Article | Mr. Li from Financial Street
The past two years have been a low point and reshuffling period for the autonomous driving industry, but the entire industry has not slowed down in its pace of researching and developing technology and products.
Despite recent negative news such as start-up companies closing or downsizing, industry giants have been completely unaffected by the winter chill of capital and have accelerated their progress. On November 29th, Baidu announced during Apollo Day, its technology open day, that it will continue to expand its business scale, and aims to build the world’s largest fully unmanned autonomous driving operation service area by 2023.
Many people say that Robotaxi is the key factor to propelling the autonomous driving industry from the “Spring and Autumn Warring States” to the “Warring States Hegemony”. Today, Mr. Li will discuss with everyone why Robotaxi has blown away the madness and sand to strike gold, who the leader of Robotaxi is, and what challenges Robotaxi faces.
Blowing Away the Madness and Sand to Strike Gold
There is a consensus in the capital market that the industrialization of autonomous driving is much more difficult than electrification. To this day, the autonomous driving industry is still seeking a way out. Robotaxi is a path that can simultaneously cater to market size and be put into practice, a road paved by funding.
Over the past seven to eight years, the development of autonomous driving can be divided into several stages, each with its own characteristics, but overall, Robotaxi is the application scenario developed through several stages of exploration:
The first stage is from 2015 to 2017. At that time, autonomous driving was a new concept in China. Industry giants and start-up companies sprouted up like mushrooms after rain, and everyone was exploring their own technology route and application scenario, lacking a systematic understanding.
2018-2019 was the first bottleneck period in industry development. After development on a “coarse” scale for three to four years, with the cooling of the capital market, everyone became more rational, and the industry’s “hot money” decreased. At this stage, everyone was “refining” practical scenarios, and Robotaxi began to become popular in the industry.
From 2020, both the industrial and capital ends have received good news. On the one hand, the commercialization progress of autonomous driving such as Robotaxi has accelerated, and whether in specific scenarios or open roads, everyone has seen hope for practical application. On the other hand, electrification companies have prospered in market value, with companies such as Tesla, NIO, and XPeng achieving immense success in the capital market. Many funds have flexibly exited in B/C round financing of start-ups, allowing the capital market to enter another investing boom. At this stage, everyone realizes that the most important thing about autonomous driving is still “practical application”, which cannot be separated from Robotaxi.
From 2021 to now, with the slowdown of investment pace in the capital market, the development of the autonomous driving industry has once again slowed down. This is a normal phenomenon; any industry must inevitably face reshuffling. At the current stage, the autonomous driving industry has appeared polarized: powerful players who have achieved full unmanned landing and are steadily expanding their areas and promoting the commercialization of scale will usher in new development opportunities. On the other hand, companies that lack sufficient accumulated technology to cross the threshold of unmanned landing may face being left behind as the tide recedes.
There are many scenarios for the landing of autonomous driving, mainly divided into closed scenarios and open scenarios. Many companies focus on the closed road scenario to achieve autonomous driving. However, the problem is that the size of the domestic closed autonomous driving market is not large, whether it is mining, hotels, ports, or distribution. There are many start-up companies and products competing in each scenario, and none of them are profitable.
Capital and industry are most concerned about open roads for several reasons: First, the market ceiling of this scenario is high, and the Robotaxi alone has a market space of trillions of dollars. If the scope is switched to the entire passenger car autonomous driving, it would be trillions of trillions. Second, the industry barriers are high, including technology barriers, capital barriers, and ecological barriers. The leading companies are most likely to emerge in this direction.
In the medium to long term, open scenarios are the trend of development, and Robotaxi is one of the few scenarios that can take care of both market prospects and landing applications.
Who is the leader of Robotaxi?
From the perspective of business models and technological development, autonomous driving and ride-hailing are indeed a “good CP” that will promote the development of the travel era with their perfect integration.
There are two ways for travel companies to improve profitability: either by improving technology to release productivity or by finding and opening up new markets. Currently, whether it is in first- and second-tier cities or third- and fourth-tier cities, market share has stabilized. Therefore, everyone can only approach the issue from the perspective of increasing productivity, in which autonomous driving is a good means.
For autonomous driving companies, the development of technology must rely on data. In cases where open road testing is not feasible, taxis in specific areas are the best source of data. At the same time, operating autonomous taxi services not only demonstrates the capabilities of the technology but can also generate revenue, thus forming a commercial cycle.
Beginning in 2020, autonomous driving companies such as Baidu, XiaoMa Zhixing, and Wenyuan Zhixing have been developing Robotaxi initiatives, which many people have already experienced. However, Robotaxi has not yet achieved large-scale growth in the past two years due to two primary reasons: First, the technology is not mature yet, and second, large-scale development requires significant capital investment, which start-up companies cannot afford, and rental rideshare companies and other partners view as too risky.
Within domestic enterprises, Didi and Baidu are the most likely to develop Robotaxi on a large scale. Different opinions exist regarding Didi, but the advantages of Baidu are their strong technological and capital strength. Currently, they are expanding the commercial operations of full unmanned autonomous driving in Chongqing and Wuhan, followed by further expansion into Beijing.
Baidu’s Q3 financial report data shows that as of the third quarter of 2022, the total number of LRB taxi orders exceeded 474,000, which represents a YoY growth rate of 311% and an MoM growth rate of 65%, with an accelerating growth of orders for full unmanned autonomous driving in Wuhan and Chongqing. In addition, in the three first-tier cities of Beijing, Shanghai, and Guangzhou, the average daily order volume per car is over 15 times, approaching the daily order volume for traditional rental rideshare services. From various perspectives, such as the scope of coverage, order volume, and user stickiness, LRB has formed a momentum of “connecting points, accumulating lines, and forming a surface.”
During the Apollo Day technology open day event, Baidu also announced the expansion of business scale in 2023 to launch full-scale operation of full unmanned autonomous driving services in more regions, striving to create the world’s largest full unmanned autonomous driving service region.
Moreover, Baidu mentioned that the progress of the autonomous driving technology’s generalization ability exceeded expectations, and the technology delivery time for landing in new cities only requires twenty days, which indeed exceeds industry expectations.
Mr. Li recalls the Shanghai Auto Show last year, where Huawei and various car makers demonstrated their autonomous driving technologies. Especially Huawei, which under the halo of a full-stack solution, laughed at the industry with shiny and beautiful moves. Will Baidu in 2023 stir up the autonomous driving enthusiasm of the industry, capital and users like Huawei? We shall wait and see.Developments in autonomous driving have led many to explore the path of integrating L4 and L2+ technologies, with the former providing technological transfer to enable the delivery of safer and more convenient urban L2+ autonomous driving products. In turn, data from L2+ products can help enhance the generalized capabilities of L4, promoting the inevitable trend of technological and commercial integration.
As Huawei’s friend often says, Huawei’s autonomous driving business model is “borrowing chickens to lay eggs, and then focusing on ADAS.” In contrast, Baidu’s model is to “focus on Robotaxi and explore and promote mass production applications.” This approach is also adopted by foreign autonomous driving giants such as Waymo, which recently obtained approval from CPUC in California to provide driverless ride-hailing services to the public in certain areas, including San Francisco and the Bay Area.
The development of Robotaxi appears promising, but it is faced with many challenges. As in many industries, progress is not always made by first addressing challenges and then developing, but by solving problems during the process of development. As someone involved in industry investment in Robotaxi over the past few years, the biggest challenge for Robotaxi is still technical. Compared to electrification and networking technologies, autonomous driving technology requires the integration of much more fields, including software and hardware, scene coordination, and ecosystem support. In Robotaxi, these aspects have been deliberately defined as the boundaries.
In closed environments, there is a clear mass production roadmap for L4 and L5, especially in mines, ports, unmanned logistics vehicles, and other scenarios. However, in open environments, it is difficult to guarantee when L5 can be realized. Additionally, open road environments are highly complex. Aside from driving routes, Robotaxi and ordinary vehicles have little difference in their operating conditions–reaching speeds of up to 120 km/h–which poses higher requirements on the perception, decision-making, and execution capabilities of the vehicle. Currently, due to limitations in the technology of perception devices and braking devices, many problems are difficult to solve, and no one can ensure a clear node for implementation, hence the high level of attention given to Baidu’s Robotaxi.Consider the business model when Robotaxi is landing on a large scale, and see if the economic model can be successful. The core business logic of Robotaxi is that autonomous driving replaces human operation. Only when the operating cost of Robotaxi is lower than the labor cost, can it have the opportunity for mass promotion. Currently, the software and hardware costs of Robotaxi are decreasing while labor costs are continuously rising. Before the intersection point arrives, Robotaxi is mainly self-operated by autonomous driving companies. For Baidu, funding is its strong advantage.
Finally, there is the issue of road rights. Although the government has a positive attitude towards the roadworthiness of autonomous driving and is actively promoting the legislative process, Shenzhen has already pushed for it, and Beijing is also issuing licenses. Overall, the introduction of various regulations for opening roads is still lagging behind. However, compared to technology and business models, policy is not the main factor affecting the development of the industry.
Many friends ask, if Robotaxi has so many problems, what is its development value? Fundamentally, Robotaxi sets boundaries for open scenarios, reduces technical difficulty, and then achieves autonomous driving. In terms of vehicle speed, Robotaxi can flexibly set the maximum speed, greatly reducing the technical difficulty. In terms of roads, Robotaxi can choose the easiest landable sections of the city for pilot testing. Within the open demonstration zones, Robotaxi can carry out commercial operations of online taxi through man-machine collaboration.
The current landing of Robotaxi is not from 0 to 1 in the autonomous driving industry, but from 0 to 0.5. However, during the process from 0 to 0.5, by iterating and evolving the technology to gradually expand the scenario, it can drive the autonomous driving industry from 0.5 to 1.
Currently, the automotive industry has shifted from electrification competition to intelligent competition, and will ultimately move towards service competition for shared mobility. From a business logic perspective, if a company does well in the Robotaxi scenario, its scalability will be very good, such as supporting advanced autonomous driving or even building cars, and Baidu is laying out for these areas.
Autonomous driving competition is shifting from “Spring and Autumn Period” to “Warring States Period”.## 《Zhengming》 moves forward to “Warring States dominance”, and who can lead the Chinese autonomous driving industry to the best solution, the answer still needs time to unfold.
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.