Line as much as you want. All right. Good. I guess it's still morning for us out here. Thanks, everybody, for coming back. We'll get started with our next fireside chat. We're very happy to have Nakul Duggal, who is the Group GM of the Automotive and Industrial Business.
That's right.
At Qualcomm. Qualcomm, as you all know, is heavily into the handset side of things, but is over the last five years plus has heavily invested in diversifying their business. Nakul is the head of some of the biggest portions of that. Probably, at least in my estimation, I don't think the company has really said it, but probably runs somewhere upwards of a quarter to nearly a third of the company. We're very happy that you could come up.
Thank you for having me.
Why don't we jump into a little bit about the diversification strategy at the highest level? How are you? We'll dive into the individual businesses here in a bit. At the highest level, the goal to diversify, the resources you're given, the strategic imperative, all of that, whether it's the board or Cristiano or Akash, how are you tasked for running that?
Thank you again for having me. I've been with the company for, you know, I celebrated 30 years in June. I understand the history of the company. I understand how we invest. I understand how we build ecosystems. One of the things that became pretty clear as we were focused on diversification about 10 years ago was what is the value of the technology stack that we build in markets outside of our mainstream markets? What became pretty apparent was our understanding of markets outside of our core mobile business was fairly limited. The approach that we took when we were starting to build out our automotive business was how do we build a much more intimate, much more proximate business that is aligned around the needs of the auto industry? This was a decade ago. Automotive was obviously very different from what you see today.
It allowed us to appreciate the complexity involved in quality, in safety, in supply chain resilience, in the long lifecycle that is required, the software differences. We have since taken automotive as a business that we manage pretty much completely independent of the mobile business. It gets not only the resources and the investments needed, but it's a much longer view on how are we going to see success there. The approach that we took in automotive helped us in formulating a lens through which to look at other markets, like industrial markets that, you know, have a slower moving cadence, but obviously much more committed to the transformation that then comes along. Those approaches have certainly helped us cement the way that we then invest as well. The bets that we make, whether M&A or, you know, strategic partnerships, that's all part of that same strategy.
Let's dive a little bit into the automotive side of things. I think you guys are doing, you're on track for nearly $4 billion in revenues in this calendar year. You're a bit ahead of what your targets were. I think over the last five years, the CAGR of that business has been about 40%. Incredibly impressive, especially at a time where many of the auto semicompanies are having some challenges just cyclically. Talk a little bit about what has driven that growth looking backwards. Then we'll get into some of the forward growth drivers as you drive towards your goal of having, I think, $9 billion towards the end of this decade.
Yeah. I think one thing that automotive requires, apart from everything else that you have to do, is a tremendous amount of predictability that the supplier has to provide to their customer. It's become a very complicated market in the last five years or so. Because of all of the complexity that automakers already deal with, electrification, geopolitics, supply chain resilience, when they look to us, one of the things that they find is a global player that invests heavily in software, has a very broad portfolio of semis, is able to think five years out, make those investments on behalf of the automaker. That just makes us a very stable, reliable supplier that is focused on the transformation that's going on in the automotive space, central compute, SDV, driver assistance, automated driving, all of that. A lot of those bets were made by automakers five plus years ago.
As they make those bets, they look at what bets are working, which ones are not. I think the revenues that we delivered a year in advance are really an indication of winning at the right time, those programs essentially launching at the right time, and frankly, being very broad-based. We supply into pretty much every automaker globally. We are not as affected by changes in the type of drivetrain that is selling or what happens in one part of the world versus the other. We are fairly insular to that.
If we just took the $4 billion in revenues now and the $9 billion that you have as your target, I think at fiscal 2031, talk about how the mix changes within that. I know you have even, just to keep throwing around large numbers, a $45 billion design pipeline, I believe. How important is that transition from probably the connectivity side to the ADAS side, et cetera?
I think connectivity is a very predictable, I guess, is the most predictable part of the composition of the revenue. It's mostly our modem business, the RF business, some Wi-Fi attached alongside it, some Bluetooth, some GPS. We have a power line communications business. The remainder of the business is actually very proportionate to the rate at which cars are moving towards central compute, bringing in a tremendous amount of integration of a number of different ECUs that were previously discrete and disparate. Now these are all getting integrated. It's happening across multiple automakers and multiple generations at the same time. To give you an example, there are more conservative automakers that are launching products that we introduced in the 2020, 2021 timeframe this year on a very large global footprint.
They will essentially take some of our Gen 3 cockpit products and deploy them across their global fleet, and those will last for seven, eight years. There are others who are taking our Gen 5 platforms, which we only sampled beginning of 2025, and those will go SOP first quarter of 2026. I won't get into the content delta, but the performance delta between Gen 3 and Gen 5 is probably 15x. We have this very interesting trend that is now happening in the auto industry where everybody's moving towards their next generation architecture, but the rate of adoption is not the same. We benefit from being able to see really fast-moving companies adopt new technology and rush to compete, while at the same time, we have a very stable, very predictable foundation on top of which previous generation silicon is getting deployed.
To add to that, as we start to see ADAS wins coming in, and ADAS on the silicon side really started to scale for us, early part of, I guess, middle of 2024 is when we started to actually get designs to scale up with our Gen 4.5. We started to build our ADAS silicon about three years ago. We have announced 20 OEMs. This number will just keep growing because everybody's going to start to kind of move in that direction. Because architecturally, we have aligned the cockpit architecture and the ADAS architecture, it allows automakers to significantly save in software cost. They have to build once, and then they can deploy based upon the workload that they have. You will see a lot more of the compute concentration over the next five years.
I think it'll be a combination of high performance, high value compute with a foundation of a lot of the legacy programs that we have on that will continue.
In some ways, it's analogous to Qualcomm going from a connectivity leader to a compute leader across the entire company, whether it even be in handsets. The same thing is happening in automotive.
I think that's right.
Is the majority of, and you might even have had charts in this at your last analyst meeting, so forgive me if I can't recall, but the majority of the $4 billion today, is it safe to say, is more the connectivity and cockpit? Then towards the $9 billion side, the ADAS side will be a significantly larger portion than it is today. That's the direction of travel?
I think you will start to see a blend of ADAS and cockpit just in terms of architectures. We introduced the Flex architecture, and that blends it. I think you will see a lot more of the compute portfolio in the next five years.
When we think about the ADAS side of things, what's the primary differentiator that Qualcomm has?
Yeah. I think there are a couple of things. In markets where the ADAS stack is well understood and you know what you're looking to deploy, whether it's an urban navigator on autopilot or highway navigator on autopilot, where the number of sensors are well known, the stack is fairly robust. We excel at being able to then optimize that for silicon, optimize it for power, optimize it for performance. That is something that is there in the history of the company. It's there in our architecture. That's where we are really able to extract a significant amount of performance on a per millimeter square basis on the thermal envelope that we have provided. That's helping us win quite a bit. The other piece that helps is our tiering.
We have built a portfolio, especially for ADAS, that allows us to go towards the, I won't say the real base tier, but entry and then above. We are very well positioned with the portfolio on entry, mid, and then premium. Customers who are building their stacks and are starting to mature them, harden them, they want to be able to have a common portfolio that allows them to be able to think between tiers in the same generation and also across generations. That I think is making a huge difference in terms of the rate at which we are winning in ADAS.
That scalability that you're mentioning, how is that applicable on both the hardware/semiconductor side as well as on the software side?
Yes, absolutely. The software architecture that we have is the same software platform that travels from a premium tier SoC all the way down to an entry tier platform. That saves OEMs a significant amount of switching cost. To the extent possible, we also try to maintain that across generations.
Please discuss on the software side the importance of the Arriver acquisition from a few years back.
Yeah, so you know, we decided perhaps a bit late that the stack business was something that we had to get into because it was going to have high attach. It was going to be global. It was going to become part of every vehicle. I think, you know, clearly that's how it's turning out. Rather than doing it organically, and we've had a team that has been working on the stack for quite some time, we decided we needed to have a team in-house that had real-world experience in actually deploying this. It was the right idea. We did this right in the middle of COVID, so it was not the best time to get into the middle of a complex acquisition.
What helped us was BMW was very interested in working with somebody that they could have a very close partnership with in terms of driving their next generation stack requirements, ADAS requirements. It became this, you know, three-way and two-way once we acquired Arriver, where we've built out our stack team over the last three years or so. We've built the computer vision stack in-house. It's going to be commercial by the end of the year across 60 countries, 100 countries next year. You know, if you think about it, this is something that we did from a standstill. In 2022, we had no stack. We will have this global stack launching next month. With BMW, we partnered to develop the drive policy stack together.
The Arriver team was very instrumental in not just bringing in the expertise, the IP, the experience, but also creating a foundation inside the company on how to go build a stack business, a stack product, which is something that Qualcomm obviously has not done before. I think it was at the right time and has been hugely instrumental in the success.
When you put together the hardware and the software side, what is the primary differentiation that you offer in the ADAS realm versus, say, the Mobileye side or the NVIDIA side, which tend to be the two primary competitors? If there's other competitors you wanted to highlight, go ahead, but those are the ones I think of.
Yeah. I think there are multiple layers to a stack. I think the active safety stack, which is basically the foundation for making sure that you are able to be compliant with all of the global requirements, is something that I think we are now working to get to par and perhaps even ahead of some of our competitors. Mobileye obviously is the gold standard, has been doing this for a long time. The goal is to be able to certainly meet and eventually get ahead of those types of requirements. I don't really look at many other stacks as really at that scale because, if you look at the data, if you look at what's out there, other than stacks that have been built vertically by OEMs, there isn't really anybody else apart from Mobileye that I would say is at the same level.
As you start to go to features and go higher up the stack, that then comes down to the OEM, what type of sensor set they're deploying, what type of features they are developing. We actually welcome, for those of you that are interested, to come check out the Neue Klasse in Munich, and we'll have some here in San Diego as well, starting next month. You can experience the customer function. You can actually see for yourself the type of performance that we are very proud of because that is something that not only have we built together, but we have, so that is a stack that we can take and extend over to other OEMs, which is what we're in the middle of doing.
You talked a little bit about the importance, or we talked about the scalability side of things. When you talk about Snapdragon Ride, and then you have the Flex side of things, talk about the Flex platform and the importance of it. I get the sense that that is a unique ability that you have on the scalability front. Just kind of expand on that if you could.
When we decided to start building safety-grade chips for automotive compute, it became pretty clear to us that this whole separation in a car between what is the infotainment domain and what is the safety domain, this obviously historically made sense. Going forward, especially with AI, especially with physical AI, these lines will start to blur. Architecturally, we decided that we will allow for an architecture that allows us to be able to run a software stack that doesn't differentiate between infotainment and driver assistance. We did that with our Gen 4.5 products. Now with our Gen 5 products, the level of performance that we are adding, you know, it's 3-5x more powerful than the Gen 4.5. You just have plenty of headroom to be able to run two stacks concurrently and still have room left over for more.
What we have started to notice is OEMs that own their own stack want to be able to run the stack as an application as opposed to a dedicated domain. Especially when you start to go mid-tier where you want to be able to get cost optimization, where you want to be able to go from two separate ECUs to a single ECU and potentially a single SoC, the flex architecture makes a lot of sense. We introduced the flex architecture for the first time. I think it was maybe at Snapdragon Summit in 2024. We will launch with Leapmotor in China, our first flex platform in the middle of 2026. This is an example where somebody who actually owns the infotainment stack and the ADAS stack in-house, they can go deploy this on the same SoC. We're starting to see that interest from a number of different players.
Do you get the sense, we talked a little bit about the uniqueness of your stack and the gold standard from a competitor point of view. What about just on the pure silicon side of things, the competitive landscape on that? Do you tend to think of Qualcomm as coming from kind of the bottoms up and so more working your way from, you know, the L1 functionality up? Or do you have the headroom to start kind of at the top and go down in what the competitive landscape looks depending upon which side you're coming from?
I think we made a bet five years ago that when it comes to ADAS, we're going to focus on L2+. Right or wrong, that was the bet that we made. L2+ and below, and that's kind of the market that we want to focus on. We deliberately did not focus on L5 and robotaxis and all of that. The platform that we started to build was the same common platform for both cockpit and ADAS because we started to add the safety capability. I think it was a very good decision because it has allowed us to flex, quote unquote, in the direction of getting down to entry tier chips, which we've now had available commercially over the last three years or so. This next generation, we've moved up tier into premium and super premium.
In the same window of time, we actually have this portfolio that really scales across the board. There isn't anything that we do not have as far as silicon roadmap goes on our platform, except for some very basic entry commodity silicon. That makes it a portfolio that is super complex to contend with for many of our competitors because it is reliable. We have tens of millions of chips that we are deploying annually, multiple hundred millions deployed. They're all safety grade. They're all supply resilient, multi-sourced, quality proven, software proven. For somebody that is trying to do this for the very first time in an environment where there's massive risk in terms of execution, in terms of a global footprint, geopolitics, I feel pretty good that I think we are actually in a pretty good place with the strategy. I mean, it's played out well.
Clearly you're well ahead of plan and doing $4 billion in revenues already at a time, like I said before, that's been challenging. Something has gone very well for you. That's a definite compliment. You said five years ago the focus and the bet at the time was placed on L2+ and below. Has that evolved now to a higher level as you've succeeded and gotten more knowledge of the industry?
We are transitioning to an end-to-end AI architecture for our stack. That becomes very data-driven. As you have a data-driven approach, you want to be able to make sure that you can keep advancing the same stack towards L3. We feel that the majority of the market is still going to remain in that L2+ space. There are a number of reasons for that. I think one is just the liability that the automaker has to take on. The other is the cost pressures, the market adoption, consumer adoption. I think the teams that we have are developing a stack that will certainly be ready for an L3 implementation. We are thinking about this more in the context of redundant architecture. If you deploy an L3 architecture, what is the fallback? How do you deal with the safety aspects?
For us, what has become a great learning in the partnership with BMW is that you really have to focus on the safety aspects of the architecture that you're building because at the end of the day, that is what matters, right? I think these are all great features to talk about and read about. As a consumer, if the product that you're building, you can't stand behind from a safety perspective, it's not going to get all that much traction. There we feel like at least one competitor has taken that seriously. I think they do a very good job with it. That's the approach that we're taking.
When customers or when suppliers, semiconductor suppliers get design wins, everybody talks about them. You guys talk about getting, you know, 20 OEMs and a bunch of, I think, 12 new design wins last quarter, all great numbers. We hear from a lot of your competitors, the same customers, design wins with the same customers. How externally do we reconcile that everybody has the same design wins with the same customers? Are they on different SKUs, different models? Especially as you move into ADAS, it seems less and less likely that you would have a multi-source agreement. How do you guys suggest we reconcile that?
I think it varies quite a bit. I think it varies by OEM. It varies by region. It varies by the domain that you're referring to. For example, we see less and less multi-sourcing across telematics and infotainment. In ADAS, we see multiple generations of programs coexist because those programs run long. Sometimes they get delayed. The next generation program that needs more functionality will probably overlap significantly with the previous generation program. There are aspects of the car architecture that come into play. For example, electrification has a massive impact on the architecture, the EE architecture of the vehicle that has a very specific set of suppliers that have been selected. There isn't really a single SoC supplier to a single domain, especially for the mid to large size OEMs. The smaller OEMs are pretty much, you know, one or the other.
As you start to look at China, where the pace at which the market is moving is much faster, you absolutely have overlapping programs all the time.
To try to put some rough numbers around it, and your willingness to give these, if you guys don't give them, then fine, you can just say that. As investors think about the content per vehicle and how that changes over time, not necessarily today to tomorrow, you could put it in multiples. As you go from connectivity to cockpit to ADAS, how should we think about what sort of multiplier effect that could entail?
I think, you know, we have not really broken this down in terms of the actual ASPs, but, you know, I think connectivity is very predictable in terms of what are the types of ASPs, right? They're mostly double digits. I think with infotainment and ADAS, that's where I think there is a significant amount of premium that is available, mostly because we are driving a tremendous amount of integration value. We are massively integrating new functionality, coexistence, and we are reducing the overall system BOM for the automaker. That's where we are able to extract a premium. You know, we haven't really provided any specifics on the ASPs themselves. I mean, it's a range. It's a broad range depending upon which tier you're referring to.
Right. I think would it be safe to say the content per vehicle rises as you move along that range towards the ADAS side?
Absolutely. Absolutely.
Does the profitability scale directionally? You definitely don't break out what the profitability is, gross or operating margin for your segment. I'm not asking that. Does it scale upward as you follow that same trend as the ASPs or content do, or is it more based on revenue scale?
Yeah, I mean, it's very dependent upon the tier that we are talking about, which underlying chip we are talking about. The profitability absolutely does scale. Yeah.
The last question on auto, because we do want to get over to the industrial side of things. You're doing almost $4 billion a year in revenue now. Your target is to get to $9 billion, like I said, in five-ish years or so, five, six years from now. What are the biggest steps for us to monitor externally that are necessary to get from the $4 billion to the $9 billion?
I think the biggest one is execution to the pipeline that we have shared. We have internally a way to track the rate at which the programs that we have won, how well are they executing. That is a very clear leading indicator in terms of the rate at which the quarterly revenues are building up. I think the other is the win rate. Are we continuing to add to the pipeline? How quickly does it replenish? There is a lot of macro out there just in terms of tariffs, the sales of OEMs, China competition, what does that start to look like, especially in Europe, especially the rest of the world. I think the size of the market is also growing, right? The silicon content in cars is growing. Every car is contributing to an expanding SAN.
I feel like, of course, there will be competition and there will be complexity. I think cars by the end of this decade are all going to become fairly advanced products pretty much all over. That certainly plays very well to the strategy that we have. As far as the roadmap goes and the investments that we are making, those are all made very well in advance. Most of the products that we have to go build are already designed or in design right now. We are only doing this because customers have made a clear case as to what they're looking for in the next generation. We feel pretty good about that.
Good. It's going well so far indeed. Congratulations on that.
Thank you.
Why don't we pivot over to the industrial side of things? This is going to be a subset of your IoT business that you don't break out. I believe it's somewhere around $1 billion that you're responsible for of the roughly $6 billion, $6.5 billion of IoT. Is that, it's a bit round numbers? We're in the zip code?
It's a bit north of that.
A bit north of that?
Yeah.
Good. That's better than south. Talk a little bit about the strategy that Qualcomm has applied to this market and how it might differ from what you did in automotive.
Yeah. This is something that I've been managing for the last 18 months or so. A couple of things rang true about this market that were maybe a little bit different from what we saw in automotive. I think one was there is a massive interest and need for Qualcomm technology to be adopted across these very wide industrial segments. We needed to package our products a whole lot better. They needed to be made much more relevant to the specific industries because there was not a lot of familiarity with Qualcomm in a lot of these spaces. They had to be made easier to use. We had to have the right channels in place, the right infrastructure in place. That has now started to happen. We've broken this down across five major product segments.
We have an industrial connectivity product segment similar to what we have in auto, which is doing quite well. We have a camera segment that we are doubling now. We have taken the camera space, called that out separately, and focused on cameras as an area where we will see a tremendous amount of AI capability getting built out. That is actually going very well. We have what we call a consumer and commercial processor business, which is really a very broad market from retail, enterprise, home and life, appliances. That is growing very quickly. We have an industrial processor business, which is focused on all of these industrial applications, which we leverage a lot of automotive IT and products from. We have a new robotics and drones business.
We have organized the product segments to be very horizontal and very focused on a large number of verticals that can take that capability. We have revamped our go-to-market team. There is a lot of familiarity in terms of which verticals need what specific type of product. We are building solutions. We are not just stopping at, you know, think of it similar to, maybe not a great comparison, but like we felt the need to go build a stack because this was something that was going to have mass adoption. Similarly, we feel like in the camera space, there are going to be a number of solutions that will be needed because AI is just going to become the way that you engage with the camera.
We are building out solutions across a wide variety of segments because we feel like as you get to the solution level, you're just going to be much more relevant to the end customer. We've also taken an approach to be very developer-focused, which is not something that, you know, it was a muscle that we didn't really have a need to really develop in the past. In this space, because there's just such a large number of people that are now developing with AI, having the ability to be able to be accessible to an ecosystem that as a B2B company, we had never really focused on. You'll hear more about this in the coming months about what we're doing there. That's become a very urgent and important focus.
As you push in all these different bookends of the problem statement, I think overall we just become relevant to a lot of players that we previously were not.
At the highest of levels, in the two, three minutes we have left, the differentiation Qualcomm offers entering this space. I know it's probably different in each of those verticals that you'd mentioned, but if there's a common thread, please summarize.
I think our scale, right? I think there are not many companies like us who exist in the industrial space. There are not many companies that are coming to meet customers where they're at in terms of trying to understand their end-to-end problems. For example, we built a partnership with Aramco a year ago. We have two dozen projects that are going on. Just with that one customer, we've set up a team in Riyadh and in Tehran to go support those customers. Our ability to be able to engage with the customer and get to a solution, given the scale at which our technology operates, our products operate, that is something that comes pretty naturally to us once we kind of provide focus to the problem statement.
I don't think there is really any other company that does everything from Bluetooth shelf labels to edge appliances to drones and robots and industrial processors, industrial PC chips. It's a massive portfolio in terms of what we have. I feel very good about the projections that we have made in terms of the growth in the industrial embedded IoT.
Is the go-to-market direct or more distribution heavy?
It's a mix depending upon the complexity of the product. Products that are more complex require a direct touch. We are also, as I mentioned earlier, very developer-focused. We are building out some more capability that allows us to make certain products simpler and easier to use. That will make those products very distribution-friendly. There are certain products that we do sell through distribution.
Talk a little bit about the, I believe, you're north of $1 billion in revenue, as you said. I believe in fiscal 2029, Akash said you had a $4 billion target for the revenues. Much like I asked on automotive, in the last minute we have left, talk about what it takes to get from where you are today to kind of quadrupling-ish or tripling, depending upon where you are.
I think it's pretty straightforward. I think there are two types of customer profiles. I think one is the ones that we are very familiar with, who are now consuming a much broader portfolio than was exposed to them previously. I think we are seeing tremendous growth there, you know, cameras, industrial processors, consumer processors, et cetera. The other is customers that are less familiar with us and where we have to change the profile of the product for it to be much more easily available and accessible to them. The size of the market is very large. This is not a small market, and the number of products that we are putting into the portfolio are also very large. By the way, the other part that is interesting is we're not building a lot of new products for this market.
We may build one or two products for what is a very large portfolio. What that should, you know, tell your audience is we have a lot of capability in-house that we can leverage in real. That obviously also helps us significantly in terms of the expense that we have to take on to enter these markets. We are directing that expense towards software, towards expanding the channel, towards building out solutions.
We're actually out of time already, but congratulations on the diversification that you've done. Automotive has been especially impressive, and we can all see that. I know the industrial side has done quite well underneath the surface as well. Thank you for joining us.
Thank you very much.