industry. Therefore, some orders have also flowed to scenarios such as culture and tourism, scientific research, and finance. The Xingxingxia robot created by Digital China was the first to be implemented in scenarios such as banks, museums, and business halls.
Chen Chen said that compared with the industrial scenario, which requires extremely high technological investment, scenarios such as banks and business halls also have standardized business requirements, such as business consultation and document delivery, which can also match the current development level of robot intelligent interaction. No matter what the scenario is, enterprises need to accurately capture and deeply explore real needs. Nowadays, every company wants to create its own characteristic scenario and become the next "Unitree". Liu Yizhang, the person - in - charge of the joints of the National - Local Joint Embodied Intelligent Robot Innovation Center and the person - in - charge of Tianguang Robot, said bluntly that there are still obvious deficiencies in scenario implementation. No mature application scenario that can enable long - term and stable operation of humanoid robots has emerged, which has become the current main weak link.
Zheng Liyao, the general manager of Gaogong Consulting, also said that when humanoid robots are implemented on production lines, they are limited by rhythm and efficiency, and the engineering difficulty is still very high. At the same time, the value scale of the industry is rising rapidly. Simple handling is now regarded as a low - level task, and investors' expectations have shifted to more delicate and intelligent application scenarios such as folding clothes and shelf sorting.
Beware of "Hype" Behind Large Orders
Of course, the humanoid robot industry has reached the initial stage of product delivery and commercial implementation. The "associated orders" that once appeared in the new - energy vehicle industry are also beginning to emerge in the humanoid robot industry.
The so - called associated orders refer to those where some orders come from affiliated parties of shareholder enterprises or group enterprises. For example, Zhipu Robot and its shareholder, Longcheer Technology, plan to deploy nearly a thousand robots on PC and tablet production lines.
"At the current early stage of the industry, the emergence of associated orders does not necessarily mean there is false content." Chen Chen told the reporter. At present, the humanoid robot market and production capacity are both on the rise. Associated orders can relieve the financial pressure on enterprises to a certain extent, help enterprises test and quickly enter the market, and accumulate technical experience and reputation. On the other hand, it can also attract the attention of investors to obtain more financial support and accelerate product iteration and optimization. In essence, it is a relatively safe exploration method in the early stage of industrial development.
However, it should be noted that those orders with only a framework agreement but no detailed information such as delivery volume and delivery time, and with a large amount, may have "hype".
"Judging from the current development status of the industry, asking full - sized robots to perform high - load, fast - paced, and long - endurance work contains more conceptual elements." Yao Maoqing also believes that there may be "hype" in large orders. For example, a few months ago, many bipedal robots were popular for moving boxes in workshops. But after on - site industry research and understanding of customer needs, it is found that this is not a suitable solution because the robots' endurance, stability, and speed are difficult to meet the high - intensity requirements of factories.
This article is from the WeChat public account "IT Times" (ID: vittimes). Author: Shen Yibin, Editor: Wang Xin. It is published by 36Kr with authorization.