"This piece of carbon was picked up by me and my team when we were investigating the mine," said Yang Wenli, CEO of Lingjun Technology, pointing to a stone in the conference room. In order to determine the direction of the autonomous driving landing, he and his senior management team have gone down to the mine, to the port and sanitation company, and even to Mohe to investigate the application of tractors in large farms.
In the process of trade-offs, Yang Wenli did not care, he repeatedly stressed to the Beijing Business Daily reporter that Lingjun Technology should "reduce the dimensionality of the blow", and also stated when announcing financing on April 13 that Lingjun Technology is now turning to Robobus and urban logistics, and looking for opportunities to return to the passenger car market. At the end of March, Baidu Apollo launched the Robobus test in Yizhuang is also the same way, the so-called "up and down", but Baidu has a large-scale Robotaxi running on the road. As the CEO of an autonomous driving company with a team of only 50 people and the former chief architect of Baidu Deep Learning Research Institute, Yang Wenli bluntly said that "large companies are more suitable for market cultivation."
During the period of waiting for the cultivation of the large factory market, Lingjun Technology has been groping for six years by relying on "slow heat". In the past six years, the autonomous driving industry has always been above the outlet, and the story of Lingjun Technology is only one of the paths that startups can choose, but the story collection supports the future trillion-level market.
Exclusion method 10 choose 2
Yang Wenli, who came from Tsinghua Automation, is no stranger to the industrial scene, and when he was a doctoral student in the United States, he went to Arizona to intern in mines and was familiar with the internal rules of the mine and the driving rules of vehicles. Back in China, because of an autonomous driving project, he went to a mine in Tongliao, where a piece of carbon he picked up was placed on the table in the conference room for many years.
Tractor factories, farms, sanitation companies, car production lines, ports... These scenarios, Yang Wenli and the team have also personally investigated, coupled with the industry to consult, Lingjun Technology studied all the landing directions of automatic driving, a total of 10, and finally determined in the second half of 2021: Robobus and urban logistics.
Why?
"We went to inspect an automobile production line in Yizhuang. It is custom production, each car has a different model, configuration and color, a robot will follow the car to feed, the degree of automation is quite high. There are also sweeping robots, warehousing robots, it can build maps autonomously, autonomous navigation, automatic planning and obstacle avoidance, is it considered automatic driving? I don't think so. Autonomous driving as defined in the industry is related to traffic, and it is necessary to run on open roads," Yang Wenli explains.
This exclusion method is also used by other self-driving startups, such as Zhixing, who was selected as a service provider for the Beijing Winter Olympics.
"According to the industrial line, smart walkers have three directions, vehicle regulation level, commercial level and special applications. Now Zhixinger has a higher degree of commercialization in sanitation and belongs to our commercial-grade product line", as for the reason, Wang Sunan, vice president of Zhixinger Strategy, answered similar to Yang Wenli's logic: "This is determined by the market, and the maturity of the industry itself determines whether we will choose the corresponding OEMs in this direction for cooperation." ”
More coincidentally, she and Yang Wenli also mentioned the "dimensionality reduction strike", "these three industrial lines, using the wisdom walker general software bottom." According to the characteristics of the vehicles used in different types of scenarios, we will be equipped with hardware systems that need to be installed at the bottom of the software, so that we will form a shipment product of Smart Walker - unmanned brain."
There are also companies that have forced the industry to mature, and Yang Wenli's old owner Baidu is. At the end of 2013, he completed the interview within a month, resigned as the chief architect of Western Digital, and returned to China to join baidu deep learning research institute. Later, Baidu cooperated with a number of OEMs to establish an intelligent driving business group, launched Baidu Apollo, and invested in autonomous driving fleets in many places.
Yang Wenli is looking for opportunities in the small projects of Lingjun Technology, in his view, "with the current volume and resources of Lingjun Technology, it is unrealistic to make the automatic driving L4 level mass production on passenger cars in the short term, and we need some commercial landing in the short and medium term."
Start with operations
After a meal in the Wudaokou Shopping Center Building, Yang Wenli found a candidate to perform the commercial landing - Pang Dongjun, the co-owner of Lingjun Technology.
"I and the current team are acquainted through the project, when I was in a mining vehicle driverless technology company, they went to investigate the tractor factory and mine is my recommended project", after coming to Lingjun Technology, Pang Dongjun's work includes financing, brand, project, operation, foreign cooperation, one of the company's most important landing business Ganzhou Municipal Government Development Zone projects, is the focus of his work after joining Lingjun Technology, in the words of Pang Dongjun, "often travel nationwide, Ganzhou to go the most."
At present, Lingjun Technology has obtained the urban operation license for autonomous vehicles issued by the Ganzhou Municipal Government Development Zone, and set up a project company in Ganzhou. "We have three models in the Ganzhou Economic Development Zone, unmanned minibuses, intelligent network buses, Robotaxi, and two 'Xiao Ant Residence' series of unmanned buses in the park mainly do point-to-point connection operation tests, such as high-speed rail station - golden airport - automotive science and technology city, golden airport - development zone management committee, VIP transfer and visit services," Pang Dongjun said.
Generally speaking, companies that start with the universal underlying technology of autonomous driving have two commercialization models, one is to do projects to sell cars and sell solutions, and the other is to invest in vehicles and be responsible for operation. Baidu, Lingjun Technology, Wenyuan Zhixing and most other companies are doing both, specific to Lingjun Technology + Ganzhou, Baidu in Yizhuang to put Robotaxi, Robobus, Wenyuan Zhixing in Guangzhou to put Robotaxi, Robobus belongs to the latter form.
There are also companies that want to avoid operations and focus on products.
"Now the autonomous driving track company presents such a situation in the choice of development path, either controlling the operation end, or controlling the road section, transportation capacity, or controlling the scene." This leads to the company to stand on the operation side of the consideration, take this task down, and then through their own products to complete, but at this stage because the safety personnel can not be removed, the upgrade on the original vehicle must be completed through a higher cost, and can not achieve cost reduction and efficiency of the commercial closed loop. Faced with the goal of commercialization, Zhixinger chose not to be an operator or vehicle manufacturer of the concept of automatic driving, but instead required himself to focus on becoming a company with strong product strength and cooperating with various industry partners as a provider of autonomous driving technology," Wang Sunan explained.
Yang Wenli believes that "automatic driving is not entirely to save money, but to experience upgrades, and to go deep into the operation section to accumulate data and promote system upgrades." When the vehicle's lidar and parts-by-wire control system are relatively mature, the cost of transformation has been reduced, and the stability, reliability and cost of the product are guaranteed to a certain extent, and small-scale commercialization can be done."
The market is still immature, the path is not right or wrong, as He Qi, senior analyst of Analysys Analysis Auto Travel, said, "Both have their own advantages, simply providing automatic driving brains, compared to the solution is easy to copy, the operating costs are lower, but because the cost of obtaining data from non-participating enterprises will increase, training algorithms are constantly needing operational data to support."
On the other hand, looking at the vehicle + operation model, most of the cost comes from labor. "Lingjun Technology Ganzhou's technical team has more than a dozen people, mainly responsible for the development of products delivered to the site, and then do test operations", in Pang Dongjun's work list, there are also Suzhou government investment projects that are being approached, which is also based on Robobus.
The abacus of "panacea"
Recruiting people needs funds and office space, financing is the first task that Pang Dongjun received after coming to Lingjun Technology, "I am also learning automation, which may be one of the most professionals from artificial intelligence engineers, I want to pull financing to explain technology and landing with investors, this is my best field, early projects and cooperation, relying on sales alone can not be taken down." If the suzhou project is finalized, it will need to recruit people."
This kind of enrollment expansion plan of Lingjun Technology is rare before 2021. When the company was first established in 2016, Yang Wenli and two Tsinghua alumni, plus several shared office seats and a whiteboard, became everything about Lingjun Technology, which still hung on the wall of Lingjun Technology's conference room. Now Yang Wenli has found a new office space, "the new people are coming, need a larger office space, decoration costs three or four hundred thousand yuan," he laughed.
In fact, compared with high-profile autonomous driving companies, lingjun technology employees are not large. "On the surface, we take less money and the team is small, but in fact, we deliberately slow down the development speed." Yang Wenli explained to the Beijing Business Daily reporter. Before Pang Dongjun joined, the core team of Lingjun Technology remained within 20 people for a long time.
However, from the results, this kind of control has caused a time difference between low-key autonomous driving companies and head companies such as Lingjun Technology.
"The time difference is very important, Baidu and Wenyuan Zhixing landed earlier, the more data accumulated, the closer the automatic driving model will be to the commercial model, the more mature." If you start late and land late, your experience is relatively lacking, which can be seen from the valuation," Zhang Xiang, dean of the New Energy Vehicle Technology Research Institute of Jiangxi New Energy Technology Vocational College, said bluntly, taking Alphabet (Google's parent company) self-driving company Waymo as an example to further explain, "Waymo's valuation is now declining, mainly because the commercialization time is longer than expected."
He Qi believes that survival and time difference are the focus, "can survive, and to the current enterprise is definitely dominant, from the perspective of the capital market, capital is obviously concentrated to the head of the enterprise, matthew effect is getting more and more intense, which proves that the time difference is very important." However, with the development of the market, many start-ups have also obtained financing with the background of large factories and the strength of technical projects, which shows that autonomous driving is a marathon, and time still needs to look farther."
Yang Wenli also wrote in the circle of friends after announcing the financing, "Automatic driving is a marathon, and it is necessary to win until the end, so pace is very important."
According to his judgment, the autonomous driving industry is divided into the core supply chain maturity period, small-scale commercialization landing period, market scale expansion period, international expansion period, "now in the second stage of 2021-2025, unmanned minibuses, intelligent network buses, regional unmanned logistics vehicles and other technical products can be applied on a small scale, and enterprises can obtain certain income."
Beijing Business Daily reporter Wei Wei