Autonomous driving has gone through a painful period, why did Zhixing catch up later?

In 2022, the technology industry entered a period of contraction, but the automatic driving industry shined alone, and many autonomous driving start-ups such as Wenyuan Zhixing and Xiaoma Zhixing obtained a new round of financing, and some media interpreted it accordingly that the industry's "pain period" has passed. In fact, the current round of financing in the autonomous driving industry has already shown signs at the end of 2021. At the end of December 2021, Zhixing announced that it had obtained nearly 1 billion yuan of Series A financing, including Meituan, Hillhouse Venture Capital, and Great Wall Motors.

In the view of Luo Chao Channel, the fundamental reason for the heating up of the automatic driving industry is that more and more commercial models are loaded with more and more intelligent auxiliary driving technology, and the once seemingly distant automatic driving is getting closer and closer to people.

Step by step, the automatic driving steadily landed on the ground

A few days ago, Elon Musk said that Tesla is committed to providing a fully autonomous driving beta version (FSD Beta) to customers in the European left-hand drive (LHD) market in the summer of 2022, and the commercial landing of domestic autonomous driving is more frequent news, He Xiaopeng just said at a public event that Xiaopeng will launch an advanced version of the intelligent auxiliary driving system XPILOT3.5 in 2022, and Xiaopeng's goal is to "transition to unmanned driving in 2026." The ideal L9, which was just exposed some time ago, is equipped with a fully self-developed central domain controller and a full set of intelligent driving perception systems, and in the just released earnings report, Weilai revealed that its newly delivered ET7 will be equipped with the latest autopilot technology NAD.

From the perspective of technological maturity, it will take time for "fully autonomous driving" to fully land, and cars that do not require steering wheels and brakes and no longer have drivers on the road on a large scale will not become a reality in the foreseeable future. The new forces rely on the differentiated ability of intelligent driving to be fierce, and now the promotion of "fully autonomous driving" is inevitably suspected of "drawing cakes".

In the view of Luo Chao Channel, the pragmatic approach of the automatic driving industry should be step by step, first let the level of assisted driving automation be raised little by little, so that consumers can experience a safe, convenient and worry-free driving experience, and this is what many traditional car companies are doing, such as Great Wall Motors, which has a development history of nearly 40 years.

On March 18, the Great Wall's new off-road vehicle tank 500 was listed, the official guidance price of 335,000-395,000 yuan, with four major selling points of powerful platform technology, strong power performance, strong intelligent technology and powerful luxury experience, not only equipped with a 3.0L V6 twin-turbo engine and a hard-core powertrain of the longitudinal 9AT transmission, but also matched the HPilot 2.0 auxiliary driving system independently developed by Millizo Zhixing.

After opening the HPilot 2.0 auxiliary driving system, tank 500 can complete the adaptive follow-up on the highway, realize automatic control of the speed and the distance between the car in front, greatly reduce the fatigue of the driver, and improve the driving happiness; in the case of a clear high-speed lane line and the vehicle driving in the middle of the lane, the tank 500 can achieve automatic lane change function in the high-speed scene, the driver only needs to dial the turn signal lever, and the vehicle will smoothly and smoothly complete the lane change action when confirming safety.

The Tank 500 equipped with the HPilot 2.0 assisted driving system completely breaks the stereotype that "off-road vehicles are not smart". Before the Tank 500, many models of Great Wall Motors' brands have already had auxiliary driving capabilities, such as Weipai Mocha, Tank 300 City Edition, Weipai Machido DHT, Weipai Latte DHT, and Haval Divine Beast have all been equipped with the Mil.D. HPilot 1.0.

Different from the third-party technical service model of some traditional car companies that "the body is responsible for itself and the brain is handed over to others", Great Wall has incubated the head autonomous driving enterprise of The Great Wall with the intelligent driving forward-looking department as the starting point.

As an autonomous driving company relying on the main engine factory, Zhixing relies on the small magic box product line to continuously iterate HPilot, and lands commercially available and mass-produced intelligent driver assistance solutions on passenger cars.

HPilot 1.0 has advanced driving assistance capabilities in high-speed scenarios, which is based on subsystems such as HWA (Highway Assist) Highway Assist System and NOH (Navigation on HPilot) High-speed Automatic Pilot Assist Driving System, which can realize auxiliary driving functions such as ACC + automatic lane changing with lights in high-speed scenarios. The HPilot 2.0, which debuted in the Tank 500, introduced new capabilities such as the integrated approach to the berth, enhanced side perception, recognition of traffic lights and small obstacles, and enhanced lane line recognition.

HPilot3.0, which is under development, will expand advanced assisted driving from high-speed scenarios to urban scenes, after Gu Weihao, CEO of Miller Zhixing, revealed that Miller Zhixing is expected to launch L2 assisted driving services in urban scenarios in the middle of the year, which should be carried by HPilot3.0. Previously, according to media reports, the NOH (Navigation on HPilot) automatic pilot assistance system for urban scenarios has invited core guests to take a test ride in Shunyi and is expected to be listed this year. At the AI DAY held on April 19, there may be more big moves about HPilot.

In the passenger car scenario, the idea of Zhixing is very clear: on the one hand, it constantly polishes the software and hardware technology of the advanced assisted driving system, upgrades the radar configuration of the small magic box, the basic configuration of the computing power module, etc., improves the intelligence of assisted driving, and introduces more functions such as the integration of travel and berthing; on the other hand, it vigorously promotes the application of mature advanced assisted driving solutions to more commercial products, which can not only obtain income, but also accumulate data and feed back technology.

As of March 2022, the driving range of the user assisted driving of TheLmot has exceeded 6 million kilometers, and it is expected that in the next two years, the total number of passenger cars equipped with the HPilot Assisted Driving System will exceed 1 million units, and if this goal is achieved, it means that the company will become the AI autonomous driving company that will enter the "Million Mass Production Club" in the industry in the shortest time.

Instead of drawing a pie like "fully autonomous driving", Zhixing let the advanced driving assistance system land on mass production vehicles, including The Wei brand Mocha DHT-PHEV and tank 500 and other Internet celebrity explosive vehicles, which really improved the user driving experience, continuously upgraded technology, accumulated data, iterated on products, and improved functions. It can be said that Mi Mo Zhi has walked a very unique road to automatic driving landing.

Less than three years old, Zhixing has already joined the head camp

Since around 2015, the self-driving entrepreneurial boom has been in China for at least 7 years. Wenyuan Zhixing and other entrepreneurs have already got C/D round financing, and zhixing has only completed the A round of financing at the end of last year, Pre-A was also completed in this year, led by Shougang Fund, meituan, Hillhouse Venture Capital and other capital investment of hundreds of millions of yuan. The reason why the financing round is so early is that Zhixing is a "latecomer".

Founded in 2019, Zhixing has not yet reached the age of three, although it is a new recruit in the industry, it has successfully entered the center of the autonomous driving stage and become a relatively mature autonomous driving company.

First of all, the commercialization process is fast, and the fastest realization of million-level mass production of passenger cars and earn real money and silver.

In the autonomous driving industry, commercialization is more difficult than the technology itself, because commercialization is not only a technical issue, but also involves market cultivation, business models, policies and regulations. Because of this, it is a "legend" that Zhixing was able to commercialize technology in such a short period of time.

Up to now, HPilot has been successively applied to Weipai Mocha, Tank 300 City Edition, Weipai Machido DHT, Weipai Latte DHT, Haval Divine Beast and Tank 500 and other mass production models, after which Zhang Kai, chairman of Mo mo zhi xing, once revealed that "mo mo zhi xing undertakes the development task of high-level auxiliary driving of 34 models to be listed by great wall motors in 2022, accounting for nearly 80% of the models to be listed by Great Wall Motors in the whole year, and it is expected that the penetration rate of high-level auxiliary autonomous driving of Great Wall Motors will exceed 40% in 2022." ”

With a turnover of hundreds of millions of yuan in 2021, Zhixing is one of the few autonomous driving start-ups that have achieved large-scale revenue. At the 2021 Yunqi Conference, Zhixing joined the Ali Little Wild Donkey Autonomous Driving Ecology, and Zhang Kai spoke about a judgment when he gave a speech: "2022 will be a watershed year for the commercialization of AI autonomous driving, and for startups in the track of automatic driving, it will be a very fatal thing if it has not been able to solve the problem of large-scale deployment of autonomous driving in 2022." He believes that the process and rhythm of large-scale commercial landing of autonomous driving will determine how fast an autonomous driving technology company can go; and the continuous and stable acquisition of high-quality scene data will determine how far an autonomous driving technology company can go.

Zhang Kai's confidence in saying this is because Zhixing has gone far in mass production and commercialization. In stark contrast, many self-driving startups, including unicorns that have been raised into the D round, have not yet crossed the mass production difficulty.

Secondly, the technical hard power is strong, not hardware stacking or algorithm iteration, but to create a data intelligence system.

Automatic driving is a technology-driven industry, the bottom line of Zhixing in commercialization is the technical hard power, which has formed three major product lines of small magic box, small magic disk and small magic camel, which belong to the three major areas of passenger car automatic driving, unmanned vehicle general line control chassis, and low-speed unmanned logistics vehicle, and all have mass production capabilities and have landed commercial scenarios. Core systems, including HPilot, have undergone several major iterations of iterative upgrades, and assisted driving functions are becoming more and more perfect.

In March 2022, with its outstanding performance in independent innovation, patented technology research and development, and commercial landing achievements, Zhixing was recognized as a "National High-tech Enterprise", which further confirmed its technical strength.

For Zhixing, commercialization and technical hard power are a game of chess: the former can accumulate massive data in different scenarios, and then feed the latter to form a stronger technical strength, including the degree of algorithm intelligence, scene adaptability and hardware mass production costs. Based on the research and development and landing practice of AI autonomous driving technology, Zhixing summed up a curve of automatic driving capability development: F=Z+M(X). Where F stands for product force, Z stands for millimae first-generation product, and M is a function that transforms data into knowledge, including data acquisition, expression, storage, transmission, computation, verification, and impact on cost and speed.

Based on the development curve of F=Z+M(X), which is an autonomous driving capability, Zhixing created the "Data Intelligence System MANA", which is the core driving force of its riding dust. In the impression of many people, Zhixing is an autonomous driving company, but its official positioning is "an artificial intelligence technology company committed to automatic driving", that is, AI is its core label, and automatic driving is only a focused scene, just like Tencent is a social entertainment group but it is also an Internet company.

Data is the cornerstone of AI technology, and the essence of the data intelligence system MANA built by Millima Zhixing is based on the company's massive data, forming data processing tools, algorithm models, test verification systems, simulation tools and computing hardware, including TARS (data prototype system), LUCAS (data generalization system), VENUS (data visualization platform), BASE (underlying system) four subsystems, covering perception, cognition, simulation, A technical system that includes dimensions such as computing.

Based on the MANA data intelligence system, Zhixing gives full play to the advantages of "fast, wide and deep commercialization", mines value from the massive high-quality data accumulated continuously, and promotes the maturity of AI technology and automatic driving technology.

Finally, the ecosystem is becoming more and more mature, and more and more "head and face" players are playing with Mo Mo.

Zhixing itself does not do C-end business, but does industry "helpers", focusing on providing automatic driving solutions, and the downstream is its service object, including terminal logistics needs customers such as Meituan and Ali, as well as OEMs such as the Great Wall, as well as online ride-hailing platforms that carry out Robotaxi's autonomous driving passenger business. Upstream is a provider of AI infrastructure such as chips and radar. Because of this, the business model of Zhixing is inseparable from upstream and downstream partners, which is the so-called "ecosystem".

Although it is a company that has been established for less than three years, it has now attracted the olive branch of leading players in various industries, and a luxurious autonomous driving ecosystem is emerging.

Zhixing has formed a deeper cooperative relationship with some enterprises through capital cooperation. Qualcomm Ventures is one of its investors, semiconductor giants are betting on AI chips, and Millima Zhixing can provide Qualcomm with competitive autonomous driving scenarios, which in turn can also make Qualcomm have a long-board advantage in the computing power link. The NOH one-shot video in the city scene shows that it will join hands with Qualcomm's world premiere 360Tops large computing power platform. In addition to Qualcomm, Zhixing and Meituan and Great Wall are similar "capital + business" cooperative relations.

The in-depth cooperation with strategic investors does not affect the ecological openness of Zhixing. Across the upstream and midstream of the autonomous driving industry, Zhixing insists on not doing downstream business, not doing consumer business, and not competing with customers, which allows it to be a "helper" and attract downstream customers. Therefore, in addition to investor partners, Zhixing has attracted hundreds of partners such as Ali, Wumart Duodian, Baidu, and AutoNavi, and they are not "platform-style" formal cooperation, but have a solid business landing.

As a cutting-edge company established less than three years ago, Zhixing attracts leading players in various industries to cooperate with it in depth, which is not only a proof of technical and service strength, but also shows the confidence of all walks of life in its future development.

It's too early to say that Zhixing stands out, but it is also an indisputable fact that it has occupied a very high position on the crowded self-driving track.

Riding on the dust, what did zhixing do right?

Luo Chao channel believes that the most distinctive feature of Zhixing in a number of self-driving startups is that it is born different. Its predecessor is the Intelligent Driving Prospect Department of Great Wall Motors, and after the integration of autonomous driving talents from technology companies in 2019, Zhixing officially "flew alone". Now Great Wall Motors is not only the major shareholder of Miller Wisdom Bank, but also a large customer of Miller Wisdom Bank, and this unique "origin" determines that it has multiple "dominant genes":

1. It seems to have been established for less than three years, but in fact it has accumulated for many years.

When did Great Wall Motors lay out autonomous driving? In 2017, Zhang Kai, then deputy chief engineer of the Great Wall Motor Research Institute and director of the Intelligent Networking Department, revealed at an event, "Great Wall Motor began to conduct corresponding research on the ADAS system as early as 2010, and cooperated with some universities in 2013 to participate in the China Intelligent Car Future Challenge, and in 2015, the self-developed intelligent driving system was dynamically demonstrated at the Great Wall Internal Science and Technology Festival, and in September of the same year, the first product equipped with ADAS system including ACC and AEB was listed on the market." At the Harvard Million Celebration in February 2017, the i-pilot system (Smart Pilot System) was launched. ”

Maybe i-pilot is the predecessor of HPilot? From this introduction, it can be seen that Great Wall Motors' exploration of autonomous driving began in 2010 - earlier than baidu and other Internet platforms to lay out automatic driving, and the new domestic forces were not yet born at that time. More than two years of "fast" before, there are N years of "slow".

2, born with a scene, so it can be commercially driven.

Many companies with high-tech companies face a common problem: holding a hammer to find a nail, having powerful technology but suffering from a scene that is not adapted. Some finally found that many "nails" were pseudo-demand collectives, such as blockchain bubbles; some had not yet waited for market demand to erupt and regrettably died before dawn. The dilemma faced by autonomous driving startups is the latter, many teams do not lack technology, what they lack is the conditions for large-scale mass production. The great wall of large customers has always been born with a big customer, so there has always been a "first have a nail and then find a hammer" thinking, not lacking and attaching importance to commercialization scenarios, insisting on not doing useless technology that only lies in the laboratory, but developing technologies that can solve real problems, attaching importance to scene landing, and then obtaining data based on data intelligence system to feed back technology, forming a positive cycle.

3. Choose the right route and do not make the mistake of "strategic laziness".

Many startups make mistakes that are "tactically diligent and strategically lazy", and the more you try to get farther away from the goal, the faster you take a detour, it doesn't make sense. Since its inception, Zhixing has attached great importance to the choice of strategic routes, such as its continuous collection of scene data through the "three leaves" of the windmill strategy, that is, passenger cars, low-speed unmanned vehicle ecological platforms and intelligent hardware, and then use data to feed back product research and development and manufacturing, forming a virtuous circle.

Judging from the results, every step of the way is correct, and this is determined by the team's cognition. When some industry players aggressively said that they want to achieve fully autonomous driving in three or five years, Zhixing chose a more realistic route, firmly took the "feasible, reliable, commercial" progressive landing route, under the premise of ensuring technical safety, with commercially available as the premise, step by step, sustainable development. In the eyes of some people, the wisdom of the end is very conservative "slow", from the results, the end of the wisdom of the line but go more "fast", the key point is that it does not take the wrong road, does not take a detour, does not take a turning back.

Li Ka-shing once summed up his philosophy of doing things in 16 words: good planning, segmented governance, not hurrying and speeding, and doing nothing to cure, of which "not hurrying and speeding up" comes from "Zhou Yi", the original text is: "Only God, not fast and fast, not to do it", which means that doing things cannot be blindly fast, but also can not be indecisive, but to plan well, to do it vigorously, not to drag mud and water, in the words of Li Ka-shing: "For me, the most beautiful battle is actually a battle that clearly calculates gains and losses in advance." Su Hua said that Kuaishou "moves quickly and thinks very slowly" expresses the same meaning. The wisdom of the end of the day is a vivid portrayal of "not hurrying and speeding up", and the speed of more than two years is precisely because of "slow thinking" again and again, and the "fast action" after seeing accurately.

4. "Automotive + AI Technology" team to ensure rapid action.

The direction is right and the execution is fast. Companies incubated within traditional giants face a common problem: slowness. It is difficult to change the "slow" style of traditional giant enterprises, and it is impossible to really have the ability to be "fast", which is the competitiveness of latecomers such as startups, Internet companies, and new forces.

In 2019, Zhixing independently developed from the Great Wall, which laid the foundation for its ability to have "fast", and it invested in technology giants such as high-end talents and Meituan, and then had ai technology genes, and further had the "fast" ability of Internet technology companies, which is reflected in its rapid growth, such fast is reflected in its rapid growth, such as in technology research and development, its products have achieved the same rapid iteration as Internet products, and the intelligent pilot auxiliary driving system NOH was established in November 2020. Four months to complete the system design matching debugging, four months to complete the system engineering development, parallel eight months for road test verification and system polishing, in November 2021 direct mass production and listing, is the fastest research and development of noH intelligent pilot auxiliary driving system in China.

5, willing to work hard, do difficult and correct things.

Many autonomous driving companies do relatively "lightweight" "algorithm" layers, and do not touch "heavy" things such as manufacturing. However, Zhixing has taken a reverse path: as long as it is in line with the company's strategic positioning, the things that can better satisfy customers must be done, including the most difficult manufacturing.

The premise of the commercial use of autonomous vehicles is mass production, and mass production needs to solve many problems: manufacturing, supply chain, cost, capital, etc., many car players are stuck in the mass production link, autonomous vehicle mass production is more difficult, just because of this Apollo mass production also have to turn to car companies such as Geely, and the choice of Zhixing is to do their own manufacturing.

Zhixing has the foundation for good manufacturing. On the one hand, it has the advantages of the main engine factory in manufacturing, supply chain and industrial chain, has a stable supply of core and vehicle parts, and has a stable mass production capacity. On the other hand, serving the Great Wall and many customers has formed a strong large-scale effect and further reduced costs. It is based on the manufacturing layout that it can frequently break records in the delivery and loading speed of autonomous driving mass production, forming a barrier in the mass production capacity of autonomous driving, becoming "China's autonomous driving mass production NO.1".

The ability to do some "slow" work is also genetically determined. The safety-first automatic driving industry can not blindly seek speed, and the dual gene of "automobile + technology" makes the "light and fast" of doing intelligent technology, and also maintains the "stability" of doing the automobile industry, to B service and automatic driving, so it chooses to do some heavy work, slow work and hard work, and it is not surprising to insist on doing difficult and correct things.

Write at the end:

Less than three years after its establishment, Zhixing has rapidly grown into a mature autonomous driving company, and the smooth progress along the way is somewhat incredible, but also a little out of place.

At the beginning of 2022, Musk reflected on Lex Fridman's podcast: "I knew from the beginning that Autopilot was not easy to solve, but I didn't expect it to be so difficult. Tesla has always been known for its self-driving radicals, and former self-driving partner Mobileye has parted ways with it, but Musk himself has always been highly optimistic about the development of autonomous driving, according to Tesla's official statement, its FSD system can achieve L4 level automatic driving in 2022, and the accident rate is 2-3 times lower than that of manual driving, but Musk also admitted that the difficulty of technology development has exceeded his expectations.

Everyone realized that self-driving was a "slow" industry, and those self-driving players who had proposed mass production plans later found themselves too aggressive. However, Zhixing came quickly and ran through the road that many friends and businessmen could only walk in five or even ten years.

In many industries, we can see players like Zhixing: not the first to enter the game, or even belated when the market is about to explode, but they can be a blockbuster and invincible. In this type of player, we can see a similar way of success: in a short period of time, the rapid development of often have a longer accumulation, but only at a specific time to break the ground, combined with the direction that has long been aimed, with a strong landing ability, "plan and then move", quickly harvest the market.