Text / Electric Motor
Produced / Node AUTO
Mushroom Automobile Alliance was established just a few years ago, but it has won the first domestic and the world’s largest city-level autonomous driving project. As a dark horse in the field of autonomous driving, it competes with and Baidu . Many people have many questions about the rise of Mushroom AutoLink.
In fact, the emergence of Mushroom AutoLink has stepped on the right node. When the mass production of self-driving cars is far away, the government takes the first step and builds new infrastructure to pave the "road" for smart electric vehicles; Tencent invested in Mushroom AutoLink and used it as a bargaining chip in the "car manufacturing boom" Pay attention to compete with , Alibaba, , Baidu, etc. to avoid falling behind.
However, the new car manufacturers who are aiming at consumer-grade products and have deep financial backgrounds look down on Mushroom Auto Alliance, which is backed by the government and the father of Tencent’s financial backer. The Mushroom Automobile Alliance is not large, but it has found a glimmer of hope in the cracks. Its existence exactly reflects the multiple aspects of the development of autonomous driving in China - the government's support for infrastructure construction and the struggle with car companies to take the consumer market route.
What is the future of Mushroom Auto Connect? Let’s wait and see the direction of the tide.
/ 01 /
The battle for autonomous driving routes
Mushroom Car Linkage " China Plan "
"If autonomous driving cannot be achieved, Tesla will be worthless," Musk once said publicly.
In the competition in the field of intelligent electric vehicles , autonomous driving is an important part. According to the standards of the American Society of Automatic Engineers, automobile autonomous driving is divided into six levels from L0 to L5. Ideal Auto founder Li Xiang once called for the unification of Chinese noun standards for autonomous driving, that is, L2 = assisted driving, L3 = automatic assisted driving, L4 = autonomous driving, and L5 = driverless driving. L5 driverless driving is the ultimate goal, and the industry currently believes that the L4 system has reached the top of the technology pyramid.
Let’s talk about autonomous driving technology
The core of autonomous driving technology is reflected in the entire process of a vehicle’s “perception, decision-making, path planning and control execution”.
is different from Google 's brother company Waymo starts directly from L4 and tries to get the core technology in place in one step. Tesla starts from L2 and initially only relies on vision, that is, "perception", which is considered to be a reduced version of autonomous driving. Tesla first gets on the car and then pays for tickets, and then continues to supplement its technical shortcomings, eventually pointing to L5.
Regarding Tesla’s autonomous driving technology, Musk said, “There are currently no underlying fundamental challenges in realizing autonomous driving L5, but there are many details. The challenge we face is to solve all these small problems and then integrate the system , continue to solve these long-tail problems. "In fact, the difficulty of fully autonomous driving lies in the more than 1 million long-tail, low-probability extreme road conditions, rather than only about tens of thousands of them encountered in 90% of the journey. Normal road conditions.
In 2019, BYD Wang Chuanfu once said, "In the transformation process of the automobile industry, electrification is only the first half, and intelligence is the second half. It is the real big battle and the real big change." And autonomous driving is exactly that. One of the core elements of automobile intelligence. There are two different routes for autonomous driving.
Compared with the “bicycle intelligence” led by Tesla, there is another route called “vehicle-road collaboration”. bicycle intelligence relies on the vehicle itself to complete perception and decision-making, while vehicle-road collaboration will build smart transportation infrastructure during road testing to assist vehicles in obtaining information during driving. MOU is the representative company of this route. Currently, MOU is at the L4 level and is in the “first gradient” position in the domestic autonomous driving field.
Tesla Autopilot's basic "visual processing tool Tesla Vision" Source: Tesla China official website
Riding policy bonus
The rapid rise of Mushroom Connected Vehicles is precisely due to the vigorous promotion of "vehicle-road collaboration" by governments at all levels. appears against the background. When it comes to choosing a technical route to realize autonomous driving, the competent authorities of the Chinese government have chosen the “vehicle-road collaboration” route. In August 2021, data released by the official website of the Ministry of Transport showed that China has built 16 intelligent connected vehicle demonstration areas, opened more than 3,500 kilometers of test roads, and the total test mileage exceeded 7 million kilometers.
In March 2021, Momogu AutoLink’s first city-level autonomous driving project in China was launched in Hengyang. This is also the largest city-level autonomous driving commercial project in the world, with a total investment of 500 million. Including the 300 million cooperation project signed with the Hebi government, In 2021, Mushroom AutoLink received over 1 billion orders. In the field of autonomous driving, this is one of the landmark events. It has only been a year and a half since Mogujia launched China's first "vehicle-road collaboration" 5G commercial demonstration project in Beijing; it has only been four years since Mogujia was established.
cooperates with local governments and relies on government payment to survive, which is a typical feature of my country's "vehicle-road collaboration + bicycle intelligence" route.
The reason why this route has received strong support from the government level is that the prerequisite for its implementation is the new infrastructure construction represented by 5G and smart transportation upgrades promoted by the state. In February 2020, the National Development and Reform Commission launched the "Smart Vehicle Innovation and Development Strategy", planning to achieve regional coverage of vehicle wireless communication networks (LTE-V2X, etc.) by 2025, and the new generation of vehicle wireless communication networks (5G-V2X) in some areas It is gradually applied in cities and highways. vehicle-road collaboration is the core component of V2X Internet of Vehicles.
The government is engaged in infrastructure construction, and companies such as Momogu Auto Connect have caught on, and thus the "Chinese solution" for autonomous driving was born.
/ 02 /
Behind Mushroom AutoLink
BAT Zhengfeng Intelligent Electric Vehicle
In September 2021, the "Intelligent Connected Automotive Industry Development Report (Shunyi Index 2021)" released by China Electronic Information Industry Development Research Institute showed , Baidu Apollo and MOU is ranked in the top two and is considered to be in the “first echelon” of the autonomous driving industry.
BAT "starts a war" on smart electric vehicles
In fact, behind the competition between Mogujie and Baidu Apollo is actually the result of the layout of BAT's three traditional Internet companies in the field of smart electric vehicles.
Baidu Apollo also followed the path of Tesla and Waymo earlier. Until the end of 2019, Baidu Apollo head Li Zhenyu announced at the first Apollo Ecological Conference that "bicycle intelligence + vehicle-road collaboration" will go hand in hand, placing vehicle-road collaboration at a strategic level alongside autonomous driving. Baidu externally calls its product "ACE Intelligent Transportation Engine", the core of which is the integration of "vehicle-road cloud map" capabilities; while Momogu AutoLink has launched a "vehicle-road-cloud integration" full-stack technology solution.
What they have in common is that they package bicycle intelligence and vehicle-road collaboration into an overall solution for autonomous driving, and sell the package directly to local governments.
Alibaba focuses on the intelligent operating system, AliOS, in the field of autonomous driving. In 2015, Alibaba and SAIC Motor jointly established Banma Zhixing. In July last year, Banma Zhixing received a new round of capital increase, receiving 3 billion yuan from old shareholders such as Alibaba, and continued to receive strong support from the funder’s father.
In terms of vehicle-road collaboration, at the Yunqi Conference in 2018, Alibaba also announced an upgrade to its automotive strategy, extending from vehicles to roads and using vehicle-road collaboration technology to create "smart highways."This may be based on the previous accumulation of Alibaba Cloud providing visual AI recognition technology on the Hangzhou Ring Expressway.
Baidu personally built cars, Alibaba established joint ventures, and Tencent chose a relatively "conservative" approach. Tengteng officially stated that "Tencent will develop transportation solutions and vehicle-road coordination solutions, but it will not build cars."
Tencent’s “best partner” in intelligent network connection business
Mushroom AutoLink completed a US$120 million Series A financing led by Tencent in February 2019. Since 2018, Momogu AutoLink’s parent company “Zhidao Network” has successively received strategic investments from SF Express , JD.com , Tencent and other companies. As of July this year, Tencent held 15.63% of the shares.
The three brands in Tencent’s hands, “Tencent AutoLink, Mushroom AutoLink, and Wutong AutoLink”, all tend to be intelligent network connections. Tencent even proposed that "autonomous driving and intelligent network connection require to decouple ", which means that autonomous driving and intelligent network connection become two routes.
In the process of external cooperation, autonomous driving and intelligent network services are regarded as "products" that are not affiliated with each other, instead of packaging them together to facilitate the output of a complete set of solutions, as is popular in industries such as Baidu or MOU. MOU is positioned as one of its important partners.
It is worth mentioning that MOU and Baidu Apollo are direct competitors, but the core management of MOU are all from Baidu. Zhu Lei, the founder of MOU, once served as the chairman of Baidu’s Vertical Search Technology Committee. He has since served as vice president of Didi Chuxing, CTO of Yiche, etc., and has rich industry experience; while Guo Xingrong, CTO of MOU, once served as director of Baidu. Architect, chairman of the mobile cloud business group technical committee and other positions.
Guo Xingrong, CTO of Mushroom AutoLink Source: MOUCC’s official Weibo
/ 03 /
takes the lead in commercialization
Why is the industry not optimistic about it?
Regarding the dispute between bicycle intelligence and vehicle-road coordination routes, Pony.ai CTO Lou Tiancheng once made a vivid analogy, "Bicycle intelligence is the ability of every soldier in the army, while vehicle-road coordination is more of a military command system. , the two promote each other and are not contradictory. "
The development paradox of "vehicle-road collaboration" and "bicycle intelligence"
In essence, the relationship between the two is: vehicle-road collaboration cannot be separated from bicycle intelligence, but bicycle intelligence But you can leave the car-road collaboration.
At the 2021 World Artificial Intelligence Conference held in July 2021, nearly 50% of the exhibition area was occupied by autonomous driving, stealing the limelight. It is in such a high-profile situation that many people in the industry have questioned vehicle-road cooperation. According to Caixin report, at the autonomous driving sub-forum, almost all guests expressed their disapproval of vehicle-road collaboration. This includes WeRide , which was ranked fourth in the comprehensive autonomous driving score by the "Intelligent Connected Vehicle Industry Development Report (Shunyi Index 2021)" after Baidu Apollo and Momogu AutoLink.
On the forum, WeRide founder and CEO Han Xu said that bicycle intelligence is like a headlight that must be installed on every car, while vehicle-road collaboration is the streetlight in the city. If a car wants to drive fully autonomously, both Drivers can assist each other, but it is impossible to require a car to not have the ability to drive autonomously.
Su Qing, the former head of autonomous driving at Huawei , pointed out that the vehicle-road collaboration system cannot completely fill in the parts that bicycle intelligence cannot reach. If bicycle intelligence is highly complete, vehicle-road collaboration is not needed. Some people also question that vehicle-road collaboration cannot meet vehicle regulatory requirements.
In their view, vehicle-road collaboration seems to be "tasteless".
On one side, the government supports it in a high-profile manner, but on the other side, many car companies speak against it. Mushroom Automobile Alliance is in an embarrassing situation.
The greater confidence of the opponents may be that Mogu Auto Alliance, as a company in the first camp of vehicle-road collaboration, has an annual order volume of 1 billion, which is already rare. In contrast, car manufacturers have deep pockets and are naturally confident in their speech. Taking Xpeng Motors as an example, the full-year revenue in 2021 will reach 20.99 billion yuan. Players from the two camps are hardly the same in terms of revenue.
Is it destined to be a self-driving "ferryman"?
What the opponents fail to mention is one of the biggest advantages of vehicle-road collaboration: cost. According to Moguolian’s calculations, through the intelligent transformation of roads, roadside sensing and cloud sensing will effectively supplement the lack of bicycle sensing, and the cost of automatic driving modification of bicycles is expected to be reduced by more than 90%.
How expensive are the bicycles built under the intelligent bicycle route? The cost of an ordinary taxi is only 200,000 yuan at most, while the cost of a Robotaxi (self-driving taxi) can reach millions. A complete set of radar , cameras, sensors and other accessories are expensive. This route is still favored.
For example, Baidu’s Robotaxi business “radish run ” has served more than 400,000 users, although it currently has obvious disadvantages in terms of operating area and price. The spin-off Didi Autonomous Driving Company once stated that "by 2030, Didi Chuxing plans to operate more than 1 million autonomous vehicles through its online ride-hailing platform."
MOU's self-driving taxi Source: MOU's official website
What they value is the consumer market where costs will be reduced after mass production in the future. Facing the government's G-end market, Moguchi AutoLink is indeed "sustainable" due to policy dividends, but few B-end companies pay for it.
Before the advent of large-scale consumer-grade products for autonomous driving, Moguolian Automotive invested as a light asset, focusing on third-party services, and tasted the sweetness of the commercialization of domestic autonomous driving.
MOULIN’s large-scale autonomous driving operation projects have been carried out in Beijing, Shanghai, Henan and other places. Why can Mushroom AutoLink run faster? In September last year, Zhu Lei explained to the outside world, "What we do is actually very simple, it is a technical system plus an operating system."
The set of solutions including the operating system provided by Mushroom AutoLink is precisely It makes up for a major shortcoming of the autonomous driving government project. The Chinese government uses vehicle-road collaboration as a means to alleviate traffic congestion and help implement high-level autonomous driving projects. According to previous reports from Caixin, it is not clear to what extent vehicle-road collaboration can be achieved in the future, and many local governments hold the idea of "entering the game first and seizing the high ground." However, some projects have even become vacant due to lack of subsequent operations after the initial construction was completed.
Illustration of application scenarios of Mogu Connect products Source: Mogu Connect official website
As a kind of "transitional" project, whether it can enter the mass market on a large scale in the future has not yet been verified. However, none of these can stop the momentum of Mogu Auto Association's rapid progress. The last time Mogu Auto Alliance got out of the industry was because of its high-profile, high-paying recruitment campaign.
At the end of 2021, Mushroom Automobile Alliance launched the 2022 autumn recruitment and released the "Double Hundred Talent Plan", looking for 100 top graduates with an annual salary of more than 1 million. When major Internet companies are laying off employees one after another, Mushroom AutoLink has shown another aspect. Mushroom AutoLink’s R&D team accounts for more than 80%, and its core R&D teams come from companies such as Tencent, Baidu, Microsoft , Huawei, BMW and other companies. The 100 graduates with an annual salary of over 1 million are probably concentrated in technical talents.
Perhaps Mushroom AutoLink can establish a technical moat to buy time for future transformation.
node AUTO statement: The content of the article is for reference only. The information or opinions expressed in the article do not constitute any investment advice. Node AUTO does not assume any responsibility for any actions taken due to the use of this article.
