Ordinary engineers who have worked for 3 years "dare to" have an annual salary of 500,000 yuan, who is the driving force behind the automatic driving robbery war?

Although it is a home office, Li Ming, the project director of an autonomous driving company in Shanghai, is still busy interviewing various types of autonomous driving talents, but unfortunately not many people have joined.

"Candidates generally feel very good about themselves, ordinary engineers who have worked for two or three years dare to ask for an annual salary of 500,000, and the algorithm is even more exaggerated, and it is 800,000 to 1 million." And they keep jumping jobs, adding a lot of salary every time, and the market is very impetuous. Li Ming said.

It's not just self-driving companies that have a recruitment need. One change in the self-driving track last year was the end of China's traditional car manufacturers. Great Wall, Geely and BYD have changed their strategy of relying on suppliers and increased the capacity building of autonomous driving self-research.

By the beginning of this year, it was not only the automotive industry that was grabbing the most talent, but also the companies of meta-universe concepts, and they were recruiting algorithm engineers in large quantities. There are also some chip factories that are recruiting software talent, and they want to enter the application layer, develop reference solutions and toolchains, and package them into higher value-added SDK components.

Behind the robbery war, a large amount of capital has poured back into the automatic driving track, driving the birth of a large number of newly created scene car manufacturing and autonomous driving companies. According to the statistics of companies such as Qichacha, the total amount of investment and financing for autonomous driving will reach 57.5 billion yuan in 2021, and the number of financings will be 107 times, and the total amount of financing will be almost 6 times that of 2019.

From the cold winter to the overheating, the change at the market level is that some functions of the autonomous driving system have begun mass production on a large scale. At the end of last year, with Beijing opening up the Robotaxi commercialization pilot to Baidu and Xiaoma Zhixing as a milestone event, the trillion-level autonomous driving track was considered to have entered the first year of commercialization.

"Autonomous driving is a disruptive technological innovation with huge business benefits." Liu Tao, a venture capital partner in Beijing, told reporters.

But behind the sexy imagination, the reality is very bone-chilling. All L4/L5 autonomous driving companies are looking for high-frequency, just-needed, and mass-produced scenarios, but none of them have run through, and they have not yet achieved financing and hematopoietic "two-legged" walking. Since last year, the information of layoffs in some head autonomous driving companies has also been circulating on the workplace platform.

"The technical route of autonomous driving is not clear, the charging model is not clear, and if these two problems are not solved, it can only be said that it is still in a very preliminary stage, and these companies will rarely survive, and this road is doomed to be full of bones." Liu Tao said.

Rob the man

The automotive industry has fallen into a strange circle of "lack".

Since the end of 2020, the domestic automotive industry is experiencing a lack of cores and power, and has recently experienced a lack of various parts, which has led to the suspension of production by some car companies. In addition to the lack of production-related materials, all kinds of talents have become the hardest hit areas for the "lack" of the automobile industry.

Among them, in the past two years, the field of autonomous driving, which has been regarded as an outlet by the capital market, has staged an extremely fierce "talent war".

"The changes in these two years have been too fast and too big. Two years ago, we pushed 200 candidates to Party A in the autonomous driving industry, and probably few people could finally get an offer, and now anyone in this industry can get a lot of offers. Zhang Chao, a headhunter focusing on the field of automatic driving, told reporters that in the past 1 to 2 years, the demand for manpower in the field of automatic driving has soared, including automatic driving companies, new car-making forces and even traditional car companies that have relied on suppliers in the field in the past to recruit talents in the field of automatic driving on a large scale.

Zhang Chao said that in 2018, he began to focus on headhunting in the field of automatic driving, but it was in the "AI winter", after doing 3 months, the company even had a team of 9 people to maintain, some team members were split into other departments, and the headhunting group said that this industry (automatic driving) was not good. In the autonomous driving related enterprises, the demand industry is not strong, and there are even layoffs, and some talents in related fields have transferred to Internet companies such as ByteDance and Kuaishou.

"In essence, the core of doing autonomous driving is algorithms, such as visual algorithms, talents in related fields can do special effects and visual image-related projects, and in Kuaishou can do recommended searches, which are universal." Zhang Chao said.

In 2020, Zhang Chao's team recommended a total of 200 candidates to companies in the autonomous driving industry, of which only 10 to 15 people received offers, and even fewer officially joined the job after getting an OFFER. The change occurred in the second half of 2020, "before the middle of 2020, the relevant companies not only recruited fewer people, but also 'held', and the requirements for talents were relatively high." After this, everyone suddenly seemed to let go, the talent grab was too strong, and even Xiaoma Zhixing had to give up the high standard of recruitment. ”

Zhu Yi, the technical leader of an engineer at an autonomous driving company, told reporters that the demand for automatic driving algorithms for manpower is quite large, taking SenseTime, a leading domestic artificial intelligence algorithm company, as an example, the company's artificial intelligence algorithm team has 7,000 to 8,000 people.

According to the "2021-2022 China Autonomous Driving Industry In-depth Analysis and Outlook Report" released by Yiou Think Tank, since 2021, the autonomous driving industry has once again ushered in a boom in investment and financing, with a total investment and financing amount of 57.5 billion yuan and 107 financing times, an increase of 14.9 billion yuan compared with 2020. Zhu Yi said that during the period from 2020 to 2021, automatic driving has become an outlet, and the technical leaders of some companies have come out to start a venture capital by themselves, and newly established companies need new teams.

According to the "2022Q1 High-end Talent Employment Trend Big Data Report" (hereinafter referred to as the "Report") released by Liepin, the demand for new positions in autonomous driving in the first quarter of this year increased by 41.92% year-on-year, ranking sixth in all popular segments.

Under the demand for large manpower, the recruitment of related enterprises has also begun to "eight immortals across the sea, each showing its magic", and enterprises are staging this offensive and defensive war of manpower competition. Liu Zhi, HRBP of an auto parts company, told reporters: "Before some peers came to me, wanted me to help provide the contact information of the engineers in our company's ADAS department, and those who left before also promised to give me a benefit fee." ”

Liu Zhi said that in the past two years, the company's ADAS department has been the hardest hit area of personnel loss, and the company's non-compete agreement at the end of 2021 has also been updated, including a large number of start-up autonomous driving companies, vehicle-grade chip companies and automakers.

"Before 2020, the non-compete agreement was more like a nuclear weapon, most of the time it was a 'deterrent', even if there are employees who violate the non-compete agreement, we rarely pursue it, but in the past two years, we have sued former employees with non-compete agreements." Liu Zhi said.

However, according to the feedback of the reporter's comprehensive feedback from relevant people in many enterprises, a more stringent non-compete agreement cannot block the loss of talent. Liu Zhi said: "As long as the problem can be solved, the company will throw money. Even if there is a non-compete agreement, there is no shortage of cases where the next family pays for 'redemption'. ”

Demand and salary are rising, who is behind it?

"This year, even me (manufacturing process engineer) management trainees can't be recruited, and there has been a gap." Liu Zhi told reporters that the company's package price for pipe trainees in January this year was 120,000 yuan, and by February it rose to 140,000 yuan, and the candidates who had previously been interested in it also rose together.

Zhang Chao said that in July 2020, Ideal Auto suddenly began to expand enrollment, after March, the Great Wall's Mo Mo Zhixing began to expand enrollment, and then Weilai, Geely, BYD and other car companies, Didi and other technology companies began to enter the game in a big way, the industry's liquidity increased, and this plate came alive.

With the substantial increase in manpower demand, in the "offensive and defensive war" of talents in related enterprises, the salaries of related positions have also risen.

The reporter learned that Weilai Automobile's current monthly salary range for recruiting algorithm engineers is 50,000 to 80,000 yuan, plus stock options. But more than a year ago, NIO offered a higher salary than the industry average for self-driving engineers.

"Usually, we used to give candidates a 30 percent salary increase, which is an industry standard. However, in terms of autonomous driving talents recently, it has not come according to this standard. Earlier, a car company that made new forces in the car dug up people, and it was not calculated according to the increase at all. In 2020, the annual salary package price of P5-level autonomous driving engineers is between 400,000 and 550,000 yuan, usually digging people, giving up to 700,000 yuan, but this new force car company directly gave 800,000 yuan, beginning to exceed the industry average. Zhang Chao told reporters.

Zhang Chao said that by the end of 2020 and the beginning of 2021, grabbing people and raising salaries has entered a climax.

It is not only the new power car companies that have raised salaries sharply, but also the self-driving car companies that have significantly increased their income levels to cope with the robbery war.

Zhu Yi told reporters that around 2021, a domestic head autonomous driving company began to enter new businesses in a big way, and the business of building cars and trucks was crazy to recruit people. In the case of the robbery war, the head company also began to reduce the requirements and increase the salary level. During the period of 2018-2019, their P5-level senior engineer intelligence gave 500,000 yuan, and now can give 600,000 yuan to 700,000 yuan; P6 senior engineers can give 600,000 to 800,000 yuan in previous years is already particularly high, and now they can give 1 million yuan, and some can even give 100,000 yuan in cash signing fees; P7 senior experts can give 1.5 million yuan, and the director will be more than 2.5 million yuan.

Zhang Chao also told reporters that some companies' autonomous driving related positions are pure cash to 1.1 million yuan, which is not counting options, other benefits, etc., and some candidates have come to him to feedback, "Am I so valuable?" "In this kind of market, headhunting can also earn a full pot, Zhang Chao's headhunting company, in 2020 is basically not profitable, by 2021 profit reached the level of tens of millions, this year is expected to be more."

Some executive search company practitioners told reporters that in general, the cooperation income of headhunting companies is usually multiplied by a certain proportion of the salary after the candidate's regularization, and the profit can reach tens of millions of yuan, which can be seen in the high income and manpower demand of the industry of automatic driving.

According to the report released by Liepin, in the first quarter of 2022, the average annual salary of recruitment corresponding to the popular areas was the highest, at 390,500 yuan; followed by the Internet of Vehicles, which was 377,000 yuan; and the average annual salary of new energy/electric vehicles was 278,900 yuan, ranking fourth.

Zhu Yi believes that the pool of talents in the fields of artificial intelligence and algorithms is so large, the total number is relatively limited; in the early years, the AI winter lost some of them to Internet companies; the recent rise of the meta-universe concept, and some talents went to related companies, and the robbery war is not only between companies on the track of automatic driving, but also needs to face competition between meta-universes, Internet and other enterprises.

Huge imagination space boosts the robbery war

In 2021, there were 42 corporate financing incidents in the autonomous driving industry in China, involving 37 companies. Among them, Hongjing Intelligent Driving, Tage Zhixing, Xidi Intelligent Driving, Yishi Technology and other companies have completed two financings. In the release of financing scale data, the financing scale of Yinche Technology, Gaoxian Robot, Yishi Technology and the three enterprises is higher than 1 billion yuan, which is the highest financing scale. According to tianyancha APP data, by June 2021, there will be nearly 5,200 autonomous driving-related enterprises in the mainland that are in business, existing, moving in and out. Among them, more than 30% of the registered capital is more than 10 million, and nearly 9 have become limited liability companies.

"The imagination of the huge commercial benefits of autonomous driving is not comparable to electrification." Liu Tao said that under the huge imagination space, there are more and more players, and more and more capital investment, and it is an inevitable thing to burn money and rob people.

Research institutes have repeatedly made predictions about the huge business prospects of autonomous driving. McKinsey & Company said in a report that China will be the world's largest market for autonomous driving, with total sales of autonomous vehicles reaching $230 billion by 2030 and orders for autonomous driving-based mobility services reaching $260 billion.

Credit Suisse China believes that about 153,000 taxis in the four cities of North, Shanghai, Guangzhou and Shenzhen can be replaced by Robotaxi, and it is expected that about 200,000 traditional taxis will be scrapped and replaced by Robotaxi by 2025, and this scale will be further expanded to 1.4 million by 2030. Market research firm IHS Markit said in the "China Autonomous Driving Market and Future Mobility Market Outlook" that under the current trend, the robotaxi market size is expected to exceed 1.3 trillion yuan.

The first reason behind the optimistic estimate is the replacement of people by machines, and in the commercial vehicle space, Robotruck is thought to be more commercially available. An investment institution analyzed that each truck is equipped with an average of 2 drivers, and the annual cost of each driver is about 120,000 yuan, and only considers the part of automatic driving to replace the driver, and the market space is nearly 1.5 trillion yuan. From the perspective of operation, the market space of trunk line heavy trucks is about 5 trillion yuan.

The number and working status of truck drivers in China add credibility to such assertions.

According to the "2021 Truck Driver Employment Survey Report" released by the China Federation of Logistics and Purchasing, there are currently about 17.8 million truck drivers in China, who carry more than 70% of the logistics work in China. The number of truck drivers in China plummeted from more than 30 million in 2018 to about 17.8 million in 2021, a decrease of 13 million in three years. In addition, the overall age of truck drivers is older, and the number of post-70s drivers in the field of heavy trucks accounts for 80%. In addition, truck drivers work longer average daily hours, with 37.3% of truck drivers working more than 12 hours a day.

Based on the aging, loss, work intensity and risk of truck drivers, the research institute believes that whether it is transportation or public services, labor costs are still the main proportion of enterprise operating costs, once the automatic driving technology matures, it can greatly reduce the operating costs of enterprises, while relying on digital management systems to improve management efficiency, from the root to achieve cost reduction and efficiency.

For relatively dangerous operation scenarios such as mines and ports, the automation and intelligence of operating vehicles can also reduce the incidence of accidents and improve operational efficiency; at the same time, automatic driving technology can improve automobile fuel efficiency, intelligently share travel to optimize routes, improve resource allocation, and promote energy conservation and emission reduction.

CICC believes that the economy of self-driving trucks is significantly better than that of traditional fuel heavy trucks, which is mainly reflected in three aspects: one is to increase the running time of trucks, the other is to reduce driver costs, and the third is to improve fuel economy. According to its estimates, self-driving trucks can eventually reduce costs by 45%.

On the self-driving commercial vehicle track, there are many star companies such as Tucson Future and Wincher Technology, and new companies are constantly being born. In the second half of last year, there were suddenly 3 self-driving truck start-up companies in the industry, namely Qianhang Technology, Xingjiao Technology and Kinte Smart Card, of which SF, Baidu, and Xiaopeng Automobile are all investors in Qianhang Technology.

Under the good prospects and expectations of the industry, the entry of a number of new companies has exacerbated the shortage of talents in the industry. However, so far, neither Robotaxi nor Robotruck has achieved commercial landing.

In April last year, Tucson Future landed on the U.S. stock market, becoming the first stock of autonomous commercial vehicles in China. According to Tucson's future prospectus and 2021 financial report, Tucson's future revenue for 2018, 2019 and 2020 was $0.9 million, $710,000 and $1.843 million, respectively; net losses were $45.03 million, $84.88 million and $178 million, respectively. Tucson Future said it had accumulated a loss of $405.2 million by the end of 2020. In 2021, Tucson Future had total revenue of $6 million and an operating net loss of $411 million.

Li Lin, an executive at an investment institution in Shanghai, is not optimistic about autonomous trunk logistics, in addition to regulatory and ethical issues, he believes that the overall size of China's freight fleet is small, and the ability to replace autonomous vehicles with internal drive is insufficient. In contrast, the commercial prospects for applying some of the functions of autonomous driving, such as AEB, to trucks, the commercial vehicle ADAS (Assisted Driving System), are clearer.

The commercialization of self-driving trucks in closed parks such as ports is faster than that of trunk logistics, but it is still dominated by test operations and has encountered a situation of competition with traditional technologies. Many ports have previously achieved partial autonomous driving and cloud control through technologies such as magnetic spike navigation sensors, but the perception system does not use lidar and cameras. Although the technical aspect is cheaper in comparison, it can also play a role in ports with relatively simple traffic environments, and the technology has matured.

True and false tracks

Among the popular applications of autonomous driving, Robotaxi is considered to be the closest track to commercialization. At the "Chief Intellectual Officer Conference" recently organized by Auto Byte, Xiao Jianxiong, founder and CEO of AutoX, said that AutoX has been focusing on removing the L4-level driverless RoboTaxi of the safety officer, which believes that this is the only way to truly commercialize the route: only to achieve the same practicality of existing online ride-hailing vehicles, completely remove the safety officer, unlimited destination, unlimited area of autonomous driving, is the real commercialization.

Xiao Jianxiong has publicly stated that if the current Online Ride-hailing market in China is completely replaced, then Robotaxi will be a $6 trillion market. Tesla CEO Musk has also said that a Robotaxi could cost less than $0.18 per mile, well below the average cost of Uber's traditional ride-sharing service ($1-2 per mile). In addition, Robotaxi can increase the basic utility of electric vehicles by five times.

But within the industry, there are also huge divisions in The outlook for Robotaxi, with one side arguing that this is a distant dream that may never be realized, and the other side believing that the commercialization of autonomous driving is just around the corner, and may be achieved in 2025.

"Robotaxi didn't mention the safety officer when telling the story to the outside world, and although the cost of the driver was reduced, the cost of the safety officer was added, and the process of removing the safety officer would be very long." Liu Tao said.

On April 7, Toyota Motor's self-driving company Woven Planet announced that it will use low-cost cameras to collect data and develop and upgrade its self-driving systems through effective training. The company says this helps reduce costs and scale up the technology. The incident has attracted widespread attention in the industry, and Toyota motor vehicles are also believed to embrace Tesla's "visual system" self-driving technology route.

Prior to this, only Tesla in the whole industry adopted the vision department, and other companies used the camera + lidar fusion system solution. In addition to not relying on lidar, the visual department does not rely on high-precision maps.

In the eyes of investors, Toyota's sudden choice of the vision department reflects that the technical route of autonomous driving is not clear. "The technical route is not understood, the cost of mass production cannot be reduced, the charging model is not clear, and the regulations are not broken, which is equivalent to these things being empty talk." Liu Tao said.

However, Liu Tao does not think that the essence of this round of autonomous driving investment boom is a capital carnival, because although autonomous driving has not been realized, some of its functions such as AEB and ADAS auxiliary driving systems have achieved mass production, and have driven a large number of industrial companies in different segments, which is the biggest difference from the AI winter in 2019.

In the self-driving track, the most difficult is the L4/L5 self-driving company with the brightest halo. Including Pony Zhixing, AUTOX, etc., they do not have any commercial landing projects. An analyst believes that autonomous driving is a "bottomless pit" in the true sense, with high research and development costs, difficult technology landing and unclear commercial monetization paths.

Although Baidu and Xiaoma Zhixing were approved to pilot commercial operations in Beijing, both companies tacitly chose a model that was close to free. "To study Robotaxi, you only need to look at Didi, Didi only made scheduling software can not be profitable, no company can do both scheduling software and automatic driving can also be profitable." Li Lin thinks.

"L4/L5 self-driving companies are embarrassed that they can't finance and hematopoietic walk on two legs. If mass production is pursued, the largest market value of Desay SV (about 60 billion yuan) is placed here, and their commercial value will be very clear, and it will not be able to support the current valuation of tens of billions of DOLLARs. Liu Tao said: "Autonomous driving feels very sexy, business expansion is very difficult, this matter has just begun, and now these autonomous driving companies can survive less, and even fewer can live in the midfield."

(At the request of the interviewee, Liu Tao, Zhang Chao, Zhu Yi, and Liu Zhi are pseudonyms)