In recent years, industrial robots in the field of intelligent manufacturing is becoming more and more dazzling protagonists, on the one hand, the application field of industrial robots gradually from handling, welding, assembly and other operational tasks to more processing tasks, human-machine collaboration is becoming an important direction of industrial robot research and development;
China has long been the world's largest industrial robot market, and the density of robots in the manufacturing industry has also caught up with developed countries. According to the International Federation of Robotics (IFR), the density of robots in the EU in 2022 was 208 per 10,000 employees, 285 in the United States, and 392 in China.
So, is the development of industrial robots almost at the top? Not really. IFR data also shows that the countries with the highest density of industrial robots in 2022 are South Korea (1,012 units/10,000 employees), Singapore (730 units/10,000 employees) and Germany (415 units/10,000 employees), and the density of industrial robots in China is still far from South Korea, Singapore and other countries. Especially in South Korea, South Korea's industrial robots are mainly distributed in 3C electronics and automobile manufacturing, while China's two major industries are rising rapidly, and South Korea is competing globally, and China's industrial robots are far from reaching the ceiling.
China's 3C electronics, new energy vehicles and other fields need more domestic industrial robots. So, what stage has the domestic industrial robot industry progressed to? Recently, Yiou conducted an exclusive interview with Suo Liyang, general manager of Peitian Robot.
As a company founded in 2010, through the ups and downs of domestic industrial robot entrepreneurship, and positive research and development of industrial robot technology for nearly 14 years, Peitian Robot has developed a number of core technologies and has rich experience in the industry. Suo Liyang, general manager of Peitian Robot, also shared many industry characteristics, common problems and future prospects in the current field of industrial robots from the perspective of long-termism.
The following is the original text of the dialogue, with some modifications.
Through the cycle
The industrial robot industry must precipitate the technical background
Yiou: Since its establishment in 2010, Peitian Robot has also experienced more than ten years of vigorous development in the field of industrial robots in China. In the past ten years, has there been any phased changes in the entrepreneurship of China's industrial robot industry, and have you observed some cyclical trends or obvious entrepreneurial booms?
Suo Liyang: Around 2010, there were not many companies involved in the field of industrial robots across the country, and Peitian was one of them, and there were some players incubated within the car factory.
My feelings are not necessarily right - we, as early entrants, actually didn't feel the entrepreneurial boom of industrial robots in 2010. The main reason why some car factories make robots is that they have a large number of demand in the car workshop, and they did not go out independently at the beginning, but one of their departments is doing it, and they do not consider other external market opportunities, that is, they do it for their own use.
At the beginning, we made robots, although there was no particularly complete market research. But we didn't take much risk because we also had the support of the group behind us.
A large number of new brands and new entrants in domestic industrial robots have appeared in the three years from 2012 to 2014. At this time, intelligent manufacturing is already a relatively popular concept, and when it comes to intelligent manufacturing, industrial robots are naturally indispensable.
Yiou: So in the past three years, what impact has a large number of new entrants had on the industrial robot market?
Suo Liyang: Prior to this, we have been in the field of industrial robots for three or four years, and have exhibited our robot products in many exhibitions, and have won a lot of professional recognition.
With the beginning of the domestic industrial robot entrepreneurship tide, many brands have sprung up, some companies focused on the research and development of servo systems in the early stage, and later transformed into the robot industry. In general, compared with the "four families" of industrial robots, everyone is not much famous.
Our product features and marketing strategy actually began to take shape in 2016 and 2017. From our establishment in 2010 to 2016, we mainly focused on our own technology research and development, and the market did not know much about us, and our product sales have not yet begun.
Before expanding into the market, we have accumulated a lot of technology, and there are also some employees with entrepreneurial ambitions who start their own businesses when the entrepreneurial tide arrives. Therefore, our peers love us and say that we are the "Whampoa Military Academy" in the field of industrial robots. But on the whole, our technical team and core personnel have remained relatively intact over the years.
Family photo of Peitian robot products
Yiou: The payload range of Peitian industrial robot products covers 3 kg to 280 kg, and has been successfully applied in many industries and fields. So, what is the difference between the technical difficulty of small and large loads?
SOLIYANG: I think there are different levels of difficulty for large and medium loads. To be precise, in terms of the realization of cost-effective industrial robots with domestic core components, small and medium-sized loads are generally produced early, and large loads are produced late. Why is this so?
In fact, this is not purely a technical reason, but also related to the market and supply chain.
The most important reason for the late release of large loads is that the maturity of domestic parts is not enough, either you have to have a performance advantage, or you have to have a price advantage.
At the level of large loads, domestic parts, especially reducers, have only been reflected in the maturity of the last three or four years. Therefore, three or four years ago, the cost performance of domestic large-load robots was relatively low.
It's actually a chicken-and-egg cycle:
In the past, the market demand for domestic large-load robots was relatively small, and the corresponding models of parts were few. Due to the lack of models, insufficient scale effect, and the lack of cost advantages of domestic parts, a small number of domestic large-load robot manufacturers are more inclined to buy foreign parts, and this negative cycle has not begun to change until the last two or three years.
Regardless of the load, the challenges faced by robots are essentially the same. There are three challenges that need to be addressed: the first is performance, the second is functionality, and the third is reliability or what we call uptime. When users choose and use a bot, their primary concern is its reliability.
In terms of reliability, the reliability of the robot has been comparable to that of international giants
Yiou: Is there a suitable indicator to measure reliability?
Soliyang: The industry generally uses MTBF (Mean Time Between Failure) indicators to measure. However, there are many shortcomings in this indicator, and customers cannot verify it.
Yiou: So, if it is difficult to verify, how can the reliability level of domestic robots be measured and compared with international giants?
Suo Liyang: Actually, we still have to use three indicators to measure, performance, function, and reliability, not just one. For example, in terms of function, we are actually no different from the international robot giants, and the reliability is about the same, perhaps in terms of performance, the movement of the robotic arm is a little slower. In fact, there is no direct relationship between performance and reliability, your robot runs faster and slower, and running fast does not necessarily have more failure rate, these are two concepts.
Speaking of reliability indicators, in order for a product to achieve a long-term stable and trouble-free reliability level, it must accumulate a certain amount of user feedback and market feedback, and continuously optimize it on this basis. The process is quite long.
Because mechanical products are different from electronic products, the upgrade and iteration speed of electronic products is fast, and once the bugs in the software are found, they can be quickly fixed. Mechanical products, on the other hand, may work well for the first two years of user use, and suddenly have problems in the third year. When a problem arises, you may look back and see that the problem came from a flaw in the original design. So, it's probably been three or four years from the time the problem was discovered to the time the design was revised, and that's when you were able to make improvements to the structure. This process is difficult to compress, so in the field of industrial robots, long-term know-how accumulation is very important. I think the reliability of Peitian has been very close to the level of imported brands.
Application of SCARA in the field of 3C
Yiou: The demand of downstream customers for industrial robots must be very diverse. What are the main areas in which Peitian robots are landed, and how to optimize performance or functions according to the needs of integrators or downstream end customers?
Suo Liyang: Take cars as an example, different customers have different requirements for the performance of cars, and some customers do not pursue high performance, such as faster acceleration of 100 kilometers, but prefer to buy a reliable and long-term car. The same is true for the industrial robot market, where you may be able to build a very high-performance robot, but most customers are not very concerned about that. In many use cases, customers don't need much performance, just reliability.
Of course, there are still some manufacturers and customers whose industry characteristics require high-speed and high-precision performance, so as to improve the efficiency of the production line. Then we have to meet the requirements of customers with demanding performance requirements in terms of hardware and algorithms. Peitian does a better job in high-speed and high-precision.
There are many cases of our robots landing in the 3C industry. For example, Foxconn, Samsung, and Huawei all use our robots in their factories. Why do these customers prefer to use Peitian's products? The reason is simple: they are looking for efficiency throughout the production line. In the process of continuous optimization of the overall production line of mobile phones, the robot plays an important role in it, helping to improve the overall production efficiency.
Yiou: There are many giants in the industrial robot, what do you think of these giants?
Suo Liyang: We have also seen some industrial robots manufactured by industry giants themselves, but compared with similar products in the market, the reliability and performance often do not meet expectations. Why? Because companies ignore two things:
First, whether the enterprise has enough technical strength to do a good job in the product.
Second, if the enterprise is not fully engaged in the competition with the external market, it does not know how to optimize the supply chain to create enough competitive products, and the industrial robot department lacks initiative in the technical optimization, supply chain optimization and quality improvement of products, and this passive attitude may hinder the long-term development of the enterprise. Even if you want to take the initiative to do well, the lack of sufficient market competition will still affect the play of initiative.
As another example, we also provided OEM industrial robots for a Korean giant. In fact, the giant also has its own robot division, and has also tried to develop industrial robots by itself, but it cannot reach the level of the "four families" of industrial robots. In desperation, their robotics division eventually changed its strategy and found a supplier in China who was willing to provide them with OEM services.
Intelligent problems of industrial robots
The essence does not change its mechanical undertone
Yiou: There is a view that the 1.0 era of robots is biased towards the industrial genes of machinery and electromechanical, while the 2.0 era is mostly based on algorithms, computers, mathematics and other backgrounds. What do you think about the implementation of industrial robots in more vertical application scenarios empowered by AI and large models?
Suo Liyang: This question is actually what will happen to industrial robots after they become intelligent. I believe that the application of industrial robots and robotic arms in the future must be combined with advanced technologies in the field of artificial intelligence. But we still have to divide it into two to see, intelligence, is to establish an artificial intelligence control system, including visual recognition, decision-making, so that the robotic arm of the industrial robot to perform specific actions. As for the specific operation of the industrial robot body or robotic arm, it only needs to receive instructions sent by the artificial intelligence system. In fact, we are also doing technical research and development of large models of industrial robots.
In other words, the robot in the future AI era, its executive parts are still no different from the present, and the robotic arm is still a robotic arm, which needs the genes of the electromechanical industry. The difference is artificial intelligence perception and control systems, but these AI technologies do not actually belong to robotic arms.
Yiou: For the upstream of the industrial chain, Peitian has both industrial robot ontology and core components (servo drive, servo motor, etc.) business. What is the relationship between the two? Is it that if an industrial robot company wants to gain a firm foothold, the three core components must be self-developed?
Soliyang: This question is the same logic as whether OEMs should produce their own robots. There are a few dimensions to consider here. First of all, if you produce your own parts, can you ensure that the parts business grows healthily and without relying on blood transfusions? If not, then for a long time, this component will be eliminated from the market.
It's not that you can't do it, but you have to think clearly about whether the parts you produce can continue to participate in market competition and maintain competitiveness in terms of technology and cost. If these conditions are met, then it is feasible.
We didn't make reducers, we used to want to do it ourselves, but then we gave up. The main reason for abandonment, first, is that the additional investment is particularly large. Second, the technical personnel required for the reducer are very different from the needs of industrial robots.
Equipment can not be shared, talents can not be shared, technology can not be shared, you have to look at this matter according to the independent company, do I want to set up an independent subsidiary to do this? But if you set up an independent subsidiary to make reducers, and you are still doing industrial robot ontology, then due to the existence of competition, your parts sales will be affected.
Medium load robot welding applications
Yiou: You have pointed out in previous interviews that Peitian is "not only an industrial robot ontology manufacturer, but also an industrial robot system integrator", and also provides customers with a complete set of system integration solutions. This is for the downstream, so how to balance the roles of the ontology business and the integrator?
Soliyang: Actually, there are relatively few cases where Ontology's customers compete with our integration business.
First of all, integrators are usually industry-specific. The main industries we serve as an ontology role are not involved in our role as an integrator. In this way, at the industry level, we avoid direct competition with our customers. For example, the application of ontology is very common in the 3C industry, but we as integrators basically do not set foot in the 3C industry, but focus on the pre-assembly field of the pharmaceutical industry, the chemical industry, and even the construction industry.
Yiou: How big is the difference between the size of the two markets of industrial robot ontology and integrators, five times and ten times?
Soliyang: More than that. I think the market size of integrators is at least 100 times that of the ontology market. Therefore, even if the integrator wants to do the 3C industry, the probability of encountering a customer of our ontology is very small.
Ontology and integration, it's not easy to compete directly, which is one reason why we have both.
The second reason is that the gross profit margin of integration is higher than that of ontology, so choosing ontology in parallel with integration is also a good supplement to the cash flow of the enterprise.
Yiou: Do you think the concept of humanoid robots, which has become popular recently, belongs more to the category of robot ontology manufacturing, or does it belong more to the category of integration?
Suo Liyang: From the perspective of the humanoid robot product itself, the two core problems have not been solved at present, which limits their application in reality, whether it is industrial or non-industrial scenarios.
One is the reliability of humanoid robots, and the other is the high cost of humanoid robots. These two problems directly hinder their widespread application. At present, the application of humanoid robots is still very early, and even the work of screwing is not as good as that of humans or ordinary industrial robots.
The reason why this industry is hot is because it is widely believed that one day humanoid robots will reach the level of high reliability, low cost and high intelligence in the future. Since all the facilities and tools in society are designed based on human use, humanoid robots will have the opportunity to do a lot of delicate work instead of humans. This is the judgment of capital and the technical prediction of many experts.
Although it is believed that this day will come sooner or later, it is not yet possible to determine the exact time, and if you want to get a head start in this area, you must start planning early. That's why many companies are now actively researching and investing capital, making humanoid robots a hot topic. But before that, robots with simpler structures may be commercialized sooner. For example, robots that use AGVs to move and have no legs but have hands and vision systems are more reliable and costly, and easier to promote.