In April, Shenzhen-based autonomous driving startup Yicheng Zhixing (Shenzhen Yicheng Autonomous Driving Technology Co., Ltd.) delivered more than 100 vehicles. In the definition of "commercial landing" of autonomous driving, the delivery capacity of "100 units" is an unwritten "golden line" in the industry. After several years of policy, technology and market preparation, the commercialization of autonomous driving has taken a turn in 2021.
Intuitive feelings from enterprises support this trend judgment, including Yicheng Zhixing. Song Chaozhong, founder and CEO of Yicheng Zhixing, told us, "Since 2021, there have indeed been some changes. From the perspective of Yicheng's customer base, the number of program users has decreased, and the number of batch users and standard parts users has increased significantly. This means that customer characteristics have gone from the verification stage to the batch promotion stage. This is not very obvious in 2019 and 2020. ”
Founder & CEO of Yicheng Zhixing Song Chaozhong
The "Independent Market" of the Autonomous Driving Industry Chain
In fact, in the past three years, many industries have not had a good life. The epidemic in Shanghai in April has nearly interrupted the automotive supply chain, although the Shanghai Municipal Commission of Economy and Information Technology has launched the first batch of "white lists" for resuming work and production in key industries such as integrated circuits, automobile manufacturing, equipment manufacturing, and biomedicine, but the epidemic is still continuing, and the production capacity of enterprises cannot be restored immediately.
Song Chaozhong said frankly that autonomous driving and the traditional automobile supply chain have some overlap, such as the required auto parts parts, which have been greatly affected by the current round of Shanghai epidemic; coupled with the fact that Shenzhen has just survived a round of epidemic in March, either the starting place "with the star" or the destination "with the star", whether it is procurement or visiting customers, the Yangtze River Delta region "can't go".
Despite this, Song Chaozhong still believes that as an emerging comprehensive category, automatic driving has its own uniqueness, and the overall dependence on the existing industrial chain is not high. "The industrialization of automatic driving has just begun, a complete and mature industrial chain has not yet been formed, plasticity and resilience are very strong, like controllers, chips, sensors, etc., are interchangeable." It also means that it's easy to find new partners. Participants (within the industry) can constantly adjust and revise to establish a supply chain model that is more suitable for the current environment. ”
To some extent, the autonomous driving industry chain or the intelligent automobile industry chain should optimize the reorganization and re-anchoring of the nodes in the traditional automobile supply chain, and this process will be realized simultaneously with the large-scale commercial landing of automatic driving. In this process, those who have the ability to run long distances in the domestic autonomous driving industry will win.
Yi Cheng Zhixing was founded in 2017, the core team, Song Chaozhong graduated from Tsinghua University, is a typical "Tsinghua department" autonomous driving entrepreneur, while among the other members, or returnees, or senior practitioners of traditional automakers, the team configuration is luxurious. For example, in the anecdotal passage, "five years of entrepreneurship, three years of epidemic", in the context of general pessimism, hard technology industries such as automatic driving have been considered strong. During the special period, Yicheng Zhixing still maintained a compound annual growth of orders and revenue, which was not easy.
"Brain (solution) + intelligent hardware + scene products", which is almost the standard mode of the current autonomous driving startups, if the same materials and applications as the two ends of the autonomous driving industry chain, the former two are in the middle position, the latter is in the market terminal. The same is true of The Business Structure of Yicheng Chi Heng.
The application of Yicheng Zhixing products in different scenarios mainly includes connection, logistics, cleaning, etc
So, how can a startup dig out its own "moat" in the Red Sea Arena? From Song Chaozhong's elaboration, we may be able to observe the unique methodology of the industry in 2022.
Streamline data and "get closer" to the market
"Brain + intelligent hardware + scene products", the corresponding is actually an algorithm-centered solution + terminal market application. Algorithms are the underlying technology, solutions are industry possibilities, and applications are the future of the industry. In other words, under the same industry logic, finding a better solution to the algorithm, empowering different industries, and opening up a niche application market with a high share are the three decisive factors for how fast and how far the autonomous driving company can go.
Instead of the "brain", Song Chaozhong likes to use another more intuitive expression to illustrate Yicheng Zhixing's algorithms and solutions - GEMINI scalable automatic driving kit. Based on the GEMINI OpenPilot autonomous driving platform with independent intellectual property rights, Yicheng Zhixing can provide a full set of technologies including intelligent algorithms and domain controllers, etc., and at the same time, emphasizes the integration of software and hardware, and the value foothold is "industry empowerment".
For a long time, the algorithm iteration of autonomous driving companies has emphasized sufficient data, but how to solve the data source is also a common difficulty of each company. When talking about how to solve the problem of data sources, Song Chaozhong emphasized the ability of "small data" more. He believes that the use of big data to train a mature AI model is a more consistent practice in the industry, but AI in the context of automatic driving has its particularity, "even emphasizing the number of kilometers does not make much sense, in the same road conditions, you run 10 kilometers of data and run 100 kilometers of cumulative data will not change much." ”
In the differentiated competition of autonomous driving, if the "big data" ability cannot be pulled apart, the one who can take the lead in moving towards "small data" will win. Song Chaozhong believes that streamlining invalid data is another important connotation of automatic driving algorithms. This means more advanced algorithmic capabilities. "How to make the algorithm itself smarter, this is what we value more." Song Chaozhong said.
Precise empowerment, adhere to the profit closed loop
Returning to the scene, Song Chaozhong's view is also very clear, "closer to the commercial market wins." Therefore, in the exploration of many scenes, Yicheng Zhixing appears to be quite "stubborn" - with a clear goal, focusing on unmanned functional vehicles in commercial and engineering scenarios, mainly including three series of connection, logistics and cleaning. Equipped with the GEMINI autonomous driving platform independently developed by Yicheng Zhixing, with various sensors, it can realize the L4 level automatic driving function. One of the most critical values of the product is to achieve low cost and low power consumption based on the Yicheng Zhixing autonomous driving domain controller, that is, Song Chaozhong's most important thing, "industry empowerment, cost reduction and efficiency increase".
"Yixing" series of shuttles, Yixing S3 is a classic product of Yicheng Zhixing, which applies L4 level domain controllers to realize APP calling, automatic parking, parking and other functions
The closed loop of data and the closed loop of profitability are two topics that self-driving companies cannot avoid. Subject to the objective conditions of policies and road conditions, in the semi-closed engineering scenario, compared with the open traffic environment, commercial profits may be realized more quickly. In the battle between Robotaxi and Robobus, YiCheng Zhixing hopes to go out of the third way, "For subdivided tracks, the greater significance of autonomous driving lies in industry empowerment." ”
"Autonomous driving does not exist independently in application scenarios, but is one of the superimposed functions. For example, no one sweeps, in addition to realizing no one, we must also ensure the basic function of cleaning. So we are optimistic about the value of 'autonomous driving +'. In different industries, through automatic driving to reduce costs, increase efficiency, liberate people from the heavy environment. New technologies should empower the industry, and it doesn't make sense if it becomes a high-cost center in itself. ”
Perhaps related to the experience, throughout the interview, Song Chaozhong has always emphasized the meaning of "technology to people", and he also specifically mentioned his childhood memories, "When I was a child, it was very tiring to do farm work, so why not liberate people from the heavy environment?" Including the Shanghai epidemic, those truck drivers who are sealed on the highway are too hard. For example, if through driverless technology, reduce the work intensity of truck drivers, or improve the operating comfort of an industry, this is what we want to do. ”