After Jiang Zheyuan shouted, "We've only run 10 meters," CATL invested 1 billion yuan, but the story for the C-end market might still be a trap.

A bionic robot with Cai Ming's face chatted with the real Cai Ming, who played a robot on stage 30 years ago, on the Spring Festival Gala stage. Before the spotlight had completely dimmed, the feedback from the capital had already arrived.

On March 2nd, Songyan Dynamics announced the completion of its Series B financing, with a cumulative scale of nearly 1 billion yuan. The lead investor was Chendao Capital, a subsidiary of CATL. So far, this company, which has been established for only two and a half years, has completed 9 rounds of financing and carried out a shareholding system reform.

Jiang Zheyuan repeatedly emphasized a metaphor during an interview: The global embodied intelligence industry is currently in its early stage, "like running only ten meters in a marathon." This metaphor is quite ingenious - it not only admits that the industry is still in its infancy but also implies that they have already taken the lead.

However, the problem lies in how to run these "ten meters" and in which direction. The choices of various companies are diverging. Jiang Zheyuan is betting on the consumer - grade market. The "Xiaobumi," priced under 10,000 yuan, targets families with children, trying to leverage scale with a low price. This strategy sounds attractive, but as of 2026, the author tends to believe that consumer - grade humanoid robots are a false proposition in the short term. The industrial scenario is the real touchstone, and the fact that Songyan Dynamics received investment from CATL precisely indicates that its real destination is not the family but the factory.

Industrial capital enters the market, not for the imagination of the C - end

Let's start with the financing structure. The fact that Chendao Capital led the investment is more worthy of consideration than the 1 - billion - yuan amount itself.

Chendao is an industrial investment platform of the CATL system, deeply involved in the fields of clean energy and intelligent manufacturing for a long time. What does CATL aim for by investing in robots? It aims for production line automation, warehousing and logistics, and even the extension of future energy storage scenarios - all of which require "hands" and "feet" to complete the final link. Industrial capital is not as easily impressed by stories as financial investors. They focus on production line adaptability, supply chain integration capabilities, and the possibility of large - scale mass production.

This is not an isolated case. Looking through the financing cases since 2025, Qianxun Intelligence has CATL behind it, Zibianliang Robot was led by SAIC Financial Holdings, and Zhifangping attracted Baidu Strategic Investment and CRRC Capital. Industrial capital is flowing into this track with an unprecedented density, and they are all investing in the same type of companies - those that can be implemented in industrial scenarios.

Songyan Dynamics currently has two product lines: "bipedal + bionic." It has accumulated more than 30 patents. In Jiang Zheyuan's plan, bionic robots are for strong interaction and strong emotional resonance, aiming for the final C - end. However, what CATL values is obviously not the face that can imitate Cai Ming, but the motion control ability and mass - production foundation of bipedal robots.

 A simple logic: If Chendao really believes that the consumer - grade market will take off soon, it should invest in channels, brands, and content ecosystems, rather than a hardware company. Its investment in Songyan Dynamics can only indicate one thing - The certainty of the industrial scenario is much higher than that of the family scenario.

"Xiaobumi" is priced at 9,998 yuan, but the family scenario lacks a "brain"

What really made Songyan Dynamics well - known is "Xiaobumi" - priced at 9,998 yuan, 94 cm tall, and weighing 17 kg, targeting "families with children." Jiang Zheyuan compares it to a "high - end BBK learning machine," a product that can enable children to have an embodied intelligence enlightenment.

This positioning is very smart because it avoids a core problem: Current humanoid robots simply cannot do household chores.

Jiang Zheyuan himself admitted that there is a core obstacle for robots to enter families - data. Although simulation data can make the rendering very similar to the real world, the physical engine is not accurate enough, and there are deviations between the modeling of collisions, soft bodies, and fluids and the real world. He prefers to use real - machine data. As long as the data problem is solved, the "brain" problem will naturally be solved.

But the question is, where does the real - machine data come from? It can only come from actual scenarios. This is a typical "chicken - and - egg" dilemma: Without enough data, robots cannot do their jobs; if they cannot do their jobs, they cannot be deployed on a large scale; without large - scale deployment, enough data cannot be collected.

 Jiang Zheyuan's attitude towards this is practical: In the short term, let "Xiaobumi" enter families to play the roles of companionship, interaction, and education, such as accompanying children to learn English, do exercises, run, and learn to program robots. In essence, this is using the expansion of the C - end market to drive the optimization of supply chain costs in reverse, while accumulating data and scenario awareness.

Can this logic work? Maybe, but it takes time. And the patience of the capital market is limited. Jiang Zheyuan himself also said that if the data bottleneck is broken through, every family may have a humanoid robot in the future, which can do laundry, cook, and clean at home. But that is the finish line, and now they have only run ten meters.

In these ten meters, it's okay to sell a few thousand units of "Xiaobumi," but to sell hundreds of thousands or millions of units, it requires not only a low price but also the ability to do something useful.

The industrial scenario is the real touchstone, but Songyan Dynamics has not proven itself yet

Now let's look at the industrial side.

Jiang Zheyuan revealed in an interview that currently, the order volume of Songyan Dynamics is in the thousands. The main implementation scenarios are scientific research, education, commercial performances, and cultural and tourism exhibition halls, with the greatest demand coming from education and commercial performances. He also mentioned that there are some new library customers, "Compared with wheeled robots, these customers prefer bipedal robots."

These scenarios have one thing in common: They are all "display - type" demands, not "production - type" demands. Buying robots for commercial performances and cultural and tourism exhibition halls is essentially to attract attention, to "sell traffic." Some analyses point out that from the perspective of brand placement, buying robots has a certain cost - effectiveness ratio. The price of a robot may be the budget for a consumer brand to hold an offline marketing event, but the freshness of C - end users towards robots and the online popularity can last for a long time.

However, although the traffic business can be done, for robot brands, it is more of a transitional commercialization plan in the early stage. Once the supply in the market increases and people get used to humanoid robots in shopping malls, the market demand based on the "sell traffic" logic will no longer exist.

What can really test the ability of robots is the industrial scenario. Zhao Tongyang, the founder of Zhongqing Robot, frankly said that although the industry is developing rapidly, the relevant products are still in the small - batch trial - production stage, and there are blockages in the supply chain, and problems such as consistency need to be solved. Chen Jianyu, the founder of Xingdong Jiyuan, shared a typical case: A batch of humanoid robots had inconsistent walking postures. After investigation, it was found that the glue was unevenly applied during motor assembly.

What does this indicate? It indicates that there are countless details between "being able to make it" and "being able to mass - produce with good consistency." And these details can only be truly exposed and solved in large - scale industrial scenario deployments.

Songyan Dynamics has production bases in Changping, Beijing, Changzhou, Jiangsu, and Dongguan, Guangdong. Jiang Zheyuan said that the main pressure now lies in production and delivery. This pressure is exactly what the industrial scenario needs.

So the question is: After receiving investment from CATL, will Songyan Dynamics tell stories in the C - end market or tackle tough problems in the industrial scenario?

The author's judgment is the latter. Because only the industrial scenario requires CATL's industrial resources; only the industrial scenario can verify the ability of large - scale mass production; only the industrial scenario can keep the "marathon that has just run ten meters" going.

Instead of asking who can enter families, it's better to ask who can survive in factories

2026 is regarded by the industry as a crucial year for the commercialization of humanoid robots. A research report from Guojin Securities predicts that the shipment volume of domestic leading manufacturers is expected to increase from thousands to tens of thousands. TrendForce is even more radical, predicting that the global shipment volume will exceed 50,000 units, a year - on - year increase of more than 700%.

However, shipment volume and delivery volume are two different things. Jiang Zheyuan defines 2026 as a crucial node for "scale - up," believing that the industry will break through the "thousand - unit - level" delivery bottleneck of last year and officially aim for the "ten - thousand - unit" mark this year. The current order volume of Songyan Dynamics is in the thousands, which is not small in the industry. But to truly gain a foothold, it needs to prove two things: First, it can continuously obtain orders; second, it can deliver products with high quality and quantity.

Jiang Zheyuan's goal is to become a leader in at least one industry segment within three years and then gradually expand into new scenarios. This goal is very practical, without shouting slogans like "let robots enter every household."

Looking back, Jiang Zheyuan's statement that "the marathon has just run ten meters" is indeed true. But how to run the next ten meters is much more important than the previous ten meters. Will it continue to tell stories in the C - end market or tackle tough problems in the industrial scenario? To some extent, the fact that Songyan Dynamics received investment from CATL has already given the answer.

As for whether "Xiaobumi" can really enter families, that is a matter at the finish line of the marathon. Now the most important thing is to reach the next supply station.

This article is from the WeChat official account "Cai Neng Quan", author: Sijiali, published by 36Kr with authorization.