In the long run, a technology-based automobile company should have a high gross profit.

With the help of intelligent technology, XPeng Inc. is rapidly exploring brand boundaries.

On the first day of the Guangzhou Auto Show, XPeng Inc. launched a sedan model, XPeng Inc. P7, with a scissors door equipped with millions of supercars, with a price as high as 409900 yuan. "this is the first time XPeng Inc. has sold it for more than 400000 yuan. XPeng Inc. CEO he Xiaopeng said at the scene.

Different from the brand strategy of NIO Inc. and Tesla, Inc., XPeng Inc. 's first product, G3, sells only about 150000 yuan, neither pursuing extreme dynamic performance nor setting up offline stores in the heart of first-tier cities. But he Xiaopeng still has the strength, "some of us think (P7 Peng wing version) more than 400000 very geek will buy, I think (sales) will be a lot, look at next year's data." "

He Xiaopeng's self-confidence has a foundation. He formally joined XPeng Inc. in August 2017 to set up a self-research system, and now it has begun to bear fruit: automatic parking, full voice interactive cockpit, self-driving XPliot have been delivered, and high-speed navigation-assisted driving NGP has also gained a good reputation. Tesla, Inc. Model 3 successive price cuts, is regarded as a "new car killer", but close to Model 3 XPeng Inc. P7 not only did not disappear, the monthly delivery also exceeded 2000.

This also drives other members of the newly built car, NIO Inc., Li Auto Inc., and so on, to increase their own research and development system this year. NIO Inc. was personally led by Li Bin to catch up with the R & D progress, and Li Xiang also recruited Wang Kai, the former Visteon global chief architect, as CTO to promote intelligent self-research.

First, he Xiaopeng, who started his own research, told 36Kr Holdings that there was a lot of hard work. "if you really want to do it well, in my experience, they may still have two to three years, a lot of time to do this." "

In the face of the increasingly urgent competition pattern, XPeng Inc. is also accelerating the hardware iteration. At the auto show, XPeng Inc. announced four capability upgrades to the next-generation intelligent hardware platform.

First, the control domain is highly integrated: XPU autopilot intelligent control unit realizes 4-in-1, which integrates the intelligent control of driving and parking, making decision-making and execution more efficient, and at the same time reducing system weight and cost by reducing the number of controllers and wire harnesses.

Second, the improvement of computing power: the computing chip has been upgraded from NVIDIA Corp's Xavier 30TOPS to 200+TOPS.

Third, high-precision positioning promotion: RTK terminal is upgraded from 4G to 5G, Gaojing map not only optimizes high-speed roads, but also complements urban roads, using the new location scheme of "IMU+GNSS+RTK (5G) + high-precision map (high-speed & city)", covering occluded scenes such as tall buildings and underground parking lots.

Fourth, the ability of perceptual fusion is improved: lidar is installed on the production vehicle to improve XPeng Inc. 's horizontal and longitudinal position accuracy and spatial resolution (ultra-high resolution).

According to he Xiaopeng, XPeng Inc. 's production car will be equipped with lidar next year.The biggest deficiency of lidar is its high cost, so it was denounced by Tesla, Inc. CEO Musk that "any self-driving company using lidar is doomed to failure."

However, he Xiaopeng's many years of research and development experience in the Internet and automobile industry is that software must rely on the underlying hardware, especially security, and rely more on hardware redundancy. "in the first few years, we will put security in a very important perspective. I'd rather have hardware redundancy and software redundancy. "

On he Xiaopeng's R & D list, in addition to the software system, there will be a series of intelligent related hardware in the future. "hardware and software is a spiral iterative process, and now the core of the hardware we solve is still integration, quality and safety, but gradually we begin to innovate, such as autopilot-oriented electronics, embedded large screens and other hardware. as well as the audio and controller-related innovations in the intelligent music cockpit. "

Research and development continues to be invested, but the rise in market volume may not be synchronized, which makes he Xiaopeng has no intention of participating in the gross profit and cost-effectiveness competition between the other two companies, NIO Inc. and Li Auto Inc.. The former has a third-quarter gross profit margin of 14.5%, while the latter is even more so, reaching 19.8%.

"in the long run, a technology-based automobile company should have a high gross profit, but in the short and medium term, it is wrong to have too high gross profit. "He Xiaopeng said bluntly, "the industry is more to use a mature system to think, I can understand, but I want to take my special path." "

The following is an excerpt of the interview with 36Kr Holdings and other media and he Xiaopeng:

Q: how much do you expect to account for the P7's Pengyimen version of XPeng Inc. P7?

He Xiaopeng: to tell you the truth, there are many differences within us. At the beginning, I was a primary school student in the field of cars. I was constantly learning. I had some wrong judgments and a lot of intuitions were right. For example, I have always thought that high range is the most sold, everyone thinks that low range sells more, and in the end, I win. I think those with autopilot will sell a lot, they think it will not be so much, and I also win.

From the point of view of Pengyi Gate, we have two kinds of interiors. One thinks that more than 400000 may be bought by geeks. I think there are a lot of them. I think I am quite confident when we wait for next year's delivery data to see.

Q: from G3 to P7 to today's Pengyi Gate, the shape and positioning of our products have changed greatly. What is the path of XPeng Inc. 's definition of future products?

He Xiaopeng: I think we have several characteristics that are quite different from other car companies.

First, I do not follow the characteristics of the platform, why do you say so? I think the advantages of the platform are scale, low cost and reusability. We will do a lot of reuse in part of the component system, but it is easy for the platform to be reused, which leads to slow change of innovation.

Second, we try not to do small positioning, small positioning is in a period of time, there is a group of users have such a demand, so to make such a car, we are more willing to walk in a straight line in technology, intelligence and innovation.

Third, the sales price of our main force is less than 15-300000, which is what we have always said, that is, do the middle and high end, do the national tide. Such a group of customers they have purchasing power, and need in this purchasing power to get a better combination of functions, performance, capabilities, this is XPeng Inc. to continue to explore.

Q: when did XPeng Inc. plan to use lidar?

He Xiaopeng: it is certain that we will have a lidar next year. We were already doing this research in the first half of last year, so basically we have to go back almost two years before it is possible to land this matter. in particular, a lot of research and development work has to be done.

We are different from the traditional L4 self-driving or self-driving companies, they use lidar as the core to draw lidar high-precision maps, we are different, we use vision as the core, visual high-precision maps as the core. Because we think that from the point of view of the trend, all the roads, traffic lights, and laws and regulations in the world look at the world to drive from the perspective of human eyes, and we think this is the most efficient, closest to real person and cheapest, and can be combined with lidar to do a good job safely.

To be honest, in the first few years, we will put security on a very important point of view. I would rather have hardware redundancy and software redundancy than do a good job of security.

Q: XPeng Inc. has a lot of self-research achievements on the ground, and NIO Inc., ideal and other companies are also investing in self-research. What do you think are XPeng Inc. 's barriers in this situation?

He Xiaopeng: first, I think our direction is the same, on behalf of XPeng Inc. first said this matter, and now we all agree. But a lot of them are hard work, from having ideas to forming a team to being able to research and development, and then to being able to do research and development well. I think with my experience, they may still have two to three years, and they need a lot of time to do this. Because the logic of the whole stack will get deeper and deeper.

From XPeng Inc. 's point of view, we were the first to have this idea and have already worked it out. in the future, first, we will do it; second, we will make the technology in depth; third, we will bring this technology to the world. there is still a big difference between the thinking logic of globalization and the thinking logic of Sinicization. I believe we will lead all automobile companies in China in terms of intelligence.

At the beginning, many people said that smart cars do not need to be developed by themselves, but rely on integration. The core of smart car is the operation, OTA, service, software, iteration, data driver, and now a lot of people are starting to think about it. I think some car companies will act slowly.But this action requires a suitable gene and a large enough team. We have nearly a thousand intelligent R & D personnel, and I believe most of the companies are still stuck in the self-research team, which may be about 100 people, so I think they need more time to catch up.

Q: you mentioned that there were nearly a thousand intelligent R & D personnel. I remember it used to be hundreds. Is this only available within a short period of time?

He Xiaopeng: no.We have intelligent cockpit, autopilot, help intelligent cockpit, intelligent hardware related to autopilot, and some people, and we will do all these electronics and electronics. including the next generation of domain controllers are all developed by ourselves.Engineers without this piece of hardware, to put it bluntly, it is also very difficult for us to make software.

Why do I keep telling my friends in the Internet circle that if you only make a software, I think it is very difficult, because it requires the support of an architecture system with a lot of hardware at the bottom, and after the hardware and software are combined together, it also needs a lot of hardware architecture of the whole vehicle to be combined, so you will find that you will do more and more deeply, more and more like a car factory.When I had an idea in 2013, when I went to invest in such a company in 2014, I firmly said that we must do something like Apple Inc in order to be efficient and barrier in the end, and it is very difficult to use an integrated approach.

Q: Apple Inc designed the chip herself. Does XPeng Inc. have such a plan?

He Xiaopeng: we are working very well with NVIDIA Corp now. I believe we will have our thoughts on this in the future and how to grasp the whole balance.

Q: can you share what kind of mechanism is needed in the organizational structure of the management R & D team? Li Auto Inc. has CTO. Do you need such a role?

He Xiaopeng: at present, we don't need a CTO role. If we need someone to combine the three architectures of strategy, product and technology, I think it's me.

Because we used to manage the R & D team of two or three thousand people (in the UC browser era), the whole UC had 4000 people at that time, and BABA had 10,000 people for entertainment. I don't think it was a challenge to manage a R & D team of 2000 people. We set the goal, define the positions and boundaries of different centers and different departments, and even different groups, and then use a good project organization way to pull its business up and pace up, which I think is very good.

I think an enterprise, especially an enterprise created by science and technology, may have the boss and experts of the technology line, but the real meaning of CTO should be the integration of products, strategies and technologies. Otherwise, there will be some challenges for the sake of technology. I think different companies have different organizations.

Q: I am very concerned about whether XPeng Inc. will invest in the field of more hardware in the area of intelligence.

He Xiaopeng: I always believe that software and hardware is a spiral iterative process. Now the core of our hardware solution is to solve integration, quality and safety, but gradually you will find that we begin to innovate, such as autopilot-oriented electronics, embedded large screens and other hardware innovations. to intelligent music cockpit audio, controller-related innovation. Therefore, behind us, a lot of software and service innovation revolves around the big innovation with the hardware, first small innovation, then slowly big innovation, around this ring.

I would like to share a story with you. I went to recruit autopilot in 2015 and increased it in 2017, and I told them that the autopilot team was not a great team 10 years later, around 2030. Because 10 years ago, the big data team and the cloud computing team all felt great. Today, this team is a very basic team, a basic technical department in a big company, just because you left early at that time, so you feel sophisticated.

Q: NIO Inc. and Tesla, Inc. will also enter the European market in 2021. What I want to ask is that 2021 will be a big year of competition in both the Chinese market and the European market?

He Xiaopeng: I don't think so. The three years 2020, 2021 and 2022 are all the first stages of the first year, and more money, more people, more research and development and innovation should be the stage of accumulated strength. Including our entry into Europe and more countries and states, we have to go back to many basic roads of China, such as local organization construction, sales system, after-sales system, policy and regulation system, and product adjustment.

Q: domestic competition is also quite fierce, why do new forces have to go out?

He Xiaopeng: I think today's competition or today's entrepreneurs must think about a problem, you want to walk steadily, but also want to go fast, is impossible. To walk steadily is to walk slowly, because stability comes first. For today's entrepreneurs, what I want to say is that, to cite two examples, the number of headlines and bytes six years ago is only about 50% of that of today, and the income may only be 70% of that of today. This is my nonsense. The data is roughly like this. Tesla, Inc. 4 years ago was only 10% of the number of people sold today.

What I want to say is that today's entrepreneurship and global competition, it is not realistic to go fast on the basis of steady progress. The first is to run fast enough, and the second is to strive to be stable and balanced in the middle of running fast enough, that is, balance in fast running, which is what XPeng Inc. will focus on next.

Q: judging from everyone's recent financial reports, the cash reserves are basically 20 billion yuan, but they seem to have entered a state of gross profit and cost-effectiveness. what are the principles and strategies for XPeng Inc. 's next expansion?

He Xiaopeng: in the long run, a high-tech automobile enterprise should be an enterprise with high gross profit. In the short and medium term, I have actually always believed that it is wrong to have too high gross margins in the short term. I take my technology path, take my position, I think the industry is more with a mature system to think, I can understand, but I take my special path.

Q: you said that technology car companies have high gross profit margins. In your plan, what is the maximum gross profit margin limit for XPeng Inc.?

He Xiaopeng: we'll talk to you when we do it in the future.