On August 17, the autopilot demonstration application of AutoX, an autonomous driving startup, was officially opened to the public, and the public can conveniently call Auto through the Gaode Map App. X self-driving vehicles, unlimited times, users can take a taxi for free. ?

Although there is still a long way to go before the "commercialization within three to five years" promised by autonomous driving companies, driverless taxis are entering our daily lives in Guangzhou, Shanghai and Changsha.

Under the epidemic, the self-driving industry went against the trend. In the past few months, there have been many large-scale financing events. Ma Xiao Zhixing received an investment of $400 million from Toyota, and Didi announced that its self-driving company had completed the first round of financing exceeding $500 million. According to the statistics of China Electric Vehicle 100 Committee, in the first quarter of this year, the total financing of self-driving enterprises at home and abroad exceeded $3.5 billion, up by 34. 1% year-on-year. ?

The capital market is hot again because I saw the possibility of the unmanned taxi scene landing. According to McKinsey & Company's forecast, China may be the largest autonomous driving market in the world. By 2030, the total sales of self-driving cars will reach 230 billion US dollars, and the travel service orders based on self-driving will reach 260 billion US dollars. ?

Robotaxi's appearance is the technical competition of "showing muscles" of autonomous driving companies, and it is also the competition for the most valuable scenes in unmanned landing. Xiao, the founder of AutoX, said in an exclusive interview with a travel reporter: "At present, the four companies in the first echelon of domestic autonomous driving companies can still be harmonious, but if one company takes the lead in crossing the starting line of large-scale commercialization, it will completely change the pattern."

The technology is mature and the policy restricts the leader.

Traveling as a guest: during the test ride, I feel that the display interface in the back of AutoX is not very clear, and the green belt and rain water are displayed as a collection of points. Will the interaction with passengers be improved?

Xiao Xiong Jian: Our focus is not to make the interface look good. AutoX's positioning is to do Robotaxi? The driver is equivalent to "the person behind the automatic driving". How to display the central control screen is a matter for the partner. We don't do taxi apps, we all use third parties, such as cooperating with Gao De. ?

Our business model is not the direction of ToC. AutoX will not get closer and closer to consumers, but will go further and provide technology for car companies and operating platforms. ?

Take a trip: AutoX got the second California driverless license (manned) in July. What does this permit mean?

Xiao: At present, we are the only self-driving company in China that has obtained this license. Waymo got the first one on 20 18, and we were the second one.

Autopilot in the United States started relatively early, and advanced faster than China in law and policy. However, the threshold of the license itself is particularly high, and no one can reach its standard after Waymo takes it. The mainstream autonomous driving companies in the world are doing road tests in California, so the competition in California is very fierce. The California government is also the most stringent government in the United States. For a long time, it has been dealing with various companies. They know how to drive automatically, so they are very knowledgeable. ?

Traveler: Will there be security officers in the background for real-time supervision and remote control?

Xiao Xiong Jian: At present, due to the epidemic, employees in California can only work from home, but after obtaining this permission, they can not be equipped with security officers on the road. ?

Both 4G and 5G may be disconnected from the network, so the requirement of this license in California is that although you can monitor in real time in the background, you can't control it remotely, and the vehicle can handle the problem independently under any circumstances. ?

Take a trip: Does this mean that AutoX has achieved L4-level automatic driving?

Xiao Xiong Jian: This is L4, so it's difficult to get a driver's license in California. It is different from ordinary exams or operating licenses. It is signed by the governor of California. No one was seen driving behind the steering wheel in the street. If the technology is not enough, they can't just go on the road.

Traveler: How far do you think it is from the large-scale commercial landing of fully autonomous driving?

Xiao Xiong Jian: The technical problem has been solved, but the business problem has not been solved, which determines how much money you can earn and how many cars you can allocate. We can operate now, but this is a pure loss.

At present, the regulations in the United States and China are that autonomous driving companies cannot charge fees. Waymo has already obtained all licenses for driverless testing, but California's policy is free of charge, and it cannot be commercialized without charge. It is no problem to put a few cars on the street casually, but it is definitely a loss to let users ride for free in a large area. ?

Take a trip: So the biggest restriction now comes from government policies?

Xiao Xiong Jian: Policy is a very big restriction. Getting a license means that it is technically possible, but whether it can be charged is a policy issue. It is even farther in China, and there is no driverless license at all. Then this standard must be established before the license can be officially issued, and whether the operation can be charged is also a problem. ?

If there is no charge, the autonomous driving company will never dare to spread the fleet on a large scale. The development of technology is changing with each passing day, but the progress of laws and policies is not so fast.

As the first person to eat crabs, we now have the first-Mover advantage, and only we have obtained the California driverless test license. We hope to make use of leading technology, but without a more open policy, our first-Mover advantage will be useless. We can be completely unmanned, but other enterprises can't, but the law stipulates that security officers must be equipped and no fees can be charged. Wouldn't that make technological advantages nowhere to play? Two years later, companies in other countries can also be completely unmanned, and the time advantage we have worked so hard to develop will not exist. ?

Waymo's car may not drive well in China?

Take a trip: Is AutoX ahead of other domestic autonomous driving companies?

Xiao Xiong Jian: Now getting a California driverless license is a huge official recognition of our technology. All autonomous driving companies operating in R&D and California want to take over, and everyone is applying. AutoX can be obtained as the second one, which only shows that we have technical advantages.

Take a trip: Waymo is widely regarded as the leader of the whole industry among autonomous driving startups. What do you think is the difference between AutoX and Waymo?

Xiao Xiong Jian: Waymo may not be as good as us in China. The road conditions in China and the United States are very different, and people's driving styles are also very different. We don't feel any difference between us and Waymo just by car experience. ?

Waymo has done better in some places, but the first-Mover advantage is not so great now. Of course, at present, no one in the autonomous driving industry is perfect, as are AutoX and Waymo cars, but this imperfection does not necessarily require manual takeover, and there is no safety problem. But judging from the feeling of riding a car, you will feel that the car is not as "smart" as a real person, for example, the traffic jam will not be blocked, and it will only be avoided by other cars. And there is room for improvement.

Traveler: Autopilot relies heavily on data accumulation and deep learning. Waymo's data volume is unmatched by other companies. Why do you think the difference between the two is not so big?

Small: the data is not based on quantity, but on quality. The density of roads in the United States is very low, and you can't meet a car for a long time, while the density in China is very high, and you can meet hundreds of people in a few minutes downstairs. These are opportunities to train machines. In the United States, it is also difficult to encounter complicated road conditions, such as traffic jams and traffic jams. Without such real data, it is difficult to improve the system. In China, the opponents of self-driving cars are all kinds of irregular cars, retrograde electric cars and pedestrians crossing the road at will. , so the data quality will be higher.

We also have a late-comer advantage in tools. Although Waymo started the road test ten years ago, the driving simulator only took shape on 20 15, and it also experienced many stages of exploration in the early stage. As a late-developing company, the whole academic field and industry knew what to do when it was established, so it quickly built a simulation platform and used data and tools to the extreme. ?

It's like Tesla was the first to build an electric car. It's very difficult to start from scratch, but it's much easier to make electric cars now. So is Waymo. At first, it was to be an assisted driver. 20 13 decided to switch to L4 development. 20 15 also tried to build a car. The ever-changing business model also consumes a lot of time and cost. As a pioneer in the industry, Waymo actually helped latecomers avoid many pits.

Take a trip: You started as a fresh-keeping logistics company in the United States, and you are more focused on carrying people in China. From the perspective of commercialization, which is more promising, freight or manned?

Small: it is necessary to adapt to the loading and carrying of goods in different regions. Once combined with business, policies and specific implementation conditions become very important. In the early days of the United States, we did fresh food and take-away. On the one hand, because of the license, there was no manned condition at that time; Another reason is that it is difficult to achieve automatic driving when making fresh food or take-out in China. People in China order takeout and ask for it to be sent upstairs. It is a very real problem that self-driving cars can't go upstairs. ?

Logistics, such as light trucks on urban lines, can be done. At present, our domestic routes are mainly light trucks with goods and Robotaxi with people.

Holding the thighs of car companies is not the best way out for autonomous driving companies?

Take a trip: How big is AutoX in China at present?

Xiao: At present, most of our vehicles are in China, with nearly 100 vehicles, mainly in Shenzhen and Shanghai. The cooperative automobile companies include BYD, SAIC, Dongfeng and Chery. ?

Take a trip: Many autonomous driving companies will choose a car company as the main investor and take advantage of traditional car companies, such as Cruise and General Motors, Argo? Ai and Ford. Will AutoX go the same way?

Xiao: AutoX is a technology platform company. We hope that the more car companies we cooperate with, the better, instead of choosing a company to cooperate exclusively. Our goal is to become a giant of autonomous driving technology, and let our achievements be installed in all kinds of cars, not just one or two cars. ?

Some companies have taken money from big car companies and seem to have found a backer. They don't have to worry about money, but from the perspective of funds, other investors will be under great pressure. The autonomous driving company has become a subsidiary of the car company, losing its independence, and the investors of the car company have a dominant position and power in the board of directors, and even have a veto power. Autopilot companies want an IPO, and car companies may object, saying that you don't need an IPO and you have no money to spend. So this will put a lot of pressure on other investors who want the company's IPO. ?

Traveler: So you don't think autonomous driving companies need to hold the thighs of car companies? You have no financial pressure?

Xiao Xiong Jian: AutoX still has enough money to spend. We don't want to find a car company to be a "father", and we don't want the company to become a pawn and department of a giant enterprise. This is the ceiling of future development. Like some Internet companies in China, ByteDance has developed into a giant because it didn't take becoming a subsidiary of Tencent or Baidu as its goal from the beginning.

We are also ambitious and hope that AutoX can grow into a technology giant independently. If the development of the company is not very good, it is also a good choice to find a giant holding company, but we still have a good life. ?

Traveler: Will there be exclusivity after the cooperation between autonomous driving companies and car companies? For example, if you talk to customers, they will ask that your system can't be installed in other cars.

Xiao Xiong Jian: If the car company makes such a request, it will be very difficult. The current situation is that there are not many companies that do well in autonomous driving in China, and the range of customers' choices is not large, just AutoX, Ma Xiao Zhixing, Didi and Baidu. Didi doesn't sell systems, so there are only three companies that can provide technical empowerment. Foreign car companies want to enter the field of autonomous driving in China, so it is unrealistic for every car company to find an autonomous driving company. ?

Many people think that car companies will be very strong when negotiating with autonomous driving companies, but they are not. Waymo and Argo? The valuation of Ai is higher than that of many car companies, and who buys it is still a question mark. As a traditional industry, the gross profit of automobile industry is very low, while the profit of technology industries such as autonomous driving is very high, with low income but high market value.

We hope to become a technology platform, just like Windows operating system, which can be used in various cars. To be big, autonomous driving companies must dock a number of car companies instead of bundling one.

Demand determines that autonomous driving can only be installed after.

Traveler: If the autopilot company cooperates deeply with a car company, the autopilot system can be installed in front. Now you can only do post-installation. Is this a restriction?

Xiao Xiong Jian: Whether it is front-loading or rear-loading, whether it is binding a car company with an autonomous driving company or cooperating with many car companies. The key point of front-loading or rear-loading is not whether it is deeply bound with the car company, nor what technical problems, such as more reliable front-loading and unsafe rear-loading. The containers are all the same, only in this factory or that factory.

There are two main points. First, have technical car companies considered the modification of autonomous driving companies? BYD has made some designs for autonomous driving. There are no wires on the ceiling. Because of automatic driving, the wiring will be different, including the wire-controlled conversion system, brake and throttle. BYD has provided us with hardware. ?

The other point is quantity, that is, the size of production scale. If the demand is large enough, of course, it can be installed before, on the contrary, no matter how good the technology is, it can only be installed after. If a factory produces 200 cars a day, the production line will not stop. Stop for a minute and lose tens of thousands, and stop for a day and lose tens of millions. If there is little demand for autonomous driving and you can't go to the production line, the factory will produce the quantity you need within two days. What should we do on the third day? Can only dismantle the production line, heavy losses. ?

If an autonomous driving company says that it cooperates with car companies for pre-assembly production, this is impossible at present. It has been hundreds of days since it was announced, and there should be tens of thousands of unmanned vehicles, but I have never heard of any one with such a large scale. ?

Traveling in the same car: If a car company and an autonomous driving company are deeply bound and trust each other, they will use autonomous driving technology as their own parts and enablers in the early stage and join the production line to provide consumers with more high-tech products. Will this be an advantage?

Xiao: At present, all the demands of unmanned driving are incompatible with the production capacity of car companies. The demand for quantity determines the form of cooperation. People usually think that car companies will pay for autonomous driving technology, but cars will eventually be sold to people and consumers are paying the bill. ?

Car companies always produce only cars. Robotaxi, a popular passenger car with autonomous driving technology, is paid by taxi companies. Robotaxi's biggest problem is that it is too expensive, basically twice as much as a car. Ordinary consumers can't afford it, so they can only sell it to operating companies, which can make money by operating unmanned taxis to recover their costs. ?

So the quantity depends on the needs of the operating company. If a car company produces tens of thousands of cars at a time, it can't deliver the goods at all. Therefore, whether it can be mass-produced before installation is not determined by car companies, nor by the needs of ordinary consumers, but by operating companies. ?

Tourist cameos: judging from the scale of Robotaxi bought by operating companies, it is not optimistic to achieve mass production.

Xiao Xiong Jian: This is actually a question of policies and regulations, because the policy stipulates that unmanned driving must be equipped with security officers, and there is no charge for trial operation. Taxi companies can't charge Robotaxi, they won't buy it without profit drive, car companies can't mass-produce it, and driverless systems can't be sold. This logic line is fundamentally a policy restriction, so it is necessary to promote the introduction of laws and regulations.

Whoever crosses that line first wins.

Guest Travel: It seems cold in the field of autonomous driving at 17 and 18. Recently, Robotaxi has been put into trial operation frequently in China, and it has once again been sought after by capital. How do you understand the current industry situation?

Xiao: From the perspective of the whole industry, the technology is very mature, but many people didn't know how mature it was before. In the first two years, most people were pessimistic, although they had never done research on autonomous driving technology. Without knowing it, both the media and capitalists think that autonomous driving is difficult to really land. ?

The public's perception of the industry is actually somewhat out of touch, and the industry has been making progress. But if you don't reach the critical point, it's hard for outsiders to see the difference, except that you are still writing code. A few years ago, everyone was in the laboratory by themselves, and there was no trial operation. Others could not see the results of autonomous driving, so the heat was relatively low. But now we can gradually see some progress, such as getting a driver's license in California, or inviting investors to experience the trial operation, so they can feel the potential of this field for themselves.

Take a trip: Does AutoX have a specific time plan for the large-scale commercialization of Robotaxi?

Xiao: At present, Robotaxi's business model is reasonable, very clear and has a big market, but everyone is worried about when it will be realized in ten or twenty years or in the next two or three years. We are not completely pessimistic about policies and regulations. If there is a big breakthrough in policy, Robotaxi will be fully implemented in a certain city in China in two or three years, and it will be charged. As long as one or two cities can make a breakthrough, other cities will also promote it. ?

Some companies will be more cautious, saying that it may be five or six years, but the specific scale is different. Landing in five or six years 10 city, and so on. But at present, our judgment is in line with the judgment of most people in the industry. ?

Traveler: In recent years, the competition among autonomous driving companies in Silicon Valley is fierce, and the capital is tilted towards several head enterprises, and the Matthew effect is becoming more and more obvious. Is there a similar trend in China?

Xiao: I think China has four obvious advantages at present: AutoX, Ma Xiao Zhixing, Baidu and Didi. We and Ma Xiao may have obvious technological advantages. Baidu's advantage lies in its large volume, while Didi has natural application scenarios and commercial empowerment. Toyota is behind Pony, after all, it has invested a lot of money, but our side is actually SAIC, Dongfeng, Ali and other local forces, investing in AutoX to make a longer-term strategic layout. ?

There will be competition in any industry, but at present, the competitive pressure in the autonomous driving industry is actually not so great, and the Hundred Regiments War and the Money Burning War are unlikely to appear. Now there are only four first echelons, because the technical threshold is too high and there are not many participants themselves. And this market is very big, which may be saved by the last four players. Everyone is a technology giant. ?

However, similar to the United States, domestic second-tier autonomous driving companies are really hard to make money now, because the strength of head players is too strong. Didi may not be the most advanced in technology, but its commercial strength should not be underestimated. Many self-driving companies founded two years ago are now silent, some are directly out, and some are transformed into port logistics or mining vehicles, but the biggest "fat" in the driverless field is actually Robotaxi. ?

Take a trip: Will these four companies in the first echelon in the future exist in harmony?

Xiao Xiong Jian: I don't think there will be any change in the next two or three years. None of these four companies will be eliminated or merged. ?

But whether it can survive for a long time is another question. After three years, there may be a finish line or a new starting line. At present, no one has realized the commercialization of autonomous driving, so although everyone is chasing each other technically, it is still at a stage.

Now everyone has gone through the initial stage and entered the second stage of development. All four companies have established the "Jianghu status" of autonomous driving. Everyone wants to squeeze into the third stage, which is large-scale commercialization. Once the first company truly realizes the unmanned operation of a city and crosses that "line", it will definitely use its first-Mover advantage to grab market share and replicate in more cities. ?

The third stage is the market? Sharing. 25% or 50% determines whether the industry will eat meat or drink porridge in the future. The first cross-line player will definitely spread a lot of products. If the remaining companies enter the market in two or three months, maybe these four companies will stay for a long time; But if it is too far from this finish line, the pattern will change greatly.

This article comes from car home, the author of the car manufacturer, and does not represent car home's position.