Wall Street Insights: On July 9, Swancor announced that Shanghai Zhiyuan Xinchuang Technology Co., Ltd. (controlled by Shanghai Zhiyuan Hengyue Technology Partnership and its affiliate Shanghai Zhiyuan Xinchuang Technology Equipment Partnership) will acquire at least 63.62% of its shares, with the transaction valued at no less than 1.996 billion yuan. Upon completion of the transaction, the controlling shareholder and actual controller of Swancor will change to Zhiyuan Hengyue and Deng Taihua. If this deal goes through, Swancor will become the first company under a humanoid robot enterprise listed on the STAR Market.

As an important part of China's capital market, the STAR Market has limited financing capacity and market scale. If Zhiyuan Robot chooses to list on the STAR Market through a "backdoor listing," it may indicate that its future development is constrained by the capital scale of the STAR Market. Could this strategy limit the growth space for the humanoid robot industry in China?

Humanoid robots, as an emerging technology, are still in the exploratory stage regarding their application scenarios. According to relevant research, fields such as hazardous chemical explosion prevention, mining, smelting, nuclear power, power grids, firefighting and emergency response, and rescue operations are considered the best initial application scenarios for humanoid robots.

However, do we really need humanoid robots?

From Iron Man to Real Construction Sites

Tesla's Optimus is "practicing running" and "cooking," Boston Dynamics' Atlas is organizing documents and doing somersaults in factories, and Figure 02 is sorting packages in logistics centers.

Many domestic robot companies have also made appearances, from performing with handkerchiefs and martial arts at the Spring Festival Gala to running in robot marathons and fighting competitions.

These scenes look a lot like the future we imagined as children: robots integrating into human society like humans, working in factories, helping humans move things, cook, walk dogs, patrol... as if once this bionic revolution succeeds, human physical labor could completely exit the stage.

But stepping back and returning to the essence of the matter, there are two unavoidable questions that need to be answered:

Why must robots "resemble humans"?
Does China really need to follow this path?

Before answering these questions, let's look at some comparative data between humanoid and non-humanoid robots:

Comparison of Financing Scale: Capital Concentration vs. Total Investment

Humanoid Robot Field:
- Figure AI
  - A single round of financing reached $2.34 B (2025), with a valuation of approximately $2.6 B (Business Insider, Wikipedia).
- Currently raising a new round of $1.5 B, targeting a valuation as high as $39.5 B (Cinco Días, ipo.club, Reddit).
- Although capital is concentrated, there are only a few leading companies, and the overall financing scale is far smaller than that of industrial automation.
Traditional Robots (AGV/Robotic Arms):
- Global Industrial Robot Market
  - The 2024 market size is about $20–27 B (MarketsandMarkets, Precedence Research, Cognitive Market Research), expected to exceed $50–84 B by 2030–2034 (thelogisticsiq.com).
- AGV/AMR Market
  - About $2 B in 2024, expected to grow to $6 B by 2033 (LinkedIn, Business Research Insights, thelogisticsiq.com).
- Investments come from countless enterprises and small manufacturers, with broad coverage and high market maturity.

Despite the astonishing single-project financing for humanoid robots, the overall market size is far smaller than that of industrial and logistics robots; the two are not comparable.

Unit System Cost and R&D Threshold

Humanoid Robots:
- Complex structure: High degrees of freedom, thousands of components, extremely high system coordination difficulty (Business Insider).
- Sensors, actuators, control systems, etc., are relatively expensive, with unit costs typically in the hundreds of thousands of dollars—component costs alone can reach $35k–60k (China vs. Tesla advantage).
AGV/Robotic Arms:
- Stable unit prices: AMRs generally cost $20k–150k, single robotic arms cost tens of thousands of dollars.
- High integration, high standardization, clear ROI, short system deployment cycles, and low maintenance costs.

Humanoid robots have high R&D costs and complex systems, while traditional robots offer high cost-effectiveness and mature engineering paths.

Market Adoption Speed and ROI Realization Rate

Humanoid Robots:
- Most are still in the R&D or testing phase, not yet mass-produced.
- Although Figure AI has validated through cooperation with BMW, large-scale commercial returns have not yet materialized (Wikipedia, Reddit).
AGV/Robotic Arms:
- Widely used in manufacturing, warehousing, logistics, welding, and other scenarios.
- ROI can be recovered within 1–3 years, with rapid deployment and full market validation.

Traditional robots are adopted quickly, yield fast returns, and have clear commercialization paths; humanoid robots are still in the experimental stage. Humanoid robots are not mainstream, not even "second best."

The U.S. Bets on Humanoids, a Result of Structural Concessions

From the comparative data, non-humanoid robots are the market mainstream. The so-called "humanoid robots" seem more like a capital story for the future. However, a key issue cannot be ignored here—labor.

Even with highly advanced automated production lines, European and American countries still cannot completely eliminate their reliance on human labor. A large amount of mid-to-low-end manufacturing still requires industrial workers. The cost of the cheapest automated production equipment can easily reach hundreds of thousands, which in high-income countries is still dozens of times the cost of human labor.

Moreover, this situation has spilled over from manufacturing to the service industry. The U.S. has managed to stand firm amid high inflation, thanks to a large immigrant population supporting basic positions across various industries.

The Guardian: "US undocumented farm workers feel ‘hunted like animals’ amid Trump’s immigration raids"

According to a report by The Guardian, industries in the U.S. such as agriculture, construction, sanitation, catering and food processing, manufacturing and meat processing, and transportation have a high proportion of undocumented immigrant employees in low-wage, low-skill threshold positions.

U.S. manufacturing has gradually hollowed out over the past few decades, lacking people, production capacity, and supply chains—it has a vast military-industrial complex and the world's most advanced automated production equipment, but it cannot produce enough basic components to supply its domestic needs. Its industrial upgrade only stays in high-end advanced industries.

On the surface, humanoid robots are the ultimate form of "intelligent terminals." But from the perspective of politics and industrial reality, the U.S. choosing this path is actually a passive choice. Under the dual impact of labor shortages and immigration on local socio-cultural values, "humanoid robots" seem to be the optimal solution to all problems.

Why Can't They Rebuild the "Ecosystem"?

Three constraints trap the U.S. in the "humanoid" path:

Labor costs are too high: Shortage of mid-to-high-end industrial workers, high wages, making frequent trial and error difficult.
Reliance on imported raw materials and components: Sensors, motors, batteries mostly come from Asia.
Environmental and geopolitical constraints: High barriers to local manufacturing, and decoupling from China makes it difficult to connect with China's vast production capacity.

The biggest problem troubling the U.S. is human resources. High labor costs, powerful unions, and a severe shortage of engineers force the U.S., including European countries, to bet their future on humanoid robots. Unless they can introduce more skilled industrial workers to complete a full industrial chain upgrade—to solve the "lack of people" problem, they need more "people" to support it.

At the same time, with the U.S. domestic supply chain, it is difficult to create a robot ecosystem driven by efficiency. So they can only create an "electronic human" that can use shovels, tighten screws, and hand over goods—essentially adapting to the old physical world, not reconstructing a new system.

Having failed to upgrade mid-to-low-end industries, the U.S. can no longer support a fully localized industrial ecosystem based on unmanned equipment.

They can only think this way:

Since I can't create a bunch of small tools, I'll create a person who can use tools.

This is a seemingly advanced but actually somewhat helpless substitution logic. It is also the inevitable result of the U.S.'s failure in industrial upgrading and the global subcontracting of production and manufacturing.

This Is More Like a Capital Dream Than an Industrial Revolution

Figure AI's valuation keeps inflating; no matter how clumsy Optimus's movements are, each new video release elicits a stock price response. In the U.S., humanoid robots are more like a future story told to the capital market.

They are in no hurry to commercialize, nor do they have the capability to do so in the short term, with no clear C-end or B-end demand scenarios. On the contrary, their capital logic is very clear:

As long as the story is good, someone will buy it.

Similar to autonomous driving, the metaverse, and brain-computer interfaces, humanoid robots are becoming the next baton in the relay of "technological singularity." But what truly lands and commercializes are ecosystems like Amazon's logistics robots—those non-human machines that follow first principles, with form determined by function.

Humanoid ≠ Efficient. Making Tools Use Tools Is a Reversal of Efficiency
Essentially, humans invent tools to replace part of our capabilities, thereby improving efficiency.

Hammers replace hand hammering.
Washing machines replace hand washing.
AGVs replace manual搬运.

But when we create a robot that resembles a human and then have it operate a wok, push a cart, or tighten screws, we are doing something absurd: making a tool use another tool.

Humans invent tools to extend the body and improve efficiency. Creating a "human" to operate another tool may be less efficient than manual operation—this is anti-efficiency in engineering.

This is a typical case of "taking off pants to fart"—unnecessary, high cost, low efficiency, and adding countless maintenance and failure points.

The U.S. lacks the full industrial chain capability to complete the intelligent upgrade of production and manufacturing, so they choose the humanoid robot path. But that doesn't mean it's the right path.

China Doesn't Need This Romance

Unlike the U.S., China's manufacturing industry itself is a high-frequency, distributed, flexible, and intelligent system.

We can mass-produce搬运 AGVs, smart trash cans, delivery robots, automated sorting centers... even in residential areas, factories, and buildings in first-tier cities, various "non-humanoid" smart devices are already in use.

We don't need a humanoid terminal to adapt to the environment; we have the ability to transform the environment.

Blindly Following the U.S. Industrial Path Is a "Copying Homework" Path Dependency

In the past, Chinese startups were accustomed to "copying U.S. homework": the U.S. proposes a new concept, we bring it back, and commercialize it locally—from early search, internet, e-commerce, to the mobile internet era's sharing economy and SaaS, almost all succeeded.

This approach was once very efficient:

The U.S. validates demand, China achieves scale and profitability, and finally cashes out by returning to the U.S.

But today, the technological gap between China and the U.S. has significantly narrowed. Even on the application side, China, with its vast user base and engineering capabilities, has developed a unique ecosystem.

We have the world's strongest manufacturing capabilities, the most complex user demands, and the most flexible application scenarios. Continuing to "blindly copy" at this point is likely inertia from path dependency, not judgment based on reality.

What we need is not to "build humans with the U.S.," but to develop our own technological path based on our own structure and market demands.

The STAR Market Is Not a Springboard, But a Ceiling

On the surface, Zhiyuan's backdoor listing seems like an "industry breakthrough signal"; but from a capital structure perspective, it is actually a warning:

The STAR Market has limited financing capacity and can hardly support companies of Figure AI's scale.
If the first company in the industry goes public through a "low-valuation backdoor listing," subsequent companies will use it as a "valuation anchor."
Investors are eager to cash out rather than continue deep cultivation, indicating they no longer believe this story can continue.

This is not a signal of "takeoff," but a posture of "preparing to land." Capital is ready to exit.

Don't let romance replace reality. Humanoid robots are not a false proposition, but they should not be China's main path. The U.S. takes this path because it has no other choice; we don't take it because we have choices.

What should truly be pursued is not the shell of "resembling humans," but the core of "efficiently completing tasks." An automatic cooking machine is far more reliable than a robot that can't cook well. A useful cleaning robot doesn't need to walk, dance, or practice Tai Chi like a human.

What we need is not to bring "future fantasies" into reality, but to transform reality with engineering capabilities.