China EVs & More

MAX Episode #9 - Dr. Jun Pei, Cepton Co-founder/CEO

October 12, 2022 Tu Le & Lei Xing
China EVs & More
MAX Episode #9 - Dr. Jun Pei, Cepton Co-founder/CEO
Show Notes Transcript

In this latest MAX episode, Tu & Lei welcome Dr. Jun Pei, Co-founder/CEO of Cepton, one of the world's leading LiDAR companies. Jun talks about the personal journey that led him to co-found Cepton, how they were able to get that design win with GM and how Cepton's LiDAR technology is different from their competitors. 
 
**We apologize for some of the audio issues with portions of this recording.**

CEM MAX #9 Dr. Jun Pei, Founder & CEO of Cepton
Recorded on 9/15/22


Tu Le:
Hi everyone, Tu Le here, one-half of the China EVs & More duo. Lei and I have been brainstorming about different ways to bring you, our audience, relevant and compelling content about the China EV, AV and mobility sectors. Especially now that a number of companies that we’ve tracked over the 80+ China EVs & More episodes have become global phenomenon. 

 So I’m very pleased to introduce our latest China EVs & More MAX episode, where we bring you conversations we’ve had with special guests from the EV, AV and mobility space. 

 In this episode, we speak with Dr. Jun Pei, Co-Founder & CEO of Cepton, one of the world’s leading LiDAR companies. Cepton revealed in late 2021 when it filed its papers for its SPAC merger that starting in 2023, it’s going to supply up to nine GM models with its LiDAR technology. With ADAS features becoming ubiquitous now on even mass market low-priced vehicles, it’s no surprise that some research firms are forecasting a compound annual growth rate of 19% over the next several years, growing the LiDAR space to an almost $7 billion market by 2030. The future is truly bright for Cepton, here’s our sit down with Dr. Jun Pei. Apologies for some of our audio issues.  

Tu Le:
Jun, thank you for having us. And I know Cepton is on the verge of gaining some traction with some clients, and meeting you in person, it's a pleasure, so thank you. Can you please tell us a little bit about yourself and your company Cepton.

Dr. Jun Pei:
Thank you for coming to our lovely Detroit office. In fact, this is my first time to be in this office space. We used to have somewhat a smaller office in the past a few years, and we just had a significant expansion. You see, this lovely place. In fact if if you look at all the LiDAR companies out there, I believe we have the biggest presence in Detroit, which is the car place in North America. In fact, I was actually having a few meetings yesterday and with obviously very notable OEM customers and they were all commenting on this support infrastructure that we have in the Detroit office. This speaks volumes for Cepton, which is, we started as a LiDAR company 6 years ago, was the main focus on making LiDARs for the ADAS industry, which is very different in those days. If you recall, 6 years ago, there was a huge bandwagon of autonomous vehicles, that fully autonomous vehicles will be everywhere in a couple years, 2018, 2019, right around the corner. Maybe I’m too old. I was just saying that I came from China many many, 30 some years ago. I was, my name's Jun, Jun in Chinese, that's very much of a cultural revolution name.

I was born in the 60s. So that gives you a little history of my background with the age and with my 30 years of experience as an engineer. I didn't believe at the time when we started the company that the fully autonomous vehicles will be any time soon. So I actually still believed then and now that LiDAR will actually become a safety device first for vehicles and that the application will be championed in the ADAS industry. That's where we set up the company to do, the focus. The thesis of Cepton itself is actually to do a LiDAR for everyday vehicles, for consumer vehicles. Actually, I want to correct your pronunciation of Cepton and it's actually not too much of a correction, just explanation of what Cepton is. At the time when we started the company, we needed to come up with the name. And the purpose for us is to come up with something that will perceive the world with a 3D fidelity. And that's part of perception. “Cep” is part of the word “perception.” And “on” is this Latin ending like electron, photon, atom, an element. So Cepton really implies an element of perception. That's what the company name is. So that's how we started that we want to create an element of perception for the ADAS industry, and that his LiDAR, that's what we focused on working on, and put all our money behind this effort and eventually capturing the Ultra Cruise program from General Motors. That’s our flagship program. We have publicized that and it will be deployed next year with. 

Tu Le:
That's really cool. 

Dr. Jun Pei:
With Cepon LiDAR inside. Kind of in a nutshell, I grew up in China, came to the States in 1989, which is a very notable year, for many reasons. Came here as an undergraduate, in fact, that studying physics and computer science. Eventually I graduated from Brandeis in the Boston area, but really hated the weather here, I’m afraid that California has this attraction to me, so I went to Stanford for graduate school, got my Ph.D. there in optical & electrical engineering. That's where I call home Silicon Valley now, there's a lot of Stanford lineage in our company. My co- founder Mark McCord actually when I was a student at Stanford, he was a professor at Stanford. I was in a way student and later on, we worked, I worked for him, he was my boss at a semiconductor company (KLA-Tencor) working in the electron optics area. One way or the other, after many years, we converged and started Cepton together 6 years ago. Kind of in a nutshell, very quickly recapping a lot of pieces of history.

Tu Le:
The weather, the time I spent in Silicon Valley, the weather is perfect, 300 days a year, it's between 60°  and 90° and yeah, so I missed it a little bit too. 

Dr. Jun Pei:
There's the weather, but I, this is my very much personal feeling that there are only two places in the States that I call melting pot, and that's New York City and Silicon Valley, where I, obviously I grew up in China. I speak with a very different accent than many other of my colleagues and friends. But whether you're from China, you're from Japan, you're from India, you're from Europe. Nobody, when we work together, when we see each other, whether it's in a working place or outside, nobody considers you as a foreigner. You are a member of Silicon Valley, same thing in New York City, I feel the same thing.

Lei Xing:
Tell him about the American pie, right? 

Tu Le:
So, very quickly. My father was in the South Vietnamese Airforce, so we were sponsored by a family in Michigan. I'm the youngest of eight, and so I came over when I was pretty young. I'm old now. But and so I didn't become a U.S. citizen until I turned 19. That was right in Detroit at the Federal Building. And so I did my pledge and the judge was like, now you're as American as apple pie. He said, don't let anyone tell you any different. And I think that's really important for immigrants to really understand that it doesn't matter where you were born and if you're American, you're American. And it's just because that, anyways, but so I thought that was pretty important for me to hear, at that time.

Dr. Jun Pei:
Yeah, it's actually a rather touching point as a company founder, as a business leader in the industry. For me, personally, I felt this obligation to support the immigrants, to support the spirit of people coming to the U.S., coming to Silicon Valley, especially for technical people to realize their dreams. So that's actually continuing, if you look at our company makeup, but is very supportive for high tech immigrants, especially. And that goes back to my Stanford days. You were just mentioning, actually, my professor who actually assigned my Ph.D. thesis. He was originally from Lebanon. He’s already a retired professor at Stanford. It's a multicultural place that fosters growth, both in technical and culturally. I think Silicon Valley is this wonderful place to develop technologies. And for me, after I graduated, I spent a decade in a semiconductor company, equipment company to do electron optics, that those are the years I get sharpened my skills are both in technology as well as management. Then as many, many graduates from Stanford almost everybody, especially electrical engineering, had the spirit of starting something of your own.

Lei Xing:
So Cepton maybe it's decades in the making. If you think about your past, right? Because you have to build up that understanding and experience, and then you figure out what to do with it, right?

Dr. Jun Pei:
Actually, yes, on the other hand, it's also everybody's path becoming an entrepreneur is slightly different. For me, I was very much down to earth in terms of working for a company, trying to grow my career path, climbing ladders and what not. But then when I realized that I can explore something else, I actually started by exploring small companies first, started a small company that actually specialized in optical instrument. Actually Cepton is my second company, so I had a first company that ran for quite a number of years and reasonably successfully, made some money for my family with a very small skill. It's a mom and pop shop. There was no investment, no VC involvement at all. It was just me and my wife.

Tu Le:
Sometimes that's better because the pressure from external money is…

Dr. Jun Pei:
It's a different type of pressure. You're basically working with your own family money. Obviously, you don't have 9 to 5 mentality anymore. Every day is a working day, every hour, every waking hour is a working hour. So it's a different pressure. You only need to answer to yourself, which is somewhat, it's an interesting time. So we actually did quite well for 4-5 years and accumulated some money, family money, that actually enabled the me and my partners and co-founders to cobble up the initial funding for Cepton. So the very first batch of a small amount of money came from friends and family started six some years ago. And then we actually realized, the LiDAR industry is going to be big. So let's take in some real serious money and expand the team. So we grew. I still remember the day we had our first meeting of Cepton was four people. Now, we're actually moving towards 200 people now. So it's with presence obviously in Silicon Valley as the headquarter, but in Michigan here, in Germany, in India, in China, in Japan. 

Tu Le:
Truly global.

Dr. Jun Pei:
Yeah it’s a global presence now. Actually our biggest office outside of Silicon Valley is here in Detroit. So it's quite a journey, although if you consider six years until today, we're a public company now, listed in NASDAQ, we actually captured a very big program from GM that is sourced for practically the next decade. And we are actually sole-sourced for that program. All of these seemingly is very fast for six years, but six years is actually also pretty long period by Silicon Valley standards.

Lei Xing:
Let's get into LiDARs. How would you describe the current state of the LiDAR sector? And it's forecasted to grow substantially toward the latter half of the decade, right, and we already know competitors, they are also in different programs from different global OEMs. And when do you think LiDAR will become ubiquitous technology on a smart EV and at what price point?

Dr. Jun Pei:
Let me basically talk about my view of the current state of the LiDAR industry first, that's kind of your first part of the question. I think I carry more or less of an insider view of this entire business sector. There's a history. I actually spent quite a few years of working as a VP of engineering at Velodyne a dozen years ago. That's my early career was LiDAR.

Tu Le:
I read that you were part of the DARPA Challenge team? Tell me more about that. Because I’m a Carnegie Mellon grad, for business school. So we learned about that, Red Whittaker and those guys. 

Dr. Jun Pei:
Yes I was very much.. Actually, it was interesting story that I was working in semiconductor industry, trying to explore some other interesting career paths. One of my hobby interests was looking at the DARPA program and how they competed. And then through grape vines I got to know Mr. David Hall, the founder of Velodyne, because he was actually with his brother Bruce Hall playing around with the LiDAR for that program. And they had a Velodyne truck. And somehow we got together and talked about, and basically I end up working for him. And I actually had an acoustic background from school. When I started working for Velodyne, a major part of my responsibility was actually working on loudspeakers. If you didn't know, Velodyne for many years was a loudspeaker company.

Lei Xing:
I did know that.

Dr. Jun Pei:
So my engineering effort was a lot of it was on acoustics. Then LiDAR started more or less as a hobby project. And then eventually, I was involved in developing the HDL-64, the big…

Lei Xing:
The bucket, right?

Dr. Jun Pei:
Yeah, as we call it the big spinning doorknob on top of cars. Very popular product that was surprised how long lasting that product has been. Dozen years of continuous production of that thing David basically championed, and I worked for him for that process, was quite a history, but then I actually left Velodyne many years ago to start my own optical company. That's my first company and ran on that. Only 2016, I came back to this industry, starting Cepton in a way actually competing with Velodyne on many fronts. 

But the LiDAR industry coming back to our topic, has morphed into, I consider three phases. There will be a 4th phase that I can also envision and talk about it. The first phase. It's actually around 2015, 2016 when we started the company. And there's just bunch and bunch of LiDAR companies popping up with all the investment into this industry. That's when I call it, is a technology competition phase that everybody is trying to come up with a new technology and saying that we're the best. We're no different. We came up with something called MMT (micro motion technology) that I’m sure we're going to talk about it. And then we all kind of said in front of the investors saying this is going to be big. Please give us money or develop this technology, and so on, so forth. So that's kind of a massive, scattered technology competition phase. But what is a win for a technology? It's not me pumping my chest saying we have the best nuts and bolts or best optics, because nobody really understand what that is. The only validation for a technology is a commercial design win. It's a commercial validation. So that's where people started to enter into the second phase. They were about 60, 70 LiDAR companies by 2017, 2018ish, those two years just popping up really like crazy. 

Tu Le:
With Velodyne really being a leader back then.

Dr. Jun Pei:
 At the time. Yes. The second phase is really started from 2018, 2019, especially highlighted by the end of 2019 where Cepton was awarded the GM contract. This is you suddenly go from 70 LiDAR companies down to a handful, how many companies would actually have a design win? That would be Innoviz, for example, with BMW at the time. And then Luminar with Volvo and then Cepton with GM. So of course GM has the biggest conglomerate of car models and car platforms that covers, basically it's an Ultra Cruise platform that for existing many cars, so we enjoyed that quite a bit. So that got us into at least a small part of companies in the second phase, I call this commercial validation, commercial design win, validation of technology. But after that, I think it's still kind of morphing. There's still some design wins being awarded here and there. The big ones are very rare and far in between. We entered into a third phase. You had your technology. We have entered into the third phase, we had the technology competition phase, we had the commercial validation phase. Now, the third phase is what we call execution phase. You have your design win. Like for us, we had the design win awarded to us in December 2019. What is it now? Almost 3 years? So with all these time past, is the execution, having a good technology, a commercial design win award, we need to actually finally make this product, so that it actually exist on cars. They can roll out of production line. That is certainly not a trivial process, we've gone through the various sample stage. There's A sample, B sample, etc., then we're actually way into the product intense samples. I just came back from Japan a few weeks ago. You probably are aware…

Lei Xing:
There was a conference, right? Or were you speaking?

Dr. Jun Pei:
You're probably aware that our design win with General Motors is with a Japanese tier-one company called Koito Manufacturing. So it's actually a huge backbone for us, Koito Manufacturing partnered with us to capture the GM design win. So we are actually the tier-two company supplying the tier-one company Koito, and Koito as a tier-one company supplying General Motors.

Tu Le:
That was a direct buy from GM?

Dr. Jun Pei:
Correct. This whole structure is very classical in the automotive industry. I do have to mention a few words about why it needs to be a tier-one, why can’t Cepton by itself supply to General Motors. We just make some parts from our contract manufacturers and assemble them in Silicon Valley, and maybe ship to GM, Dearborn, or other places. Why can’t we do that? Apart from all the other logistics and production matters, the simple reason for the liability of a safety product, you just cannot be a small company supplying a major OEM.

Lei Xing:
But the Innoviz Volkswagen relationship is a tier-one. That's interesting, at least they claim to be a tier-one.

Dr. Jun Pei:
That is a different model. You may actually look at their previous model with BMW through Magna, that seemingly have disappeared from the radar screen. And they are becoming tier-one themselves. So it's actually in this industry it’s very challenging to work with, more challenging to work with tear-ones than with OEMs, I have to say, because tier-ones are finally responsible for the liability of the product, for the cost, and for basically overall delivery. So that's a challenging process. Coming back to the execution phase, as I have already mentioned, entire after you have the design win to make it into a car production line is a 3-4 year process. Not to mention all the details in terms of component procuring, you know given all the challenges we have, so having the tier-one being at the frontline managing all this. And Koito has been a supplier for GM for many decades. And that actually established a path for success. And then you still have to work through this validation of your final product from your supply chain, all the way to technical viability. These are just kind of a wheel turning for a good part 3-4 years. Then you have a solid product coming out. The vehicles that we're aiming for is rolling out next year. You probably heard models like Celestiq. So these are the execution phases. My career is really interesting when I graduated from Stanford, sorry, to digress for a little bit, I was given a title: senior research scientist, how glorious is that? It's just so fantastic. I was so in love with my business card. And then after a couple of years in the semiconductor equipment industry, I became a system engineer. And then I realized it's much harder to be an engineer than a research scientist. Because one is actually you publish a paper, the other is, you make a product.

Lei Xing:
It's like a downgrade, right? 

Dr. Jun Pei:
And then another couple of years, guess what, I became a manufacturing engineer. It looks like my career went downhill, but god, it was so much more difficult to be a manufacturing engineer to actually make the delivery. So the analogy is, there is very much that I realize having the technology, invent something, patent it at the early phase of the LiDAR industry, or specifically for Cepton, it's not that difficult. You just write things and you publish and you file your patent,  and then you have a design win. Ok, you need to get into some fighting mode to actually compete with your competitors and really come on top. We did that. That was pretty hard. What's harder is actually the execution phase, is the manufacturing of these things that seemingly my career is going downhill again, but it's not, is really the delivery. So these three phases, I call the technology competition, commercial validation, and execution, one is harder than the other

Tu Le:
Say what you will about Elon, love him, hate him. No one really appreciated how difficult mass production or mass was or is, until he said building one prototype is easy, building the same product, thousands of times with the same quality and reliability is really, really difficult. And so that's where I don't think Silicon Valley appreciates the legacies very well, because in Silicon Valley you contract manufacture that out. So you don't actually get your hands dirty generally speaking, right? The contract manufacturer manages most of that stuff for you, and so, totally agree.

Dr. Jun Pei:
So there's a saying in our smaller circle that the collaboration or the partnership between Cepton, which is a Silicon Valley company of six years old, with Koito Manufacturing of Japan, a tier-one company in Japan of 110-year-old, is a very old company marrying a very young company. Actually, this kind of collaboration is indeed a marriage of two fronts. One is the frontier in technology development. The other is actually a solid anchor in manufacturing, knowhow and capability and all the financial backings to make this program successful. So this is where we shine is these three phases. If you look at the other equivalent design wins whether it's the Volvos, BMWs, BMW announced their design win in 2017. What is it? 2022 now? And I still don't see a BMW car with a LiDAR out there. They will come out, but it's much longer than the three years or three or four years we have gone through. So the execution phase is much harder, it's much longer. We certainly feel very confident, comfortable almost at this point seeing as I went to Japan a few weeks back, and actually saw what they called the Cepton production line in Koito Manufacturing. And this is in Shizuoka Prefecture in Japan, South of Tokyo. And you'll see huge production line for our LiDAR just churning all the parts.

Tu Le:
So they are your CM? 

Dr. Jun Pei:
It's not CM, it's actually, CM is a contract manufacturer. They are the tier-one. They are actually our customer. Contract manufacturer would be our vendor. We actually do use contract manufacturers. So we had some announcement earlier on that we're using Fabrinet from Thailand to make some components as well as sub-modules, some assemblies. Cepton would put our stamp on them, whether it's software, or some knowhow in the assembly, and then we provide these over to Koito. Koito is the final assembler. They are responsible for putting these parts together and get them qualified, and vouch for reliability and the liability overall and deliver to GM. So initially, all the LiDARs for the Ultra Cruise system will be manufactured in Japan.

Lei Xing:
So it's like an assembly line with your Cepton part. 

Dr. Jun Pei:
Correct.

Lei Xing:
So like a normal supply, tier-one, two relationship.

Dr. Jun Pei:
Like a normal tier system. So this is where I think if you look at the industry right now, we probably enjoy the closest relationship with a tier-one, a traditional tier-one company. We certainly didn't venture with all these risks and this complexity. We just talked about the manufacturing. We didn't think it's best use of our investment money to become a tier-one manufacturing company. 

Lei Xing:
So the Cepton LiDAR itself will be manufactured at the Koito…

Dr. Jun Pei:
In Shizuoka factory in Japan. It will become a Koito product with Cepton’s parts and technology, obviously. And then eventually delivered to GM worldwide.

Tu Le:
So they are licensing your technology.

Dr. Jun Pei:
More than just licensing, not only the license. They actually have to purchase critical components and sub-assemblies from us so we are the tier-two, a very good example is that our LiDAR has the MMT technology. All of these things are that we publicized. But what we said a little bit less about is there's this critical component, ASIC chip that actually is the soul of our LiDAR. It captures all the signal categories and coordinate the structure and eventually pump out the point cloud, is all because of this little chip. Obviously, that chip is a Cepton product, and we sell it to Koito as part of the assembly that happens over there.

Tu Le:
So you guys are chip designers to them.

Dr. Jun Pei:
We are very much of a chip designer, actually, or you can say the chip designer nowadays is more of an intelligence designer. You fuse all of your knowhow into software, firmware, and eventually, it becomes a chip that can carry out these tasks in an intelligent way. We certainly don't fabricate the chip ourselves. So we have a foundry that eventually realize our design and the chip that I talked a bit about this chip was really fascinating that has all these functions of analog front end all the way to, we have ARMS core processors at the back end, when these things all come together, it is a sesame-sized chip, a millimeter by 2 millimeters, that cost a little over $1 apiece. This is where we actually can capture the GM win, we can have a price advantage. This whole LiDAR design thing, we will dive into that, I'm sure. It's all about how you balance things. If you don’t have, the best image, the best technology the shuttle, there is no use for cars if it costs more than a few hundred dollars.

Lei Xing:
I think that was also Omer’s point as well. You can’t have the best or the perfect product right?

Tu Le:
We also interviewed Maxwell Zhou, who is the founder & CEO of DeepRoute. And he has an entire hardware/software stack, he was telling us the solid-state LiDARs. I think he put three of them. And they are probably less than $800 each. They're trying to squeeze that down even further to make their hardware software stack affordable to many OEMs, so that it becomes standard, eventually, on smart, even the mass market priced EVs.

Lei Xing:
Which is the next part of the question that I asked, that the 4th phase, the commercialization, the volumes, the ubiquity. How long is that going to take? What's your prediction? 

Dr. Jun Pei:
My view on, you know we talked about the first phase, second phase, third phase, which is the execution phase, is still ongoing and they're only going to be one or two companies really that can be successful in the execution phase. So there's another phase. I simplify it, you talk about ubiquitous or what price range and all that. The 4th and I feel is the final phase of a LiDAR program is you got to make money. How about that? I just simply put it. At the end of the day, you know if you're not making money, it's not successful. 

Tu Le:
It’s your science project.

Dr. Jun Pei:
Yeah after all the hardship you’ve gone through, the initial technology invention, capturing design win, execute the program, you still have to make sure that things are balanced out that you have some profit to be made at the end. That will be proven to be even more challenging. It's not only unique to LiDAR industry, even the car OEMs, at the end of the day, they need to make money. So, how much your LiDAR cost? Is there? I can tell you today, if you have over $1,000 LiDAR, you have no market. So can you actually get the reasonable, say, also with $500 LiDAR? The answer is yes. But then the challenges are how are you going to make, how much money to get to that point. Of course, everybody has a dream of $100 LiDAR. So that actually is even more challenging in terms of how to draw profit out of that. 

Lei Xing:
I remember two years ago or in 2020 the CES that I went, the former CEO of Velodyne, he held up a $100 LiDAR. I thought that was one of those moments, right? But what happened afterwards, you know.

Dr. Jun Pei:
Well Velodyne story is a completely different topic that I’m not going to talk. 

Lei Xing:
But it just makes think to the…

Dr. Jun Pei:
I’m still very much a friend of David Hall, but you know.

Tu Le:
There's a lot of drama.

Lei Xing:
But that's another story. 

Dr. Jun Pei:
So it is interesting industry I’m sure you're familiar with China EV industry…

Lei Xing:
Sure, that's why we call China EVs & More.

Tu Le:
That's our sweet spot. We hear a lot of things. But I think what's important is you're the first person, Jun, that has really talked about how LiDARs are for ADAS. I think most LiDAR companies try to just glaze over the ADAS bridge to Level 3/Level 4 autonomy. And so I feel the longer that runway is, the better chance of Cepton has to be very profitable. Because we're still 15 years away from complete Level 4, point to point, at least 15 years.

Dr. Jun Pei:
Why do you say that? Actually, let's spend a few minutes of that 15 years. What makes you think it’s going to be 15 years? I don't disagree. I'll tell you my reason in just a little bit, but I’d like to hear yours.

Tu Le:
Well, for me, I won't say it's a random amount of time. But there's regulations, and it's more about, you're saying everybody two years ago, three years ago said it's right around the corner. And how I follow the sector. There's many different use cases and the ultimate is the robotaxi. But that also means there needs to be plenty of data for every single edge case and use case. And as we're still looking at the pilots in cities like Arizona, where and we live in Michigan, right? And so there are going to be many edge cases that still need a ton of data that aren't even being looked at right now.

Dr. Jun Pei:
I agree with you 100% there. But I’ll elaborate on the data part which at least to my reason of coming up with more or less the same, my time frame is 16 years.

Lei Xing:
So that's very specific!

Tu Le:
You are an engineer, I’m not an engineer.

Dr. Jun Pei:
On the reason of 16 years, and you'll see there's a logic to it. And I started with a fundamental hypothesis, which I’m sure you agree that when there’s a Level 4 or Level 5 vehicle out there, the fundamental requirement is actually this vehicle is safer than a human driver. So this is a very much obvious to everyone that if it's less safe than you drive yourself, then you actually drive yourself. So everybody agrees with this hypothesis. So I start from there, is the safety level of a fully autonomous vehicle has to be higher than a human driver, then you can start look at the data. You look at 2015 or maybe even a little earlier. The human safety level is here, your autonomous system is actually down there. 2016, 2017, every year you can get a little bit data and I did this exercise and look at oh, it's actually the trajectory is getting better and better is aiming towards that threshold that is actually going to exceed someday, that will become a better driver than human. Then you plot this line. You extrapolate where that is, that's 16 years. 

Tu Le:
I didn't do that.

Dr. Jun Pei:
So I think that's the logic I have, and you can, extrapolation is very dangerous. So you can have errors, but it's not going to be tomorrow, I can say that.

Lei Xing:
I think one example I wanted to throw out there is if you look at the example of the Xpeng P5 that was announced as one of the first vehicles with the Livox LiDAR, right? They had the city- NGP function announced in April 2021, and that still hasn't become available. They're still doing testing. And right, it's going to be probably around 2 years after they announced that this thing will actually, what they call the all-scenario door to door, but it's still L2, right? Because you have to pay attention. That's still not available.

Dr. Jun Pei:
So let me tell you about the other side of this perspective that is Cepton, as I have mentioned, at the very beginning of this conversation, we started with the aim of delivering LiDAR as a safety device for the ADAS industry. Obviously, our design win, General Motors’ Ultra Cruise is an ADAS function, it's not Level 3, it's Level 2+.

Lei Xing:
I think it's similar to the all-scenario…

Dr. Jun Pei:
So what's the aspiration there? I’ll give you one example from a very notable OEM, just don't want to disclose the name, that they have the aspiration of the next 5-10 years, they didn't want to accomplish the goal of point to point transportation, they want to accomplish the goal of reducing the accident by 90%.

Lei Xing:
Very specific again, right?

Dr. Jun Pei:
So isn't that a nice goal to have? If you can reduce the accident by 90%, that's one out of 10 times from the current status. That's a great thing. So with that, they did a study of what we need, what kind of sensors or software and other stacks that you need to do. Obviously, you need cameras, you need the radar, and they studied to death, I think it's related to one of your questions: LiDAR or no LiDAR. So they studied in-depth of these functions and realized with camera and radar, you can only reach 50% of reduction in accident, right? The rest of the 40%, you have to fuse it with something else. And it's just like you're, as a human, your sensor, you have nose, you have eyes, you have ears, these are all your sensory system. Can you…

Tu Le:
Redundancy.

Dr. Jun Pei:
Yeah, can you live without your eyes only your ears? You can, to a certain extent. You want the whole sensor to be functional, to get to the 90% safety range. and that's where LiDAR kicks in, is to get seemingly a trivial thing, like the camera and the radar will take the majority of the thing, but it's the minority that get you to the safety level, that gets you the adoption of such a system. So as you said, it's the edge cases. Those are the most important cases, not the everyday…

Tu Le:
And I think, the other part of the edge cases in the automotive space is that I always look at it as an upside down normal distribution curve, whereas the edge cases have the highest potential for fatalities or serious injuries, because driving in snow, not everybody does it, but guess what, if you lose control, you can hit 3, 4, or 5 cars and it could be potentially be fatal. And so I think that's an important distinction where most product managers, when they look at edge cases for the iPhone or something like that, it's a nice to have feature that you accommodate this edge case, but it's not going to endanger anyone, right? So I think many folks don't have an appreciation for this edge case is actually very, very serious and it needs to be taken seriously before we say, there's a point to point Level 4 robotaxi that you can just get in, order it any time you want and it will drop you off anywhere, any time in any weather or…

Dr. Jun Pei:
So looking back to the original question, I think LiDAR will be ubiquitous, it will be there for the future vehicles, there is no doubt about that. Not so much for the grand aspiration of fully autonomous systems, just as an additional safety device that makes your driving experience much more pleasant and much more secure, that device will be there. It will be ubiquitous.

Tu Le:
It's funny that there is one automotive manufacturer that doesn't think they need LiDAR.

Lei Xing:
This is another debate.

Tu Le:
Yeah that’s another conversation to have.

Dr. Jun Pei:
I think there are many reasons for Tesla to speak, discredit LiDAR more than one occasion. I'll just illustrate one. I think there's a lot of non-technical reasons for it. I got my Tesla, the first batch in 2013. It was not, it was the very first batch and it was good, it was a love and hate relationship.

Tu Le:
When the S came out, right?

Dr. Jun Pei:
Right after it came out. But that love and hate relationship is a different story, but 2015 is the year I realized Tesla is going to do full self driving, when I get the pamphlet from Tesla, as a Tesla owner, and I was told if you buy a Tesla, 2015 onwards, all the hardware are there. You just need in the future upgrade to have fully self driving. 

Tu Le:
That will charge you for it.

Dr. Jun Pei:
Yeah, sure. But all the hardware, I don't have to bring the car back to the dealer and install a new device to make it self driving. All the hardware is there. So that was actually edged in stone. So I’m afraid LiDAR was not there as part of the hardware, right? So it's a little bit boxed in for that purpose. I think there is a scenario that when our LiDAR comes out with GM Ultra Cruise with this large volume deployment, and it cost only a few hundred dollars apiece, it actually can reduce the accident rate by 80-90%. Would you rather pay an extra few hundred dollars for it? 

Tu Le:
Sure.

Dr. Jun Pei:
I think that question is actually is not even a question. So.

Tu Le:
I think that's the actual head scratcher, why one person or one company can just say you know what, we changed our opinion on this. But that would also mean that they either need to bring in these vehicles and get them upgraded or change their entire process for autonomous driving.

Dr. Jun Pei:
I actually wouldn't be surprised if one day that is the reality: hey, we have change, a change of thought process, because LiDAR is so prevalent now, we're going to adopt it as well. So I think I wouldn't be surprised if one day that's happened. One of the privileges I realized that being a founder and the CEO of a company is you're free to admit your mistakes. This is the last worry I have is, to claim that doing something wrong, and let's change it, because nobody is going to fire me, because I admit my mistake. So yeah, I think I do that often. I’m not going to be surprised if others do the thing.

Lei Xing:
The execution with GM, so there's the ADAS, the safety with the passenger vehicles, and just to confirm you also have the Cruise Origin, is that correct, or no? Are you also doing the robotaxi? 

Dr. Jun Pei:
No, we are not.

Lei Xing:
So, you are not, right now.

Tu Le:
So Cruise they actually act like separate entity. 

Lei Xing:
So you're solely focused on ADAS.

Dr. Jun Pei:
Solely working on the project that will eventually lead us to making money.

Lei Xing:
So what do you think currently is the biggest hurdle or biggest bottleneck in this execution phase moving onto the next phase? Maybe we'll get into the technology route a little bit, there is OPA, there is MEMS.

Dr. Jun Pei:
Yeah, so it is actually always good to emphasize the spirit of this company that the thesis was on ADAS. So unlike many technology companies in Silicon Valley that you start with a new invention, a technology looking for application.

Tu Le:
Frontier tech.

Dr. Jun Pei:
This is more of ok, we saw LiDAR has a market for ADAS industry, let's invent something that's pertinent to be deployed in that industry. That's how MMT, the technology for our scanning was developed, is actually there's a huge focus on a balance between performance, cost and reliability. After all, we make an everyday consumer vehicle component. That's the spirit of this invention, MMT. To make LiDAR, you make your decision of what light source you use, what detection source you use, and eventually how you make imaging. So these are the three categories of decisions that you have to make. And for us, for me, who has been studying optics all my life, the first two: illumination, the laser source, the detection, are the two wheels that's already invented. It was invented in 1960s, right after laser was invented, you don't have to worry about them anymore. There are still some, this and that work going on there, but for us and the most practical way is to pick up the most mature components from these two categories, specifically take the 905 nanometer as our working wavelength. This is similar to many other LiDAR companies, because it's a very cheap dollar part, automotive-grade component, you don't have to requalify anything. The detection, the same thing, we use APD: avalanche photodiode as our detection, it’s a silicon diode that's already automotive qualified, mature components. It's less than a dollar a piece. So you pick those and focus everything on how to make an image. That's where you mentioned about MEMS device, rotating is another scanning like what I did with Velodyne many many years ago, and OPAs championed by Quanergy many years ago. So all these would have benefits and deficiencies one way or the other. And we came up with something called MMT: micro motion technology. It works like a loudspeaker, your voice coil over a magnet, that once you run current, it will actually move slightly. That's what we call micro motion. It's not, if you look at a loudspeaker or subwoofer, you see the motion when you're actually making sound. But instead of putting a paper cone on the voice coil to generate sound, we put in an optical array on top, so that starts moving, and just like your loudspeaker, there's no friction, nothing rubbing anything. The lifetime of this is, it’s still a mechanical device, just like your loudspeaker. But you know your loudspeaker actually, in your cars outlast your car. It's a very, very reliable device and we capitalized on that and employed that technology, which is well over a century old to the LiDAR industry. That's the essence of MMT, is a voice coil driven optical array that formulates the image. 

Yeah, with that, it was a very unique, Cepton has this patent, worldwide patent over this technology, it’s very unique. That you don't see any other company do. RoboSense is doing MEMS, Innoviz is doing MEMS, and rotating LiDAR is still quite popular in many applications. So all of these are, there are many companies competing on the same thing. The MMT is very unique. And that's one of the fundamental reasons we captured the design win, is with this, with this technology that is really low cost. You look at buying a loudspeaker, dollar stuff, $10 stuff, then couple that into a LiDAR, then you suddenly have a balance between reliability, performance, and cost, right? That's another reason Koito can partner with us to manufacture this LiDAR. They are after all a tier-one manufacturing company. Every penny is gold to them, they would not venture into manufacturing some component that is actually low yield, difficult to procure. MEMS device is always a custom piece that's sometimes gets into a lot of difficulties. Certainly, we didn't have the courage to go into that route. We want something simple, something robust, something low cost. So that's where we shine is we came up with this LiDAR that actually has the capability to keep the sweet spot. So in a nutshell, we certainly feel confident where we are today, especially we have the almost the shortest execution phase, even though it's 3 years in the making, is still the shortest in this industry compared to some of the earlier announcement. And having a tier-one from Japan, which is a very conservative traditional company backing us, vouch for our technology, back up this production process. Yeah, I think I sleep very well at night.

Lei Xing:
That's good to hear!

Tu Le:
A couple of things. So I was the commodity manager for graphics for Apple, so I understand the chip making. But I’m also a mentor for some startups and a lot of founders tend to try to reinvent the wheel for everything, like they need to be innovative in everything. And my basic questions are, why wouldn't you buy that bolt off the shelf? Why wouldn’t you try to find something off the shelf and then really focus on what you're trying to innovate. And so I think that's a terrific strategy that you're like, this is a dollar, this is a dollar, and we're going to focus on chip design. We're going to focus on these differentiators that really make a difference. And the second part is, are you finding that you're pushing Koito and GM, or are you guys working together pretty well from a speed standpoint? Because, yeah, the capital-intensive industries tend to be more deliberate with their decision making, because their decision making cost a $100 million, cost $200 million, cost $500 million. So are you finding that the speed at which things are getting done is pretty equal among the three partners?

Dr. Jun Pei:
It's not equal for sure, it's not equal. Silicon Valley companies are very much high speed, high intensity, and almost at times, or most of the times, so rogue operation, where Koito and specifically GM would be more of a methodical…

Tu Le:
Koito is, I mean the Japanese is even more conservative.

Dr. Jun Pei:
Oh, by far, Koito even among the Japanese auto industry, it’s considered as a very methodical, very conservative company. So how did we get married? I think, no, it's not equal at all. But the past 3 years of this, not only collaboration, the deep partnership between us making the Ultra Cruise program successful, has been very satisfying. I emphasis again, it's a satisfying experience that we actually trade these different thought process, different geographical location and different culture, all of these things, working together. It turned out to be, again, another balance and the sweet point that we actually strike. It's not only Cepton always champion something that's great and something that will conquer the world, but also a methodical deployment mechanism and well established supply chain by Koito. All of these are just wonderful elements are coming together. I'm actually happy that we're so different, that gives us more advantage as compared to others. As you see, there are companies that are working with tier-one, startup companies working with tier-one companies. Eventually, the relationship would actually fall apart. For Koito, we came much closer. You probably realized that Koito invested $50 million into Cepton in the Series C phase, this is before we had the GM contract. They invested another $15 million not too long ago through our IPO process. Recently, they just announced the indication for another $100 million investment into Cepton. So we are getting closer and closer as a very traditional company and a very frontier Silicon Valley company. You don't find another example. We enjoy this very much, even though we can be as different as anybody can tell a technology company in Silicon Valley, a classic manufacturer in Japan. So and GM is this another, it's our mutual customer that they actually hold both end, they're very traditional, very process oriented, but they at the same time, GM is considered as a frontier technology company as well. They are the most active in adopting the latest and greatest technologies. And that's why they have the Super Cruise. That's why they will have the Ultra Cruise. They are the first one in the massive adoption of LiDAR, it's a very enjoyable ride so far.

Tu Le: 
Good. Let me make sure that they're learning as much from you and you are from them, and kind of, because I think on the high tech side, people don't appreciate what automotive grade, right? Because you're not going into a consumer product, you're going into a vehicle that vibrates, that shakes, that has these sudden, I think a lot of the Silicon Valley companies that want to get into the automotive space don't understand why there's so many qualifications that need to be done for their product to work on a vehicle.

Dr. Jun Pei:
I want to just chip in an example for that automotive-grade component. If I were asked 3 years ago, okay, what does it take to make an automotive component? I would say ok, you got to have a wonderful design that balance between cost and reliability and performance. And then you got to study your temperature sensitivity, you got to do shock vibration, all of these things are, that's my answer 3 years ago. If you ask me the same question again today, what does it take to make an automotive part? My answer is, you got to have a design win and go through it. Seriously, you don’t know what you don't know, the answer that I had 3 years ago was so bookish, so classic from an engineering perspective. You actually, that indicates you didn't know what you are talking about. After 3 years, what we have gone through, I start to get a sense of it.

Tu Le:
So a quick story, first week on the job at General Motors, out of university undergrad, I go into the Warren Tech Center. There's a four-man cube, and I'm the 4th person, and they had, this department hadn’t hired anyone for 15, 20 years. And so each person in that cube, besides me, had 35 years with the company. And one of the older gentlemen, we called him Ollie, he asked me to come over. He's like I can't figure out this email thing, he’s like, what's going on. He had a fatal error because we were using Windows machines on his computer, I was like, you got to turn it off and turn it back on. He’s like, you know, I don't get these tech companies, because if we built a car that only worked half the time, we'd be out of business. I've really gotten appreciation for the automotive space at that point, because it just didn't make sense for him that something wouldn't just work. And so I think we share an appreciation for how difficult it is to enter the automotive space. And that's not to say that there are some unnecessary things going on that probably hindered the speed or the agility of companies. But there's a fine balance.

Dr. Jun Pei:
Exactly. That's the balance. And I think this three-way working together with GM, Koito and us is finding that balance and hit the sweet spot.

Lei Xing:
Maybe a couple more questions. I think we covered pretty much. Obviously, we can't talk about China EVs & More without talking about China. So what are the business opportunities there for Cepton? What is your take on these very aggressive, RoboSenses, Hesais, Livoxes, Huaweis, you know. What do you think of this competitive landscape?

Tu Le:
Let me massage that question. Is the GM contract a global contract or the Koito contract a global contract?

Dr. Jun Pei:
It will cover more than just here, so yes.

Tu Le:
So, but then it’s still…

Dr. Jun Pei:
It’s still to a pertinent question. With Cepton, China market, Chinese competitors, and what's it like for us out there? We have a China office, and we certainly have our presence over there. And in a way we're participating in some of the evaluations over in China over there. To me, even though I grew up in China many years ago, I already kind of very much localized here in the States. So China market to me is a very peculiar place. It’s unlike Japan market, European market or the North American market. With the recent years, there are also some, there's for obvious reasons, there's geopolitical environment…

Lei Xing:
The business environment part of it.

Dr. Jun Pei:
Yeah, it’s starting to shift. What I’m trying to arrive at is right now, to U.S. or Western companies into China, there are uncertainties. That's what I’m trying to arrive. There are uncertainties ahead of…

Lei Xing:
Same the other way around.

Tu Le:
You are saying it very nicely.

Dr. Jun Pei:
And with that said, I think nobody can ignore the China market. It’s a huge market. Everybody would like to have a piece of the pie and us included in there, but the uncertainties are bringing a lot of considerations. So we are, maybe just quickly jump into the summary, we're moving forward in China, but we're moving forward cautiously.

Tu Le:
I think, so in 2022, I think the global NEVs, new energy vehicles, let's just loosely say, new energy vehicles, I think we're going to see China leading the way around 6.2, 6.3 million, 6.5 million?

Lei Xing:
6.5, 6.7 million probably, will be very high up there.

Tu Le:
So the global market is probably going to be around 10ish, so China will hold about 60 or 65% of that market. I do think Ford was pretty intelligent by going F-150 first. But I believe I’ve been kind of bringing in some new train of thought. I think GM is going to be the leader because they did clean sheet and said we're going to go EV platform with all of our new vehicles. So that's why they're launching the Blazer, they're launching. So I think there's a huge opportunity for Cepton to grow with GM. I have no idea how GM's going to get to that $30,000 price point yet, because battery prices in the short term are actually increasing. And so I think there's going to be a squeeze on their ability to fund this EV future. But I think GM is in the best positioned here, aside from Kia to really take advantage of the recent excitement about the sector. And it's messy. So I think that's where startups have a really good opportunity when things are really messy, right? Because when things are pretty predictable, it's the legacies to tier-ones that really really shine, right?

Dr. Jun Pei:
Yeah, I hope the pandemic will actually taper away. Even I say that in China case, we're still pretty intensive over there. I'm sure you share your experience.

Tu Le:
Yeah, he just came back, so he knows, I haven't been back since, early 2020. So I don't know.

Dr. Jun Pei:
I’m sure having this type of meeting without masks on, it's probably very enjoyable for you guys. Certainly, for me.

Tu Le:
It is really, what’s the right word, jarring. Because living, sitting in Beijing just a month ago, and watching sports and seeing thousands, we just went to the Lions game.

Lei Xing:
60,000 people, right? 

Tu Le:
So my sons were at a local school in China, they are still wearing masks, and we have QR code, we have an app, so.

Dr. Jun Pei:
I’ve heard of that. So I'm still hoping that this pandemic will leave us eventually and become a non-issue. And then what you said excitement will come. We'll get promoted in China, and there's just too much uncertainty at this very moment. I hope they will actually taper away in the coming years, everybody can benefit from the excitement of technology development.

Lei Xing:
Yeah, I don't think pandemic is the biggest uncertainty. I think it's other things.

Dr. Jun Pei:
Side effect.

Tu Le:
The one thing that we can tell you is that, China EV Inc. is coming to the U.S. So I think if you have an opportunity to get some more design wins, it's going to create a formidable defense for you, because and what we're hearing is it's a little more delayed than it was. But NIO has already announced they are coming by 2025. It's going to really drive a lot of UX redesign on the U.S. legacies and the technology stacks, how the Chinese EV companies deal with data privacy, data security. The chips is going to be interesting, because is that going to be driven by policy? Is that going to be driven by the sector, the GMs and the Fords? Probably a little bit of both. And we'll probably see a lot of investment from China EV Inc. in North America. So maybe Canada and Mexico. But I would say within the next 3 or 4 years, there's going to be at least five or six Chinese companies that have a small, mid-sized presence in the U.S.

Dr. Jun Pei:
I can't wait for that to happen. For the foreign carmakers to come the States, there are many success stories, especially like Hyundai, or earlier on the Japanese entering the U.S. market.

Lei Xing:
Now the Vietnamese, VinFast, they're very aggressive.

Dr. Jun Pei:
Then they can become very popular and products for…

Tu Le:
They learned, because the Koreans learned fast, right? And let me tell you this. So I don't know if you still drive a Tesla, but the new NIO ET7, ET5, beautiful, beautiful cars. So.

Dr. Jun Pei:
No, I don't drive a Tesla anymore. 

Lei Xing:
BMW.

Tu Le:
That's right. I could totally see you in an ET7, that’s just say that, because it's very contemporary, but because the design studio is in Germany, so.

Lei Xing:
There's a big cult of NIO investors in the U.S.

Tu Le:
I'm pretty much done.

Lei Xing:
How, one last question, how's, you want to, how's the balance sheet? Do you need more money? Are you? Because, right? You want to make money, but first you got to put in money.

Tu Le:
Manufacturing is the most expensive part, right?

Dr. Jun Pei:
So we kind of touched on that point a little bit, there is a very recent announcement for Koito investing $100 million.

Lei Xing:
Right, you mentioned.

Dr. Jun Pei:
We believe we're quite well funded for the GM deployment program. So with that we're already at the tail end of the deployment and with this additional infusion of capital we should be very good, but we're not going to stop there, right? We will capture additional design wins. And we'll have another, one or two big programs that go through the same process. And with that, we'll need more money to do that, but that's a good problem to have. And I'll deal with that when that comes.

Lei Xing:
The announcements are forthcoming. Is that what you're saying?

Tu Le:
The weird thing is that kind of looking at valuations and, Luminar is kind of worth stupid money right now. They're the only company that's, the LiDAR company that's like I have no idea how there was so much money. But they're moving aggressively into in China.

Lei Xing:
In China, they have this SAIC commercial (win), the Feifan Auto.

Tu Le:
But they're going to feel when they enter the China market, that squeeze too, because the Hesais are going to be really aggressive…

Lei Xing:
Sure that's what we were talking about.

Dr. Jun Pei:
Yeah, I think there's a huge push for localization of technology in China. And so those are challenges for foreign companies. Again, us included in there, compete in that market, we shall see, there’s still some time. But coming to valuation for Luminar, that's we have a whole session over that. The stock market is a very different animal. I have, since we became public, there was a few weeks, I had the habit of watching our stock price every other hour. I have weened myself out of that. It's actually totally irrelevant because I see completely irrelevant to my daily operations. So stock market is its own different animal at works. So I try, of course I have to pay attention, but I only look at it once a day now after closing. 

Tu Le:
I just think that it’s pretty unpredictable because there's these retail investors that kind of get attached to certain stocks that create these artificial valuations, including Tesla, so.

Dr. Jun Pei:
Well, when you were in China these years, there was this Game Stop, that was paid a lot of attention.

Lei Xing:
Yeah, early last year, right? 

Tu Le:
Even BYD lost 30% over the last month, because Warren Buffet sold 0.6% of his holding. So, Jun, thank you for spending the last hour plus with us. I learned a ton. I admittedly didn't know a lot about Cepton before sitting down and doing a bit of research the last couple of days. But I think Cepton does have a very unique approach to the LiDAR space. And that ADAS, the focus on ADAS is a lot different than what I’ve heard from other LiDAR companies. So.

Dr. Jun Pei:
Yeah, thank you again for coming to our office and I really enjoyed this conversation. I guess if there's a takeaway from this, I hope you have an impression that Cepton is actually founded by engineers, run by engineers, and doing practical projects.

Tu Le:
Yes, which is pretty unique. So.

Lei Xing:
Thank you for your time!

Lei Xing: This is your co-host, Lei Xing. As you’ve heard from Dr. Jun Pei, Cepton, unlike any other LiDAR supplier, has a unique go to market approach by being a tier-two supplier supplying to a very traditional tier-one supplier in Koito, which manufactures the LiDARs in Japan at a so-called Cepton production line, which are in turn supplied to GM’s upcoming Cadillac Celestiq and other models with the Ultra Cruise driver assistance feature. Dr. Pei believes that this three-way working relationship is the best at finding that balance and sweet spot on LiDAR performance, cost and reliability whereas supplying a major OEM like GM on its own will be more difficult when it comes to those metrics. For Cepton and many other LiDAR suppliers, it’ll still be some time before they move on successfully from the current execution phase to that making money phase, and China remains elusive and Chinese competition will continue to be a factor in that process. We wish Dr. Pei and Cepton the best of luck.

Tu Le:
Lei and I will be sharing more of our conversations with the men & women around the world moving the EV/AV mobility sectors forward as part of this China EVs & More MAX series. Some folks will be instantly recognizable, but some will just be people that are doing amazing things in the space that we think deserve to be highlighted. Don’t worry though, Lei and I will continue to host our live weekly China EVs & More Twitter Spaces room that summarizes that week’s most important news coming out of the China EV, AV and mobility space. For those that can’t catch the live show, you can find the China EVs & More pod on all major platforms or wherever you normally get your podcasts. 

As EV adoption reaches its global tipping point, it’ll be even more important to stay updated on everything that’s happening here. Lei and I are confident that China EVs & More is the best resource to do that. Until next time, as always, thanks for listening!