Text | Online Travel, Author | Xiong Yuge, Editor | Chen Feng
The autonomous driving company Haomo Zhixing is about to face a big test.
Recently, late auto reported that Great Wall Motors has introduced Yuanrong Qixing as an intelligent driving supplier. The latter will provide end-to-end intelligent driving solutions and plans to launch three models this year.
This means that in the context of the large-scale implementation of urban NOA (navigation assisted driving) based on the bev (bird's eye view) + transfromer model, and the accelerated penetration of the "end-to-end" model into the industry, Great Wall Motors is actively expanding outwards. Avoid the potential risks of relying on a single smart driving supplier while accelerating intelligence.
Prior to this, Haomo Zhixing became the first and only smart driving supplier of Great Wall Motors by virtue of its deep relationship with Great Wall Motors.
The origin of the two parties is that in 2019, Great Wall Motors separated the intelligent driving forward-looking department and established Haomo Zhixing. After becoming independent, Haomo Zhixing introduced some new forces from the leading intelligent driving companies in the industry to form a new company headed by Chairman Zhang Kai, CEO Gu Weihao, cio Zhen Longbao, and coo Hou Jun are the core team.
However, at present, Tianyancha shows that through layers of equity penetration, the actual controller of Feimo Zhixing is still Wei Jianjun, chairman of Great Wall Motor Co., Ltd., with a shareholding ratio of 34.9601%.
In this way, a difficult problem facing Fei Mo Zhixing is obvious:
How to get out of the comfort zone of Great Wall Motors under the trend of rapid technology iteration - it needs to avoid the potential risks of Great Wall Motors, the single largest customer, but In terms of installed capacity, customer expansion, profitability and other dimensions, we must be truly independent and respond to market tests alone.
Lianxingxing noticed that Hao Mo Zhixing is already actively pursuing changes.
In February this year, Haimou Zhixing completed a B1 round of financing exceeding 100 million yuan. The investor was Chengdu Wufa Fund. The funds raised in this round will mainly be used for the research and development of AI autonomous driving technology such as Haimou large models. At the previous China Electric Vehicles 100 People Forum (2024), Feimo Zhixing CEO Gu Weihao also said that its second-generation hpilot noh smart driving products have also entered delivery status.
Picture source Haimou Zhixing official website
For Haimou Zhixing, this big independent test has just begun. Whether it can continue to prove its ability in the future and get more orders from more car companies is whether Haimou Zhixing can develop in the long term. The essential.
When you are smart, you must learn to "walk independently"
Great Wall Motors introduced Yuanrong Qixing as its second smart driving supplier, focusing on its "end-to-end" smart driving capabilities.
Previously, new energy car companies have adopted the traditional divide-and-conquer approach for their intelligent driving systems for a long time—dividing the autonomous driving task into three independent models: perception, prediction, and planning, and then integrating the system to complete the entire autonomous driving. Task.
The disadvantage of this model is that under the rule-based modular model, the upper limit of autonomous driving capabilities is low, the system generalization is poor, and the processing capabilities of unseen scenes and complex scenes are often insufficient. Under the "end-to-end" technical path of
, the three models of perception, prediction, and planning are integrated into one. There is no need to formulate rules through lengthy codes, but to train the system through massive data, allowing the machine to have autonomy. Learning, thinking, and analytical abilities enable you to better handle complex driving tasks.
This is also the hottest concept in the field of autonomous driving this year. At the beginning of this year, Tesla launched the end-to-end autonomous driving system fsd v12 in North America, which caused a craze. The new domestic automobile force "Wei Xiaoli" followed suit and announced the end-to-end smart driving mass production time.
Yuanrong Qixing was founded in 2019. It will fully shift to an end-to-end model in early 2023 and complete road testing in August. At Nvidia's GTC conference in March this year, Yuanrong Qixing CEO Zhou Guang took the stage and announced that Nvidia's next-generation drive thor chip will be used in 2025 to adapt the company's end-to-end intelligent driving model.
However, there are also voices that believe that end-to-end is still in the concept stage of speculation, and whether it can actually be put into use remains to be verified.
Great Wall no longer has the upper hand in urban NOA (high-end intelligent driving assistance) layout. In January 2024, Xiaopeng announced that 243 cities across the country would launch urban xgnp, becoming the first car company to implement urban noa on a large scale.
In the first half of 2022, Xpeng Smart Driving and Nio nop announced that their cumulative mileage exceeded 140 million and 160 million kilometers respectively. As of March, the assisted driving mileage of Haimo hpilot users has exceeded 130 million kilometers.
Under such circumstances, Great Wall Motors obviously does not want to miss any opportunities for technological changes. "Neither Yuanrong nor Haomo has done end-to-end mass production. Great Wall is now walking on two walking sticks and sharing risks." A person close to Great Wall said this about Wawanan Auto.
It is worth noting that the support of Great Wall Motors is particularly important in the development of Hao Mo Zhixing. The latter is not only the controlling shareholder of Haimou Zhixing, but also the company's largest customer. Haimou Zhixing's smart driving product hpilot is mainly installed in Great Wall Motors, with more than 20 related models.
Now that it has lost its “exclusive supplier” advantage, it has become even more urgent for Hao Mo Zhixing to seek independence.
The suv tank 700 hi4-t will be released in February 2024. It is equipped with the Haimo high-end smart driving hp350. Source: Great Wall Motors official website
One of the very important reasons is that for Haimou Zhixing, under the past deep binding model, its market expansion In fact, it is a constraint - autonomous driving technology requires the input of massive data. However, because of its relationship with Great Wall, many car companies are "wary" to a certain extent and are unwilling to open the data to Haomo Zhixing.
As of April this year, according to public information sources, Haomo Zhixing's products have not yet been installed on models of other brands other than Great Wall Motors. According to "Wanwan Auto", Feimo will obtain two new OEM contracts in 2024, including a mass production order from a leading OEM.
This is undoubtedly a key step for Hao Mo Zhixing to expand its customer base. In April 2023, Zhang Kai once revealed that Haomo Zhixing had signed fixed-point contracts with three main engine manufacturers, and the projects were already being delivered. But so far, no new progress has been disclosed in these cooperations.
2024 is a critical year for Hao Mo Zhixing. The company will usher in the big test of its "Double Hundred Plan": to implement urban NOH (navigation intelligent driving) in 100 cities and achieve 1 million installed units. But perhaps the more critical challenge is that Hao Mo Zhixing must first learn to be "independent."
Are the technology and business layout mature enough?
For autonomous driving suppliers, there are two variables that determine their industry status. One is the quality of autonomous driving technology, and the other is whether they have mass production capabilities.
Backed by Great Wall, a large car company, Haomo Zhixing, which was "born with a silver spoon in its mouth", naturally has a certain advantage in mass production capabilities. On its official website, it is "No. 1 in mass-produced autonomous driving in China". The advantages and disadvantages of
technology stem from the strategic vision of managers. In 2021, Gu Weihao, a 16-year veteran from Baidu’s intelligent driving business department, will serve as CEO. Great Wall veterans Zhang Kai and Zhen Longbao, as well as Hou Jun from Huawei, will serve as chairman, cio and coo respectively, forming a combination of the “four brothers”. , Haomo Zhixing officially entered the industry's attention.
Haimou Zhixing core management team, picture source Haimou Zhixing official website
Before 2021, the basis for realizing most autonomous driving functions was target recognition technology, front vision camera perception and HD high-precision maps. In terms of hardware, mobileye chips that integrate algorithms and cameras are the mainstream choice. Autonomous driving under this mode not only has complex modules and huge labeling requirements, but also relies heavily on high-precision maps, which ultimately makes it difficult to scale.
cannot be programmed into the self-developed algorithm, and the mobileye chip, which is like a "black box", is gradually being eliminated by the market. In 2021, NVIDIA's orin chip will become the new favorite of most OEMs. In August of that year, Tesla founder Musk brought the application experience of bev (bird's eye view) + Google transfromer model in autonomous driving at AI Day, starting the transformation of autonomous driving technology into a 3D spatial perception + deep learning architecture prediction model.
As a result, bev (bird's eye view) + Google transfromer have become the mainstream choice for smart driving suppliers to impact urban noa.
Haomo Zhixing, led by Gu Weihao, is also accelerating its embrace of AI. It has turned to transfromer since mid-2021 and is the first domestic company to "eat crabs". On the other hand, Wawanian Auto reports confirmed that because Haimou Zhixing chose Qualcomm Snapdragon Ride chips from the beginning, it took Haimou Zhixing longer to adapt to the transformer architecture. Technical route choices such as
have paved the way for Haomo Zhixing’s most important battle in 2024—the delivery of “thousand-yuan-level noh smart driving products without pictures.” In October last year, Haomo Zhixing launched its second-generation urban noh smart driving products, hp170, hp370 and hp570, priced at 3,000 yuan, 5,000 yuan and 7,000 yuan respectively.
Haimou Zhixing’s second-generation hpilot product, picture source Haimou Zhixing’s official WeChat public account
Haimou Zhixing’s city noh reflects the company’s clear choice in the technical route. Because it was the first to enter the transfromer model, Hao Mo Zhixing's urban noh was able to take the lead in proposing "emphasis on perception, light on maps" to solve the problems of slow approval, high cost, and long iteration cycle of high-precision maps, leading subsequent manufacturers to recognize the "mapless" solution.
The first-generation urban noh in 2022 is the first urban noa product in China that is "re-sensory" and can be mass-produced. The second-generation urban noh has completed a breakthrough in cost performance, abandoning the lidar required for high-precision maps, etc. After parts, it has changed people’s impression of smart driving systems at the level of 10,000 yuan, and launched the slogan of cost-effectiveness at the level of 1,000 yuan.
The three second-generation hpilot solutions do not need to use lidar. This may be seen as Haomo Zhixing’s cutting-edge exploration of another industry issue: “Does autonomous driving ultimately require lidar?” In the long run, smart driving solutions at the thousand-yuan level will help penetrate into models in lower price ranges, and may also change the cost calculation method of smart driving solutions.
From this point of view, from the perspective of mass production, technology, price and other dimensions, Hao Mo Zhixing is not weak, but the key question is, can the commercialization prospects based on these products be guaranteed?
In January this year, He Xiaopeng, the founder of Xpeng Motors, published an article summarizing the ten-year development history of Xpeng Motors and mentioned its experience in the field of autonomous driving: The most difficult part of intelligent driving is the selection of technical routes. No company dares to determine which one. The road can be successful; secondly, there are too many "uncertainties" in the mass production practice of smart driving technology.
Currently, there are two challenges facing Haomo Zhixing.
First of all, Haomo Zhixing needs to give results in the competition for the launch of urban NOA in 2024 and get rid of the single "Great Wall Series" label in order to truly establish its brand influence;
Secondly, and more related to the future, the industry has begun to The end-to-end model helps realize the discussion of urban map-free NOA. In the face of a new round of technological evolution, Haimo is also making technical judgments.
html In March 43, Gu Weihao publicly stated: "End-to-end must be a very important direction in the future, but it will not come so soon. I think it will still take several years of development." Recently, media reports stated that Haomo Zhixing is conducting end-to-end research and development.
At present, autonomous driving in the industry is still "divided". OEMs are exploring self-research, and suppliers are competing for the top spot. Wrong choice of technical route will not only waste a lot of money, but also lose the opportunity to sit at the poker table. Smart driving suppliers must choose carefully.
At present, Haomo Zhixing’s business layout is gradually maturing. As technology evolves and computing power and data become more and more important, Hao Mo Zhixing may continue to explore technology and commercialization. How long will
continue to rely on financing to advance?
The development of autonomous driving technology is inseparable from “burning money”. Since its establishment, Hao Mo Zhixing has completed 5 rounds of financing, with a public financing amount of about 1.3 billion yuan, and the remaining amount has not been disclosed.
’s most recent financing occurred in February this year, when Hao Mo Zhixing received over 100 million yuan in Series B financing from Chengdu Wufa Fund. Wuhou District State Investment Corporation stated that "investment to promote recruitment" is to empower new productive forces with capital, and Hao Mo said that it will build a robot demonstration zone in Wuhou District. Regarding the funds raised, Haimou said it will mainly be used for the research and development of AI autonomous driving technology such as Haimou large models.
Previously, in addition to the angel round investment from Great Wall Motors, Haomo received an investment of 300 million yuan from Meituan Hillhouse Beijing West and others in the pre-A round in February 2021, and received another nearly 1 billion yuan in the A round in December 2021. In 2022 Hundreds of millions of dollars were obtained in April.
It is not difficult to find that the last two rounds of financing of Hao Mo Zhixing have been two years apart. After the popularity of in 2021, the industry has gradually formed a consensus that in the face of technical difficulties and mass production difficulties, the possibility of profitability for autonomous driving suppliers in a few years is slim, and investment enthusiasm in the autonomous driving industry has plummeted.
it Orange data shows that the number of investments in autonomous driving tracks in 2022 will fall by nearly half compared with 2021, and will continue to decrease in 2023.
The number of autonomous driving investments in recent years, source it orange
In 2021, Haomo Zhixing once revealed that it achieved revenue of 100 million yuan that year, but there was no further disclosure of profits and capital investment.
In response to the issue of profit and loss, BiMo Travel responded to Wired Travel: "BiMo's revenue is not convenient to disclose to the outside world. BiMo is a company that has mass production and implementation very quickly on the smart driving track and is leading in commercial implementation. BiMo will stick to technology. long-termism”. When
was founded, Hao Mo Zhixing proposed the "windmill strategy", using the three business lines of passenger cars, low-speed unmanned vehicles and intelligent hardware as the "three blades", thereby collecting data to drive the evolution of the company's autonomous driving capabilities, and ultimately forming a feedback system. cycle.
At present, in addition to passenger car smart driving products, the business that is most promising to take on some of the profit pressure is the terminal logistics unmanned vehicle distribution business that Hao Mo Zhixing will achieve breakeven from the fourth quarter of 2023. Meituan, as an investor in Bimo, has previously made Bimo the sole production partner of Meituan's new generation of unmanned delivery vehicles. Bimo has previously become the sole production partner of Meituan's new generation of unmanned delivery vehicles. In addition, Feimo has also reached cooperation with Wumart Duodian, Dada Express, etc.
Regarding the future commercialization prospects of the low-speed unmanned vehicle business, Feimo Zhixing told Wired Travel that in the field of terminal logistics and distribution, Feimo first focuses on the fulfillment scenario of supermarkets. First, the scale of supermarket fulfillment continues to expand; second, distribution The unit price is high enough. "After the verification of this model is completed, we will work with customers to copy this model to mature areas in major domestic cities to expand the scale and increase revenue."
However, in the long run, whether Hao Mo Zhixing can achieve self-generation as soon as possible depends on whether it can seize opportunities in the field of autonomous driving.
In the first half of 2022, the penetration rate of domestic L2 functional passenger cars in my country reached 32.4%, a year-on-year increase of 46.2%. It is expected that the penetration rate of all levels of autonomous driving will reach 80% in 2025 - in other words, this will still be a blue ocean in the next few years. market.
For Haomo Zhixing, which must move towards independence, this is its best opportunity now, but whether it can seize this critical window period and realize self-development as soon as possible depends on Haomo's own exploration and operation.