The Truth Behind the "Gold Rush" in the Humanoid Robot Industry: Manufacturers of the Complete Robots are Losing Money, While Suppliers of the "Hands" are Making Big Profits
Hello everyone, I'm your financial journalist. Today, we're going to discuss a topic that's currently very popular but perhaps not fully understood by many people – humanoid robots.
If you've been following the news recently, you might have noticed a paradox: on one hand, tech companies are racing to develop robots, claiming they will "change the world"; on the other hand, financial reports show that the vast majority of robot manufacturers are in huge losses, with the situation getting worse the more they sell.
So, where is all the money going?
It's similar to the California Gold Rush of 1849 – it wasn't the people who dug for gold who made the real fortune; it was the ones who sold shovels, jeans, and water. In the current humanoid robot industry, it's the suppliers of the "essential components" who have become the real winners, especially the "hands" of the robots – the dexterous hands.
Let me break down this phenomenon in simple terms and show you the underlying business logic.
---
The Grim Situation for Robot Manufacturers: Increased Sales, but No Profit
First, we need to face a harsh reality: most manufacturers of complete humanoid robots are not making money right now.
In the past, people thought that as long as you built a robot and sold it, you were successful. But the market logic has changed. Domestic manufacturers are competing fiercely by cutting prices to gain market share. Sales have indeed increased from 20,000 units to 70,000 to 80,000 units, but this is achieved at a substantial loss.
Here are some real examples:
- New Song Robotics: A well-established state-owned company that sold over 13,000 robots last year, a 65% increase in sales, yet it lost 398 million yuan – and the loss was even greater than the previous year. This is a classic case of "economies of scale not working in their favor"; the more they sell, the more they lose.
- EFT: Listed on the STAR Market, it has been in the red for six consecutive years, with a cumulative loss of nearly 1.4 billion yuan.
- Yibixuan: Although its loss margin is declining, it still lost 790 million yuan last year.
- Other players: Companies like Xianggong Intelligence and Ledong Robotics either see no increase in revenue despite increased sales or lose as much as what other companies earn in three years.
Why is this the case?
Manufacturers are facing three major challenges:
1. Inadequate technology: Robots are not yet efficient enough to perform tasks effectively, and their versatility is limited (they need to be retrained for different applications).
2. High costs: To mass-produce robots, prices must be lowered, but the core components are very expensive.
3. Lack of viable use cases: No clear, large-scale profit-making applications have been identified yet.
Therefore, manufacturers are in a phase of "burning money to secure the future," with their profits squeezed to the minimum.
---
Who Is Making the Money? The Surprising Profits of Upstream Suppliers
Since manufacturers are not making money, the suppliers of components are doing much better. Especially the dexterous hands, which are the most valuable, technically challenging, and profitable parts of humanoid robots.
Why are the hands so expensive?
- High cost proportion: According to Morgan Stanley, the material cost of the dexterous hand in Tesla's Optimus robot accounts for about 17.3% (approximately $8,650 to $10,380). A teardown of Yushu Technology's G1 robot shows that the dexterous hand also accounts for around 18% of the cost.
- High unit price: A high-end dexterous hand with 22 degrees of freedom can cost 20,000 to 30,000 yuan to purchase. With additional research and development and assembly, selling them for tens of thousands or even hundreds of thousands of yuan is quite normal.
- Robust profits:
- Weihong Co., Ltd., which produces joints for dexterous hands, saw its net profit increase by 48% in the first half of the year.
- Leisai Intelligence, which focuses on robotics, had a 62% increase in net profit in the same period.
- Fulai New Materials: Its orders for tactile sensors jumped from a few thousand sets to 100,000 to 200,000 sets, indicating imminent performance growth.
Why the hands, and not other components?
The hands are the key to a robot's interaction with the physical world. Previously, robots were impressive for being able to walk or dance, but now the market demands they can perform practical tasks (such as tightening screws in factories or sorting packages).
- Human hands account for 60% of the body's functionality, and fingers account for 90% of hand functionality.
- A robot's ability to use tools flexibly and perform precise tasks depends entirely on its hands.
- Therefore, the dexterous hand is considered the "last millimeter" towards making robots more versatile and the "decisive factor" in achieving embodied intelligence.
---
Why Are Dexterous Hands So Difficult to Develop? High Technical Barriers
You might ask: If they are so profitable, why don't manufacturers build their own dexterous hands instead of buying them from others?
The answer is that it's extremely difficult, and the technology is still not fully established.
There are mainly three types of dexterous hand technologies, each with its pros and cons, and there's no clear standard:
1. Linkage type: Similar to mechanical arms, simple in structure, low cost, and durable, but not very flexible (like a pair of pliers).
2. Tendon cable type: Works like human tendons, using cables to move fingers; very flexible and biomimetic, but with a short lifespan and complex maintenance.
3. Direct drive type: Motors directly control the fingers, providing precise control and high integration, but requiring high-quality motors and being the most expensive.
The key point is that the industry has not yet reached a consensus on a standard technology. This means that companies with the core know-how in any of these areas have significant influence.
- There are many factories that can produce basic joint modules, but only a few can stably mass-produce 16-degree-of-freedom dexterous hands in China.
- This scarcity gives upstream suppliers strong bargaining power, leaving manufacturers in a weaker position and dependent on these specialized suppliers.
---
The Hidden Cash Flow Generator: Dexterous Hands as "Consumables"
Many people overlook this: dexterous hands are not just installed and then forgotten; they are consumables.
- Short lifespan: Current tactile sensors usually fail after around 100,000 uses, and the overall lifespan of the robot is only 2 to 3 years.
- High wear and tear: In Tesla's Optimus, a dexterous hand costs over $6,000 but lasts for only about 6 weeks during parcel sorting.
- Doubling demand: A robot typically comes with two main hands, and additional hands are needed as backups. This means the demand for these parts doubles immediately.
What does this mean?
Even if robot sales are slow, once the robots start working, the demand for maintenance, replacements, and spare parts creates a steady stream of revenue.
- Manufacturers make a one-time profit from selling the robots (although they are still losing money).
- Suppliers of dexterous hands, on the other hand, have a steady source of income (similar to the razor-blade model).
This "consumable" nature gives dexterous hand manufacturers more stable financial prospects than robot manufacturers.
---
Future Outlook: Will It Be the Suppliers or the Core Players Who Dominate?
Let's look ahead. The competition over dexterous hands is just beginning.
- Huge market potential: The global market for dexterous hands is expected to grow from $1.16 billion in 2021 to $3.035 billion by 2030. In China, policies are clearly focusing on dexterous hands as a key technological breakthrough.
- No single company will dominate: Similar to the new energy vehicle industry, no single robot manufacturer will control all aspects of the industry.
- Supply chain leaders will emerge: Similar to companies like CATL and Luxshare in the battery and assembly sectors, there will be leading suppliers in the humanoid robot industry.
- Dexterous hand manufacturers that can enter overseas supply chains or excel in specific applications (such as industry or healthcare) will be able to sustain themselves and make profits in the long term.
- Risks and Warnings:
Despite the promising prospects, there are also risks to be aware of:
- Technical barriers: Rodney Brooks, a robotics expert from MIT, warns that robots currently have fewer than 200 tactile sensors compared to the 17,000 in humans. Lacking tactile data is the biggest obstacle to AI advancements in robotics.
- Bubble risk: If the technology cannot break through the "performance-cost-reliability" dilemma, the massive investments made so far could be in vain.
In summary:
In the first half of the humanoid robot industry, manufacturers are competing to produce robots; they are burning money to develop the technology. In the second half, upstream component suppliers are competing to mass-produce and sell these components, earning stable profits.
Currently, it's the suppliers of dexterous hands who are in the best position, standing at the forefront of this industry and reaping the benefits. For investors and industry insiders, focusing on suppliers with core technology and solutions to issues like lifespan and cost will provide a clearer view of the real profit flow in this industry.