Author: Michelin
$175 billion, that was the valuation that Morgan Stanley gave to the autonomous driving industry leader Waymo in 2018. This valuation can be said to be the “peak moment” of the autonomous driving field, and it has also become the “target” of later valuations of autonomous driving companies. In 2021, Waymo’s latest financing estimate is only $30 billion, less than one fifth of that in 2018.
This is not the fate of Waymo alone. From concept emerging, valuation soaring, to bubble clearing and valuation falling, autonomous driving has almost experienced the old road that every emerging technology industry has gone through, ultimately the bottleneck lies in the difficulty of commercialization.
In the article “Robotaxi, the new battlefield of the world’s richest people?”, it was mentioned that “Robotaxi is difficult to commercialize and land, resulting in continuous blood loss. When capital is no longer stimulated by the growth of test data, only giants can afford to play this game.” However, the giants do not want to burn money forever, but are looking for commercialization opportunities.
Enabling autonomous driving technology in various industries has become the unanimous choice of autonomous driving companies today. After persisting in the Robotaxi field for many years, Waymo recently partnered with US logistics company Robinson to officially enter the autonomous driving truck business, trying to use commercial autonomous driving to support larger valuations.
Not only Waymo, but also domestic Mushroom Car Union is creating a car-road cloud integration solution to build a car-road cooperative project; Baidu Apollo is trying to land autonomous driving technology with its own Jidu cars; Wenyanzhixing has launched unmanned sanitation vehicles; Momenta has cooperated with SAIC to apply assisted driving solutions to mass production vehicles…
When autonomous driving sets foot on the road of commercialization and landing, everyone’s expectations have turned from chasing the distant prospects to being more practical, and more emphasis is placed on the landing of each order and each project. In this “practical era” of autonomous driving, we have to ask: who will pay for autonomous driving after all? What is the new logic of autonomous driving business?
About autonomous driving: What are the real needs and pseudo-strong needs?
“Are you willing to pay for autonomous driving functions?”Previously, AlixPartners, a global consulting company, released a report on global consumer attitudes towards autonomous driving. Chinese consumers are the most positive towards autonomous driving, with 51% of them more inclined to wait and see, while the remaining 49% have a purchasing inclination but are only willing to pay an 8% premium for L4-level autonomous driving features.
If the report seems somewhat “abstract,” there is a more direct example: the Tesla FSD, the most successful commercially-driven autonomous driving product currently available. As of the end of 2020, only 1-2% of Chinese car owners have chosen to purchase it, even though a monthly software subscription option has been available since last year. The reason behind this is that the fully autonomous driving feature in the FSD service is still in “futures” and it is uncertain when it will be activated in China.
It is worth noting that Tesla’s audience is comprised of users who are relatively open to radical technology. Other autonomous driving pay-per-use services are therefore self-evident.
When it comes to autonomous driving, the “ultimate goal” we often think of is to achieve unmanned driving in our private cars. However, this places enormous demands on the ability of autonomous driving to cope with complex road scenarios. Such high difficulty makes today’s autonomous driving more like “semi-finished products” in evolution. It is almost impossible for users to pay for features that are still in the process of evolution, which has become a challenging issue for all car and technology companies. It even raises questions: do users really have a demand for autonomous driving?
When it is difficult to replace human labor with systems in the short term, or to use systems to do the work that humans find difficult or are unwilling to do, it seems more sensible to expedite the implementation of autonomous driving. Taking the unmanned sweeping vehicle launched in Hunan’s Hengyang by MogoCarLink this year as an example, the aging of sanitation workers has made it difficult to recruit younger employees, and there is a shortage of labor. In addition, older workers are faced with safety hazards when working in extreme weather conditions, such as at night, in dangerous areas, or in heavy rain and high temperatures.In this case, the combination of bicycle intelligence and vehicle-road coordination allows unmanned sweeping vehicles to safely operate in designated areas, enabling the system to do work that is difficult or unwilling for humans to do. According to Mushroom Car Link, a small autonomous sweeping vehicle can undertake the workload of 3-5 workers, solving the problem of cleaning in dangerous environments and labor shortages, and can also help some sanitation workers transition to machine operators. The situation is similar in the fields of road maintenance, public transportation, security, and distribution.
As for ordinary consumers, not everyone may choose to buy autonomous vehicles, but they certainly hope for less congestion on the road and even for their non-“smart” vehicles to receive timely intelligent reminders.
Compared to directly letting users foot the bill for autonomous driving products, these may be the “real needs” of people for autonomous driving at this stage. These “real needs” have given autonomous driving new commercial prospects, and the interconnection between vehicles, roads, and intelligent platforms has also lowered the technological threshold for the implementation of autonomous driving.
The first step in the business of autonomous driving is to discover the demand. However, it is the continuation of the business of autonomous driving to meet the demand, achieve rapid implementation, and create value.
Improving the comprehensiveness of capabilities can accelerate the implementation of autonomous driving.
The rendering of the level of bicycle intelligence has always made the public believe that future intelligent cars are like robots with three heads and six arms, capable of seeing six roads and blocking one. However, just like the carriage era needed roads and the automobile era needed operating rules on the road, in the era of intelligent cars, “intelligent” should not only refer to cars but also to the intelligent management of operating rules.
After all, even the most sensitive LIDAR cannot perceive the road conditions several kilometers away, and even the smartest autonomous driving cannot predict the fantastic driving of the driver in the adjacent vehicle, let alone the mixed-quality intelligent operating system today. For an ordinary user, if the only way to experience intelligence is to change cars, the implementation of autonomous driving will be too difficult.
“Vehicle, technology, and operation are all indispensable.” Recently, the head of a domestic autonomous driving company described the difficulty of implementing autonomous driving. Here, operations refer to the coordinated cooperation between vehicles and vehicles, roads, environment, and the entire city’s transportation system.With bicycle intelligence and vehicle-road cooperation, we can solve “corner cases” that current intelligent driving cannot handle by leveraging the “God’s eye view.” This enables non-intelligent vehicles to experience the convenience of intelligent transportation through vehicle-road cooperation, which naturally speeds up the landing of autonomous driving.
As the first municipal-level automatic driving intelligent transportation project in China, the Hengyang project covers 200 kilometers of city open roads and adopts the “Vehicle-Road Cloud Integration” system 2.0 solution from Mushroom Car Union.
In this solution, Mushroom Car Union starts with public service vehicles that people find difficult or unwilling to operate, such as automatic driving sanitation vehicles, patrol cars, and regional logistics vehicles, so that L4 autonomous driving technology is not just limited to the experimental stage, but can also create value in practical scenes.
On the other hand, Mushroom Car Union monitors road information in real-time through the layout of intelligent road testing facilities and establishes a city-level intelligent traffic management cloud platform. This way, even vehicles and pedestrians without autonomous driving capabilities can receive real-time traffic information.
Does an emergency accident occur ahead? Is there a sudden congestion during rush hours? Which route is safer and traffic-free during a heavy storm?… All of these alerts can be collected, analyzed, and issued to everyone through the city’s traffic management AI cloud platform.
When autonomous driving is only a capability of a car itself, people can only experience it by purchasing an autonomous driving car or taking a Robotaxi in a demonstration area. However, when autonomous driving cooperates with roadside facilities and intelligent transportation platforms, people can experience the convenience of autonomous driving simply by walking on the road.
Autonomous driving landing: From 1 to 100 in replication capabilityIn 2015, domestic autonomous driving companies began testing Robotaxi on highways and city roads. Seven years have passed, and while the number of test vehicles and the scope of the test sites have increased, the goal of large-scale promotion still seems far away. It remains difficult to promote L4 level autonomous driving cases in a demonstration area to the entire city and to simply copy FSD tests conducted in the US onto Chinese roads. This is the difficulty in the popularization of intelligent bicycles from 1 to 100.
Instead, the mode of intelligent bicycles + vehicle-road coordination is relatively easy to replicate with the support of new roadside infrastructure. By building smart road facilities and intelligent traffic platforms, the level of intelligent bicycles can be improved.
As mentioned earlier, since March 2021, Hengyang has begun to use the “vehicle-road-cloud integration 2.0” system provided by Mogu Car Union. The city combines urban traffic new infrastructure, autonomous driving operation services, and urban smart transportation cloud platforms to create a city-level autonomous driving project.
One year later, this model has started to be replicated in cities and regions such as Hebi, Dali, and Sichuan Tianfu New Area. Among them, the 30-billion-dollar order Mogu Car Union recently signed with Sichuan Tianfu New Area has created the highest order amount among publicly disclosed autonomous driving projects worldwide. Both parties will jointly develop autonomous driving, vehicle-road coordination, AI cloud, and other fields, and deploy various autonomous driving city services such as autonomous driving rideshare, autonomous driving shuttle, and autonomous driving delivery vehicles.
The secret to replicating intelligent bicycles + vehicle-road coordination from 1 to 100 lies in the deployment and operation of autonomous driving public service vehicles to reduce public service costs and the use of smart networked roads and smart traffic AI cloud platforms to improve traffic operation efficiency.
Finally
In the seventh year of road testing for autonomous driving in China, some companies choose to collaborate with automakers to explore the mass production of autonomous vehicles; some shift towards commercial scenarios, attempting to deploy autonomous driving through freight logistics; and some aim to implement autonomous driving through city-level planning…
Compared with a few years ago, the autonomous driving road today is more diverse and practical. The fundamental challenge lies not only in the rapid technological advancement, but also in finding a suitable business model that suits China’s national conditions and can be successfully implemented. The road to autonomous driving is still long, and only those who travel farther and longer have a chance to truly win.
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.