Great. Thank you, everyone, for being here. I have the pleasure of hosting Qualcomm for the next session. With me is Akash Palkhiwala, who's the Chief Operating Officer as well as the CFO of the company. Akash, thanks for being here.
Of course. Pleasure.
We wanted to start with automotive this time around . We've done this session a few years now in a row and started with different topics. Let's start with automotive and just outline for us the pipeline of opportunities and how should we think about those translating into revenue, particularly how to think about Qualcomm share, where it stands today on a revenue basis that we can look at versus the visibility of market share that the pipeline provides you on the automotive side.
Yeah, let me start with maybe a few financial metrics, then go to a qualitative part of the answer. As you know, we'd set a target of around $4 billion of revenue in fiscal 2026. We're very close to that in fiscal 2025, and our run rate indicates that we'll beat the 2026 target that we'd set. Executing to what we said, we have set a target of $8 billion in fiscal 2029. That's a 20% CAGR from where we are at. As we've said in the past, most of that revenue, cumulative revenue of the next four years, greater than 80% is covered under design wins we already have. There is a lot of predictability in this market. You guys all know how the auto market works. Once you want something, you know it for a while.
There is predictability and high confidence in getting to the $8 billion number. What we like about our revenue stream is we participate in the growing part of the SAM. For us, it's not as important as what is the total size of the automotive market. A lot of other companies who are dependent on the scale of the market, the conversation is, will the market grow or not? How will you grow with the market? Can you grow faster than the market? For us, it's not that. The part of the market we participate in, which is cloud connectivity, digital cockpit, and autonomous driving, just the SAM for us is growing at 15% in a flat automotive market. We're very confident we can grow faster than the SAM because we're gaining share.
New car launches are happening with our content, which puts us in a very strong position to execute on what we've laid out.
Let me follow up on that. The $4 billion versus the $8 billion, how does the composition look different when you break it down between telematics, infotainment, or driver display, and ADAS? How does the mix look from $8 billion, $4 billion to $8 billion ?
A year ago at Investor Day, when we laid out our kind of pipeline, design win pipeline, we had split it into three parts. If you just eyeball the chart we showed, about a third was ADAS. Of the remaining, the majority was digital cockpit and connectivity. The rest was connectivity. It is spread across all three things. We also showed a chart that split that across geographies. In terms of when you think about the auto market, a lot of the questions people have is, who's going to win? Is China going to take a lot of shares? Is Europe going to take shares? What are the Japanese going to do? We have wins across the globe. I think it puts us in a very strong position because the portfolio is very diversified from an OEM and a geography perspective.
It is, you play in a very strong portion of the market.
At the Automotive Investor Day, you also highlighted that total content opportunity per vehicle can be in the thousands of dollars in some cases. Is that really playing out? When you now look at design wins, are you getting certain vehicles where you have upwards of $2,000, $3,000 of content?
If you think about what is happening in this market versus two years ago today, the connectivity content has largely remained the same, but the digital cockpit content has increased by a lot. I think competition is fueling the need to have a more modern digital cockpit, more modern instrument cluster, and everything is software enabled. We are, by far, the number one supplier for cars in those. I think we're very, very well positioned. Agentic AI is going to change that further. Of all places, car is a place where your hands are busy, and being able to have a conversation with your car about any topic and being able to ask questions and do things is going to be very critical.
I think that is also going to help us both from a content perspective, from an on-device AI perspective, and I think it puts us in a very strong position going forward.
OK.
Sorry, last one I should have addressed. On ADAS, we're seeing the same thing, right? The content is growing up. Over the last year, especially, you've seen an acceleration of features that everyone's trying to deploy. Incredible trend for us because what has been happening is we would win sockets with folks, and they would say, I have a top-tier car, a mid-tier car, a low-tier car. I need three separate tiers of parts. That's how we won the design. When you get to the launch, they want to take the top-tier performance and launch it in more tiers of vehicles, which is great. The more the market moves up, the better positioned we are.
Let's talk about BMW. You've referred to them as a sort of big customer. That's maybe for us to understand sort of why that's a big milestone for you. You focused on the Snapdragon Ride chips, the co-development of the stack for this market. Why is BMW weighing such a big milestone for investors to take note of? What is the volume implication of it?
Yeah, so let me address it in three parts. When we think about our ADAS business, the first is the chip for the ADAS enablement in cars. We've said there are approximately 20 OEMs that we've won that are using our chips for enabling ADAS. This would go all the way from BMW, who are using our chip and our stack, to some of the Chinese OEMs who will use our chip, but they'll use a China stack, to another set of companies who are trying to develop their own stack, but they'll use our chip with it. For us, business number one in ADAS is to win the SoC socket. The second opportunity then is to win the stack on top of it. The stack market, obviously, very interesting stuff happening. There are a few stacks in China.
Internationally, I think we are one of two or three stacks that are merchant. Then you have certain OEMs who are working on their stacks as well. The reason why the BMW launch is so important for us is it is a jointly developed ADAS stack. We've taken the assets from the acquisition that we did, combined it with BMW's asset, and we've developed this stack together. It's being deployed at scale in new BMW cars globally. The most important thing about something like a stack is you have to verify it on the street at scale. BMW allows us to do that. Once we've done that, the opportunity is to take that stack and sell it to other folks, other OEMs. We already have a couple of wins.
I think there's a lot of OEMs who want to get a merchant stack from someone outside China looking for a non-China stack. We would become the option for those OEMs. Very excited about the launch that's coming up and very confident. Obviously, we've done a lot of testing, very confident of the product. Looking forward to seeing the launch happen. The third part that's important is if we broaden the horizon on this and think of ADAS as really physical AI, this is the first step for us. The second step then becomes taking the chip and the stack and applying it to robotics, to drones, to humanoids, because a lot of the problem statements are the same. You need a lot of sensor fusion. You need low power. You need small form factor. You need wireless connectivity technologies. You need a very good camera.
You need object tracking and things like that. You need to be able to take a lot of input that's coming in and make sense of it as central compute. In some ways, everything that Qualcomm does well in one device would be a robotics end market for us. The reason why I'm excited about ADAS beyond the direct financial opportunity in auto is that it becomes the platform for robotics going forward. Obviously, that's very exciting for us.
I was going to ask you, so what you're saying is you have the stack with your own assets, plus BMW, you're allowed to resell the entire stack? Or do you have to go add, integrate with every OEM?
We are allowed to resell the whole stack. The way the business relationship with BMW works is you develop the entire stack together. They get to use all the technology for free for their cars, and then we get to use the entire stack for free to everyone else. We can monetize it.
Got it. OK. Robotaxis, a lot of excitement, enthusiasm around it recently. I mean, the easy question is, what is Qualcomm's sort of view on the market? How does Qualcomm play into this market eventually?
I think several parts to that question. Our view is clearly it's a good thing because that drives more demand for autonomous driving features and content. In our minds, there's going to be select pockets where robotaxis will take off in the short term. It doesn't change the overall auto market for a while. I think we have a great position in it. This will drive the need for more content, more capable chips, more capable stacks. All those things are positive for us. I'm excited about what's happening. As we roll through the next few years, it'll show up in our performance.
In the announcements of the 20-plus OEMs that you won, is actively engaging with robotaxi OEMs a part of that, or will that be incremental to the 20 OEMs?
We have some content in each of the everyone uses our chips. So even if you take any robotaxi examples, there's a certain amount of Qualcomm content in those devices as well. It's really the question of how much of our content is being used in those cars. Primarily on the ADAS side for cockpit and connectivity, we still have incredible opportunity in those cars.
Got it. OK. Before we transition off from the autos topic, just relative to NVIDIA and Mobileye, which are your key competitors, where do you think your capabilities stand today, particularly if I take a further sort of five-year view and say full autonomous driving? Obviously, Mobileye has been talking about it for a while. How do you think about sort of differentiation, particularly with your Ride Flex SoC on that front?
Yeah, we don't compete with them in connectivity. We don't compete with them in cockpit. The competition is really limited to ADAS. The Ride Flex concept, if everyone's not understood it yet, I think it's a key one to understand. What we've done is we've taken our digital cockpit chip and we've enabled ADAS on the same chip. You can take the same piece of silicon, and if you're a lower-end car, you get to enable ADAS with that same chip. As you go to a higher-tier car, you can take the exact same chip, put a chip next to it, the same one, and the software is completely compatible. They can act as backup for each other. You go to the premium tier and you want a more capable ADAS chip in that case.
Flex is a unique way for OEMs to leverage their investment in cockpit and enable ADAS features, especially the low-tier cars, which is, I think, a unique position in the industry. It allows us to kind of scale up from there, and it shows up in the design wins and the traction we're getting. Overall, when you step back and look at our competitive positioning with those two companies, we think we're in an incredible place. I think the scale of our investment in cars across all different technologies is very differentiating versus what they're doing. The SoC, in our minds, it's very clear that we have the best SoC in the industry. We're going to launch this stack with BMW, which will verify our stack credentials as well. I think we're in a very strong position.
By the way, today we released a white paper, doing it specifically for this conference. We released a white paper on automotive and ADAS and our position in it. If you're curious about the topic, you can take a look at it.
Will do. Maybe just a financial question on that front. You had strong performance in fiscal 3Q in the automotive segment. The guide for F4Q implies a moderation in the year-on-year growth rate. Our impression is those growth rates should start to track higher once you have more ADAS coming through from the pipeline into the revenue. Is that sort of a fair expectation? What does the timing of that look like?
Yeah, so it is a fair expectation. We had incredible growth rates in the first half of the year. It is a difficult standard to hold anyone to. You're not going to keep growing 60% a year in this market. We're super happy. As I said earlier, kind of the key thing is, based on the targets we've set, we're looking at a CAGR of 20%-ish over the next four years. We're very confident we can well exceed that. Just in terms of timing, the way our revenue ramp is working, we started with the connectivity portfolio getting activated, and that drove the revenue in the initial years. Over the last three or so years, it's mostly been driven by cockpit launches, and that will continue. Now, starting now going forward, we're going to have ADAS launches that overlay on top, and that will accelerate our growth rate.
That's how the pyramid is playing out over time.
OK. Got it. Let me just pause and see if there are any questions from the audience. Yep, go ahead. Just wait for the mic, I think.
Thank you. Hey, Akash. Thanks for taking the question. In terms of the AI proliferation, in terms of LLMs into the car, in terms of your conversations or how you anticipate it playing out, do you think that it would benefit the connectivity portion of the business first or the cockpit portion in terms of moving to 5G faster, enabling better connectivity, and pushing to the cloud or on-device?
Yeah, yeah. So, lots of layers to the answer. I'm going to try to be succinct so I don't keep talking. Let's start with cockpit. For cockpit, it is super simple. Using agentic AI as an overlay, combine it with the cameras that are in the car, combine it with RAG of the manual, of the car manual, and then connect it to sources like Salesforce that has service data and things like that. You create an agentic front end outside of really controls of cars. There's an amazing opportunity just around that, completely transforming how you interact with your car. I mean, today, in any modern cockpit, you still have to touch things while you're driving and doing it. There's no reason in the modern AI-enabled agentic workflows for you to have to do that. That clearly, the cockpit experience is completely going to change.
You can argue whether it's one year or three years, but tremendous opportunity there. ADAS, it's interesting what is happening with the stack. The stack used to be largely based on you drive the cars, you get the information, you train the stack on it, you improve the stack based on the scenarios you're getting, and making sure that it's a stack that is highly performant. What has happened with AI is now you're creating synthetic scenarios and improving the stack much faster in the lab without any car really driving outside. Just tremendous acceleration of how you improve your ADAS stack with the use of AI. What happens is when you adopt these technologies in the car, wireless connectivity, 5G performance becomes very, very important because some of these use cases are dependent on a wireless link to the car.
It helps all three things that we do in the car. If you kind of forget the wireless link for a second, most of the processing is happening in the car. It's a great example of the vision we've been talking about for a while of on-device AI. We think it doesn't stop there. It happens in the car. It will happen in any kind of physical AI, the robotics area that I mentioned. It's going to happen in personal devices. Then it is going to happen a lot in industrial devices. I think you're seeing kind of a front end of a very, very broad adoption of on-device AI. That obviously is a very strong trend for us.
OK. Maybe moving to IoT. Let's start with what are the big group of products that are driving the growth in the segment. You had strong results in Q3 as well. XR glasses are one. What are the other product segments you would like us to think about that's driving the growth? Given the strong results, there's always this overarching sort of concern that they might be a bit of a pull forward from the supply chain and trying to get around tariffs. How do we think about whether you're seeing anything on that front or not?
Yeah, yeah. As you know, what we call IoT is very diversified within itself. You kind of have XR, and let's categorize all wearables as a part of personal devices. You have PC, there's industrial, and there's networking. Those four are what are driving the IoT revenue stream. Let's talk about each one. Personal AI for us is kind of we are at this place where there is a massive transformation happening. What used to be the scenario is you put a phone in the middle. All wearable devices, including XR, were just an add-on to the phone. Increasingly, what is happening is, one, there is a very strong push from several players to make personal devices a separate ecosystem.
That is, and then glasses are the best example of it, that can see what you can see, that can hear what you can hear, and that can be your gateway for personal agentic AI experiences. It's the most logical form factor you can think of. There are people who are making lockets. There are people who are making a box that sits on the table with you. There are people who are making a watch. In each case, what you're trying to solve for is how does this become the personal device that knows everything about you, that becomes the gateway for personal agentic AI experiences. It's a device that can hear what you can hear, see what you can see. That's kind of a gold mine of information, user data, services opportunity, just tremendous.
Everyone who has a model, an LLM model, is trying to make that device and win that device. The fortunate place for us is they're all using our chip. We are the underlying common thread across all those devices. It puts us in a very, very strong position. We very much believe in this vision. The reason why we like this market is this is a market that requires very, very, think about glasses. For someone to make 50 g glasses, it needs to be super low power because the battery life has to be strong. It has to be lightweight. It has to be very small. It has to be integrated with memory and other things like that so that you can further shrink the form factor. You need very good wireless connectivity. You need on-device AI.
We showed a 1 billion-parameter model running on the glass, which is required, at least that much is required to do at least the bare minimum of an agentic AI interaction with the glasses. We're very excited about what we have there. It's a new device category. You shouldn't think of this as this is a watch market which has 20 million units, or this is a glass market that has 5 million units. It's really a new category of device that we'll see how big it gets. It has the opportunity to be very significant, and it's not in any financial forecast for Qualcomm. The second category I want to talk about is industrial AI.
One of the things that has happened over the last several years is, especially with phones, a little bit with PCs, entire companies got created by monetizing data that was created at the edge in those devices. Obviously, Google and Meta are the best examples of that. In industrial AI, a lot of data gets created, a tremendous amount of sensors. Camera is the easiest sensor to think of, and that data doesn't become actionable. If you think about security cameras, it's still someone sitting in a room staring at 10 TVs trying to figure out what they should look for versus just running an AI model on the camera. We're kind of at this very interesting place where the industrial market was already transitioning from microcontrollers to microprocessors. AI completely changes what you can do with it now.
In our mind, industrial AI is kind of a next 10-year trend, tremendous opportunity. As that happens, the market will automatically move from people who make microcontrollers to people who excel in edge AI. This is an example of an area where edge AI is important. It's not just like ADAS. You need to do it in the device, in the factory for privacy, security, latency, cost, all those reasons. We think we have a great position there. Finally, on PC, this is a market that we said, OK, there is a disruption in the device because of what Apple did. They brought the power of a cell phone chip, a smartphone chip into the PC and changed the performance metrics. We are doing the same for the Windows ecosystem.
We are very convinced, especially with everything else that's going on in the market, that this is a sustained trend that is going to happen, whether with Qualcomm or not, it is going to happen. We are best positioned to monetize it. What we've achieved so far, the step number one for us was to come in and establish ourselves as the best performing chip for the Windows ecosystem. We think we've achieved that, and we're going to double down on that in a month when we announce our next set of chips. Very excited about establishing ourselves as the performance leader. The second step was then to start building the go-to-market muscle, which, as everyone knows in the room, this market is not easy to get into. It's a very, very complicated go-to-market motion by design. What we wanted to do was establish ourselves very strongly in the U.S.
and Western Europe because a lot of the market sits there in retail. Enterprise, obviously, is going to take longer and kind of overlays over a period of time. We've done a very good job at that. I think we've talked about 9% share in the part of the market that we play in those geographies. Very excited about that as well.
And any pull forward from a perspective of timing, with OEMs trying to get around?
Not really tariff-related. I think as new launches of devices happen, and XR being an example where there is obviously a big launch that happened, the new Ray-Ban glasses, but also the Oakleys, there are builds that happen for launches. You see the benefit of that show up. It's not necessarily tied to tariff. It's just tied to the normal launch cadences in the market.
OK, got it. On the PC front, one of the questions I had for you was you've been referring to this 9% share, excluding the emerging markets. Some of the lower price points, I think below $600, is sort of what you exclude. Is that sort of what we should think is embedded in your $4 billion revenue guide as well? Do you incrementally cover the white space in that emerging market as well as that lower price point over time?
If you look at the dollar SAM, there is a unit SAM and there is a dollar SAM. As you know, the way we think about the cell phone market is you have to win in the dollar SAM. We're by far the leader there. Our view here is the same. You have to be the performance leader, and you have to win in the dollar SAM addressable market. Very much everything we're building is focused on $500 and above. Of course, there is a very easy opportunity for us to just take an existing phone chip and apply it to the $300, $400 PC market. We will do that eventually, but that's kind of later in our priority list. We want to make sure that we establish ourselves at the performance end of the metric.
OK, great. Let me transition to the data center opportunity that you've talked about. Maybe starting off, how should we think about how much of this opportunity for Qualcomm would be CPU-related versus NPU-related versus maybe even connectivity?
Let me step back and talk through how we're thinking about the opportunity, why we think it's an opportunity for us, and then kind of stay away from specifics for now because I'll stick with what we've already disclosed. We have, in our minds, the best CPU on the device, best custom CPU. We've taken—we have an architectural license. We've built our custom CPU. We've proven it in phones. Now we've done it in PC. Data center is the next logical step. The idea is to take that, scale it up, and make it available for data center. As all of you guys will recall, when we bought NUVIA, the market they were actually focused on was the data center market. We redirected them to phones because that is what was important to Qualcomm overall.
Now, since we've executed on that, we're going back to the opportunity that they were focusing on. That's very logical. The foundation of it is the performance of our CPU. It's that simple. If we can go into positions where we have a performance advantage, then the incremental cost to us is limited. When I think about it financially, the CPU is already being done for other reasons. Then we're taking that and applying it to a different market. It's an incremental cost metric for us rather than an absolute cost metric for us when we think about it. The CPU market is split into the classic CPU market. Now with AI, you need the head nod for that as well. That is a market that is expanding by a lot. It's not a mature market anymore because of multiple reasons. We have the opportunity to play in both.
That's opportunity number one. The second one is the AI opportunity. Inference is what we're focused on. You should just think of it as we have core assets that apply to the data center market. If you're looking to do things where power is the absolute requirement and you want to be very low power, several use cases are about power performance. We can bring our NPU from the edge to the cloud. That's the play. Clearly, one of the important assets in that area is having connectivity in the data center. Especially for inference, that's important. Our acquisition of Alphawave will add to the capability that we already have on the NPU side.
The focus is, can we combine those, work with data center first-party workloads where we can focus on a specific problem statement and build a chip that is highly performant versus the alternatives that they would have? That is the opportunity we're chasing. As we've said, we have kind of one large customer that we are very engaged with. There are a couple of other conversations, one with Humane that I'm sure you saw, saw that we have an MOU. There is work to do in really kind of translating that into a product and revenue. We also had a press release on NVLink as for us being able to integrate the other side of the link into our solution so that it becomes relevant to the Nvidia ecosystem of products. There is a lot of stuff we're doing.
The core premise is relatively simple. The core premise is we have a very strong technology base from the device side. How do we leverage that technology into specific problem statements in the data center? How do we look at it as incremental investment but then exploiting a very large available SAM for us?
Yes, sir. Just for me to be clear on what you said. Our focus in the data center would start with ARM CPUs. Yeah, thanks for taking my question. In the data center, we'll start with ARM CPUs, and then we will work on the full ASIC solution, all under the NUVIA team. Is that correct?
The team is not really relevant, right? It's a Qualcomm team, and the team that came from NUVIA is a CPU team. We obviously have a lot of accelerator expertise, our NPU on the device side that is independent from the NUVIA conversation. That's in every chip that we make, right? A very, very broadly used asset in our portfolio, which is relevant to the data center as well. I don't know if it is this one and this is two. I wouldn't think of it like that. I just think of it as two separate opportunities. Both are relevant. We are in discussions with a lot of customers. As we said before, when we have news to share over the next several months, we're going to, we'll communicate that.
Have you talked to when we should expect some revenues?
Yeah, we said fiscal 2028 is what we've said for now.
OK. Did we say whether it's CPU or ASIC?
We did not.
OK, thank you.
Thank you.
OK. If I can just follow up to the revenue timing question, I understand that a large hyperscaler engagement takes longer. In some of the other engagements that you're highlighting, is there opportunity to see revenues earlier? Is that just the roadmap in terms of your stack being fully ready to sort of commercialize? Even with the other customers, it's probably going to be fiscal 2028 or beyond because you have a stack.
I think for now, I'll stick with the timeline we've talked about. Of course, we have chips that are already relevant today. We've been in the accelerator market for a while, so that could be a way to accelerate it. I think for now, let's focus on what we've said.
One of the other questions I've got from investors recently is how dependent is Qualcomm on the IP from Alphawave to pursue the data center opportunity? As in, if there was to be a roadblock on the acquisition, how can Qualcomm still pursue this opportunity independently?
At this point, we expect to close the transaction in the first quarter of next year. No change to the assumption we've laid out. As you guys know, Alphawave is a licensing company. They have licensed their IP to various folks. That's always been an alternative for anyone who's playing in the market. Qualcomm wouldn't be any different. The key asset for us is think about the number of chips we have, our expertise in process, low power. We have an NPU. The incremental asset would be the connectivity asset. I think there's a lot of assets that we have independently. The CPU is obviously a kind of a separate conversation. Of course, it leverages connectivity as well. The product there is obviously more concentrated on assets we have today.
OK. One of the questions I had for you is, and this is my view looking from the outside in, is most of the semiconductor companies today playing in the AI space are ones that are proving their compute capability or leadership in training and then looking to extend that into inferencing, whereas Qualcomm's strategy being we don't want to really go after the training market. We will intersect the market right at inferencing. Does that not pose certain hurdles in terms of seeing adoption? Do you not have to prove your leadership in training to get access to the inferencing market?
Yeah, so think about what we do on the edge today. A lot of things that are run on the edge, pretty much all the workloads are inference. They are trained on NVIDIA silicon, and they are run on our silicon at the edge today for inference. I don't think that necessarily, I think in my mind that carries over to the cloud as well. You're going to have workloads that are trained on NVIDIA silicon and running on other silicon. We'd be one of them. We don't necessarily think we need to go into training first to be able to go to inference. It's really inference is our strength, and we're going to play to it.
OK. Any other questions before I switch to handsets? OK. Let me continue here. Let's switch to handsets for the last few minutes. Your guidance for fiscal fourth quarter, that's mitigating your growth, that did come in better than expected by investors and ahead of what seasonality expectations from investors were when you adjust for the share loss with Apple. Just help us understand the drivers there, what's embedded in sort of the fiscal fourth quarter or September guide, and maybe why isn't there a bigger headwind from the Apple share loss?
Apple is what it is, what we've already said. This is really a kind of an Android conversation. As we've said in the past, there are three big things that are happening in the Android market that have helped us over the last, I'd say, two, three years. Content growth. We've seen a lot of increase in content, especially at the premium tier, which is where we are strong. The second is mix shift up. People just continue globally to continue to buy more expensive devices when they buy devices. I mean, Samsung just launched the Fold, very strong initial traction for the device. It's just an example of what is happening in the market. The third thing is, especially in China, the Android ecosystem has gained share versus Apple. That's highly accretive for us.
We've said in the past that the content opportunity for us in Android is multiple, several multiples of what it is in Apple. Those three things are strong. Those are showing up in our numbers every quarter. That's just a reflection of the market that we have. The one other factor that did help us is we are launching our new premium-tier chip sooner this year versus last year. OEMs are going to launch phones right away, Xiaomi being the first one right away with our chips. You're seeing the benefit of that happen within the quarter also.
OK. Got it. On the Xiaomi front, you have the long-term agreement that you've signed with them. You have a long-term agreement with Samsung as well. That assures you a leading share position on the premium end. Is that something we should expect to become standard practice for you across all OEMs that you engage with, to have a more longer-term contract, to have more stability in terms of share?
To be honest, especially at the premium tier, the engagement with OEMs is always long-term. On the technology side, it's always been you have to plan two, three, four years out. There are times when we have a contract that I think makes everyone more secure. Even when we don't have a contract, the engagement is several years out. I'd love to have contracts, and I'd love to have predictability in the business. I don't necessarily see it as a required part of how we operate and something that we want to actively pursue with every OEM. There's a different conversation with every OEM, and we'll adopt accordingly.
OK. You have customers like Xiaomi still pursuing in-house solutions at the same time while they're assuring you higher revenue year over year. How do you see that playing out? Do you see them sort of pursuing those in-house solutions more at the premium tier, or is it more about addressing some part of the low end that you really didn't sort of anyways were not going to be interested in longer term? How do you think about that roadmap playing out for some of those OEMs?
You know, it's no different than competing with internal solutions in the past, right? Like we've done it with Samsung for several years, and what we've ended up with is a majority of the share in the premium tier, and the minority has been for their internal platform. The reality is in this market that is so competitive, there are two things that are very important for every OEM. First is to keep up with Apple and have the best chip in the market. Second is to not just grow within China, but to export globally. Both those things lead OEMs to Qualcomm, and we like our position. For us, it's relatively simple. You have to lead in technology, and then the rest happens as you want it to happen.
OK. Let me do a quick check before moving to the final questions here if you have any here. OK. Let me wrap it up with the final question since you have a flight to get to. Margins for QCT going forward. One, how sustainable is this sort of 30% level in QCT, particularly with Apple revenues coming off? If I add one more nuance to it, you have highlighted that once Alphawave comes in, there's an incremental OpEx associated with that. How should we think about 30% being sustainable in those two scenarios?
Yeah, so if you look at the scale of Apple revenue and look at the scale of the growth opportunity that we've outlined in IoT and auto, forget handset for a second. We think handset continues to grow very strongly as we have over the last several years. The opportunity for growth is multiples of the revenue we have at Apple. I think when you kind of step back and look at the next three, four-year horizon, we expect to have strong growth rates even with the ramp down. In the end, that is what matters for an operating margin target, right? Timing across quarters and across months may move around. I think when you step back and look at the broader spectrum, I think we have an opportunity to continue to do very well on margins as we have done over the last year.
Does Alphawave-related cost add any sort of near-term headwinds on that front?
I mean, on the cost side, as we've said in the past, our framework is pretty simple. We are reducing investment in our existing mature markets, moving that investment over to diversification areas. We're adding people where we have new skills that we need, right? Alphawave would be an example of new skills to get into data center. Of course, as we do the data center chips, that increases our OpEx a bit as well. That's the framework, right? We're not looking to really kind of the core technology scale of the company exists regardless. It's really these diversification initiatives that add OpEx on top.
I will wrap it up there. Thank you. Thanks for making it to the conference. Thank you, everyone.
Thanks, everyone.