Original link: https://www.latepost.com/news/dj_detail?id=1758
“Later Auto” exclusively learned that the autonomous driving supplier Heduo Technology has recently completed the last round of financing of the C round, with a financing amount of hundreds of millions of yuan. The investment was jointly led by Yueke Finance and GAC Capital. Subsidiary financial investment institutions. Horizon, an autonomous driving chip company, also invested in Heduo.
In March last year, GAC Capital announced its exclusive investment in Heduo’s C1 round of financing. In November last year, Heduo completed another C2 round of financing of hundreds of millions of yuan, led by GAC Capital, followed by Zhidu Investment and Chaos Investment. The total financing amount of Heduo’s C round is RMB 1 billion.
Ni Kai, the founder of Heduo, told “Later Auto” that Heduo expects to narrow its losses significantly in 2024 and achieve breakeven in 2025, when Heduo will start its IPO.
Established in June 2017, Heduo Technology provides car companies with integrated solutions for advanced assisted driving and parking. At the same time, it also designs domain controllers by itself, and can sell products integrating software and hardware to car companies. Since its establishment, Hedu has completed more than one billion yuan in financing. Investors include financial institutions such as IDG, Sequoia China, and Yueke Finance, as well as auto industry companies such as GAC Group and NavInfo. Ni Kai, the founder of Heduo, was a senior scientist at Baidu Deep Learning Research Institute and one of the important early technical leaders of Baidu’s autonomous driving business.
It is understood that GAC, which has invested in Heduo many times, currently accounts for about 13% of the shares of Heduo, and is the largest external shareholder; GAC is also the largest customer of Heduo. The assisted driving solutions of GAC’s Trumpchi, Aian and Haobo brands are all jointly developed by GAC and Heduo.
In the first half of this year, GAC Group sold a total of 1.163 million vehicles, including 236,000 new energy vehicles, ranking third in China’s new energy passenger vehicle market, second only to BYD and Tesla, with a market share of 6.8%.
With the support of major customer orders, Holdo is now one of the leading smart driving suppliers in mass production. Ni Kai told “Later Auto” that by the end of this year, the cumulative number of mass-produced vehicles equipped with Heduo’s pre-installed assisted driving solution will reach tens of thousands. He believes that 200,000 vehicles is the threshold for Heduo to survive, which is also the prerequisite for Heduo to achieve profitability and start listing.
Accelerate mass production, become more standardized and cheaper
The demand for advanced assisted driving was initially initiated by Tesla, which is known for its intelligent capabilities. In 2020, Tesla will sell well and become the car company with the highest market value in the world. This has stimulated various car companies to imitate Tesla’s products and technologies and increase investment in advanced assisted driving.
A group of new car companies such as Xiaopeng have been developing advanced driver assistance functions by themselves since 2016, while other car companies have chosen to cooperate more with external suppliers after 2020. A group of third-party smart driving solution suppliers, such as In-Vehicle and Shangtang Jueying, have created a market window.
As it has evolved, the competition around ADAS is now centered on features such as Highway NOA and Urban NOA. NOA (Navigate on Autopilot) is pilot assisted driving. It means that within a certain area, the vehicle can complete functions such as road recognition, lane change, and overtaking by itself, which is close to “driving by itself”. High-speed NOA is relatively simple and mature, and urban NOA is more complicated . Gaogong Intelligent Automobile Research Institute predicts that the number of NOA pre-installed standard equipment may sprint to 500,000 vehicles this year; if optional models are included, it is estimated that the potential NOA installation volume for the whole year will exceed 1.5 million vehicles.
For suppliers like Heduo, the key to whether they can seize the current market window is whether they can help their car company customers get more sales and share with their intelligent capabilities, and at the same time obtain sufficient income for their own survival and development. profit.
There are two challenges in terms of capabilities: one is to improve the success rate and efficiency of delivery; the other is to improve the cost performance of the solution and maintain a certain profit in the game with car companies.
In terms of delivery speed, the self-developed camp is currently one step ahead: Xiaopeng has supported NOA in five cities, and Huawei plans to support urban NOA in 45 cities by the end of the year; Ideal claims to support 100 cities by the end of the year.
The cooperation model between suppliers and car companies generally starts later. A group of top suppliers are in the transition stage from serving the first big customer to serving more car companies in batches. For example, after serving GAC, Heduo announced that it has won orders from Chery and Dongfeng this year; Momenta was the first to serve SAIC Zhiji, and later won customers such as Great Wall, BYD, and Lotus; DJI first cooperated with Wuling Baojun kivi EV, After that, it won orders from brands such as Baojun Yueye and BYD.
In the process of expanding models, different brands and different models have differences in functional requirements, chips and sensors used. This brings short-term challenges to suppliers serving multiple car companies at the same time, how to complete customization more efficiently Delivery; their long-term topic is: how to improve the degree of reusability and standardization of products, and reduce customized components.
According to “Later Auto”, during the cooperation process between Heduo and GAC, there were more than 150 R&D personnel stationed in the factory at most, and Heduo has more than 500 R&D personnel in the whole company.
Ni Kai explained that the reason why so many engineers were needed to serve GAC was that when cooperating with the first model of the car company’s new platform, it would involve platform-based R&D and engineering issues, and the manpower and time costs were higher. When serving subsequent models on the same platform, the required resources will be greatly reduced. For example, he said that it took them two and a half years to serve the first car on the new platform of a certain car company, but it only took 4 months to make the second car, and the number of engineers needed was only the first one. About 1/10 of the model.
Ni Kai believes that advanced assisted driving will eventually become a relatively standardized product, because its goal is to become an “veteran driver” and its functions will eventually converge. As more and more mass-produced cars equipped with advanced assisted driving capabilities, the current Many path divergences will converge. Just before that, it will go through a process of accumulating product definition capabilities and engineering capabilities through customized services. Hedo is also ensuring the standardization of the delivery process as much as possible through product definition, system development, and system verification systems, which is the “basic skill” of every autonomous driving supplier.
Ni Kai said that Heduo will launch the first version of the urban NOA solution this year, and in the second to third quarter of next year, there will be mass production and launch of models equipped with the Heduo urban NOA solution.
The second challenge in the volume-up process is to reduce costs and improve cost performance. According to comprehensive industry information, for models below 150,000 yuan, the cost of intelligent driving software and hardware acceptable to car companies needs to be controlled at around 5,000 yuan, for models priced at 150,000-250,000 yuan, it is 5,000-12,000 yuan, and for models above 250,000 yuan, the cost can be relaxed to more than 12,000 yuan.
Ni Kai told “Later Auto” that Hedo’s current cost reduction method is to improve the overall development efficiency and reduce internal costs; the second is to optimize the algorithm and reduce the dependence on hardware performance, such as using fewer sensors and reducing the need for high-precision Dependence on maps, etc., to reduce the cost of the entire solution.
Two modes of competition
In addition to capability challenges, the greater uncertainty facing suppliers such as Hedo comes from the undecided market structure: that is, how many car companies will really need suppliers to provide intelligent driving capabilities in the future?
Ni Kai’s judgment is that in the future, the ratio of car companies’ self-research and intelligent driving suppliers’ services will be 1:9. Only leading car companies with both sales volume and R&D strength will persist in self-research for a long time. This is because car companies that sell too few cars can hardly bear the cost of continuous development and iteration of intelligent driving systems, and the economic accounts cannot be settled; the second is because intelligent driving capabilities will eventually become standard products, and the differentiation is not obvious, nor is it strategically The car companies have to do it themselves.
Car companies that firmly follow the self-research route hold the opposite view. Before the end of “standardization” arrives, the consensus of many car companies is that whoever can come up with better intelligent capabilities faster will help increase sales, seize market share, and win the war. Self-developed by car companies, with higher communication and collaboration efficiency, theoretically faster R&D and iteration.
Another benefit of self-research by car companies is that they can innovate business models across the board. For example, Musk said in the earnings conference call in the first quarter of this year that Tesla will even be able to sell cars at zero profit in the future and rely on autonomous driving software services to make money.
What affects the market share ratio of self-research and suppliers is how concentrated the smart electric vehicle market will be in the future. One possibility is that due to the increase in “silicon content” and software capabilities, smart electric vehicles can rely on the mutual drive between data and software experience, and obtain a certain network effect in addition to the scale effect of traditional fuel vehicles, that is, buying a car, More people using cars → more data → better intelligent driving experience → attracting more people to buy cars. A small number of top companies will eat more market share, they have both the scale and ability of self-development, but also willingness.
At present, the two largest brands in the new energy vehicle market are Tesla and BYD. The former will continue to conduct self-research, while the latter is also aggressively advancing its self-research plan, and has organized a self-research team of 3,000-4,000 people for intelligent driving.
In the first half of this year, these two brands together accounted for 46% of China’s new energy market and 10.9% of the entire passenger car market. The market share of leading car companies is continuing to increase. From 2020 to 2022, the total market share of the top five car companies in China will climb from slightly more than 45% to nearly 60%.
Ni Kai’s observation is that the market concentration is indeed higher than in the past, but it will not be so high that only a few companies dominate the entire market. Considering consumers’ diverse needs for cars and local governments’ demands for supporting local auto companies, He judged that there will be about 8 mainstream brands left in the Chinese market in the end, and the annual sales threshold will be above 1 million to 1.5 million vehicles. Ni Kai believes that all brands that adopt supplier solutions can support three major suppliers: “the three will not be too monopolistic, but they can also have sufficient scale.” The pattern of solution providers will be settled earlier than that of vehicle brands .
The founder of another assisted driving solution provider also told “Later Auto”: In the future, assisted driving suppliers will be very concentrated. In the end, there may be two in the Chinese market, and a total of three in the world. He also believes that autonomous driving is not as differentiated as the car itself, “it is only good and better.”
The general view in the market is that the structure of China’s new energy vehicle market will be initially determined in 2025, and the window period left for suppliers may not exceed 3 years.
This article is transferred from: https://www.latepost.com/news/dj_detail?id=1758
This site is only for collection, and the copyright belongs to the original author.