How will the $10,000 L4 autonomous driving solution impact the industry?
On December 8, Yuanrong Qixing officially released DeepRoute-Driver 2.0, an autonomous driving solution for front-loading L4 levels. It is reported that the program uses 5 solid-state lidar and 8 high dynamic range cameras, which can be adapted to different types of sensors and models.
Notably, the solution costs less than $10,000, significantly lower than the L4 level autonomous driving solutions on the market that start at $50,000.
In fact, in September this year, Yuanrong Qixing received industry attention for obtaining a $300 million Series B financing led by Alibaba. So far, Yuanrong Qixing, which has been established for less than 3 years, has raised about 400 million US dollars in total, and its valuation has exceeded 1 billion US dollars, becoming another unicorn in the field of automatic driving.
Yuanrong Qixing CEO Zhou Guang
In the exchange with Forbes China, Zhou Guang, CEO of Yuanrong Qixing, mentioned cost down more than once. He believes that for the mass production of autonomous driving, cost is always a key factor. Ultimately, autonomous driving must be a low-cost, highly automated solution.
"Someone else's child"
When it comes to entrepreneurial stories, Zhou Guang started from his childhood. "I don't think I was the same as other children when I was a kid. When other people's children are immersed in learning, I will soon be able to complete school assignments, and I have a lot of spare time to spend on my hobbies, especially hands-on, that is, DIY. For example, dismantling and repairing electronic components and equipment, making aircraft models, and so on. ”
Zhou Guang recalled that he had been exposed to programming since junior high school, and had the opportunity to contact various precision instruments and equipment, and may have opened up a new world of his own at that time.
In his freshman year of high school, he represented the school in the "Robot Olympiad" by chance. "At that time, there was no experience, not even a supervisor, relying only on some of the school's equipment, from beginning to end, it was all self-made, including studying mechanical structures and programming," he said. In the end, I actually got the gold medal. I think this has a lot to do with the fact that I like to do things since I was a child. ”
After that, Zhou Guang entered the "Basic Science Class" of Tsinghua University, in addition to robots, he also began to participate in physics competitions and won the first prize at the international level, so he laid a solid foundation in mathematics and physics. While studying for a Doctorate, he chose the direction of artificial intelligence, and the previous mathematical foundation played a huge role, and he was also very interested in robots. He believes that true creativity is not just about being able to write a few papers, but needing to be really hands-on.
Autonomous fleets
Start again
Talking about his years at Baidu American Research Institute, Zhou Guang was full of gratitude, he said: "This is my best choice. "Although Baidu is not the first company to do autonomous driving, Baidu American Research Institute is in the 0-1 stage, which is also the most valuable stage. Thank you very much to Baidu for providing rich resources and a relaxed environment, and for assembling such a good team. At that time, he also made a lot of friends, and now this group of people has become the backbone of the domestic autonomous driving field.
Zhou Guang was blunt about the failure of his first venture: "From a technical point of view, Roadstar is still very successful, and at that time it was already at the first-class level in the field of autonomous driving in China. But the problem is at the non-technical level, especially at the level of equity allocation and interpersonal relations. Now it seems that the failure of the last business is a very valuable lesson for yourself. ”
At the same time, he is grateful to his team, in this case, still maintained a lot of passion, most of the engineers want to start over with me. In February 2019, Zhou Guang participated in the founding of YuanRong Qixing, and the meaning of the company's name is "the army starts".
Zhou Guang said that in order to avoid sensitivity, at the beginning of the company's establishment, he only grew up with the team as a consultant, and only recently officially became the CEO. "With the lessons learned from the past, our team today is very cohesive and start over. At present, the company has achieved rapid development. At the same time, I have also been reflecting that entrepreneurship can not only focus on the matter itself, in addition to technology, but also need to deal with interpersonal relationships, formulate incentive systems and promotion space, as well as the distribution of benefits, equity distribution and investor relations. ”
In-vehicle sensors
Endgame thinking
When it comes to the company's core competitiveness and advantages, Zhou Guang bluntly said that it is a technical barrier. Citing the recently announced $10,000 L4 autonomous driving solution as an example, the company is currently the only company in the industry that can achieve L4 autonomous driving at such a low price, which shows the company's leading technology and underlying competitiveness.
Since autonomous driving is difficult to achieve absolute safety, some potential contingencies are still difficult to avoid. In fact, at present, the L4 level automatic driving solution competes with a "takeover rate", that is, the "takeover mileage" that determines the realization of highly automatic driving (L4 level) in the road in the core area of the city, and one of the indicators is whether it can reach 10,000 kilometers.
At the same time, he mentioned "endgame thinking" (which refers to considering the problem from the end point to determine the current choice when facing many choices. Regardless of the layout of the technology and the future commercial layout, the early layout is achieved through end-game thinking.
The first is the technical level, Zhou Guang said, as early as 2017, he proposed the concept of "pre-integration". In 2019, at the beginning of the establishment of Yuan Rong Qixing, he proposed the new concept of "reasoning engine". He said: "At the beginning, many people did not understand and questioned. But since then, I've proven that the technology is a forward-looking layout. At the end of the day, it's all based on thinking about the endgame. Because, what autonomous driving ultimately requires is a set of low-cost, highly automated solutions. ”
Zhou Guang believes: "Today, many L4 level autonomous driving companies are even overtaken by Tesla's route (pure vision) due to cost control problems. Therefore, self-driving companies must reduce the cost, otherwise even if the algorithm is leading and the architecture is leading, it is equally difficult to win. ”
Comparison of technical routes
Yuanrong Qixing has long been based on reducing costs, trying to achieve first-class results with lower-cost, average-performance sensors, which has taken a lot of time and effort, so that the price of a complete set of L4-level autonomous driving solutions can be controlled below $10,000. "That's our technology-centric endgame thinking." Zhou Guang said.
The second is the product level, which also relies on "endgame thinking". At present, many autonomous driving companies are two or more business routes coexist, for example, Yuanrong Qixing is mainly Robotaxi and the same city freight two major businesses. Zhou Guang said: "Initially, we chose this business with a deeper consideration. Because for start-ups, they can only do well if they focus on one thing. ”
If you are involved in two completely different scenarios, it means that the consumption of resources and the use of talents are twice as large as usual, and it is more difficult to manage. However, the two major businesses selected by Yuanrong Qixing are based on the automatic driving of urban roads, and only a set of technical systems and a technical team can support two businesses and products at the same time, avoiding waste of resources and internal contradictions. Therefore, Yuanrong Qixing only made a set of automatic driving systems, focusing on urban roads and urban areas.
Zhou Guang said: "In this way, we can use funds and talents very efficiently. At the same time, Robotaxi and Same City Freight are very large tracks, not like closed scenes, and both have very large commercial value. Therefore, I think that whether it is technology, product or market, the company's grasp of the 'end game' is quite accurate. ”
Autonomous driving will become standard in smart cars
As for whether automatic driving will become the standard for smart cars, Zhou Guang believes that back in 2000, mobile phones were not standard; in 1990, telephones were not standard. Therefore, the wheel of scientific and technological development cannot be stopped. Today, autonomous driving may not be the standard, but I think that over time, autonomous driving will no longer be science fiction, but will become the standard for smart cars, and even become infrastructure, become a part of people's lives.
When it comes to business models, Zhou Guang believes that the form of charging for autonomous driving still needs to be determined by the market, one may be a one-time high fee, and the other may be a subscription system, that is, monthly payment. He said that he prefers the subscription system, with the maturity of technology, the entire automatic driving research and development investment and cost will be lower and lower, I believe that the subscription system will be accepted by more users.
At the same time, autonomous driving can be understood as an "AI driver". The landing of the L4 level autonomous driving solution means that the labor cost of the next driver will be saved, and the overall cost will be greatly reduced. In the future, the cost of autonomous driving will certainly be much lower than the cost of labor, thus promoting its widespread application.
Zhou Guang said: "Autonomous driving may be difficult to achieve profitability in the short term, but in the long run, a healthy business model and a certain profitability are indispensable. So, if you want to be profitable, you have to be cheap, that is, reduce costs. "At present, the general L4 autonomous driving solution is more than $50,000, which is obviously difficult to be mass-produced. In the future, after mass production, the cost of this solution will further decline, and the final price may be 20,000-30,000 yuan.
In July this year, Robotaxi, which started from Yuanrong, has launched trial operations in Shenzhen's Futian District and achieved a certain degree of commercialization. Zhou Guang stressed that Futian District is the core area of Shenzhen, which is no longer a closed environment, but a regional operation attempt. He believes that the future of Robotaxi is an inevitable trend, especially for L4 autonomous driving, an all-weather, all-terrain scene is particularly important.
Finally, Zhou Guang introduced the product planning of Yuanrong Qixing: from 2022 to 2023, Yuanrong Qixing will expand in-depth technical cooperation with the main engine factory and develop a vehicle-grade front-loading mass production program. It is expected that by 2024, cars equipped with L4 autonomous driving systems will begin mass production and enter the market on a large scale.