In 2024, the penetration rate of China's new energy vehicle market exceeded 50% for the first time, which means that for every two cars sold, one is new energy, and intelligence has undoubtedly become the core battlefield for car companies to compete. In this wave, Huawei Qiankun injected a shot in the industry with its impressive report card. In April 2025, Jin Yuzhi, CEO of Car BU for more than a year, introduced the Huawei Qiankun Intelligent Technology Conference with HUAWEI ADS 4. During this period, he fully explained Huawei Qiankun's vision in an exclusive interview with Phoenix Finance "Cover" for the first time - "bring intelligence into every car."

This "technical" executive who has been rooted in Huawei for more than 20 years has a distinct Huawei gene: pragmatism, long-termism, and loyalty to technology. It is also the same technology that supports Jin Yuzhi's silky transformation from the optical product line to the smart car track. "We are just entering the intelligent part of the automotive industry, and the foundation is still ICT technology. If we have to say the difference, it is that in the past, it is more targeted at the B-end, and now the C-end users are added."

According to the public annual report, Huawei's R&D investment in 2024 reached 179.7 billion yuan, accounting for about 20.8% of the annual revenue. Jin Yuzhi revealed to the "Cover" that the direct investment of Car BU "has been over tens of billions". Faced with the question of whether there is profit pressure, he calmly responded: "Huawei Qiankun's consistent strategy is to help car companies build and sell good cars. Profits are not a question we care about, and sticking to the strategic direction is the core."

After ten years of hard work, Jin Yuzhi said that Huawei Qiankun Intelligent Driver ADS will fully cover full-price models of 150,000 to 1 million, and will cooperate in depth with more than 20 car companies such as Cialis, BYD, GAC, and Audi. The new capacity is expected to exceed 1 million.

Speaking of the pain points of the automotive industry generally using high-end versions to promote the ability to assist in intelligent driving, Jin Yuzhi repeatedly emphasized the principle of "safety is the highest priority". "We adhere to the veritable product naming logic, from ADS SE to ADS Ultra, different configurations correspond to clear capability boundaries." For users, Jin Yuzhi also gave clear judgment criteria: brand R&D investment, auxiliary intelligent driving data, real user reputation, and hardware configuration capabilities.

"Without lidar, cars cannot move towards L3 and L4." Jin Yuzhi suggested giving priority to products with lidar capabilities. Behind this conclusion is also the confidence of Huawei's Qiankun fusion sensing technology route: the combination of lidar and millimeter wave radar can extend the obstacle identification distance to 200 meters in extreme weather and complex road conditions, and strive for "golden reaction time" for safety.

"L4 level autonomous driving in 2030 will be technically achieved." At present, Huawei's BU's phased goal has been locked in L3 - in the second half of 2025, conditional autonomous driving in high-speed scenarios will be implemented. From optical product lines to smart cars, Jin Yuzhi, who has crossed the border, always believes that the logic of "good quality, good experience, and good service" is a common language that penetrates industry barriers.

At the end of the interview, when talking about the balance between work and life, Jin Yuzhi said that "work is a part of life". He should use self-discipline to balance high pressure as much as possible with exercise frequency twice a week. In this interview, we saw that Jin Yuzhi outlined the stars of Huawei Qiankun with the rationality of technical personnel and the passion of idealists. Huawei Car BU polished technology with ten years as the scale and defined the future with safety as the bottom line. Perhaps, it is this "turtle spirit" that makes every step forward in technology closer to the ultimate imagination of mankind.

The following content is compiled from the dialogue record:

01

"Our consistent strategy is to help car companies build and sell good cars."

"Cover": Welcome Mr. Jin to be a guest on Phoenix.com's "Cover". I noticed that you have been in charge of Huawei Car BU for more than a year. Last year, Huawei Car BU turned losses into profits for the first time in history. Are you satisfied with this result?

Jin Yuzhi: Actually, there is nothing to be satisfied with. Our strategy is very clear. During an interview with the media in April last year, I asked, "I heard you want to make a profit this year?" I said that profitability is not a matter of concern to us, but the question we care about is whether we are moving forward in our strategic direction. As long as the strategic direction is right, one day you will make a profit.

Jin Yuzhi: We have always adhered to long-termism, not to look at relatively short-term fluctuations. Huawei is not a listed company and is not affected by too much public pressure. It has been like this for so many years.

"Cover": But with such achievements, you must have done something right?

Jin Yuzhi: That's certain, it's mainly two aspects. One is that we have long-term and in-depth investment in technology. Vehicle BU was established in 2019. It has been more than ten years since the initial investment in the 2012 laboratory. At the same time, these technologies are based on Huawei's entire ICT technology platform. Last year's public financial report showed that Huawei's annual R&D investment was 179.7 billion yuan, and Car BU has invested for more than ten years, so there is also a lot of indirect investment in it, which is actually invested by Huawei's platform.

Another factor is the business model. Our smart car selection model and HI (HUAWEI Inside) model help car companies build and sell good cars, so that business can be closed-loop and drive the sales of intelligent components. The core is the dual-wheel drive of technology and business, which allows us to make profits.

"Cover": Judging from Huawei's investment in car BU, what is the approximately magnitude of the investment in ten years?

Jin Yuzhi: Direct investment is definitely hundreds of billions of dollars, and indirect investment plus the quantity will be even larger.

"Cover": In fact, it also helps cooperative car companies save a lot of costs?

Jin Yuzhi: There is a huge investment in intelligent solutions, especially intelligent assisted driving. If everyone uses our public intelligent platform solutions, it will save a lot of money on R&D and long-term investments of car companies.

"Cover": Is the current Yiwang company clearer of Huawei's strategy of not building cars?

Jin Yuzhi: Our consistent strategy is to help car companies build and sell good cars, and to build an intelligent open platform. The automobile industry chain is very long and requires the participation of industrial chain partners. In the era of intelligence, software and intelligence are the core, which requires in-depth cooperation between platforms and car companies, and even continuous operations covering the entire life cycle.

"Cover": That is to say, you all form an ecosystem.

Jin Yuzhi: Yes, we will listen to users’ voices together.

02

"Correct use of 'man and car driving together' is safer than pure human driving"

Picture-in-Picture
PlayCurrent Time 0:00
/
Duration 0:41
Fullscreen
Mute
Automatic playback
"Cover": I am also a user of Huawei's auxiliary intelligent driving system. I have seen many users comment that "Huawei is the first in assistive intelligent driving, and there are no other brands." What do you think?

Jin Yuzhi: Whether it is good or bad is just like what you just reported, but the user actually has the final say. During use, users' perception of system security and experience are the most realistic, because this is a product that users strongly perceive. This is even more true for the future of autonomous driving. Users will directly feel whether the safety is good and whether the experience is smooth.

"Cover": So what is your future prospect for car BU?

Jin Yuzhi: Our goal has always remained unchanged, that is, to truly realize autonomous driving and unmanned driving. So we will continue to iterate along this strategic goal to help car companies achieve this goal together.

In this process, we must strengthen users' awareness of the auxiliary intelligent driving system and make more publicity that allows users to truly know how to use it safer. Especially at this stage, regulations have not yet allowed full autonomous driving. At present, it is essentially still in the "human-car driving together" stage, and users need to know more clearly where the safety boundaries of the system are. If you can use it correctly, it can actually be achieved in the "human-machine driving together" state, and it (assisted intelligent driving) is definitely safer than pure human driving.

"Cover": In fact, when promoting the auxiliary intelligent driving function, many car companies talk about high-end or top-level versions, which leads to many problems found in actual use.

Jin Yuzhi: Yes. Therefore, companies must objectively present the ability of intelligent assisted driving in publicity. For example, we released ADS 3.0 last year and also released ADS SE, which allows consumers to clearly define the differences in hardware configuration and capability boundaries through name distinction. In fact, there were different opinions when naming ADS SE at that time. Some people think that calling SE will look low-end, and should it be called Pro? But we finally discuss that the name must be true. In the future, when L3 is truly autonomous driving, we will also define different sensor configurations and capabilities from ADS SE, ADS Pro, to ADS Max, and ADS Ultra. Different names correspond to different sensor configurations and capabilities will be clearly presented to consumers.

"Cover": The overall name system will be clearly defined.

Jin Yuzhi: Yes. We must be responsible to consumers and must clearly convey three points: first, how to use the system better; second, accurately publicize the system capabilities; third, help consumers distinguish between the advantages and disadvantages of the system - after all, many people do not know what a good system, a general system or a poor system is.

03

"In the future, towards L3/L4 autonomous driving, cars must have lidar."

Picture-in-Picture
PlayCurrent Time 0:00
/
Duration 2:18
Fullscreen
Mute
Automatic playback
"Cover": As an ordinary consumer, what I am particularly interested in is how to simply judge whether an auxiliary intelligent driving system is good or not?

Jin Yuzhi: First, look at the brand, and check how long and how much money the system developed by this company has invested; second, look at the data, for example, by the end of March 2025, the cumulative mileage of assisted intelligent driving exceeded 2.5 billion kilometers, and the number of assisted intelligent parking exceeded 160 million times. These data will be published to consumers through reports every month; third, look at word of mouth, you can ask people around you because the real experience is the most credible; finally, look at the configuration, some manufacturers have great computing power, but the actual operation may be discounted, so consumers should not only focus on computing power, and there are other reference values ​​in configuration, such as whether there is lidar and millimeter wave radar. Lidar is necessary for the future direction of L3 and L4 autonomous driving.

"Cover": Do you mean lidar is indispensable?

Jin Yuzhi: It is impossible to achieve L3 and L4 without lidar, so I suggest that consumers should be clear about the above basic concepts, and also look at the parameters in a comprehensive manner, and make comprehensive judgments from these aspects.

"Cover": You have led Huawei's optical product line to be the world's number one and are a top expert, so you can answer from a professional perspective why lidar must be used to achieve L3 and L4?

Jin Yuzhi: Technically speaking, lidar can detect any obstacle, and its principle determines that there is no need to identify what it is, just avoid obstacles directly. Pure vision solutions require a lot of training to recognize objects, but there will always be extreme situations that have never been seen to bring risks. In addition, pure vision cameras will also significantly reduce their performance in bad weather or dim environments, so we insist on integrating the sensing route to be safer.

"Cover": Can lidar and millimeter-wave radar leave car owners with longer reaction time?

Jin Yuzhi: If you find the target 200 meters away, you will have plenty of time to react; if you find it 20 meters away, you will have almost no time.

"Cover": So under normal circumstances, is the recognition distance of the lidar farther than that of the visual solution?

Jin Yuzhi: It cannot be generalized. The visual view can be seen very far in sunny days, but the gap is obvious in bad weather.

"Cover": So it can be understood that the fusion solution of lidar plus millimeter wave radar is better than pure vision?

Jin Yuzhi: That must be better. Normal road conditions may not differ much, but the real road scene is complex and accidents may occur at any time, such as falling objects from vehicles in front of the highway, or bulging or water horses on the ground. These situations may be missed in pure visual conditions.

04

"Automobiles are related to life safety, and their quality requirements must be far higher than consumer-grade products."

Picture-in-Picture
PlayCurrent Time 0:00
/
Duration 2:33
Fullscreen
Mute
Automatic playback
"Cover": Compared with the era of traditional fuel vehicles, the combination of software and hardware is now emphasized. What do you think is the key point of combining software and hardware?

Jin Yuzhi: Taking sensors as an example, there are differences in hardware such as lidar, millimeter-wave radar, and cameras. For example, lidar is not easy to use.

"Cover": So not all lidars are at the same level?

Jin Yuzhi: Surely not. We have used industry general lidar and found that many noises cannot be processed, resulting in false triggering. Therefore, we must optimize lidar data from the bottom to solve problems such as false triggering, accuracy, and distance. Many companies in the industry do not use lidar well, which also leads to the lidar on the roof of some automotive products being useless, either unable to use it or not. This is a typical case about the combination of software and hardware.

For example, from perception to decision-making and then execution, deep integration with vehicle-side actuators (such as steering and braking systems), requires actuator manufacturers to open the interface, shorten decision-making delay, improve response speed, and thus gain more reaction time. This is a system project.

"Cover": What you mean is that consumers often only focus on the advantages and disadvantages of the auxiliary intelligent driving system. In fact, there are also differences in the hardware?

Jin Yuzhi: Of course. Hardware is also divided into automotive grade and consumer grade, and there is a big difference here.

"Cover": What are the specific differences?

Jin Yuzhi: Consumer-grade hardware is designed according to consumer electronic standards such as mobile phones and TVs, such as limited operating temperature range; while automotive-grade hardware needs to adapt to various extreme environments around the world, and there are more applicable scenarios.

"Cover": There is a trend in the field of new energy vehicles now, which is to 3C and home appliances. The advantage is that the product iterates quickly and pays more attention to user experience, but the problem is that cars are vehicles after all, which are directly related to the safety of consumers' lives. Many companies may ignore this.

Jin Yuzhi: My understanding is that cars are related to life safety, and their quality requirements must be far higher than the consumer level. From the perspective of level division, the telecom level in the middle is the telecom level, the automotive level up, and the aerospace level up. Moreover, the service life of the car is far beyond that of fast-moving consumer goods, so the hardware reliability requirements must be higher. We often see some news that some automobile products suddenly interrupt during the upgrade process and become "bricks". So whether in hardware, software or system, users must choose truly capable (products). Otherwise, you will face the possibility of buying a system that cannot be upgraded and cannot obtain sustained value.

Therefore, whether we are car companies or as smart solution providers, we must put safety at the highest priority. We also use this "Safety First" as our quality culture, and we place safety at the highest level in any link.

"Cover": What specific efforts has Huawei Car BU made in terms of safety?

Jin Yuzhi: We have built a complete quality system: the bottom layer is the basic quality of the system, the user experience and solution quality upwards, and the quality culture and process upwards. From product definition, design and development to delivery, the entire process ensures quality priority.

05

“Real autonomous driving is expected to be achieved before 2030”

Picture-in-Picture
PlayCurrent Time 0:00
/
Duration 0:59
Fullscreen
Mute
Automatic playback
"Cover": Recently, the Ministry of Industry and Information Technology released new standards and requirements on driving assistance. What do you think?

Jin Yuzhi: We firmly support these standards. This time, it was made clear that the L2 stage is the assisted driving stage, and it is not until L3 or above that it moves towards autonomous driving. This means that in the L3 stage, users know that they are still the driving subject. I think this is a very clear transmission, and it will also make the publicity more standardized, so we firmly support it.

"Cover": So what issues do you recommend users pay attention to when using driving assistance?

Jin Yuzhi: First, it is necessary to make it clear that this is a "human and vehicle driving together" state; second, as time of use accumulates, the ability boundaries of the system are clear, and the gap between a good system and a poor system is very large.

"Cover": In fact, the original intention of assisted driving is to help humans be safer. The advantage of intelligent assisted driving over humans is that there is almost no blind spot in the field of vision?

Jin Yuzhi: It can achieve a 360-degree view without blind spots. If equipped with lidar and millimeter wave radar, it will not be affected in dark scenes, and you do not need to adapt to light changes like the human eye, and you can also avoid such accidents.

"Cover": So do you believe that one day in the future it will be possible that humans do not need to drive themselves, or even need a steering wheel?

Jin Yuzhi: It will definitely come true. Our goal is to achieve true autonomous driving by 2030, at least reach the L4 level.

"Cover": Do you think it will be realized in 2030?

Jin Yuzhi: Yes, you can technically have such capabilities. However, whether users can experience it in the end depends on the follow-up of supporting measures such as policies and regulations.

"Cover": When is the expected L3 commercialization?

Jin Yuzhi: L2 is assisted driving, and L3 is self-driving with conditions. L3 will break through from two dimensions: one is the commercialization of L3's conditional autonomous driving, and the other is that L2 assisted driving will cover more models, and even fall to the 150,000-level market. From a technical perspective, we will be able to achieve L3 capabilities in high-speed scenarios in the second half of this year, but whether we can go on the road requires gradually relaxation in accordance with laws and regulations.

"Cover": What is the biggest difference between the ADS 4 released during this auto show and the previous ADS 3.0?

Jin Yuzhi: Simply put, the architecture and security have been upgraded, and the experience has also gone further. The core of ADS 4 is to move towards L3, making driving safer and smoother.

"Cover": After entering the automotive industry, you have made many accurate predictions. What are your predictions for the future? Jin Yuzhi: The next step is very clear - it is expected to move towards L3 commercialization in the next one or two years, that is, autonomous driving with conditions will gradually move towards commercialization.

06

“We have a greater vision to bring intelligence into every car”

"Cover": How do you balance work and life?

Jin Yuzhi: Work itself is a part of life, and it is a meaningful thing. The only problem may be the high work intensity, how to balance life. My method is to keep exercising, such as running and playing football two to three times a week.

"Cover": How long does it take each time? What is the pace?

Jin Yuzhi: It’s about three to five kilometers, the pace is OK, 6 minutes per kilometer.

"Cover": Do you feel that running will help your career?

Jin Yuzhi: Sure, very helpful. Especially after every run, you will be very energetic for a day and can keep your energy in your work, that is, exercise is to work better.

"Cover": Speaking of running, there is a very popular robot marathon recently. Have you followed it?

Jin Yuzhi: I saw the relevant news.

"Cover": I'm curious how you evaluate the development of the humanoid robot industry? This industry has also faced controversy recently, such as well-known investors who said they may withdraw from investment in this field.

Jin Yuzhi: I think so, the direction of the robot industry is fine, but robots do not necessarily have to be "humanoid". For example, in the topic we are talking about today, the future self-driving cars are also a kind of robot - robots that operate independently in complex environments with four wheels.

"Cover": From this perspective, smart cars may actually be one of the best carriers for humans.

Jin Yuzhi: It should be said that it is a robot that can operate well in very complex environments. Many of its technologies, as well as their capabilities in perception and motion control, are connected at the bottom with many technologies in the robot industry. If "autonomous driving robots" can eventually be commercialized on a large scale, it will lay a better foundation for robots to be more intelligent in various scenarios in the future.

"Cover": In other words, the technology of smart cars may inversely promote the development of the entire robot industry?

Jin Yuzhi: Yes.

Picture-in-Picture
PlayCurrent Time 0:00
/
Duration 1:13
Fullscreen
Mute
Automatic playback
"Cover": I saw that your past experience was mainly to create optical product lines, and later crossed the smart car industry. How did you achieve such cross-industry success?

Jin Yuzhi: When we enter the automotive industry, we are actually focusing on the intelligent part, and the foundation is still ICT technology. So from the perspective of technical compatibility, this is consistent with the logic of past optical communication and transmission networks. The same is that we always need to provide users with high-quality and high-experience solutions. The difference is that in the past, the ICT business was more aimed at corporate customers (To B). Now, it is engaged in the "three intelligence" solutions (intelligent driving, smart cockpit, smart car control) in the automotive industry. Although it is also sold to car companies (To B), the final experience will be directly passed to consumers (To C). Therefore, our requirements have changed from the past "good quality, good service, and high efficiency" to the present "good quality, good experience, and good service, while maintaining high efficiency".

"Cover": Now you have to face C-end users directly. Have you encountered particularly great pressure during your career transformation?

Jin Yuzhi: It’s okay. As long as the strategic goals are clear and continue to progress along the goals, the final result will naturally be achieved. In fact, what we see often is the result, the result determined by the process. Our company used to talk about the "turtle spirit", just like chasing the dragon-chasing spacecraft with the speed of a turtle. As long as we make a little progress every day, we can continue to surpass ourselves and move towards our goals.

We also have a greater vision to bring intelligence into every car and truly achieve unmanned driving. It is a very inspiring thing for our team to create something that did not exist in the world, and toward this goal together, bringing value to humans and all users.

"Cover": It really keeps people passionate.

Jin Yuzhi: Yes, turning non-existent into reality will bring you a sense of accomplishment.

"Cover": From another perspective, automotive intelligence is indeed a huge trend and trend.

Jin Yuzhi: The trend of intelligence is irreversible. Huawei saw a clear direction of automotive intelligence more than ten years ago, which is why we started investing so early.

"Cover": But there will definitely be many competitors in the process.

Jin Yuzhi: Competition is the norm. As industry leaders, we focus more on how to surpass ourselves than on our competitors.

"Cover": Have you set a goal for yourself? For example, in 2025, Huawei Qiankun Intelligent Driver should achieve?

Jin Yuzhi: The efforts of the past two years have determined the results of this year. Taking Huawei Qiankun Intelligent Driving ADS as an example, the number of new cars this year should exceed 1 million, which means that more users can experience our technical capabilities.

We now publish Huawei Qiankun assisted driving user usage reports every month, and partner car companies will also disclose user usage data. These reports will display real data such as mileage, usage proportion, and usage duration to help users establish a clearer understanding based on the real experience.

"Cover": As an industry leader like Huawei, how do you think China's transformation from a major automobile power to a strong power?

Jin Yuzhi: I think any industry or product needs to go from big to strong, and the core is to return to quality and service. If the automotive industry wants to go global, it cannot just do sales. It must build a complete after-sales service system, including accessories, maintenance and long-term service support. Only when these abilities are in place can we truly become a powerful automobile country.

"Cover": Thank you Mr. Jin for accepting our exclusive interview today.

Jin Yuzhi: Thank you.

[For more exciting content, please pay attention to Phoenix's "Cover" dialogue with Huawei Car BU Jin Yuzhi]

Special link:

https://finance.ifeng.com/c/special/8itbaiNUeK8

https://ishare.ifeng.com/c/s/8itbaiNUeK8

"Special statement: The content of the above works (including videos, pictures or audio) is uploaded and published by users of the "Dafenghao" self-media platform under Phoenix.com. This platform only provides information storage space services.

Notice: The content above (including the videos, pictures and audios if any) is uploaded and posted by the user of Dafeng Hao, which is a social media platform and merely provide information storage space services."

[Editor in charge: Xie Wei PF123]