Author | Leng Zelin

Editor | Wang Pan

On June 27th, Alibaba announced that it has obtained the L4-level “driverless” automatic driving truck public road test license issued by Deqing, Zhejiang.

It is reported that this city is the first city in the country to issue such a test license. The first batch has only two licenses, one of which was given to Alibaba DAMO Academy, and the other was given to PlusAI.

The former merged the small G team under Cainiao and the AI Lab team under Alibaba Innovation Business Group in 2018 to establish DAMO Academy’s autonomous driving laboratory.

In September 2020, the team released an unmanned vehicle called “Xiao Manlu” for the logistics distribution scene of the last 3 kilometers. As of the end of March this year, the cumulative delivery logistics orders have exceeded 10 million. The license obtained by Alibaba DAMO Academy this time is actually another product, a light-duty truck for urban logistics called “Da Manlu”.

PlusAI was established in 2018 and initially aimed at the Chinese trunk logistics scene. It is a technology supplier and operator of autonomous driving trucks. On December 23, last year, it completed the domestic first batch production-capable L4-level unmanned driving heavy truck test on the closed high-speed test field in Laiwu.

It is worth noting that behind the two parties who obtained the “driverless” automatic driving truck public road test license this time is not just competition between two autonomous driving truck companies. Several giants in the domestic logistics market are standing behind them.

Alibaba DAMO Academy can speak for itself, as it has the largest e-commerce platform in the country. Its Cainiao’s “third-party logistics and fourth-party logistics” covers most of the express delivery business of Taobao. PlusAI’s B/B+ round of financing has the shadows of JD Logistics, Meituan, and Deppon Express. It seems that a battle from logistics distribution to trunk/city logistics is gradually unfolding.

E-commerce and logistics converge on different paths

JD and Alibaba have different views in the e-commerce and express delivery fields and have looked down on each other for a long time.

In the book “Alibaba: The House That Jack Ma Built,” the author recorded that Jack Ma said in 2015: “JD will be a tragedy in the future.”

This tragedy refers to the fact that as the total volume of Chinese express delivery continues to rise, JD’s labor costs will also rise sharply, high costs and heavy asset models will drag down JD’s e-commerce; while Alibaba, through the light-asset network built by Cainiao, can better radiate most groups and handle larger packages.

Similarly, Liu Qiangdong also predicted: “In the future, only JD and SF Express may win in China’s logistics industry.”

In Liu Qiangdong’s view, Cainiao Network is essentially constructing a data system on several express delivery companies, and most of the profits of several express delivery companies will be absorbed by Cainiao.

Up until today, JD (京东) and its self-built logistics have not collapsed, while Alibaba and Cainiao are still occupying the majority of the market and even catching up to JD.

At the beginning of this year, Tmall (天猫) started building self-operated flagship stores from the 3C category and upgraded corresponding logistics experience. Some Maocai (猫享) self-operated products can now enjoy services such as home delivery and appointment delivery.

During the 618 Festival, Cainiao expanded its scope of delivery. Cainiao Direct (菜鸟直送) which previously only delivered for Tmall supermarket, now provides delivery services for Tmall Global for the first time. Cainiao Post (菜鸟驿站) and Cainiao Supply Chain (菜鸟供应链) also started to provide services such as appointment delivery and home delivery for large items.

However, the express delivery industry itself is a labor-intensive industry. At present, the growth of the number of couriers is not proportional to the growth of the number of parcels, resulting in a rapid increase in labor cost due to personalized delivery at the end.

In 2010, the total domestic express delivery volume was about 2.3 billion items, and the number of employees was about 542,000. However, as of the end of last year, the former has grown about 47 times, the latter has only grown about 7 times, and the workload for each courier has naturally increased. In April last year, the Ministry of Human Resources and Social Security released the "Top 100 Shortage of Professions in the First Quarter of 2021", where couriers ranked eighth.

Self-built logistics and direct sales systems have brought JD good reputation, but they also come with high operating costs. JD Logistics has over 300,000 frontline employees, and its accumulated losses for nearly four years have exceeded 20 billion yuan.

As the largest player in the takeaway industry, Meituan revealed in its financial report in the fourth quarter of last year that the income from delivery is not enough to cover the delivery expenses, and that delivering takeaway per order loses one yuan.

But why are many internet companies still trying to perfect the "last mile" delivery despite the loss? The reason is that with the internet traffic stagnating, service quality and efficiency are obviously the next stage of competition.

To find a balance between labor costs and service quality, it may be necessary to introduce unmanned delivery vehicles as the ultimate solution to improve the end delivery. It is clear that simply stacking "human head" is not enough.

## The epidemic accelerates the outbreak of unmanned delivery vehicles 

"At present, the development of autonomous driving needs to comply with the three laws of autonomous driving: first low speed, then high speed, first carry goods, and then carry people, first commercial use, then civil use."

This is how Hou Jun, COO of WeRide (毫末智行), explained at the "2022 China Low-Speed Autonomous Driving Scene Ecological Building and Expansion Conference" held in June.

Low-speed autonomous vehicles

Low-speed autonomous vehicles are considered the project that will be the first to achieve commercial landing because of their lower driving speed, which provides more decision-making and reaction time, as well as high degree of autonomous decision-making rights that do not involve the division of driver and vehicle responsibilities.

The three links of autonomous delivery industry

The autonomous delivery industry is divided into three links. Upstream suppliers provide key parts such as wired control chassis, LiDAR, and computing platforms. Midstream is the technical solution provider, mainly technology and Internet companies. Downstream is the distribution service provider and demand side, generally mainly e-commerce, offline retail, and express companies.

Role of enterprises

A single enterprise does not necessarily play a single role. For example, Maodou Zhixing, which independent from Great Wall, provides both parts and single technical solutions, and also has its own autonomous delivery vehicles. Alibaba, JD.com, and Meituan have two roles, in the midstream and downstream, and their autonomous vehicle projects are mainly for internal supply.

Early entry of players

In fact, the entry of players such as JD.com, Alibaba, and Meituan may be much earlier than the public imagines. In 2016, Meituan established the W project group to carry out related research and development work on autonomous delivery methods for takeaway and errands. In the same year, JD.com renamed its logistics laboratory, which was founded in 2013, as the X business unit, undertaking a series of research and development tasks including autonomous vehicles, fully automatic logistics centers, and warehousing robots.

Achievements

Before the establishment of the autonomous driving laboratory at DAMO Academy, Chen Junbo, the current technical director of the autonomous driving laboratory at DAMO Academy, formed a small team of seven or eight people and developed ‘Xiao G’ in 2015. In 2016, Waymo became independent from Google, but people’s attention was still drawn to Robotaxi.

In 2018, Zhang Chunhui, the director of Alibaba Cainiao ET Logistics Lab, stated that the company’s unmanned equipment would reach 100,000 units in the next three years. However, as of October 2021, only 350 units had been sold. If it weren’t for the pandemic that began in early 2020, the landing of autonomous vehicles might have been even slower.

In September 2020, in addition to Alibaba, JD.com also jointly released the DIDO autonomous driving logistics vehicle with Jinlong Auto. The following year, the Meituan Mobag 20 and Xiaoman Donkey produced by Maodou Zhixing were put into use.

Especially during the Shanghai epidemic this year, Internet companies such as Meituan and JD.com, as well as technology companies such as Maodou Zhixing, Baixiniao, and Xinshiqi, have successively invested dozens or hundreds of unmanned delivery vehicles in relatively closed areas such as communities, supermarkets, hospitals, and campuses.Can “on-site services” be completed in enclosed spaces? Actually not. Apart from buildings, most unmanned delivery vehicles face the problem of being unable to reach upper floors, requiring personnel or couriers to “relay”.

This approach only reduces personnel contact during the pandemic. When the demand for the pandemic decreases, unmanned delivery vehicles that cannot reach upper floors will face various problems, such as whether the efficiency improvement is proportional to the cost, including but not limited to the costs of research and development, maintenance, depreciation, and operation. For example, the electric vehicle fire incident that occurred nationwide in China makes it difficult for unmanned delivery vehicles with power batteries to convince most property management companies and parks to allow them to ride elevators. Secondly, there is a problem of pedestrian and vehicle right-of-way when the elevator is crowded.

During 2016-2020, some companies discovered similar problems and put the commercial scenarios of unmanned delivery vehicles on more open public roads.

However, once they are on public roads, unmanned delivery vehicles need a suitable “identity”.

Chentao Capital pointed out in a research report that the category of unmanned delivery vehicles is still unclear. If it is managed according to robots, the law does not allow robots to be on the road; if it is managed according to vehicles, it does not support the mass production of “unmanned driving vehicles” on the road. At the same time, the definition of non-motor vehicles and motor vehicles also puts forward different requirements for unmanned vehicles.

Last year, Beijing issued the “Implementation Rules for the Management of Unmanned Delivery Vehicles,” which awarded Meituan, Jingdong, and New Shiqi the qualifications for road use. In the rules, unmanned delivery vehicles are allowed to refer to non-motor vehicle rules for managing road rights. In addition to the refinement of standards such as size and load, one point also requires the provision of on-site drivers and remote drivers.

At this point, the difference between unmanned delivery vehicles and autonomous driving vehicles seems to be only the difference between carrying people and carrying goods.

It seems that Alibaba isn’t particularly focused on unmanned delivery vehicles on the road. Alibaba’s Xiaomanlv has remained on campus and in communities and has used large trucks such as Dalanmu to complete the “road” tasks.

Escalation of Warfare, Focus on Unmanned Trucks

Among the three major scenarios of autonomous driving (robotaxi, unmanned truck, unmanned delivery vehicle), judging by the “Three Laws of Automated Driving” mentioned earlier, the landing of unmanned delivery vehicles is faster than the first two, while the landing of unmanned trucks is faster than robotics.

In terms of unmanned trucking, almost all downstream scenario parties are currently laying out through strategic investments, while Alibaba stands out.

In addition to Vwxyz Technology, which has investments from Meituan, Deppon, and JD Logistics, and Qiangua Technology, which has an investment from SF Express, Full Truck Alliance has participated in the A/A+/C/D rounds of financing for Driven Logistics.While direct entry of level 4 autonomous trucks is not currently realistic, SL4/L3 autonomous trucks are already capable of reducing costs for enterprises, such as by improving fuel economy through fuel-saving algorithms. The autonomous driving systems of Zhi Jia and Yinchu can increase fuel efficiency by about 10\% on average, and Yinchu’s L3 autonomous driving trucks reduce the driving burden on drivers, thereby requiring only one driver.

However, there is still a distinction to be made in autonomous trucks. The aforementioned tech companies are almost entirely focused on heavy-duty trucks, and the investment partners involved are also mostly companies that require improvements to trunk logistics.

Although there is not much information revealed by Alibaba’s Big Mule, former Alibaba Group Vice President and Head of Damo Academy’s Autonomous Driving Laboratory, Wang Gang, mentioned last year that the Big Mule is an L4 autonomous driving truck for urban open road delivery scenarios.

According to the information regarding Alibaba’s acquisition of a public road test license, “it can conduct road tests (including some sections of expressways) in designated areas of Deqing County,” which also confirms that the Big Mule is not primarily designed for trunk logistics.

In fact, during the 2017-2018 period, JD.com once revealed their joint development of an unmanned truck with SAIC-GM-Wuling and a self-developed unmanned heavy-duty truck, the former for urban areas and the latter for expressways. According to JD.com’s plan, L4 autonomous heavy-duty trucks will complete domestic commercial deployment trials in 2020.

Clearly, it seems that news on these two trucks has disappeared. This also indirectly explains that a technology-based company that is able to handle small autonomous delivery vehicles may not necessarily be able to handle L4 autonomous trucks.

However, it is currently difficult to fully implement large-scale deployment of unmanned delivery vehicles, and without first establishing a delivery network, the uplift for companies is somewhat trivial. Moreover, the issue of road rights for public roads seems to have fallen into a similar quandary as autonomous cars in the short term, so it is particularly important to lay out an array in the field of high-capacity trucks.

The competition in the logistics industry is also gradually transitioning from people and small cars to trucks.

This article is a translation by ChatGPT of a Chinese report from 42HOW. If you have any questions about it, please email bd@42how.com.