Text丨Wang Hailu

Source丨Investment China Network

In the Huangpu district of Guangzhou, dozens of “autonomous taxis” with radars, sensors, and other equipment of various shapes are mixed with ordinary social vehicles and run on busy open roads.

The steering wheel rotates automatically. The safety officer in the cockpit placed his hands on his lap and allowed this “smart” Nissan Sylphy to automatically change lanes, steer, and avoid. The driving experience can't be described as comfortable. For example, if a social vehicle gets too close, the autonomous vehicle will brake urgently and give way politely. Social vehicles can be blocked and not abide by traffic rules, but autonomous vehicles never violate regulations.

In September 2018, autonomous driving company Wenyuan Zhixing joined Guangzhou Baiyun Taxi and Guangzhou Science City Investment Group to establish a joint venture “Wenyuan Guangdong Tour” to jointly operate autonomous taxis (Robo-Taxi). This company has obtained autonomous driving road test licenses from Guangzhou's Huangpu District and Guangzhou Development Zone, as well as taxi operation qualifications. Up to now, it has renovated dozens of Nissan Sylphies, with a cumulative test mileage of over 1.15 million kilometers. Local passengers can call an autonomous taxi at any time by downloading the app.

Currently, Wenyuan Zhixing's autonomous taxis are still in the stage of accumulating data and improving algorithms, and do not charge passengers any fees. However, Han Wook calculated an account internally: if the machine replaced the driver, the cost per kilometer would be cut in half. If the total price of the sensor plus the car can be kept within 350,000, a profit and loss balance can probably be achieved for four years of operation.

This is why Wenyuan Zhixing chose Robo-Taxi as its main line to enter the autonomous driving market: in the future, autonomous taxi operators will have the opportunity to compete with online car-hailing platforms for the travel market. Compared with providing autonomous driving solutions to car companies, this is more certain, and the implementation cycle is not that long.

Of course, achieving complete driverless driving also depends on the relaxation of policies. However, according to Han Xu, China's policy on autonomous driving is still very forgiving. Currently, safety officers are required because the technology itself is not mature enough, and “it has not reached the level of urgent requirements to remove safety officers.” Han Wook said.

Han Xu was a doctoral supervisor and tenured professor at the University of Missouri, studying computer vision and machine learning. In 2014, he wrote a letter to Wu Enda, Baidu's chief scientist at the time, saying he wanted to do autonomous driving. Then, at Wu's invitation, I joined Baidu North American Research Institute. At the beginning, Han Xu was responsible for deep learning. At that time, Baidu still only had a team under the Deep Learning Research Institute to do autonomous driving. Until 2016, Baidu focused on autonomous driving, drawing talents from various organizations, and Han Xu became the chief scientist of the autonomous driving division.

In 2017, Han Xu and Wang Jin, the general manager of Baidu's autonomous driving division, left Baidu and founded Wenyuan Zhixing (still called Jingchi Technology at the time). Wang Jin was the CEO, and Han Xu was the co-founder and CTO. Everyone knows the later story. Baidu sued Wang Jin for violating the competition agreement, Wang Jin resigned, and Han Xu took charge. After several months of turbulence among executives, Jingchi Technology changed its name to Wenyuan Zhixing and set sail again.

He Xiaopeng, chairman of Xiaopeng Motor, who once invested in Wenyuan Zhixing in his personal name, commented that Wenyuan Zhixing is a company that has “experienced ups and downs” and is “resilient.” He Xiaopeng, who is also a serial entrepreneur, can understand the ups and downs of starting a business.

Due to the team's experience in the industry, Wenyuan Zhixing successfully received 57 million US dollars in the Pre-A round led by Qiming Venture Capital and co-invested by Innovation Factory and Nvidia. In October 2018, Renault-Nissan-Mitsubishi Alliance (AlliancerNM) strategically led the A-round financing of Wenyuan Zhixing. This is the first time that an investment group with a background in a car company has set up the Chinese market.

Getting investment from car companies is of great significance to autonomous driving companies. GM and Honda invest in Cruise, Toyota invests in Uber, Ford invests in Argo.ai, Honda invests in Aurora, and Fiat Chrysler and Waymo cooperate. Han Xu believes that there will also be a lot of strategic synergy between Wenyuan Zhixing and Renault-Nissan-Mitsubishi in the future.

In 2020, Wenyuan Zhixing plans to increase investment in Huangpu to 200 units, and plans to expand to 5 cities in 2023. There are not a few companies that are developing Robo-Taxi at the same time as Wenyuan Zhixing; among them, there is no shortage of giants. At the end of 2018, Google Waymo began operating Robo-Taxi in Phoenix. In August of last year, Didi spun off its autonomous driving business and said it would operate a Robo-Taxi service in Jiading, Shanghai; in September, Baidu's Robo-Taxi passenger service was put into trial operation in Changsha.

According to Han Xu, the Chinese market has huge potential, and competition will allow startups to run faster and accelerate technological progress and implementation. He believes that what is really worth fearing are opponents who have made rapid gains in technological development and have ignored safety. So far, there have been no safety incidents with Uber or Tesla's autonomous driving in China. If it raises public concerns about safety, it could be a fatal blow to the entire industry.

It seems that autonomous driving will take time from technology maturity to commercial implementation. Capital's enthusiasm for this track is also gradually returning to rationality from excessive optimism a few years ago. Han Xu believes that the arrival of winter will squeeze out the bubble, but there is still a chance that good projects will be favored by capital. Overcome this stage, overcome technical problems, and make the team and investors more confident. This is the CEO's job.

“For a really good company, capital will be difficult in the winter, but it won't die.” Han Wook said.

The following is a summary of Han Xu's interview with China Investment Network:

Removing security officers within 2 or 3 years will definitely make a profit

China Investment Network: Wenyuan Zhixing began operating autonomous taxis in Guangzhou's Huangpu District and Guangzhou Development Zone at the end of last year. How many taxis are currently on sale?

Han Xu: Our goal is 100 units in 2019 and 200 units in 2020. Of course, we will gradually increase our investment because increased capacity and technology complement each other. In my opinion, if safety officers are not removed, the scale of thousands of cars is meaningless. You see, Waymo is also a few hundred cars, but when the safety officer removes it, tens of thousands or hundreds of thousands of cars can be paved.

China Investment Network: In the Huangpu area, the launch of 1,200 vehicles will probably have an effect?

Han Wook: Actually, nothing. Wenyuan is in an area of 144 square kilometers where it can operate. If you operate taxis, it is normal to have more than 1,000 cars. But for autonomous vehicles, if we can actually run to 100, I think it would be a great feat. There is no place in the world where autonomous driving runs 100 cars.

Investment.com: How do you view the imagination of autonomous driving combined with mobility?

Han Wook: If you rent a car every day and rent on call, I don't think many people would like to drive their own car. Why isn't there an increase in rental volumes? Because it's more expensive. The reason why prices cannot be reduced is mainly due to labor costs. After machines replace people, if the hardware is cheaper, this thing (autonomous taxis) will definitely come.

Investment.com: Will Robo-Taxi need to participate in the travel market war in the future?

Han Xu: The travel market is very vague. We mainly use technology to replace people.

With the Chinese market being so big, in terms of autonomous driving, I'm looking forward to a respectable rival even more. What I'm afraid of is the kind of rival that doesn't consider safety. The advantage of a market economy is competition, making everyone stronger and providing the best service to consumers.

China Investment Network: What do you think of the previous incident between Uber and Tesla?

Han Xu: Any technological advancement must ensure safety. It should not be said that sacrificing safety for the sake of the speed of technology is a basic act of conduct.

But on the other hand, any technology, from immature to mature, cannot be 100% problem-free. Various tragedies have occurred in the history of human development, including aviation being able to carry passengers for civilian use. Every technical improvement is followed by an air disaster. This is a very sad thing. So is space. So many astronauts have died; why don't we start the space business? Definitely something to do. We must ensure safety to the greatest extent possible, but at the same time, technology has not progressed 100%.

China Investment Network: What is Robo-Taxii's biggest technical challenge?

Han Wook: I think it's the stability of the entire system. For example, the perception module can do a better job than a human. But the point is, how far can it be done in the worst case scenario? It is often the worst situation that causes an accident in the end. When the stability of our entire system is such that the worst situation is better than the worst situation for people, then we can say that cars are safer than people and cars.

China Investment Network: In the future of autonomous taxis, will Didi be the biggest competitor?

Han Xu: I think Jack Ma is right in saying, focus more on the market and less on competitors. We now have a bunch of problems in the market. If you focus on being a sophomore, there's probably something wrong with being a sophomore. As the old three becomes the second, you have to pay attention to 2345. Actually, there's no need; you just want to know how to run faster.

China Investment Network: You mentioned earlier that Wen Yuan wants to find a profit point within two or three years. How does he plan to make a profit?

Han Xu: Autonomous driving, removing safety officers within 2 to 3 years will definitely be profitable. We can connect to Didi or Gaode. If it can achieve profit, this company is already in the top 0.1% of the country.

China Investment Network: When can security officers be removed, and how are policies stipulated?

Han Wook: Policies are being relaxed all the time, and I think 2022 will move people. China's policy has always been very encouraging experimentation. The reason why the policy is uneasy now is that technology has not reached the level of urgent requirements for the elimination of security guards.

Han Xu, CEO of Wenyuan Zhixing

People always think that university professors are lecturers

It's not like that

Investment China Network: You were a professor at a university in your early years. Why did you choose to work as a teacher instead of going to a company after graduation?

Han Wook: At that time, there was a family reason. I grew up on a college campus. My grandfather was a college professor, and my mom also worked at a university. My grandpa was a middle school principal. In the 1960s, he built more than 20 middle schools in Hohhot. At the time, I also longed for that kind of life as a professor. I was more free to explore something new.

China Investment Network: What is life like as a professor and researcher at the University of Missouri?

Han Xu: At the beginning, the research team had 3 or 4 people; later, there were probably about a dozen people. Our main research direction was object detection, recognition, and tracking. Every day we have some new ideas, then we do experiments, improvements, and publish papers. At the same time, I also bring students and have teaching tasks.

China Investment Network: Which part do you enjoy the most?

Han Wook: What I enjoy the most is scientific research. I think in an ideal state, if you don't have to apply for funding, think directly about these issues, solve these problems, and finally achieve results, it's very, very interesting.

China Investment Network: Why did you go to Baidu later?

Han Wook: I got a lifelong teaching position and had an academic vacation at the time. I wanted to go to the industry at the time. I'm very interested in autonomous driving. Waymo was my first choice at the time, but it didn't work for various reasons. I later wrote a letter to Wu Enda. Wu Enda said I don't have autonomous driving, but you are welcome to watch it. I started deep learning. Later, Baidu had autonomous driving, so I moved on.

Investment.com: What research was done on autonomous driving during Baidu?

Han Xu: Before, I only read a few articles, but since I arrived at Baidu, I have actually been working on projects. At first, I was responsible for sensing, and later I also covered various fields such as sensors and hardware. Before I arrived at Baidu, Baidu had a very small team working on autonomous driving. It was a research group under the Deep Learning Research Institute. Later, Baidu started autonomous driving on a large scale and drew people from all over the world, so I left.

China Investment Network: Why did you choose to start a business?

Han Xu: Small companies are faster, more flexible, and have more dominance.

Investment Network: From academia to entrepreneurship, you need to deal with management, financing, and handling government relationships... are these things challenging you?

Han Wook: I think it's a new challenge for everyone, as long as you haven't started a business.

China Investment Network: Are you interested and passionate about doing these things?

Han Wook: Yes. People always think that university professors are lecturers; that's not the case. A university teacher wants to bring in students; this is a management team; to recruit new students, this is recruitment; to apply for a research fund from the government is no different from our current financing; he wants to introduce its organization at the conference, that is, it engages in PR and marketing, which is very similar to running a company.

China Investment Network: Is there any part you have to do, but you probably don't really like to do it?

Han Wook: The state of people also fluctuates. Many times, as an entrepreneur, you can't rely on your nature. I think restraining my emotions and overcoming fatigue is probably already very depressed, and I also need to show a state of full energy. This is a necessary quality for entrepreneurs.

If this thing itself is done well, there will always be capital favors you

China Investment Network: You got financing from Nvidia in the first round of starting a business. How did you get it?

Han Xu: At that time, I had only been working for a month or two. I had a little bit of technology, but it wasn't important; it mainly depended on the team's qualifications.

But later, for example, when Nvidia CEO Jason Huang recently looked at our car, we unveiled WeRide Smart Suite 3.0 at the GTC conference hosted by Nvidia in December last year, and placed the self-driving car roof integrated kit on the roof of the car. I think after so long, Nvidia is also very approving of our technology development and operation. Otherwise, it wouldn't have put our logo on GTC, in the center of the stage.

Investment China Network: Are you also collaborating on products?

Han Wook: We are always collaborating. For example, our newly released WeRide Smart Suite 3.0 application is the DRIVE AGX Pegasus platform developed by Nvidia for L4 autonomous driving. We are the first platform in the world to achieve full sensor access, placed on the top of the car, and completed object detection. I haven't seen any other company do it before.

China Investment Network: How was the A round of financing, Renault-Nissan and Mitsubishi Alliance invested in Wenyuan Zhixing?

Han Xu: They watched it several times, came to 6 teams, and compared them in China and even the world before deciding to vote.

Investment China Network: This is strategic financing. Will your products be used in their cars in the future?

Han Wook: There are lots of development opportunities, but I can't easily reveal too many.

China Investment Network: Why did He Xiaopeng invest in Wenyuan Zhixing?

Han Wook: It's a personal investor. He feels that this team is very tenacious. After going through ups and downs, they still work hard to move forward. He places great importance on this point.

China Investment Network: Since he himself is a person who has founded a business, can he understand the ups and downs of starting a business?

Han Wook: I think so.

China Investment Network: What do you think of Xiaopeng and the new car-builders?

Han Xu: Building a car requires 200 to 300 suppliers, with 20,000 to 30,000 parts. It is admirable that such a complicated matter can be brought to the point it is today.

China Investment Network: Xiaopeng and many other car companies are also developing autonomous driving. Have you compared the technology of car companies?

Han Xu: Car companies are mainly L2 and L3 levels, which are fundamentally different from L4 levels.

China Investment Network: Why haven't you done L2 or L3?

Han Wook: L4 really represents the future. We expect one person to sit in a car without a driver's license. I think L2 and L3 are car companies' business. Its kind of advanced assisted autonomous driving is closely related to the car. For example, for controlling this car, let's take lane line maintenance as an example. It is necessary to integrate with this car's system; it is not enough to simply know its trajectory. Moreover, it has a lot to do with your viewing angle. Where the camera is installed and how many cameras you use, the effect will be very different. Only car companies will decide on these things, so it's better for car companies to take the lead.

Investment China Network: There is a saying that car companies are doing L2 and L3 only to increase the selling points of cars; in fact, they are not what users just need?

Han Wook: I don't think L2 and L3 are just what car buyers need, but they are a great feature.

China Investment Network: Wenyuan Zhixing has recently carried out Series B financing. What is the current financing environment in the field of autonomous driving?

Han Xu: It is true that entrepreneurs have a cold capital winter now, and everyone's financing situation is not optimistic.

China Investment Network: How long will this cycle continue?

Han Wook: I don't think anyone has any clear predictions. The economy has this kind of cyclical nature. In a cold winter, a number of companies that are not suitable for the market will be eliminated. For a really good enterprise, capital will be difficult in the cold winter, and it won't die. If a company's capital comes in the cold winter and then dies, I think it may also be because the company is not adapting well.

China Investment Network: What is the biggest challenge in winter?

Han Xu: Actually, my biggest concern is whether technological development is too slow. If the matter itself is done well, there will always be capital that favors you. In the process of development, safety may be affected, and the trust placed on employees and investors is misplaced. These are my biggest concerns. Employees are watching every day. Your CEO leads us. Is your vision correct, and your decisions are right? Are you working hard enough, they watch you every day.

China Investment Network: You are targeting Waymo. Does Waymo seem to be slowing down in terms of technology too?

Han Wook: Overall, I'm quite optimistic; I think they are making pretty good progress. I think autonomous driving is coming soon. If Waymo is deployed on a large scale and in all directions next time, everyone may be excited again. In fact, now is the time to stick to it.