地平线创始人余凯访谈：车载芯片、解放大脑与中国创新速度

Kai Yu: In some sense this is terrible business you know, heavy investment, very slow return. But in my view sometimes terrible business means a great business, because that's the entry barrier for all the other newcomers is extremely high.

Nicolai Tangen: Everybody, I'm Nicolai Tangen, the CEO of the Norwegian Sovereign Wealth Fund. 10 years ago, we all assumed self-driving cars would come out of Silicon Valley. But today, the most advanced cars on the road are Chinese and Western car makers are going to China to buy the technology. Nobody sits closer to the shift than my guest Yu Kai, who's the founder and CEO of Horizon Robotics, which supplies the chips and software that let Chinese cars drive themselves. He was one of China's leading AI scientists before he became an entrepreneur and few people can explain better how China got there and where he goes next. So Kai, welcome.

Kai Yu: Hi Nikolai.

Nicolai Tangen: Um let's just start with a simple question. What does Horizon Robotics do?

Kai Yu: Well essentially what we are doing is the computing platform for all generic robots. So including self-driving cars, but of course our current business is probably, you know, 95%, nearly 100% focusing on cars, passenger vehicles.

Nicolai Tangen: So if I sit in a car that you supply into, what does your chip do?

Kai Yu: Well, essentially it's running the software for all the way from the camera perception to trajectory planning and control. So essentially, it's kind of like the brain for all the cars.

Nicolai Tangen: Now your chips run roughly one out of four cars in China. Just tell me about the—

Kai Yu: One out of three. I think that's our latest track record.

Nicolai Tangen: Thank you. Well done. Who are your customers?

Kai Yu: Well, you know, the major Chinese OEMs like BYD, Geely, Chery, Changan, Great Wall, and also some of the global OEMs like Volkswagen, Toyota, Suzuki. So quite some local domestic and global players.

Nicolai Tangen: Who do you compete with?

Kai Yu: Oh, very good question. I think like Mobileye, you know, and also Huawei and quite a few startups, right? And also global players like Qualcomm, Nvidia, you know, in a sort of different scenarios.

Nicolai Tangen: Why are you the—why are you the best?

Kai Yu: Oh well a very good question. I think it's about the focus, you know. I think since the day one of the inception of Horizon Robotics, we believe computing for automotive requires a dedicated computing architecture. So we don't believe, you know, the architecture designed for smartphone or designed for PC or designed for cloud server will really naturally fit for automotive computing. So therefore we started to design the whole computing architecture from ground up for automotive intelligence.

Kai Yu: Right, and we had this vision which was quite unique. Now looking back 10 years ago in 2015 when we started the company, Mobileye had their chip architecture for traditional computer vision and they believe deep learning is useless, right? And Nvidia started their architecture for graphic processing unit for cloud computing, and Qualcomm for smartphone, and Intel for PC. So we are probably the single company in the world from day one having this belief that edge computing is a dedicated architecture, especially for automotive and robotics. So that's the reason why we are very unique.

Right and we had this vision is which was not a existing opportunity. So we envision in the future the world will be full of all kind of different robots you know working in a farming field, manufacturer assembly line and the logistic warehouse you know, every house home service and elderly care and also of course self-driving cars. So we envision there should be a unifying platform just like, you know, Windows and the Intel did for personal computing and just like, you know, ARM and Android what they are doing for mobile computing. So we envision in the future they're going to be such a kind of like computing platform which is generic for all kind of, you know, robotics working in different across different vertical domains.

Nicolai Tangen: When did you—when exactly did the vision hit you?

Kai Yu: Well, that's a very good question.

Nicolai Tangen: Were you in the shower or were you walking in the mountains or just what?

Kai Yu: Well, that's a very good question. I think exactly you are right. I think the vision actually hit me not I find this opportunity or this vision. Wow. Um I don't know you know, I have been working in machine learning for over 30 years.

You know since my early days as a junior student in college and by chance I read a book about machine learning and I was so amazed that it's so fantastic to develop a computer algorithm one day it can be like a human brain getting smarter and smarter over experience, yeah, by learning more and more from data. So then I decided to devote my whole career into this single subject machine learning, yeah. And that was for my thesis for my undergraduate and then my thesis for my master and also for my PhD and also for my first job back to my days in Siemens and later to Silicon Valley joined NEC Labs in Silicon Valley and then later joined the Baidu and then later founded the Horizon Robotics. So I changed my locations and I changed kind of different roles from you know individual researcher to a team leader to a head of research labs and later become a founder of a technology company. So I changed my roles but there's one thing you know remains never changed which is my mission and my vision to develop a machine learning algorithms to make machines getting smarter and smarter.

Nicolai Tangen: Two companies that we've had on podcast, XPeng and Xiaomi, they also make phones, cars, in addition to chips. So why did you decide never to build a car?

Kai Yu: I always position ourself going to be like a Microsoft and Intel for not for PC but for robots and I think a platform, a enabling platform I think it's a much more interesting positioning not just about like a business model but also about in terms of the impact you're going to make to the whole world, you know. For example, back to the days of a personal computing era, you know Windows and Intel they are not just, you know, help or enable like Lenovo or Dell, they actually help the entire mankind, you know, get into the new kind of a sort of a technology civilization or new area of a technology which is personal computer to enable all the white collar people to increase dramatically their productivity, right? What an amazing impact. So XPeng, they actually maybe help one brand, you know, to become very very appealing, very very attractive to consumers, but I want to enable and help all the brands, all the cars.

Nicolai Tangen: Is it difficult to build a car?

Kai Yu: Yeah of course.

Nicolai Tangen: So but you don't have—you have the technology, you have the chips. So if I ask you to build a car, how long time would it take you?

Kai Yu: You know, I wanted to achieve like a "Horizon Inside", you know, that's I think that's much bigger dream, much more interesting for me. I don't want to be a car maker, you know, and I want to help more car brands to to become more attractive.

Nicolai Tangen: Now the Chinese car market is kind of brutally competitive. What is driving the Chinese car market just now? Is it price, intelligence, just what is it?

Kai Yu: Well the current Chinese car market is really crazy. I think you don't see any market like here. There are you know I don't know how many, probably 60 or to 70 different car brands, and for the first half of this year more than 580 new car models launched. And it's extremely tough market, but I think the market is really driven by competition. This is but on the other hand is also driven by consumer demand. So consumers are really expecting, you know, more and more smart features, you know, automatic parking, autonomous driving and the more fancy infotainment systems. And I think all of those factors driving the market not just the competitive, but also very innovative, very innovative.

Nicolai Tangen: What what does it do to Chinese manufacturers when you have such strong competition at home that when you get out to let's say Europe is it just like easy peasy to compete relative?

Kai Yu: I think we position ourself not first we are not competing with any car makers. We are actually enabling them. We are helping them. We are building strong partnerships with all the car makers and all the ecosystem partners including the automotive suppliers for example Bosch, Denso and many other global and local suppliers. And our philosophy is really make friends and make a long-term trust with our customers and partners. Essentially, this is fundamentally in the gene of our company. So we don't have too in a sense of we don't have too much culture from the competition perspective but more from, you know, what kind of value we can create.

But luckily if you look at our business I mean fundamentally compared to like those car makers we are not so crowded, right? And if you look at China I think pretty much the three top players for the automotive computing I think namely nowadays I think really consolidated to Nvidia, Huawei and Horizon Robotics. Not very crowded.

Kai Yu: Yeah, partially because we choose the position where this business is having extremely high entry barrier, you know. It's not like okay a bunch of guys having VC funding and then easily get into the business, because this is automotive. Automotive means mission critical and it has to having the highest standard for the quality, for the safety, you know. And let's say we are doing is a chip design for automotive, right? It's not like a consumer type of silicon, and it has to go through all the very rigorous and tough qualification and testing process.

So normally if we initiate a new project, a new generation of product, so it will take three to four years to get the product, the chip and the solution ready for car makers. And then if luckily enough we won a new nomination from one car maker, and then we need to also go through the cycle of the production for the car platforms, and that's going to be another three to five years — in China two years, right, for global OEMs they are slower than typically five years. So then think about it's going to be at least from five years to eight years for the cycle.

So in some sense this is terrible business, you know, heavy investment, very slow return. But in my view sometimes terrible business means a great business, because that's the entry barrier for all the other newcomers is extremely high. And then I think a good business is something well you can sleep well overnight, and that's my opinion.

Nicolai Tangen: Talking about sleeping well and the autonomous journey. So you split between hands-off, eyes-off, mind-off. These are kind of the three steps that a car will go through. And in January last year, you put dates on them. So we'll have hands-off in three years, we'll have eyes-off in five, mind-off in 10. Just where are we today?

Kai Yu: Well, we are in a very good progress. I think you know what in the industry there's sort of like standard from like a Level 2, Level 3, Level 4, Level 5 is really from the engineering standpoint. For the consumers, you know, they have no clue what is Level 3, you know. So I always try to connect our technology and the product to something where my mom at least she can understand. So then that's my way to interpret the technology stages from you know hands-free hands-off, eyes-off and minds-off.

Nicolai Tangen: So what do you tell your mom just now?

Kai Yu: Right now we are in a very good progress to achieve fully hands-off. Yeah, I think then I said that I made that statement last year, right? So that means by end of 2026 we're going to achieve fully hands-off, and then by the end of 2030 to achieve fully eyes-off, and then by the end of 2035 to achieve fully mind-off.

Yeah, it's just like a new version, a smarter version of automatic transmission. You know, if you remember over the last 20 years so the whole car industry evolved from like the manual transmission and you need to control all the gears and later getting into the automatic transmission. So I think hands-off is just another version of that. But you still need to pay attention, right? You still need to be aware of what's happening around you and ready to take over. But already it's a huge relief.

Nicolai Tangen: What is still lacking? What part of the technology are you still—what's the critical element which is still left here?

Kai Yu: Well, I think we need a much bigger knowledge to test and we need to have a better numbers for statistics and then to prove that the technology is safe enough to claim, you know, eyes-off, right? So I think already we are in a very good progress.

Nicolai Tangen: Do you build your own AI models?

Kai Yu: Sure, sure, the self-driving model, yeah. You remember we position ourselves as a "Wintel", right? Not just the Intel, we position ourself as a Wintel. And I think Wintel is really an amazing business model.

Nicolai Tangen: And where is the kind of the future breakthrough going to come? Is it the model, the data, the chips?

Kai Yu: Well, it's a combination of three factors together. I think this is a typical formula of today's AI business. I think you need to have a much bigger compute running locally on the vehicle, and also you need to continuously collect data from the fleet, and continuously train your bigger and bigger models, and deploy the updated model back to the car. So it's compute, data, and the model architecture working together.

Nicolai Tangen: How many people are you now in the company?

Kai Yu: Around 2,000.

Nicolai Tangen: What do they do?

Kai Yu: I think we are very technology savvy company. So most of our employees they are engineers. They are either software engineers or hardware engineers and they are working extremely hard.

Nicolai Tangen: Because you don't produce the chips. You outsource the production, right?

Kai Yu: Exactly. For the hardware side, we are typical Fabless company. We design the chip but we don't manufacture the chip, outsource to our partners. And also we develop the software, the end-to-end full stack software.

Nicolai Tangen: What does working hard mean?

Kai Yu: You know we are kind of like a typical technology culture. We don't have the rules or the policy you need to come to work in 8:00 AM. Actually, typically most of our engineers, they come to work around 10:00 AM.

Nicolai Tangen: What do you think on average how long days do you think on average they work?

Kai Yu: Actually I have no clue, and quite some of our engineers they even working from home, right?

Yeah we don't measure that by like working hours. I do know some of the China companies measure by working hours. We don't, we never measure that. We just caring about, you know, how far we can achieve and what kind of innovation we can make.

Nicolai Tangen: When you say that the model time in China is, you know, the cycle time is two years versus 5 years in Europe. What explains the difference?

Kai Yu: As far as I know, because I work with so many Chinese OEMs, they are working extremely hard, they are innovating, they are not really afraid of making mistakes, and they are kind of continuously learning from users and continuously improving. So they are very fast.

Nicolai Tangen: What's the key to innovation? How do you create the really innovative culture?

Kai Yu: I think first of all we are a company driven by the vision, and we are not like opportunity-driven. From day one, we had this big dream: computing platform for generic robots. Many opportunistic companies just jump into whatever looks hot or profitable today. But we stay focused on our long-term mission and willing to take the long road. Real innovation comes from that kind of patience and deep commitment.

Nicolai Tangen: The fact that you have all these companies jumping into these sectors, what does it mean in terms of the supply chain and the various choices you have as a manufacturer? How does it all go together?

Kai Yu: For us we just choose the best to work with.

You know I work very closely with many global automotive companies like Bosch, Denso, Volkswagen. I talked to them about that. It turned out actually over the last maybe two decades or so, China became a country where the automotive supply chain has become incredibly complete, vibrant and agile. Not only domestic players are pushing hard, but global giants are also deepening their R&D integration in China. This collaborative ecosystem makes the whole industry move forward at unprecedented speed.

Nicolai Tangen: Your company is called Horizon Robotics. How far is a car from a robot now?

Kai Yu: Still a little bit long way to go, but not that far. I think probably as I said, in two years by the end of 2026, the passenger vehicles in China will be largely hands-free. So it's already kind of a semi-robot, right?

Nicolai Tangen: Right. So it's already kind of a semi robots, right? And what is your—what is your business going to be in the robotics business?

Kai Yu: Actually, we are already a leading player in China right now, one of the top two players for generic robotics computing. We are enabling a lot of robotics companies.

Yeah, that's a very good question and I believe eventually the two business technologies will really converge and eventually become a unifying computing platform, right? So cars will become just one type of robot with wheels, while other forms of robots share the same underlying brain.

Nicolai Tangen: Tell me about the speed of development of humanoid robots in China now.

Kai Yu: Well you know what, it's crazy again! There are like hundreds of such companies. That's so intense! But I think all of them are still really a long way to go. If you look at the automotive domain, it's taken years of rigorous validation. Humanoid robots are even more challenging.

Nicolai Tangen: What do you do inside the robot? What is your—what does your software do?

Kai Yu: We supply chip solutions. Yeah, and for robots, we are not so much into the software. We are supplying chips to the robot companies.

Nicolai Tangen: When do you think we'll have meaningful volumes of humanoid robots coming?

Kai Yu: Five years if optimistic? Yeah, I know I'm a little bit conservative compared to many other guys, and I try to be objective.

Nicolai Tangen: What will be the number one scenario?

Kai Yu: Well, I think the number one scenario is home service. For each family, we enjoy all the fun things, but we don't enjoy many of the tedious works.

Nicolai Tangen: I have a theory that divorce rates will go down when you don't have to do boring things anymore. You know, taking out the rubbish or walking the dog.

Kai Yu: Exactly! There's another thing which is, you know, China is facing very terrible this aging society issue. I mean it's coming so quick, quite unexpectedly. The issue is getting worse actually compared to Europe and even Japan. So what about the elderly care? What about the lacking of labors?

Humanoid robotics is not just a technological gimmick; it is an essential necessity for society to address the elderly care and labor shortage challenge.

Nicolai Tangen: So if you win the EV race, you'll win the robotics race.

Kai Yu: Yeah, that is fundamentally what I believe. My argument is if you are a winner for automotive supply chain, you're going to be a winner for robotics. If you look at the fundamental components of a humanoid robot — battery packs, electric motors, sensors and cameras, thermal management, and the high-performance computing brain — all of them are nurtured and scaled by the smart EV industry.

In China today, almost all the major OEMs and tier-1 suppliers are naturally extending their capabilities from automotive to robotics. Robotics is the natural beneficiary of this immense supply chain.

Nicolai Tangen: Where does that leave the race between China and the US in robotics?

Kai Yu: You know what, I don't like this kind of view to look at this game. It's not a zero-sum game. I'm not sure this is the right framework. Looking at history, from PCs to smartphones to EVs, real technological progress always thrived on open collaboration, global supply chains, and shared scientific knowledge.

We shouldn't view technology through the lens of zero-sum competition. Global collaboration is the best way to move human civilization forward.

Nicolai Tangen: So what is a city going to look like in 10 to 20 years time?

Kai Yu: You know, I'm a huge fan of the ancient Greek culture and history. In July, I've been to Athens, walking among those incredible monuments of human civilization.

When you look at modern cities today, people are constantly stressed, rushing, trapped in heavy traffic and endless physical chores. In 10 to 20 years, with full autonomous driving and pervasive robotics, the physical city will be transformed. Tedious transportation and physical labor will be handled by autonomous systems. Humans will finally be liberated from mundane toil.

Nicolai Tangen: And what's the town going to look like physically? What's the city going to look like physically?

Kai Yu: The environment will be much nicer, you know, with much less noise, much less stress, and much less anxiety. You won't hear honking horns and roaring engines. Traffic will flow smoothly. Vast parking lots can be reclaimed as green parks, walking plazas and cultural spaces. The city will become far more livable and truly centered around human well-being.

Nicolai Tangen: Take me back one second to your Athens visit. When you're in Athens, what do you do? What do you do there?

Kai Yu: Museum, yeah. And also I visit all the islands.

Nicolai Tangen: So you go up to Acropolis and you stand there at the Parthenon temple, right? What do you think when you're there?

Kai Yu: I was stunned. I was so amazed. I think culture means everything. Human thinking is unbounded. Essentially, humans can achieve unimaginable wonders when driven by spirit and intellect. Over two thousand years ago, ancient philosophers laid the foundations of science and philosophy. Today, our pursuit of AI and robotics is just another chapter of humanity exploring the boundaries of mind and seeking spiritual freedom. Technology is the instrument, but human culture is the ultimate soul.

Nicolai Tangen: Personal question. What's your biggest personal failure in life?

Kai Yu: Personal failure, to be honest, I'm not a very pessimistic guy. If I ever encountered any frustration or failure, I always turn into the new page and look forward. I never dwell on past regrets.

Nicolai Tangen: What's the key to the Zen philosophy for you?

Kai Yu: You know, the happiness is so simple. According to one of the great monks in the history of Zen, he said actually many people have a miserable life just because when they are eating, they are not eating — they are thinking about thousands of other things; and when they are sleeping, they are not sleeping — they are troubled by anxieties and fears.

Zen tells us: when you are hungry, just eat; when you are tired, just sleep. Live purely in the present moment. Be completely focused on whatever you are doing right now. That simplicity brings profound inner peace and strength.

Nicolai Tangen: What's something about you that most people don't know?

Kai Yu: Oh there's one thing I'm not very satisfied about myself is about my weight! I'm just having too good appetite. I love food so much! So this is really something not so disciplined.

Nicolai Tangen: What do you read?

Kai Yu: Well, recently I'm reading one book telling about the story about three companies: Microsoft, Intel, and Apple. It's about all the critical strategic decisions and historical turning points they went through. Highly inspiring.

Nicolai Tangen: What's your advice to young people?

Kai Yu: You know, doing what you love and loving what you're doing probably are equally important. Many people always dream to do what they love. But in reality, you are not always so lucky to start with your ultimate dream.

If you can learn to love what you are currently doing, dedicate your full heart to it, master the craft, and find joy and beauty in the process, you will excel and eventually build a deeply fulfilling life.

Nicolai Tangen: Well, Kai, I think that's a wonderful place to end. Big thank you! What an incredible story, what an incredible company! Love your life philosophy. There is so much to take away from this conversation.

Kai Yu: Thank you, Nicolai.

Nicolai Tangen: Really hoped you like the episode. If you did, please like it, subscribe, and put on the notification.