Great. Good morning, everyone. I'm really glad you're here. I'm Jeff Palmer, SVP of Investor Relations at NXP, and I'll be your host for today's event. We're really glad to have you here at the NXP 2021 investor event. First time in about three years we've been with you, and the first time in about two years that we've been at an actual in-person event. We're really excited to be here with you today. I've got a few housekeeping items, and I know I'm not gonna read this slide, so don't worry, but I've gotta make a few comments. First off, this event is being simulcast on our website, both audio and video. After the event today, I will post the slides to the website so you can download them for reviewing later on. Today's event will contain forward-looking statements.
Those forward-looking statements do reflect NXP strategy, NXP's new products, the revenue and financial performance we anticipate from those new products and our strategy. Remember, NXP undertakes no obligation to revise or update any of our forward-looking statements. Today's presentations and Q&A will also reflect non-GAAP financial metrics. In the appendix of the presentation is a reconciliation of these non-GAAP metrics to their closely correlated GAAP metrics. You can find that in our appendix and more details on the NXP investor relations site. We've got a good agenda today, and it's kind of a little, kind of packed tight, but I think you'll enjoy it. I hope you're excited to be here. We're gonna kick off today with Kurt Sievers, our CEO, who will give you a sense of how we anticipate to accelerate profitable growth over the next number of years.
For the morning, we'll have the general managers from the automotive team come up and give you their insights to the major drivers in their businesses. We'll do a group Q&A with the auto team. We'll take a short break, and then we'll come back, and we'll review our industrial and IoT business, our mobile secure ranging technology, and our communications infrastructure business. Bill Betz, our new CFO, will come up and give you a new financial model, which I'm sure you're all excited and can't wait for. We'll have some closing remarks and some more Q&A. We'll break for an informal buffet lunch, and we'll hope you stay and join us and do some dine-arounds with everyone. With that, let's start the event.
At NXP, we believe in a world that anticipates and automates to meet our needs.
Where we can realize our goals and pursue our dreams.
Enabled by technologies that don't wait for us to ask for help or feedback.
Sense us, support us, respect us.
We believe in our technology.
Technology that is aware, smart, connected, efficient and safe and secure.
Tested by our decades of experience and leadership.
We believe it.
Because we have seen it firsthand.
The opportunity for technology innovation to make an impact in our lives.
At home or at work, often called at the edge of computer networks.
In our cars, trucks, and mass transit.
In our wallets and pockets.
In our retail businesses and industrial factories.
We believe in delivering electrification, sustainability, and AI ethics.
We believe in what we do.
Building the processors and MCUs that enable smart devices.
To sense, think, and act.
Building the Wi-Fi, UWB, and other solutions that allow them to connect.
Building smart cities, smart homes, smart transportation.
Building smart lives.
We have a vision of a world that anticipates and automates to meet all of our needs.
We believe our people help make it real.
For everyone.
We are NXP.
We are NXP.
We are NXP.
We are NXP,
And we welcome you to our Analyst Day 2021. A very good morning and a very, very warm welcome. I can tell you I'm truly excited to be here because I'm here live and on stage. It's a bit special because it was just Monday this week that the travel ban for E.U. citizens to the U.S. was lifted. My colleagues and I, we were jumping on a plane on Tuesday, and we are here today. I'm really excited. It makes such a difference to not do this over Zoom. Still, I'm happy that the room here is full, and I know we have a couple of hundred more participants remotely and virtually joining the meeting. Also a good morning, good afternoon to all of you seeing and hopefully hearing us well on the phone and over video.
When I say we, I am very proud to have the complete management team of NXP here, including Peter Kelly. I know that Peter has been a good friend and a great CFO for many years, more than 10 years actually with NXP. Just a few days back, I announced that Bill Betz will take over as CFO from Peter Kelly. Now, it's nice that Peter is here, so you will have time in the breaks to shake hands and chat with Peter. It's even nicer that Bill is here, and Bill is, as the new CFO of NXP, is gonna speak about our financial model in a few minutes later. Now, what are we gonna talk about today? It's all about accelerating profitable growth for NXP. Profitable growth doesn't drop from the sky.
A lot of that has to do with what we've done, what we've created, what we've architected through our history. Just a little reminder, we are a leveraged buyout from Philips in 2006. That is the year when NXP was founded, and actually, we celebrated our 15th anniversary this year. We did IPO here at Nasdaq in 2010. Ever since then, we've been extremely focused and very busy with working our portfolio, optimizing our technologies, improving the team performance in order to become more profitable and to create the foundation for growth. A key step on that journey was our merger with Freescale. Back in 2016, we put Freescale and NXP together to create what we used to call the powerhouse in high-performance mixed-signal.
The real idea was to combine the processing leadership of Freescale with the RF analog and security leadership of the legacy NXP. You remember well, a little bit later, we had the attempt of Qualcomm to acquire NXP, which we exited in summer 2018. I was actually here on stage in Wall Street in September 2018, so just a little more than three years ago. I was named president of the company, and we laid out the future for NXP as a standalone company. The one piece which we were missing out of the failed deal with Qualcomm was connectivity. In order to build complete edge processing solutions, and you will hear much more about complete edge processing solutions, we needed connectivity, which is why we did close the deal of acquiring the connectivity assets from Marvell in 2019.
Now let's have a look on how we've done since that September 18. I would call that solid performance. We delivered growth in the range which we had talked about, 5%-7%. We improved our gross margin performance. We actually returned $12 billion to our owners. Probably most importantly for the future, we spent more than $6 billion in R&D. We spent that R&D to create our future, to create sticky, accelerated growth drivers, which we will speak about in the next couple of hours. When I think about the last three years, there is more than just the numbers. The last three years have actually exposed us to a market in turmoil. In 2019, we saw a pretty sharp and unexpected decline in car production in China for the first time since China has been producing cars.
Early 2020, we and everybody has been hit hard by the pandemic, both relative to the ways of working, but also relative to how the economy came down rapidly. Even faster it came up again in the later part of 2020. Since then, we are seeing a very fast, unprecedented rebound in all of the markets we are serving. I think this is much more than just a short-term cyclical rebound. I believe it is a transformation of the value of the semiconductor industry, of the way how the downstream industries are looking at semiconductors as an incredibly important and critical part of infrastructure. I believe it is a value reset for the semiconductor industry, which is here to stay. It is not going to go away. It's a reset, and it puts the semiconductor industry at large in a very different light.
Now, what it means to us is that obviously, given the supply shortages, we have been and we keep going through very stressful times in order to continue to optimally serve our customers. Stressful is one way to look at it. The way how we look at it is it gives us the opportunity to create relationships with our customers deeper, more intimate, and more strategic than ever before. I can tell you I've had a higher frequency of customer meetings by a factor of five, easily five, as compared to any year earlier in my career. The same holds true for anybody in NXP.
Thanks to an outstanding performance of the entire NXP team when it comes to responsiveness, accessibility, resilience, lean-in attitude, and execution capability, I think we are building out of this, we are leveraging this situation better and stronger customer relations than we've ever had, not only with our direct customers, but in many cases with the customers of our customers. We gain much more insight in their plans. We get much closer to their future designs. In the end, that comes back in a design win performance and customer loyalty, which I think is a great achievement and present out of this current situation. Now, what is that opportunity going forward when we speak about growth? I look at the semiconductor industry as an industry which has been driven by mega applications in 10-year cycles, driving the growth.
Think about 2000-2010, laptop computers, desktop computers, game consoles. 2010-2020, smartphones, tablets, and cloud computing. Now, the next 10 years are the 10 years for NXP because it is the 10 years where we will see the rise, we could even say the explosion, of smart connected devices or the secure edge. I am deeply convinced that edge computing is gonna be that mega trend for the next 10 years, growing the overall semiconductor market. It's not just a future projection, but we see it live and happening right now. There are projections about 75 billion smart connected devices by 2030. Now think back to what I said about the merger with Freescale, the acquisition of the connectivity assets from Marvell. All of that was geared to build the best portfolio in industry for connected edge computing.
Now let's go a little bit deeper into this edge computing. What is it? It is all of these applications which are connected to the cloud, but close to people, spanning across the smart home, industrial automation, personal electronics, and certainly last but not least, transportation, formerly labeled automotive. You might wanna say the car is probably the ultimate edge device when it comes to complexity. One of my colleagues is gonna speak later about a factory on wheels, which I think is a very nice way to speak about cars. Now, what all of these edge devices and edge applications have in common is they collect data, they process and take decisions, and then they do some form of actuation, because that's the bridge to the human and to the physical world, while they are connected to the cloud.
There are lots of great numbers how things are evolving. We will speak about electrification in cars, where this year, I think we come close to 20% of the cars being electric, going to, like, 60% in 2030. Automation levels of cars. 5G is spreading. The smart home volume is finally really taking off with a 20% volume growth. What matters to NXP here is the fit to our capability. Here is the result of the portfolio which we have crafted, architected over the last five to 10 years. All of the edge applications have a signal path which starts with some sensing of the environment. Don't think now about historical sensors. Think more about voice recognition, face recognition, secure ranging, that kind of sensing.
It takes a great deal of processing performance, local processing performance, at the lowest power levels possible, given that a lot of these applications are wireless. Machine learning is gonna play an ever-increasing role to take the right decisions. Connectivity is clear. Connectivity to the cloud, I mean, there is no debate about this. Finally, actuation, because that's the bridge to the physical world. All of these edge devices, they will actuate, they will do something in the end. Those four pillars are not enough, and now comes where NXP has an unrivaled precision. It also takes cybersecurity and functional safety. Cybersecurity, I guess, is intuitively very clear since all of these applications are connected to the cloud. For privacy concerns, for making sure they cannot be intruded, they need to be secured at the highest levels.
Functional safety might be a little bit more less known to some of us. What this is all about is resilient system performance. Resilient system performance for robotics applications. Many of these applications go into higher levels of automation. It's not just the automatic car, the autonomous car, which is always the prime example. It's also about the vacuum cleaners, which are doing the job on their own. What they all need is, since they are robots, they have to be designed on a system level in a way that nothing can go wrong. Now, cybersecurity is in the genes of NXP. You know our history of mobile payments, of eID technologies, but it's the same for functional safety. We have been leaders in automotive safety systems like ABS systems, airbag control systems for many, many years.
That know-how, that engineering know-how at NXP, we are spreading now to those applications. There is a last element, which is a significant differentiator in doing this right, which is scalable system solutions. We are serving 10,000 of customers with these applications. Many of these customers have absolutely no way to cope with the enormous complexity of these applications, connected applications, cybersecurity. This is truly complex stuff. The best thing we can do as NXP is give them turnkey solutions, reference designs, and a perfect software enablement. It is those six factors which you find on this slide, which put NXP in a very prominent and special position relative to differentiation to our competitors in this rising edge computing opportunity. We have the muscle, we have the capability, and again, it didn't fall from the sky.
We've been building towards this for many years. Now, the other lesson I have learned in semiconductors is the opportunity is always bigger than what you can actually do. The name of the game is focus and sharp choices. We have a very rigid system in the company how to prioritize our R&D investments. It's following something which in principle is simple. We want to get as much as possible from our R&D investments on high-growth segments which are offering an opportunity which is above market growth. Secondly, and equally important, we have to have a tick mark on both of them. We have to have a fair chance in our plans to hit a relative market share leadership position in those segments.
Relative market share leadership, meaning we don't wanna be just number one, but we wanna be number one by a distance to the next competitor in order to be financially able to out-innovate our competitors, to build a sticky position with a roadmap which is better than any competitor because we can spend more absolute dollars in the space than anybody else. Now, we've done that very rigidly, and it's a pretty brutal process because that also means saying no to many things. We can say today that 75% of the R&D NXP is spending currently is going to segments which offer growth opportunity, which is above market. Secondly, more than 60% of the current revenues of the $11 billion which we make this year are actually from relative market share positions above one.
In more than 60% of our business, we are not only a number one, but a number one by a big distance to the next competitor. My colleagues, they give you many great examples for this in the next hours. Now, if we put this all together, the opportunity of the rise of the secure edge, the capability from a portfolio and technology perspective, then you get to our growth plan. We project to grow to $15 billion in 2024. The way how we wanna speak about this today in much more detail is actually by segmenting it into two pieces to start with. We have a very strong set of core businesses with high relative market share positions, which we will grow at market 5%.
Then we have a very fast-scaling set of accelerated growth businesses which are growing at 20%-25% CAGR, and they will grow to almost 40% of the company by 2024. In the next couple of hours, almost all of the focus of our presentations are geared to those accelerated growth drivers. We will make sure you get a better understanding what are we doing in that part of our business, which sits on a very strong core business. Now, my colleagues will do this in the next hours, but I wanna give you a sneak preview on each of our four segments. I wanna get started on automotive, where especially this year, there is no more doubt that there is a accelerated content growth opportunity.
That accelerated content growth is a result of three megatrends, electrification, further levels of safe driving or automation, and connecting the car and achieving a software-defined car. Those three trends are massively driving the semiconductor content in cars going forward, yielding a market opportunity of 7%-13% growth. This is ahead of what the market did in the past years. Now, NXP has been for years a very strong leader in the automotive semiconductor market. We will focus today on three elements in our accelerated growth businesses, which are really leading the pack to achieve this 20% growth of what you see what is a very significant part of our overall automotive business.
Those are radar, which is a key element for safer driving, a key element of a Level two, Level three ADAS system. Torsten Lehmann, the General Manager of that business, is going to speak about how we are leveraging a 40+% market share position which we have today. We are number one, 40%. We are extending that position. Torsten is going to explain you how we grow that business to more than $1 billion by 2024. That's one of the strong pillars in the growth drivers for automotive. The second one is electrification. We've had a lot of discussions with you in the past years about battery management solutions and what success we've been driving there. Now, I kind of anticipated in earlier calls with many of you that NXP is actually having more under the kimono than just battery management solutions.
Jens Hinrichsen, is going to speak about adding e-motor control systems to our position of strength in battery management solutions. He will put the two together and will lay out a plan which gets us to $500 million revenue in electrification from these two systems alone in 2024, which is a 30% CAGR and way above market. Thirdly, and probably most fundamentally, we are helping the car industry to transform the architecture, the electronic architecture of the car. This is where we speak about our S32 microcontroller and microprocessor solutions for domain structures and zonal structures.
It's actually just this week, and I'm really proud of this, we made a nice press release together with Ford Motor Company, where we laid out, and Ford is attesting us, how they are using for the new F-150, for the new family of Broncos, for the E-Mach, how they are using our processors for their network performance and for their infotainment. This is a big business for us. It will grow above $3 billion in 2024, and we're going to grow it 10% and 12%, and Henri is going to give you more insight in a few minutes from now. All in all, the automotive opportunity is growing, the market has accelerated, and we keep outgrowing it by a factor of 1.2. The next segment in order is industrial and IoT.
It is probably the showcase segment for what I said earlier about system solutions, about leveraging our processing leadership to pull through analog attach, connectivity, and security. I mean, the numbers speak for themselves. Ron Martino, the GM of that business, he's going to lay out how we outgrow the industry here by a factor of 2.3 going forward, 9%-14% growth based on those solutions, but also based on very, very good customer intimacy. In that market, the customer intimacy is actually twofold. We have a very, very leading position in the channel. A lot of that business is going through the channel where we have a prime position. We are also building very strong relationships, and that is helped by the current supply situation, to be honest, to the big OEMs.
Ron is going to speak about recent success with Microsoft, recent success with Schneider Electronics, which are giving great evidence on what we are doing in this segment. Industrial and IoT, 9%-14%, just like automotive. Third segment, mobile. Very different characteristics. We have a laser-sharp focus in mobile. We've been building this over the years by driving, leading, and now exploiting the ecosystem for the mobile wallet and for mobile transit ticketing. The way we've done that was all about building the ecosystem based on a tripod of competence, the hardware secure element, the NFC radio, and the associated software. We've done that very successfully under leadership of Rafael Sotomayor, to an attach rate of the mobile wallet of 50% by this year.
We laid this out as a target, I think, in 2018, and we hit the 50% mark this year. Now, this is going to continue growing, but what Rafael is going to focus on today is what we call secure and mobile ranging. It has characteristics which are similar to the mobile wallet. You could actually say it's an extension of the mobile wallet. The whole idea is to just like in the mobile wallet, where we took the coins and the credit cards out of your pockets into the mobile phone, now we want to take the keys out of your pocket. We want to facilitate mobile access solutions. We are in great shape. This business is going to go across IoT and automotive.
It's not just mobile, but it's also going to sit in IoT and automotive, and it's all based on the technology called ultra-wideband. The mobile side of it is gonna grow by 70%, as you can see here. If you put automotive and IoT into it, and Rafael is gonna present it that way, it's actually an 80% growth business, landing us at $400+ million in 2024. That gets us within mobile to an 8%-10% CAGR over the next three years. Last but not least, comms infra. The focus today is gonna be on RF power. Our RF power business is under the leadership of Paul Hart.
Here it's really about a range of technologies for a range of frequencies in order to cover applications for this 5G super cycle from the macro over MIMO all the way to millimeter wave applications. 1 MHz to 100 GHz. Or from a technology perspective, LDMOS, gallium nitride, and silicon germanium. We have all of them. It's not only that we are capable to design in all of them, but we also have the manufacturing capability inside NXP to do all of these technologies, and that puts us into a number one position in the RF power for the network infrastructure market. Paul is gonna lay out how he will outgrow that market by a 15% growth rate over the next three years. In that same segment, we also have legacy businesses.
We talked about them a little bit in the past, the multi-core processors, the banking card business. If you then put it together, we project a 2%-6% growth for the entire segment. The focus today is gonna be on the RF Power part, which is growing at 15%. Now let me sum this up. If you map the four segments all together, we actually do accelerate the growth of NXP. I can go that far to say we almost double the growth of NXP going forward to a range of 8%-12% from 2021 to 2024. That is being led by automotive and industrial IoT, each of them 9%-14%, and together almost three-quarters of the company's revenue. Mobile, 8%-10%. Comms infra, 2%-6%. Again, taken together, 8%-12%.
That gets me to the investment thesis of NXP. Clearly, the core and the bottom of this is accelerated strong growth, and that's gonna be the focus of today. Obviously, we are complementing that with what I think is a proven reliable financial model, which does deliver very robust and very resilient profitability. I think the last years in a market in turmoil gave good evidence of our capability for that execution. Thirdly, we do stick to our capital allocation policy. In other words, all excess free cash flow is returned to our owners, so long we stay at two or under two net leverage. At this point, for me, it's time to actually let my colleagues speak and deep dive into the accelerated growth drivers.
I just want you to keep in mind, everything we will discuss the next few hours is focused on that portion of our business which will grow at 20%-25%, and which will be almost the 40% part of the $15 billion revenue of NXP in 2024. The first one we wanna get started with is actually inside the automotive segment. It's Henri Ardevol. That's the microcontroller and microprocessor business for automotive. Henri, I ask you on stage. Thank you very much. Here we go.
All right. Thank you, Kurt. I'm gonna take you on a short journey in terms of how vehicle architectures are advancing and what are the kinds of problems that we aim to resolve, what that means for our product portfolio, and most importantly, how we're working with customers, and what are the results that we plan to deliver out of all of this work. From all of the conversations that we have with the OEMs, clearly they are facing significant challenges, right? Significant complexity. They have to integrate 100 or more processors. The data is buried into single ECUs. They wanna have faster software innovation cycles, but it's tied to these very long hardware cycles, and so they need to evolve the architecture. It's not gonna happen just with an organic continuation of the way of working.
That evolution of an architecture, it's very costly, it's highly complex, and it's a long-term decision for any OEM. They are thinking of it in terms of platforms. The core of those platforms are the three domains that you see there at the top, vehicle dynamics, vehicle networking, and body and comfort. Vehicle dynamics means how do you organize all of the functionality around how a car moves, so your powertrain, your braking, your steering, your suspension, your chassis management. Vehicle networking is how do you organize the flow of data, all of the control signals and the connectivity to the cloud. Body and comfort is the broad interface of how the car reacts to passengers, crash detection, airbag, all kinds of motor control pumps and switches, HVAC, interior and exterior lighting.
These three domains form then a core infrastructure in this evolution of architecture, how the OEMs are thinking about it, on which you can plug additional functionality that is either with a faster refresh cycle or that is more scalable, and so things like infotainment and e-cockpit or ADAS, and indeed, Torsten will be speaking shortly about our planar radar. That core vehicle infrastructure needs to be rooted in the core attributes of automotive because it's what makes a car, an object that moves, right? It needs to be functionally safe. It needs to be secure, it needs to be reliable, it needs to be robust. That's our know-how. That's the core of what we do. However, it's also clear, and we recognized that four years ago, that the evolution of requirements of software and networking are gonna outpace traditional MCUs.
Four years ago, we started to refresh our portfolio and develop this S32 family of products, which is the core of what we're gonna discuss today. This slide kind of quantifies a little bit the challenge, right? I mean, OEMs are adding all of these boxes and laying all of these wires. You have 45 processors per car on average now. Mind you, that's a worldwide average. If you take North America or Europe or the higher-end vehicles, you easily have 100 processors. They're all driving their own network and networking requirements. You have CAN, and LIN, and FlexRay, and Ethernet, and Gigabit Ethernet, and multiple of those, and you end up with kilometers of wires inside the car. You end up with these very complex harnesses, very costly, that weigh as much as a couple of passengers.
Clearly, there's a physical side for the OEM in the requirement of evolving the vehicle architecture in terms of simplifying and reducing the cost of the topology of the architecture. On the software side, the growth is exponential, and it's very clear you're not gonna be able to follow that path if you continue with the traditional approach. You need a new category of processors. To get to this level of complexity, you need to move from the development of single functions to the development and the deployment of entire applications that cut across the vehicle. For that, you then need a platform of products which are scalable, which offer consistent services across the vehicle, so your safety, your connectivity, your security, how you manage over-the-air updates of the different systems within the vehicle.
Most importantly, that allow parallel execution of real-time applications such that you are able to not only coordinate those different domains, but you're also able to up integrate the functionality into single processors. OEMs start from this flat architecture, and then there's two transitions that they are thinking of, domains and zones. Domains is all about how do you simplify the software. It's about how do you move the management of the strategy for that set of functionality from the edge to the center. Zones is about simplifying the network. It's about reducing its cost, and it's about making the data more accessible. There is a requirement, both logically and physically, to simplify the architecture, and a good example would be the work that we're doing with Bosch on the next-generation hybrid powertrain and transmission control domain controller.
If you take this example of a hybrid vehicle, you have multiple sources of torque generation and battery storage, right? What you really want to get to is a place where you can coordinate the strategies of your internal combustion engine, your inverter, your battery management system, your chassis control, torque vectoring for each wheel. You wanna have predictive gear maintenance, predictive control. You wanna react to changing terrain and traffic conditions. You wanna have the ability to do data mining. For that, you need to have a platform that is then able to support these multiple, real-time applications in parallel that has powerful isolation features, such that on the chip, every application is accessing their own resources, but they are protected in hardware, and it has an efficient framework for data and resource sharing.
On the zonal side, it's all about simplifying the flow of all of this data. You install four zonal controllers, one on each corner of the vehicle. You deploy an Ethernet backbone. You are then able to connect your domain controllers to each of these four zonal controllers, and you can run much shorter wires from the sensors and the actuators from all of the end nodes to those zonal controllers. In this way, you dramatically simplify your network. You reduce the cost and the weight.
Again, these zonal controllers are another class of processors versus the traditional MCUs, and this is why we developed our entire S32 portfolio of products with this in mind, how to address the transition to domain, to zones, and to end nodes with the evolving requirements that OEMs are facing. By having this scalable and consistent portfolio with a consistent software architecture across the entire portfolio, we offer a lot of reuse opportunity to the OEMs, but also a lot of flexibility in how to develop their software and where to place their application at different spots in the processing architecture. How do we do this? We have this portfolio of processors that has a consistent architecture, which is the same across the entire product range, which is the same and consistent across applications and across technologies.
Whether you talk about microcontrollers with embedded non-volatile or high performance, 28, 16 FinFET, and now 5 nm . You see in this simplified block diagram, two classes of compute: application processor and real-time processing. That real-time requirement of automotive, that real-time control of the vehicle is a key differentiator of what you need to do in auto. The entire fabric of our system on chip is designed for that freedom from interference, to have virtualized containers on which you can develop your application. All of the resources, whether it's the cores, all of the IP blocks or all of the interfaces, are protected in hardware to ensure that your compute happens in a deterministic manner. We put a set of dedicated application-specific accelerators.
Whether it's a communication accelerator to virtualize 10s of CAN, LIN, FlexRay interfaces, or whether it's a high-bandwidth Ethernet, whether it's a radar co-processor, whether it's a flexible timer for motor control. Those accelerators then are able to offload the processors, the cores, such that you can have that guaranteed deterministic operation in real time. The same for functional safety. The entire fabric of the chip is built for ASIL D execution. All of the services across the chip, whether that's your reset, your boot, your clock, your memory, your buses, everything is designed to run in ASIL D. On the microcontroller side, there also, it's about running lots of application in parallel, but you have additional requirements, which are very fast wake-up time and ultra-low power.
Because on the end node, very often you have the requirement to be always on, to have an all, you know, an immediate reaction time toward the user. You can see that this S32 portfolio becomes very different from your traditional MCUs, where on the high end you have these really powerful virtualized real-time machines, and on the low end, you go to ultra-low power, very fast wake-up time, much simpler microcontrollers that then enable the proliferation of functions. On the software side, as mentioned, we have exactly the same software interface, all of the services, and the same user experience for our customers, which gives them a lot of flexibility, a lot of reuse. Software is becoming a much larger part of our product. It's, you know, giving us opportunity to add more value.
A good example would be, you take one of those powerful multi-core processors, you know. It has those multiple cores that are gonna be running some of them, you know, Linux applications, some of them real-time applications. You can have different units of those processors within your architecture. How do you coordinate all of those cores? We are developing middleware for intercore processing to allow exactly that seamless operation across cores and across chips to create a more virtualized complete unit from which, again, you can either coordinate the functionality of your domain or you can up integrate the functionality. Of course, we refer to this analogy, right?
You often hear the analogy, a car is gonna be a server on wheels because, you know, everybody thinks about the explosion of data, and that's correct, but it's flawed in depicting the challenge as being virtual, as you know, being something that you can simply resolve by software. Much better analogy is a factory on wheels because the fundamental desired outcome is actuation, and the fundamental constraint you have is real-time. A car is not an app. A car lives in the virtual world, in the physical world, not in the virtual world. It's not about how many software engineers do you have or, you know, how do you make them work better or do you have the right kind of software engineers.
For us, the kind of challenge that we are trying to solve for our customers is how do we integrate the software development with the SoC architecture and with their design of the network and of the architecture. It's really a system-level question that we are trying to address. We have such a breadth and such a long history of application and deep partnerships that we are in a pretty unique position, actually, to contribute to those conversations. This is why we developed this roadmap, which is integral across all of those transitions. This focus on software, of course, leverages also the broad strength of the NXP portfolio into complete solutions. Jens is very shortly going to be talking about electrification. This is an example on networking.
You see a board here, which is a reference design that we have shipped in the hundreds. It's a networking board. It's providing a production-grade basis for our customers to go into production. It has all of the pre-integrated software drivers for all of the elements on the chip. You know, back to adding more value, we provide concepts like a functional safety concept, which is at the entire system level. That really solves a pain point for customers in accelerating time to market. Of course, we also work with Jens in terms of optimizing the bill of material and having more of an analog companion chip to all of these processes instead of having disparate units. That represents for NXP on top of the processor opportunity, another $40, right?
Another 29 chips that we can sell on top next to the processor. Every networking board would be at around $60 opportunity. I can tell you that the first designs that we have won there, we are very close to capturing that amount of value and, you know, which is commensurate to what we offer to our customers. We also try to extend into the ecosystem. You see a set of partners there where we are working with them to go to the new use cases that OEMs want to get to. One to highlight is Amazon Web Services.
We have very strong partnership with them, where we have pre-integrated the stack, the cloud connectivity stack on our network processors, such that you can enable new applications for data integrity or intrusion detection by streaming data through the cloud. These kind of use cases are the heart of the software-defined vehicle that Kurt was referring. Hyundai Motors Group has selected the S32 to be the basis for their software-defined vehicle architecture. What our S32 product is going to be doing is managing the cloud connectivity as well as managing the networking inside the vehicle. What you want is you want to have the ability to access all systems inside the vehicle such that you can download firmware updates from the cloud and be able to securely install them inside the vehicle.
This is a process where the protocols which are being used, for example, are extremely time sensitive. You need to have dedicated accelerators in order to guarantee that over-the-air update in a fail-safe way. The same in terms of the connectivity with the cloud, you then go to Ethernet, and you need to manage the latency of those with dedicated accelerators. You need to have data compression engines, you need to have AI engines on the chip in order to do intrusion detection, for example. In this way, you are also able to manage the bandwidth, which is connected to the cloud. This data intrusion and vehicle data analysis is just one of the examples.
There's many more that we're working with partners here on the slide in terms of predictive maintenance and vehicle health. This is happening in real time as we speak. You have a few more examples here on the slide with the timelines also as to when these are coming into production. At the top, you have Hyundai Motor Group that we just discussed. By the way, Ford is exactly on the same kind of use cases of using our family of processors to manage both the networking inside the car as well as the cloud connectivity. This software-defined vehicle architecture, how does it look like on the end nodes? We can take the example of BMW.
They selected the S32, among others, to manage their lighting solutions, their interior lighting solutions based on a protocol called ISELED, intelligent, smart, embedded lighting, which allows for the creation of dynamic, ambient, and functional lighting. It starts with a BMW iX. You have four lights under each window, where they are visible both inside and outside. That creates the possibility for new scenarios, like welcoming when you get towards the car. We expect this kind of applications to have a very strong uptake, not just within BMW, but on a broader basis. What we offer here is not just the S32 micro, but it's also a production-grade software stack.
It's a complete solution, which is then managing the synchronization of all of these LEDs, the color matching, as well as the ability to cut across physical boundaries to connect one window to another window or the dashboard to a door when you open it. That's one of the examples where on these end nodes, you need ultra-low power, you need security, and you need networking. Why?
Because when you put your hand on the door of the car, when you swipe your foot under the bumper, when you wanna sit, with the seat already in the right position, on new applications like precise ranging and location that Rafael is gonna talk about with ultra-wideband, you wanna have an instant reaction, which means you need super fast wake-up time, and you need very low power consumption 'cause these devices need to be powered, and you want to save your battery life, and you want to have a cost-effective power supply and thermal management. Ultra-low power. Security, there's gonna be many more features that are gonna be developed after the car has left the dealer, and you don't wanna return there.
To do these updates, you need to have security on the chip such that you can have a consistent over-the-air architecture. Finally, networking. These updates, they need to be friendly for the vehicle architecture, right? I mean, no downtime. For this, you need to move to more advanced protocols, and that's at the core of what we do with the S32. Now, these products are also super versatile. If we go back on the processing side, right? I mean, they have this core functional safety. You have the capability of running applications. You have these powerful real-time parallel processing machines that we see them being used as a kind of a Swiss Army knife of domain control for new use cases. Audi selected the S32 for their next-generation ADAS.
There, in the higher end, the S32 is managing the communications and the safety for the big AI engines. On the lower end, it's actually running all of the assisted driving management. On the infotainment and e-cockpit, that field is developing fast, so our i.MX has a strong position. 19 out of 20 OEMs are using the i.MX in production today. Some of those are still in the early phase of the ramp, like the Ford that Kurt was talking about. It's only starting. There we manage the 3D graphical interface and all of the displays which are around the car. That trend of the proliferation of touchscreen sensitive displays to replace mechanical switches, and new applications like in-car vision, driver monitoring system, telematics, V2X.
They're a great opportunity for our portfolio of vision and display-based processes, and so they're gonna be a focus for our i.MX roadmap moving forward. Next, of course, to the continuation of low-end IVI, low-mid IVI and clusters, and will be a great fit not just for automotive applications, but also for industrial. Finally on vehicle dynamics, the Bosch example and we talked about that already. A great partnership, and they really intend to make full use of the capability of running eight parallel applications with these powerful isolation features, as well as dedicated AI accelerators that we have put on this solution for powertrain, for machine learning geared towards predictive maintenance. In summary, very healthy market growth in the coming period, 9% CAGR. Driven, number one, by electrification and hybrids.
That will still largely be a traditional MCU market in that period before the integrated domain controllers start to ramp materially. Two, networking and gateways, with both the acceleration of penetration of simple gateways to a very high number by 2024, and within the period, the beginning of the ramp of the network domain controllers, and you've seen some of the dates on the previous slide. Finally, the acceleration of electrification adoption by end users is definitely driving an acceleration of digitalization, which is driving an acceleration in end nodes. NXP with decades of application know-how, with the world's best functional safety, with a tremendous focus on execution, where we benefit not just from digital networking, but in the last three years we have had an outsized focus on bringing SoC execution capability.
We have built a scalable and complete portfolio that is compatible in software, and that really offers the opportunity to our customers to make a decisive step in these next generations. Our large size also allows us to make decisive investments for automotive specifically. This is why we've been able to transition our portfolio to 16 FinFET, and this is why we're able to now fund the first automotive-grade 5 nm in close partnership with TSMC. We have deep partnerships with OEMs and tier ones, and those then enable us to have a coverage of 75% of the projected revenue that you see here of $2.3 billion by 2024. That represents a 10%-12% CAGR.
That represents the opportunity for NXP to capture an outsized share of that significant market growth that we have here. Our portfolio, we are very confident, will continue to fuel growth much beyond the horizon that we have here. With this, I will now hand over to my colleagues who will deep dive into specific applications and start with electrification. Jens, thank you very much. Thank, Jens .
All right. Thank you, Henri. Also a very good morning from my side. I have the honor to guide you through our electrification business in the next 30 minutes, and let's get started. If you look at the light vehicle production market, you can see that the amount of electrified vehicles is constantly increasing. As already mentioned by Kurt, we expect that by 2030, approximately close to 60% of all the cars being produced have some sort of electrification. This is partly driven by legislation demanding strict CO₂ emission reductions, motivating car makers to entirely electrify their fleets. It is also increasingly really truly consumer demand driven since the performance of these cars are getting continuously better, as well as that the cost structure of the cars are getting more attractive.
If you look a little bit on the regions, you can see that explicitly China is leading the EV adoption. More than 50% of all the electric vehicles being produced are coming out of China. Europe is leading for the plug-in hybrids, but also rapidly scaling the EV adoption. In the U.S., of course, Tesla leading not only in the U.S., they lead worldwide. The large, U.S. Big Three OEMs are increasingly focused their XEV activities. If you look a little bit more detailed into the market, and basically look at the years 2020 as well as 2021, you can basically see that the EV launch did not slow down in the COVID-19 pandemic crisis.
You also do see that in the current supply environment, a lot of car makers, and this is a tendency we see, really prioritize the electric vehicles over other models. This is constantly increasing the demand and the growth of the electrified vehicles. Not only that the market is attractive and hot, no, the market is also at an inflection point, maybe even already beyond an inflection point. What I mean with this, that the market is moving from a niche low volume market towards a true high volume automotive mass market. For the carmakers to be successful in this environment, they continuously have to optimize the total cost of ownership, they have to improve the performance of these cars, they need to basically shorten the innovation cycles, as well as have to be really ready for the true high volume mass market automotive production.
If we zoom into the semiconductors behind this market, the market looks like this. As I have mentioned already in the investor day in 2018, we are not focusing on power discretes like the MOSFETs or the IGBTs. NXP is focusing on the power control solution. We are leveraging our strengths of precision analog, processing, monitoring and control, which is controlling the power of these applications. The market we are addressing is in 2021, approximately $1.5 billion in size, growing with a 23% CAGR towards $2.7 billion in 2024. The applications which are supported by these semiconductors are outlined to the right. NXP has a very focused approach.
Like mentioned by Kurt, we are starting our electrification activities, or better to say we started our electrification activities with a sharp focus on the battery management system, which has got the largest part of the market. I will show you in today's session on how we are also rolling out structurally and systematically our electrification scope along the applications outlined to the right. Let's leave the market for a minute and let's have a look into the technology. This is a powertrain of an electrified vehicle, and this is actually a hybrid car. As this is a hybrid car, it still has got a combustion engine and of course, an electric motor. Accordingly, you still need an ICE motor control unit. You also do need an e-motor control unit, which is basically the so-called inverter platform.
It has got an onboard charger, which is basically charging the car, so translating the AC into DC. Accordingly you need a DC to DC converter, which is connecting the low voltage network with the high voltage network. In the middle you find the propulsion domain controller. This is what Henri was referring to. This is the central compute engine, which is basically orchestrating the entire functionalities of the powertrain. Of course you need the high voltage battery. The high voltage battery needs a sophisticated battery management system, which is constantly monitoring and balancing the battery cells as well as the entire battery pack. Let's move a little bit more detailed in this application and see what kind of semiconductors are needed for this. You see here that across all the applications in this powertrain, a lot of semiconductors are required.
As mentioned before, there are of course also power semiconductors needed like the MOSFETs and the IGBTs or silicon carbide modules. We are not doing these activities. What NXP is doing is everything outlined here in green, leveraging our expertise in power control. These are the NXP automation, automotive electrification solution for electrification. They, across the applications, basically consist out of processors which are explicitly tailored for electrification purposes. These are precision analog front-end solutions such as the battery cell controllers. These are functional, safe power management solutions supporting the processors as well as in-vehicle networking, sensors and driver solution. NXP is providing all these solutions in a very sophisticated approach. I will update you today on the battery management system, building on what I have presented in earlier sessions. I will also outline today on how we are expanding our scope in electrification.
I will give you an idea on what we do related to the e-motor control unit, the so-called inverter. Along with the update on the battery management system, I will also show you on how we are constantly doing innovation in this field together with the carmakers as well as the battery makers and tier ones. I will therefore reflect also on the propulsion domain controller, along with this building on what Henri has presented early on. After we have looked now at the market and the basic technologies being needed, let's have a look on what kind of challenges the carmakers are facing in this very dynamic market environment. The carmakers have to constantly reduce the cost, so they have to simplify the architectures and reduce the bill of material. They need to find ways to do truly an automated assembly of the complicated battery pack.
The other critical element is they constantly have to improve the performance of these cars because the performance is increasing the consumer confidence in these kind of cars. That basically resonates in a further extension of the range and even faster charging capabilities. How do they do this? They need to put even more power into the battery and find ways to completely unleash the full potential of the battery power into the performance of the car. They need to find ways to introduce even higher voltages, and we see systems now up to 800 V , which is a critical factor for driving the charging capabilities. The faster charging is enabled by this.
If they do this, you can imagine the more power you put into it, and the more you wanna unleash this power, you need to have really increasing requirements under control regarding functional safety as well as accuracy and diagnostic features. The car makers are requested based on the strict CO₂ emission reductions to upgrade their entire fleets. To do this, they need to find ways to do this in a very modular approach or in a very scalable approach for the hardware as well as for the software platforms. Last but not least, not to be underestimated, we are expecting that there will be more than 30 million electrified vehicles being produced in 2024. The car makers needs to be ready for true high volume automotive mass production in this environment. This is basically where NXP comes into the game.
NXP is a top automotive semiconductor vendor with true automotive expertise. We do know what it means to ensure highest level of automotive quality, including longevity programs as well as high-end operational excellence. Our easy-to-use system solutions are scalable across all the electrification platforms from 14 V up to 800 V across all the brands and all the models of the car makers. On top of this, our high-precision analog capabilities as well as the high-end processing capabilities really enable highest leverage of the efficiencies of these cars. On top of it, like I have said, this is a really safety-critical application. Our capabilities and our strengths in functional safety, supporting the software as well as the hardware documentation on system level is a nice supportive element in this environment.
NXP can tick all of the four boxes you see here on the chart, and that makes NXP value proposition so very unique and very strong. Let's have a look into the battery management system, and let me give you an update on where we are. This is a typical architecture of a battery management system. Of course, it might vary car by car, model by model, platform by platform, but this is pretty much what you can find. You see here on this picture that the actual system of a battery management system is even more complex than I've shown in earlier simplified graphs. You find plenty of subsystems in it, and all these subsystems require plenty of semiconductors which do the power control I was referring to. All the semiconductors outlined there in green are provided by NXP.
These are NXP's embedded control and high-end precision analog solution for battery management. If you look at the cell module controller, that's basically the part of the system which is constantly monitoring each particular individual cell and doing the balancing. They do the voltage and the temperature measurements. The battery pack controller basically does the monitoring and control of the entire battery pack as a total and does on top of it the current measurements and the battery junction boxes, also disconnecting the battery in case of a malfunction. The central brain of the battery management system is the battery pack controller, and this is usually powered by one of the semiconductor microprocessors we have learned in the earlier session, basically based on the powerful S32K processor. They are running all the software and do the monitoring on system level.
They do the state of health check, the state of charge, and the state of function check. They also do the entire thermal management of the battery system as well as they measure the battery pressure as a functional safety requirements. It is also connecting the battery system or the battery management system to the outside world. The NXP solution is so very successful because it is scalable across all the platforms and especially the analog front-end solution, so the battery cell controllers, which you find in the battery cell module, this is basically providing highest level of accuracy and not highest level accuracy just at a certain point of time. No, really truly over lifetime without calibration. That's the key element for the efficiency of the overall system.
As we are providing the entire bill of material, we have a high level of component integration, of course, which is optimizing the system cost. As mentioned multiple times, we are providing functional safety on system level here because it doesn't bring anything if a component has got a specific functional safety requirement. It needs to be ensured on system level, and we provide the system level functional safety support up to ASIL B. We do not stop here. We work with the car makers as well as the battery makers and the tier ones to further improve and enhance the system. The architecture you have seen in the previous slide is potentially evolving to something you do see here.
There are basically three areas of innovation I like to highlight NXP is working on, and that's also the part where I'm building on what Henri has presented in the earlier session. A, we wanna connect the entire battery management system to the cloud. This is why we are introducing a connectivity unit, so basically a gateway that enables the entire system to leverage cloud-based services and does the over-the-air update of the software needed for the system. Regarding to the software, and this is what we outlined under B is, we are supporting a software app integration. We are offloading the entire software of the battery management system onto the domain controller, which is basically the central propulsion domain controller, and therefore, this one needs to be a very powerful engine. This is what Henri was mentioning.
This will be basically powered by the S32S microcontroller or microprocessor, which is doing the entire processing of the software for the entire powertrain, not only for the battery management system. They do the energy and range optimization for the entire powertrain, including basically the battery management system. This is the perfect example of how NXP is basically supporting the shift towards domain processing. As this is connected to the cloud, we can even leverage AI-based algorithms to do these optimizations. The way that we are enabling this software application integration is only possible because, like outlined under C, on the battery management system, we are basically introducing a battery gateway with a standardized interface, which is CAN FD.
This enables basically to offload the software and creates on the battery management side, the so-called software-free battery, which provides a lot of flexibility and simplifications of the car makers on the battery management side. Furthermore, we are also, this is more related to the connectivity of the cell controllers, we are working on introducing wireless communication technologies. We are basically developing multiple wireless communication technologies, which is simplifying the wiring harness and therefore enabling the car makers to do a true automated assembly of the battery pack. We have a Bluetooth Low Energy RF-based solution available, but as a leader in connectivity, we are also developing further, more robust and even enhanced communication technologies. All this will help the car makers basically to improve constantly the performance of the system, and therefore of the entire electric vehicle.
It helps to take the cost out of the total cost of ownership, and it basically also helps to simplify and to create a lot of scalability and flexibility. This basically helps the car makers to be ready for the true high volume automotive mass production. This also makes us the trusted partners of the car makers, the battery makers, and the tier ones in this triangle. We are designed in to 16 of the top 20 car makers, including SAIC-GM-Wuling, for the best-selling Mini EV. Let me explain that. SAIC-GM-Wuling is the Chinese joint venture of GM. The Mini EV, probably not so known to you or to people in the Western world, is a high-selling or top-selling, maybe even the top-selling electric vehicle in China. Keep in mind, China is covering more than 50% of the worldwide electric vehicle production.
As announced in earlier sessions, since we have won Volkswagen, where we are supporting them on the electrification of their entire fleet, this additional win of SAIC-GM-Wuling is a very important win for us because SAIC-GM-Wuling is, in the meantime, the third-largest EV maker in the world. This basically gives you also a lot of confidence that the revenue outlooks we are projecting for 2024 is underpinned solidly by serious and great design wins. Let me now give you an idea on how we are expanding our scope in electrification, even beyond the innovation part I have shown to you. Similar as to what we have done in the battery management system, we are now also focusing on, we are actually already working on this for several years, on the e-motor control unit, which is the second-largest opportunity of the addressable market.
The inverter is basically converting electric energy of the battery into motor torque. For this one or so, several semiconductors are needed, among others, also IGBT and silicon carbide power modules, which we are not providing, as mentioned now a few times. There are also a lot of power control semiconductors needed, as outlined in green. Also this one we offer in a complete system solution approach. Here, again, a powerful processor is needed based on the S32K platform, which is running all the software, which is supported by the PMIC and the IVN, as well as by several isolated gate drivers. These isolated gate drivers are not just any gate drivers. These are gate drivers tailored explicitly for the e-motor control unit because they have highest level of accuracy and special diagnostic and protection features.
You use in a system up to 18 of these isolated gate drivers. You can imagine, the more efficient and safe you can do the power conversion from the electric energy into motor torque, the better is the performance of the car, and therefore, again, the range of this particular car, while we do this in a very safe environment and ensure highest level of functional safety of the safety-critical application. Explicitly, the more accurate you can do this power conversion, the higher is the range of the car. The NXP gate drivers provide highest level of efficiency in the industry. On top of it, our system is supporting basically the IGBT power modules, but also silicon carbide power modules. Silicon carbide is becoming more and more attractive for this application because of their faster switching and their higher performance.
Faster switching also requires faster protection of these power modules, and the NXP Gate Drivers provide 20% faster protection than any other solution in the market. Of course, also here, we optimize on system level the functional safety, as well as the bill of material to make it cost efficient. More and more car makers are basically bringing the design of the e-motor control unit, so of the inverter, in-house because it becomes a performance critical application or is a performance critical application. Our ready-to-use system level tested reference design, which is also outlined here on this slide, is a perfect fit for them to shorten the innovation cycle as well as to improve the time to market.
On this one, we are collaborating explicitly also with Wolfspeed in promotion of their silicon carbide modules to bring basically jointly together the EV performance of the cars of the future to the next level. We have launched this application around 2019, and already nine out of the top 20 EV makers are using our solutions, again, including SGMW, so SAIC-GM-Wuling, BYD, and NIO. To conclude, the market both of these applications are addressing together, so the inverter is the e-motor control unit, and the BMS, is approximately $1.2 billion in 2021, which is the lion's share out of the entire addressable market of $1.5 billion I have presented in one of my first slide. This market is rapidly growing with a 20% CAGR towards $2 billion in 2024.
NXP is outperforming the market by factor 1.5. We are projecting approximately a $200 million run rate of both of these applications in 2021 sales, growing with a 30% CAGR towards $500 million in 2024. All this growth projection, as I have mentioned, is strongly underpinned with solid design wins, and this is what makes NXP so excited about the electrification business. I hope I could also get you excited about this electrification business as we are in NXP. Thank you very much. With this, I'd like to introduce Torsten, who covers the radar part. Torsten.
Thank you, Jens. Good morning, everybody. In the following presentation, I would like you to take away three main items. First, the radar market is very attractive, and it keeps growing very significantly in the coming years. Secondly, NXP is in a unique position as a clear number one within this market. Then thirdly, NXP will keep outgrowing this market significantly and thereby continue to gain shares in the years to come. I will elaborate more on all three items as we go through the presentation. Now, first of all, why do we talk about radar systems at all? Well, every year, about 1.3 million people get killed in car accidents, and another estimated around about 20 million-50 million people get severely injured in such car accidents. A horrible number.
Radar technology is a key technology to make driving safe, to avoid accidents, and to save lives. That by itself is, of course, a very meaningful mission. It drives the adoption of ADAS and radar systems, and that in turn, of course, drives a significant market opportunity. When we speak about a radar node, let me explain NXP's view on what is a scalable radar system. Every radar node consists out of a number of key radar components. At the heart of the radar system are the radar transceiver and the radar processor. Both have evolved quite significantly over the past several years. Radar transceivers can be one or multiple in such a system, contain all the receive paths, the transmit paths, the VCO, the ADC, et cetera.
Also radar processors have evolved big time over the last period, going to deep submicron technologies, integrating thereby vast amount of processing power, combined with dedicated radar accelerators and DSP cores, together with functional safety, together with security, and memory. Next to those radar core components, you have very important peripheral or support components or attached components, such as the safe power management and the in-vehicle networking components. All of this together makes a radar node, and NXP is really in the unique position to cover the full spectrum, cover the complete system in this space. Now let's zoom out a little bit on the broader ADAS evolution.
You're probably all familiar with the so-called SAE levels of driver automation, where in the past we had no automation, level zero, and step by step we are evolving to driver assistance, partial automation, conditional automation, eventually all the way up to level five to fully autonomous vehicles. As we evolve, you see on the bottom, there's more and more sensor content and more and more radar sensors, radar nodes coming to the vehicles. Now, what is important to note is there's two very distinct different segments at play here. The first segment is the classical vehicle ownership. So that's you and me buying a car, and that is where the vast majority of the volume is, and that is primarily happening in the levels one to level three. Then on the far end of the spectrum, you have mobility as a service.
This is where the Waymo, Baidu, Cruise of this world provide fleets of robotaxis. While technologically very exciting and also a great long-term revenue opportunity, it's volume-wise, not so much needle moving in the nearer term as we look to the next three years. If we look at 2024, then about 65% of all vehicles will have some sort of driver assistance or ADAS functionality embedded. This is progressing from the levels one all the way to the level three. 99% of the semiconductor value is represented by these levels one to level three. Let me zoom in a little bit why we are pushing towards these higher levels and also the interesting new segment level two plus evolving.
The radar adoption is very significantly accelerated by mandates across the different regions and regional NCAP ratings. You have a number of those as example depicted on the bottom. Many regions have, for example, issued legislation or five-star safety ratings for making features mandatory, such as automatic emergency braking or blind spot detection or vulnerable road user detection. This is really driving the adoption and driving to higher levels of automation. Within that, there's an interesting new segment of level two plus evolving.
Level two plus means it gives safety and comfort features very similar to level three, so for example, enabling autonomous highway pilot functionality, but it avoids the liability concerns of the car OEMs, because at level two plus, the driver has to keep the hands on the wheel, and the driver has to be able to take over at any given moment, while a level three system would do that fully by itself. The good news is from a semiconductor perspective, you need essentially a very similar, if not the same, perception system to enable a level two plus functionality. That means from a semiconductor content perspective, definitely from a radar semiconductor content perspective, we don't care so much if it's level two plus or level three. Level two plus provides an excellent and fantastic business opportunity for us.
Now, speaking of the perception system, every kind of automation needs a complementary suite of sensors around the vehicle. Most widely adopted are combinations of radar and camera sensors. In some vehicles, you might additionally find LiDAR sensors, although this is primarily really at the higher levels of automation to add yet another complementary sensor. Our view is that LiDAR is probably still cost-wise and form factor-wise a bit prohibitive for very high volume mass adoption over the coming years. Every of these sensing technologies have a key measurement or a key strength of what these sensors are really good at. If you take radar as an example, radar is great to measure speed, distance, and angle. A camera on the other side is great to do object detection, pattern detection, color detection.
Now on the other side, several of these sensors have certain limitations and might struggle with some environmental effects. For example, a camera can be blinded by bright light or not see at night or struggle with fog or snow. Likewise, a LiDAR sensor might have issues when it's raining or snowing. Radar, on the other side, has very insignificant limitations. A radar sensor doesn't care if it's night or bright light, doesn't care if it's fog or snow. A radar sensor works very robustly and very reliably as an all-weather capable sensor under all these conditions. Radar keeps evolving. Radar has evolved big time. In the past, in the early days, radar was essentially good at detecting big, bulky metal objects like ships or like detecting a single car.
We have long evolved from that stage, and today, radar technology is capable of drawing a precise map of the whole environment 360 around the vehicle. Some of the key technology ingredients that have enabled that were the migration from 24 to 77 GHz, from specialty technologies like gallium arsenide or silicon germanium to standard plain RFCMOS, as well as going from low channel count to high channel count advanced MIMO configurations, and from basic processing to high-performance processing with dedicated accelerators. With all of that together, we are today able to build very small, very power-efficient modules with a great range resolution and a very high angular resolution that reaches below 1 degree and LiDAR-like resolution. This right-hand side is where NXP spends its investment focus. Now, speaking of the market, I mentioned in the beginning, the market is very attractive.
Just in the past year-over-year market has grown by about 37%, expected to reach $1.6 billion in 2021. Then on the three-year CAGR going forward, we expect an annual growth rate of about 18%, reaching $2.7 billion by 2024. Within that, very important to note that the 24 GHz segment keeps declining and will phase out over time, while 77 GHz is taking over and is driving the vast amount of the growth going forward. That's where NXP anyway has a leadership position. Speaking about leadership position, looking at the market share picture on the right-hand side, NXP is the clear market leader. This is a great example of the true leadership that Kurt has mentioned in the beginning. We have a relative market share of about 1.4 in this field.
That means NXP is about 1.4 times as big as our closest competitor, a German company which we call Competitor A here. Then there's about five other companies in the market that share the rest of the market, none of which is anywhere close to NXP, nor even to Competitor A. Not only that we are very proud of our market share achievement, but it's really fundamental to our strategy because it allows us to spend more R&D dollars to out-invest our competition. That, combined with the deep technical know-how that we've built up over decades, gives us a significant competitive advantage and also a high degree of confidence in our future. Now, you might ask, what's really driving this growth, and is it all real?
The answer is yes, because there's three distinct growth vectors at play, and we call that the triple acceleration. First, there's simply the penetration of radar sensors within vehicles is growing, and that means there's simply more cars with radar coming to the streets. Secondly, you find more radar nodes per individual vehicle, and that is driven by the need for 360 sensing all around the vehicle. Thirdly, there's more semiconductor content, and thereby more semiconductor dollars, coming to the individual radar nodes. And that is driven by the increasing adoption of imaging radar and the drive towards precise environmental mapping around the vehicle. This is not an addition, it's a multiplication. It's more cars with radar, times more radar nodes per vehicle, times more semiconductor dollars per individual node, and that's driving the nice exponential growth curve.
The good news is, and that's what you see on the right-hand side, we're just at the very beginning of this S curve. If you take all automotive radar sensors produced in 2020 and divide by the global car production, then on average, we had about one radar sensor per vehicle globally. This is nicely growing towards two by 2024, and to three in 2027, and then eventually towards above five as we move forward. On top of the quantity increase comes then the additional dollar content as we move more and more towards higher level of automations and imaging radar.
Now, I spoke a little bit, I touched upon the evolution of where has radar come from, seeing other cars in the past, and I would like to also show you how we see the market evolving and segmenting going forward. As mentioned in the beginning, it was more seeing other cars, 24 GHz technology, and essentially we had 2D-capable radar sensors. With 2D, we mean the two measurement dimensions, where radar is the only sensor that can directly measure speed, and it also measures distance. That was a 2D sensor. From there, we moved to 77 GHz and eventually towards RFCMOS and advanced processing. Today's state-of-the-art are essentially 3D to 4D-capable sensors. When we say 4D sensing, then we mean measurement of speed, range, as well as horizontal angle and vertical angle. That is 4D sensing.
That allows us to not only see cars, but see pedestrians, see bicycles, see smaller objects in general. For example, seeing in low light conditions, a child standing between two cars or seeing a motorcycle right next to a truck approaching with high speed from behind and separating those objects in a far distance. Very important use cases. That is one of the directions where we see the market segmenting. One is really on this high-end side, seeing smaller objects, going to higher resolution, and going towards imaging radar for full environmental mapping around the vehicle. Then on the bottom end of the spectrum here, it's about this 360 cocoon around the vehicle, so it's seeing around the whole car. The buzzword here is corner radar. Multiple small modules around the vehicle.
As corner radar suggests, at least four, but oftentimes even six or seven radar sensors, as you might have additional gap filler radars to the side, et cetera. Also the long-range radar keeps evolving, not only front-facing, but increasingly also rearward-facing, going to higher ranges exceeding 300 or 400 m and more. It doesn't stop here. I mentioned to you we are trying to out-invest our competition, and we have the funds to do so by the revenue that we generate. On the right-hand side, just a few examples on where the journey is heading. In the bottom part, it's about higher integration, smaller footprint, one chip integration of transceiver and processor on a single chip, antenna up integration, capturing more value, integrating PCB stuff, and simplifying the application for our customers.
Then as you move up through these segments, it's about higher performance, longer range, new waveforms, advanced algorithms, applying AI and machine learning. We're by far not at the end, but radar keeps evolving and we keep innovating big time. Now let me show you how NXP maps onto these three segments. This is a simplified view of our comprehensive radar portfolio, which stretches across all of these different segments. I would say we are really unique in the complete lineup that we have, as we cover the whole spectrum. We have a full lineup of 77 GHz RFCMOS radar transceivers in full mass production and in very high volume running in 40 nm technology. We have a complete lineup of leading-edge processor processing solutions in 16 nm.
Really the leading-edge node in the market and covering everything from corner radar all the way up to imaging radar based on the same building blocks, the same Lego blocks, but different product instantiations for the different use cases. Newly coming up, you see that in the middle, we have a new one chip coming up as an addition. Taking the transceiver IP that we have in RFCMOS, combining that on a single chip with the processing capabilities for smallest footprint and highest integration level. That is coming up in 28 nm technology. On top, we have the state-of-the-art 77-GHz RF know-how, and best range, and we apply unique radar architecture and algorithms that allow us to implement solutions with minimal power dissipation and smallest footprint. Now, we have been traditionally very strong in the market.
I showed you the market share, and we are globally very strong. However, we always had a bit of a stronger center of gravity in Europe and the U.S. Therefore, looking at this one chip family, I'm very pleased to announce that here we add an additional strong partnership with leading ADAS Tier One DENSO in Japan. Very pleased with the endorsement statement of Takayoshi-san, one of the senior leaders of the DENSO executive team, where he really reconfirms the high expectation on the one chip family being part of a scalable system solution and boosting the radar performance in the smallest footprint, thanks to our state-of-the-art RFCMOS technology. Keep in mind this upcoming one chip family, it's already our third generation RFCMOS.
It's in 28 nm that enables smallest footprint and, smallest size radar sensors, and it, supports high performance both corner as well as front-facing radar solutions. We expect to see multiple instantiations of this one chip family in a single vehicle going forward. Now, coming back once more to the portfolio picture that I showed you, I would really like to, stress the message with you that we have the complete performance scaling solution. We reuse the same IPs and the same building blocks across these different segments. That in turn allows our customers to have a full software reuse, as also Henri mentioned earlier. That in turn gives fast time to market to our customers, and that translates into lowest total cost of ownership. This is a very important differentiation for our customers. This complete performance scaling up and down.
Mind you again, NXP is really the company that has the full system solution across the whole application. We have the transceivers, the processors, the one chips, the safe power management, the networking, and the software. Really in a unique position here, stretching it up and down, left and right, the complete thing. One-stop shop. Now let me briefly zoom in also into this imaging radar space at the high end. There we're really very excited to announce that we are able to boost the performance with high-resolution three-in-one radar, and we believe that this will give an acceleration to the adoption of level two plus vehicles that I mentioned earlier. Two innovations that are driving this. One is what we call maximum awareness. What we can do is simultaneous multi-mode scanning.
That means concurrently scanning the wide near field, the mid, and the far range at the same time across the four dimensions that I mentioned earlier. Measuring range, speed, horizontal angle, and vertical angle. That is one innovation. The second one is maximum efficiency that we achieve by applying our complete system solution. As depicted here on the right-hand top, we apply four of our transceiver chips, combine that with a high-performance imaging radar processor and the peripheral components. Out of that, we get 192 virtual channels, which are boosted by a factor of 12x efficiency gain of the raw radar performance. That allows us to achieve below 1 degree angular resolution and very smart and lean designs. In a nutshell, an easy way to remember this, it's four, three, two, one .
It's 4D sensing with three-in-one sensing capabilities for level two plus automation with below 1 degree of angular resolution. Very exciting stuff on the high-end side. Now, obviously, we don't do all of this in a vacuum, but we work very closely together with our customers. I mentioned the DENSO example, which is a great example of how we collaborate and co-develop with our direct customers, who are always the Tier Ones. We also work very, very closely with the car makers themselves. We co-define roadmaps, we co-specify future architectures with the car makers, and we are seen as a trusted advisor to the car OEMs. That is what we call the innovation triangle. Working closely with Tier Ones as well as with OEMs.
This has been a really great recipe for success, resulting in what we are really proud of. We're designed in with 20 out of the top 20 car OEMs globally, and that is a fantastic result and the innovation triangle is a key contributor to that. Now, wrapping it all together, I mentioned in the beginning, radar market, very attractive, growing strongly at 18% CAGR towards $2.7 billion by 2024. Remember, the driver is what we call triple acceleration. So more cars with radar, times more radar nodes per vehicle, times more semiconductor dollars per individual radar node. Within that market, NXP has a very strong and unique value proposition.
We are the ones with the leading-edge technology nodes, with a complete performance scaling, complete system solution, and complete software reuse, allowing our customers to achieve fast time to market and lowest total cost of ownership. That allows us to outgrow the market by a factor of 1.2x, resulting then in true leadership revenue growth. We expect to grow at 20%-25% over the next few years, reaching from $600 million this year to $1.1 billion by 2024. The good news is about 90% of that is already underpinned with design wins as we speak today. I mentioned we're designed in with the top 20, and we've also seen major share wins and adoption momentum through 2021, which gives us a high degree of confidence.
Summing it up, great market, great outgrowths, true leadership, revenue growth while we make a massive contribution to avoid accidents and saving lives. With that, I would like to thank you for your attention and invite my colleagues and Jeff for a Q&A session on the automotive part. Thank you.
Thanks, Torsten. Everyone, we're gonna spend about the next 15 minutes going through Q&A. I've invited up the three auto general managers and Kurt, and I will be the moderator here this morning. Let's first go with CJ. One second before. We're gonna have microphones being passed around.
Thanks, Jeff, and thank you all for hosting this live in New York. It's great to see you all in person. I guess two-part question. I guess, Henri, you spoke around system-level kind of approach. I'm curious, how important is the ability to scale down a 5 nm to that effort? And how much of a competitive advantage is that for you over time? Then I guess the second question, for all of you and maybe Kurt, bigger picture, can you speak to your ability to actually monetize software as a standalone business? Or is it, you know, something that is functional with the hardware side? Thanks.
On the 5 nm question, it's pretty essential for our roadmap. You know, people were kind of wondering a little bit the same four years ago when we decided to pivot to 16 FinFET. It is clear that these architectures are going to accelerate. It is clear that there is the need for up integration while reducing the power consumption. That is the essence of what 5 nm is gonna bring us, but by being automotive grade. We're not just planning to take the vanilla 5 nm from TSMC, which, you know, maybe you could do for some high-end infotainment system, but not for the kind of systems that we're describing here.
We already have two test chips that are out, with 5 nm, and we are in very active design and very active conversations with our customers. Now, in terms of revenue impact in the horizon, we're describing here zero impact of course, but in terms of the trajectory of our business, very, very important.
Let me take the question on software. It is an integral part of our value proposition. I think Torsten spoke very clearly what he's doing in imaging radar, which is all about algorithms. We eat more and more into that territory. Part of the solutions which Torsten's team is providing is more and more moving into software including those algorithms. Later today, Rafael will speak about ultra-wideband as a secure access solution for the car. Again, here, as opposed to the former keyless entry solutions, we are providing the software coming with it. My answer would be a clear yes, it becomes part of our value proposition, and with that, also part of our revenue stream.
Let's take the back here.
Thanks very much. Vivek Arya from Bank of America Securities. Thanks for hosting the Analyst Day. My very simple question, Kurt, is when you look at that 9%-14% growth CAGR you gave for automotive, you know, how much are you expecting from units and content and mix and pricing? Because if I look at the last few years, units did not grow. Right now, you know, at the start of this year, there was a lot of expectation of unit growth, did not happen. Content was very strong. Next year, again, there is expectation of unit growth. What if, let's say, industry units don't grow from here, right? What are the factors on mix and share gains and pricing that you think you have very high confidence in? Right.
That's one part. The B part, just to clarify. In the past, you described your auto business in terms of the 70/30, right? 70 SAR and 30 is kind of tied to the growth activities. If I caught it in one of your charts, is it now 50/50, or did I get that mix right? Just I think related questions, but
Yeah, thanks, Vivek. It is all about content growth. We will not get obsessed here about the SAR because obviously the SAR is jumping around. What we use for our forecasts on the SAR side would be IHS, and they speak about a 10% SAR growth from a very muted number this year. Maybe the year after coming back to the 2019 levels and then in 2024, probably getting above what used to be the high point in 2018. The true growth engine for us is content growth. It's really. All the drivers which we have explained in the past one and a half hours is all about content. Now, Vivek, I will not parse this into units versus pricing, et cetera. Yes, there is a component of pricing in this.
I talked about the value reset of the industry, which we are going through. We also talked about during earnings last week that the input cost increases, which we are exposed to, we are passing that through to our customers, not to pad our margins, but to protect our margins. I mean, that's the way how you should think about the pricing component in this. The overriding element in that growth algorithm of 9%-14% is content growth in the industry and share gains of NXP. I think all three examples which we've discussed showed clearly how we outgrow what is already above average growth opportunities. I mean, the radar market is growing much faster than the overall auto semi market, and Torsten showed you how he's outgrowing it.
The same on the processing solutions and actually to pretty extreme, since it's still a bit smaller on our electrification business. It is a combination of content growth in the industry and share gains of NXP.
Let's go to John right here in the end.
Yeah, thanks for the presentations. Kurt, I've got two quick questions, one for you and then one for the group. Whether it's the 9%-14% CAGR for the auto or the 8%-12% overall, the tricky thing with CAGRs is always picking the starting point. I think there's a lot of people out there would say that there's a lot of cyclical tailwinds. I'm kind of curious, as I look to how you break it apart, it seems like almost all of the growth is coming from sort of these new design product wins. I'm kinda curious, with the 5% of the core, how are you thinking about cycle within that number? More importantly, when you think about sort of the product design pipeline, what inning are you in to actually seeing those designs go into volume manufacturing?
Because it seems to me like it's hard to have inventory in future designs that haven't yet hit volume manufacturing. Then my second question for the rest of the group, just as you think about autos doing more actions, how do you think about redundancy and how is the industry sort of, you know, tackling the challenge of redundancy? And is that built into your content expectations? You know, as you think about, you know, cars, you know, driving themselves. I think a lot of auto manufacturers have said they need to have the ability that 80% of the subsystems go down, the car can still figure out how to get to the side of the road safely. I'm kinda curious how you see that dynamic playing out.
Thanks, John. Lots of great questions. I try to be very concise. I saw a lot of hands raising. Number one, I don't wanna steal the thunder of Bill later, but he will give you a better idea about the year 2022 out of that three-year horizon which we are giving. But that's what Bill is gonna do later. Secondly, we don't really think about inventory growth here at all. You know the supply situation. I explained that at large last week. We are just not in a position to build any inventory. So the numbers which we speak about here are really derived from secular growth drivers. It's not about inventory. I mean, honestly speaking, I think the industry will ask for strategic inventory at some point.
I hope that point is not so far away because of the capability of the industry. I mean, we are just not in a position to do it, certainly not next year. Maybe in the years after, but it's not part of this algorithm.
Okay.
I don't know.
I can start very briefly, and then I pass it on to you, Torsten. That's, you know, there's a lot of new mission profiles that are being adopted in the industry, and we see this as part of this integration journey. Moving to fail-safe operation, moving to fault-tolerant systems. That is core concept built into our functional safety within an integration of running these multiple applications in parallel rather than multiple chips next to one another. On the sensor side, it may be a little different, right? Because you're adding more and more.
Yeah, indeed. I mean, it's important to have multiple complementary sensors and thereby have a certain sense of redundancy. Also, as Henri mentioned, the functional safety as such is already adding silicon content and driving redundancy within a chip. For example, a lot of these any functionally safe processor that goes to ASIL D, for example, will have redundant cores. You will have, for example, two cores running in lockstep doing the same task, where one is sort of monitoring and checking if it's correct what comes out from the other one. I would say the answer is yes, in multiple ways, redundancy drives silicon content, yes.
Great. Ross had his hand up here for a few minutes.
Thanks, Jeff. Ross Seymore from Deutsche Bank. One for all of you, but probably more for Kurt. And following on John's question, I'll try to make it relatively simple. There's a lot of cyclical versus secular debates. Your team did a great job laying out the secular drivers, the content gains, et cetera. But I wanted to go back to one of the things you said, Kurt, right off the bat about the customer behavior changing. These are the same customers that wanted no inventory two years ago, and yet now they can't get enough. Can you give a little visibility, not customer-specific examples, but what gives you the confidence that their belief has structurally changed in such a short period of time?
Well, first of all, Ross, let me be just sure. Currently, nobody is building inventory. They just cannot. When you say they cannot get enough, they don't get it. On the longer horizon, the learning is clearly that the just-in-time system on the OEM side versus our reaction time, given the three to six months manufacturing cycle time, is not compatible. What gives me the assurance that the learning from this incompatibility is now sinking in is they haven't even known this before, Ross. Only through this pain period, which we are in the middle of now, I think many of the car companies have learned what is it with semiconductors.
I mean, they knew that semiconductors apparently are important for a lot of their functionality, but they, I think they've been looking at it much more as a commodity, as a very low-cost commodity. At least in the spaces where we are moving, I mean, we are the opposite of a commodity. We are a very application-specific business which needs a lot of upfront investment and has long manufacturing cycle times. Since they have understood this, I think the whole discussion is a much more educated discussion about what you could possibly do about more inventory in the extended supply chain. I just wanna be sure, Ross, it's not gonna be at us. We will speak, and Bill is gonna show this later in our future financial model.
You will not see that we plan to grow our inventory, but there will have to be more inventory in the extended supply chain at the right points. I just remind you that actually the Japanese car industry, they seem to be a bit more immune to this supply situation in the first half of 2020, and the whole reason was that they had this inventory concept already earlier since Fukushima. I mean, they learned the lesson years ago and had successfully implemented it. Now, just for half a year, not long enough, but at least you saw it was at work. That gives me the confidence that given that lesson learned and given the much bigger understanding and proximity also to semiconductor companies like NXP now, it is gonna work, but it will take time.
We'll take one last question this session. Right here on the left, sir.
Hi, thank you. The dollar content numbers you gave for the S32 were very interesting. Could you provide similar for the battery management and the inverter control? Related to that, can your inverter control solution work without Wolfspeed and vice versa? Does Wolfspeed need you, or do they have other partners? Thank you.
All right. Yeah, thank you for your question. Let me start with the last part of your question. Wolfspeed is just a collaboration, and the system basically works with any power semiconductors, with any power module, whether it is IGBT or silicon carbide, regardless of from whom it is coming from. The first part of your question was the content, and we had also several conversations in the investor teach-in on the battery management system. What is really the semiconductor content of this system? The system really significantly varies. The semiconductor content of a 12-V or 48-V battery management system is completely different than for a high-end sports car for an 800-V system. It really varies, and it's very difficult to say that. The same thing is also on the inverter platform.
You have seen that we are scaling, for instance, the gate drivers. Eight to 16 gate drivers are being used in these solutions, depending on the power requirements and performances, so also here it is scaling. However, what we were showing basically in earlier sessions is that on average, if you basically take this across all the platforms, I would say in a typical battery management system, we are providing approximately $60 content, and for the inverter, it's roughly half of it.
Thanks, Jens. Just adding and highlighting what you said, the more it goes to high voltage systems, the more cells you have in the battery and the more analog front ends will be needed-
Absolutely.
Which means it will tilt to higher dollar numbers. This average number strongly depends on the mix between hybrid electric vehicles and fully electric vehicles because this is where the move is from a voltage perspective. Optimistically looking at this is since there will be more fully electric vehicles going forward, it's actually gonna move to higher numbers for us.
Well, great. Well, listen, this brings us to the conclusion of the morning session. We're gonna take a 15-minute break. There's refreshments outside, and we'll get back in here and kick off in about 15 minutes, 15 minutes after the hour. Thank you very much, everyone.
We're gonna get started now. I'd like to introduce Ron Martino. Ron runs our application processor business for industrial and IoT, and he's gonna give you some really great insight into our long-tail business. Ron, please take it away.
Okay, thank you. Welcome back from the break. I'm excited to give you insight into NXP's full industrial and IoT edge processing business. The world's transitioning to the age of intelligent edge computing. It's a $40 billion market. It's complementary to cloud computing. There's billions of devices that are being deployed around reasons for health, productivity, and safety, and the capabilities are being enabled by embedding small computers that are low power, connected, and secure. These devices need to be easy to use. These devices need to process many different types of data sources. These devices need to become more intelligent. They need to act independent of human instruction, and they need to perform tasks consistent with human desire. Edge computing is creating data processing at the source of data generation. We are creating this local compute to seamlessly interact with cloud computing.
A very exciting moment of broad adoption with this local compute. We're enabling edge to cloud, and we're focusing on optimization on the value for the end application and for the infrastructure in our mind. There's three categories of edge processing that I describe. Networking edge processing that involves data aggregation and data access. Then there's industrial edge processing, which really is building out the infrastructure capability in many different forms. Then there's the IoT infrastructure for the proliferation of these many devices that are coming into your life, in your home, in retail. The way that NXP is servicing this is by developing processor platforms and product offerings that service multiple markets by one platform. A multiplying effect of revenue to a dollar of R&D that's spent servicing these many applications with common and some optimized capabilities.
We leverage the fact that our customers and the developers within our customers first choose processor platforms as the key to their current and their future needs. Selecting NXP with the processor platform allows us to bring in the complement of connectivity, analog, and security optimized to give our customers more complete solutions and benefit from the optimization that we enable in terms of the interoperability of these different components and the software that pulls it together, more development efficiency, faster time to market. Our customers, when looking at edge processing, benefit in many different ways, reduction of cost of ownership when working with the cloud, processing data locally and simplifying it, so you only have to send what's necessary to the cloud and not raw data, choosing which content is needed to go to the cloud, so selectively choosing.
In doing so, it also has a benefit on total power. You don't have to waste energy transmitting to a remote and then back to a local environment in order to perform tasks because local computing enables the user experience that's needed. You get faster action at the edge for critical tasks, real-time tasks, safety tasks by not having that latency. You have enhanced security capability that you can control locally, and you can combine it with different ways to work with security in the cloud and transmission of data, but you can also control it locally. Enhanced security is a benefit. Then we have quality of experience.
Systems that are working with the cloud when they're disconnected, you want to have a user experience that's consistent, and you want local compute to enable that local experience until it can reconnect for the tasks that are needed from the cloud. There we go. When we look at our edge strategy, how are you gonna grow? It starts with the breadth of our processing portfolio. We have a processing portfolio that goes from the smallest MCUs to very high-performance processors that very few in the industry have in terms of the range of compute that we bring to the market. Then we optimize that portfolio between those two endpoints, so our customers can scale in terms of the capabilities, the costs, and the performance levels that they need in order to operate their systems and the variance of the systems that they wanna deploy.
We complement this with our enhanced connectivity portfolio, our analog content, AFEs, interfaces, and our security to bring the complete solution to a broad set of customers. We then focus and target our investments around key technologies to differentiate this platform that scales. Not just a point solution, but a scaling platform. Power efficiency, whether it's in a standby monitoring or runtime or optimization of total energy usage, is a key part of where we invest and optimize the platform. Making these devices more intelligent, the deploying of machine learning in a very easy way that the broad market can enable devices and systems using technology and product offerings delivered from NXP. The highest levels of security that builds off our long history in terms of financial and banking business and what we've done to integrate security or provide it in discrete forms. Reliability, longevity, and quality.
We have unique capability demonstrated over decades servicing the industrial and automotive market that then applies to our future product offerings. We then are making additional investments in strategic OEM relationships by placing engineers locally with these key customers, working together in terms of optimizing both tactically and long-term our roadmaps to meet their needs, to build confidence that their capabilities will be met, in some cases bringing custom IP from those in small forms into our more general purpose platforms in order to get the benefit from our offerings and still allow some of our customers to have specific differentiation. I'll touch on some of those as I go through the discussion.
Beyond these big OEMs driving growth with the significant adoption that we're seeing, we're serving over 25,000 customers in the mass market with easy-to-use solutions with this compute platform and with the attach that I've described. We do this with very strong ecosystem partners that service the many smaller companies. We have active communities with developers around the world working on these platforms, learning from each other. We offer turnkey solutions, we offer services in order to enable the broad adoption across the world through the mass market. We're taking this strategy, and we're focusing it on high growth applications, and that enables us to achieve the greater than 2x growth versus the market that Kurt showed. One great example that we have is the touchless human machine interface capability growing 35% CAGR. Why is it growing 35%?
Health, efficiency, ease of use of systems, of interaction with machines, and how machines can help you in your life, work, productivity. This aligns extremely well with our investments in machine learning, in vision, in voice, as well as rich display content. We see wins in smart homes, in many devices that are coming into the home, in infrastructure around building access and industrial systems where you have machine-human interaction that needs to occur in a seamless and easy way. You combine this with our turnkey solution capability that we've already been bringing to the market. We just recently announced last week the latest turnkey HMI solution, which we refer to as the RT117F. It's using one of the industry's fastest microcontrollers in our crossover microcontroller family and building upon that a full HMI solution of voice, vision, and rich display.
Combine this with our connectivity portfolio, and we have a great home and industrial control system that can be easily deployed by small companies, as well as large companies who wanna leverage it as a starting point and then modify it to their specific needs. The second area where there's significant growth is in the smart home, where we're extending our leadership and growth in personal devices. Within this market, it's growing 9%-11%, 2x the market growth. Here, we're making investments in low power, such as taking our crossover portfolio with rich integration, putting it on technologies like fully depleted SOI and bringing leadership power capability that is winning in wearables and hearables, as well as other personal devices. We have applications that are needing higher levels of capability and performance, so traditionally 8-bit going to 32-bit.
There's a migration into our LPC800 and our low-end crossover MCUs that's continuing to get more and more traction as these devices require that higher level of capability. Within the home, you have many more systems on Linux and Android using our i.MX platform, i.MX 6, i.MX 8, and we just announced this week i.MX 9, the latest machine learning integration, the latest energy efficiency, security services, a rich content for our customers to build applications at different cost and performance levels to roll out to the explosion of devices that are coming into our home. Then there's the attach, as well as many of the customer engagements that we have, such as our partnerships with Garmin, with Sony, as well as worldwide appliance manufacturers, including companies like Whirlpool. We're also growing significantly in the industrial space, in factory automation, building infrastructure, energy.
We're working with customers to move custom IP that was traditionally ASIC, in-house ASIC, into our crossover family of processors, growing our SAM and addressing it with our more general-purpose processors, but maintaining that custom need. We have vision systems that are proliferating within the industrial environment, robotics, security, safety, leveraging both crossover MCUs with vision at a very competitive bill of materials as well as our i.MX applications processors, the 8, the 8M Plus, and the 9 series that integrate machine learning vision capability and the ability to make decisions in a very efficient way and interface with the environment around them. Then in the case of medical door access and metering, we have a portfolio of our narrowband MCUs that we continue to roll out with various features and functions to expand in applications like diabetes care, access to doors that you encounter throughout your work, retail.
We have the attach aspects of this in terms of connectivity, security, analog, combined with the partnerships with companies like Honeywell, Schneider and Zebra. The core of this is the hardware portfolio of processing. As I said, the first decision. We go from traditional embedded flash MCUs to very high-end processors and application processor space and MPUs. We go from, if you use a relative term for performance, 1 DMIPS to 200,000 DMIPS to just give you a feel of how broad this portfolio is. We optimize in between this crossover family of MCUs and application processors that have a rich integration of content. MCUs of the past had embedded flash, had real-time processing capability. MCUs now have graphics, have audio acceleration for voice.
They have higher levels of security, higher levels of capability, and then scaling to application processors that are also now integrating real-time domains so that you can choose which optimization that's needed. On top of that, we have our connectivity portfolio. We have Wi-Fi N, AC, AX. We've integrated this into our BSPs, into our development environments with MCUXpresso for ease of integration. Industrial customers, small and large, can choose our solution, quickly attach connectivity with Wi-Fi and simply get their platforms up and running. Combine that with our narrowband MCU portfolio and then our wired time-sensitive networking or more deterministic networking capability with integrated MX and switches as the industrial protocols and bandwidth increase. NXP supporting this with proprietary IP and integration to lead in the industry. In the end, it's about application performance, and the investment in our software is absolutely critical.
At the bottom of the chart, there's many base capabilities that I could talk about for quite some time. What I want to focus on is that next layer, that layer of differentiation to enable the edge. We're investing with companies like Au-Zone, and our engineer teams together are rolling out our eIQ machine learning development environment for ease of deployment of machine learning onto our scalable processor portfolio, making things intelligent in an easy way. We are investing in acquiring R&D capabilities such as voice with the latest Retune DSP integration into the NXP voice capability to make even stronger capability for HMI, both open source and industry-leading capability in triggering and processing multi-mic audio environments.
We have security services with companies like Microsoft, where we have secure enclaves and we have IP compliant Pluton capability to allow Microsoft and NXP to offer Azure Sphere security services for 10 years of support. Then from a device management point of view, all these billions of devices, putting them out there, decommissioning them when they're done. Using our history in MIFARE with our EdgeLock 2GO capability to provision devices and working with partners like Foundries.io in order to enable that management of all these devices. If we take that portfolio and we look at some of the customer examples, I'll discuss three. Schneider Electric.
We work together very closely to drive leadership in factory automation and control, taking custom IP, integrating it into a crossover MCU, where our customer benefits in terms of energy efficiency, they benefit in security, they benefit in terms of networking IP, and we combine it in a system solution with the analog front end, power management, and connectivity to create one-stop shop system solution and increase our serviceable market. ChargePoint. They use a very comprehensive portfolio of NXP offerings, and they create EV infrastructure rollout capability. Together, we address the needs of point of sale and secure interfaces. We address energy metering and safety for high voltage and thermal concerns. We address connectivity to personal devices as well as cloud connectivity. The third is Microsoft. Working together, we actually literally go from the surface to the cloud with Microsoft, where our MCUs support keyboards, docks, and accessories.
Our crossover MCUs allow for the integration of sensor aggregation modules into their entire lineup, and then we integrate their Azure RTOS onto our MCUs in a real-time subsystem or our apps processors for easy connection to the cloud, deployment of ML, and enablement of services, including services around Sphere, a win-win for both companies. This all comes together into the smart home in terms of many devices that are being enabled. NXP is serving all of these applications with its industrial and IoT edge processing offerings.
We're working together on how these devices can easily be entered into your network, how they can interoperate, working together with industry bodies around home interoperability standards like Matter, deploying products that are enabled with Matter that's sitting on top of our BLE, Zigbee as well as Wi-Fi stacks, and enabling the home to perform tasks that are useful for the owner and the occupants in a more independent way. The same technology applies to the factory in terms of machine learning, voice, vision, rich display, and then enhancing this with higher endurance, higher safety, longevity programs, more deterministic networking connections that are critical for handling factory automation, solving critical tasks as a very deterministic part of the optimization of the factory. EV infrastructure rollout with solutions such as the ChargePoint example.
Pulling this together, as Kurt mentioned, we're growing from 9%-14% in terms of CAGR in a market that's growing 4%-6% on a $40 billion base based on the strategy I outlined with a scalable portfolio for complete system solutions, big OEM customer intimacy, broad market adoption through our 25,000 customers to achieve 2.3x with optimizing our portfolio towards these growth applications. In March, in the teach-in, I discussed 2020, and I mentioned processor revenue of $1.2 billion and attach of $600 million. In 2021, we're over $1.5 billion in terms of the processors and over $800 million in terms of the attach, and that's 100 basis points higher than the growth rates that I talked about earlier in the year in terms of the overall CAGR.
That is driven by the adoption and success we see with our customer engagements, the rich portfolio of connectivity, and the adoption of the system solutions. NXP is driving a positive impact in enabling this new age. We're addressing needs around productivity, around health and safety, as well as a greener world. With that, I want to thank you for listening, and we will open up a Q&A for this industrial and IoT area.
Thanks, Ron. Appreciate it. So we'll take Q&A just on Ron's presentation right now. I know there's a few folks who had their hands up before. Does anyone want to lead us off? John?
Yeah. Thanks for the presentation. I'd be interested in kind of understanding how you've been able to take the Marvell assets into this business and kind of expand kind of your footprint there. Is that relevant within this business? I mean, if memory serves me correct, that was a business that was kind of hampered by the fact that it was in a company that had a direct business model and didn't really have a strong distribution channel. I'd be curious about what you've been able to do with that asset since acquiring it.
Sure. And I like to refer to it as NXP Wi-Fi opposed to other names. But no, that's absolutely key to the growth that we have. It's integral to what I described. We have connectivity solutions that go from Wi-Fi through narrowband. We've integrated it into our BSPs. We've replaced what we've used before in terms of module partner Wi-Fi content that was not NXP, and now it is NXP. The attach rates through some of the examples I've given you is very high, and we see a very positive growth. It allows us, as this whole strategy was set up, to have a more complete connected compute system.
It is, as expected, in terms of adoption and leveraging what we've done with a 25,000 customer base plus, understanding how to work the channel and our partners that we have in there is making for a great combination that didn't exist before, but with NXP is much, much stronger.
Gary, in the back. Right here.
Yeah. Thanks for taking my question. On IoT connectivity, other than Wi-Fi, I noticed you had all the different logos up there, Zigbee-
Mm-hmm.
I don't know if you had Z-Wave up there, but Thread perhaps. My question is, how far along is NXP in driving that short-range wireless connectivity revenue as a part of your industrial or embedded compute offering? How much further do you have to go? Thank you.
The great thing about this industry and what we're describing is there's always a long way to go in terms of innovation and creating products. To answer your specific question, we have a very strong base in narrowband connectivity in the company. We are selling into diabetes systems, we are selling into access systems. We have a family of both Bluetooth Low Energy/Zigbee-enabled processors and more that we have in development and are launching that have attributes that service multiple markets as well. The same concept of a MCU with a radio integration around narrowband servicing multiple markets is in flight today. It is a meaningful multiples of 10 millions of revenue type of capability.
Go right here to Will.
Thanks for taking my question. It's Will Stein from Truist Securities. I think we hear a lot less about relative market share in this part of your business.
Mm-hmm.
I think it's perhaps because the applications and the customer set's much more fragmented than much of the rest of your business, especially automotive. Can you comment as to whether that becomes relevant at some point at relative market share specifically? Can you drive it to a more meaningful level or are you very happy with the position that you have today? Can you talk about the role that distribution plays in this market and your connection with distributors globally to enhance the growth here? Thank you.
Sure. Two questions. First question, relative market share matters. It is company philosophy that applies to this area. You can look at it in different ways. If you try to get down to a small application in this space, you can argue how meaningful it is. When you look at NXP's position for MCUs, we are number one. From a relative market share, very strong. If you look at applications processors for industrial, we are number one. If you look at it from a product point of view, you can look at relative market share and from an overall edge processing position in terms of our total revenue in this space, we have a very, very strong position, and the whole mindset of RMS applies there as well.
Yes, philosophy applies, and it also goes to focusing on high growth areas and becoming the leader in those areas. We try to look at it from multiple angles. Specific to distribution channel, we service over 25,000 customers, so distribution channel is absolutely critical. We have relationships with all the top distributors. Obviously with our sales team, we focus on the most effective distribution partnerships in order to service the NXP business. It is an ongoing relationship where we work together also in co-development modes to enable our platforms for going to market. That's across multiple different key distributors.
Anyone else? Anyone, anyone on the left side? I can't quite see past the post. No? Okay. CJ, why don't we take your question? Great.
Great. Thank you. I'm curious, I think, you know, your relative growth to the market is the greatest here I think out of all the other business segments that you're all presenting. I think we all traditionally think of kind of this business as GDP plus. Curious, you know, what the high level drivers are. Is it, you know, having the right portfolio today? Is it having the right customer mix? The right, you know, support? You know, what is the critical driver? I'm sure you're gonna say all of the above, but would love to-
Yeah.
You know, dig a little bit deeper into that if you could.
Sure. Well, I touched, as you just said, on the combination of things. Take industrial as an example. We offer decades of proven capability in terms of longevity, in terms of quality and reliability. Okay? We distance ourselves from other competitors who are trying to come in there. Partnerships are becoming more critical, similar to what Kurt described. Here, we have more and more partnership discussions every single day. Capability needs and the complexity of what our customers need to do, where they don't have the engineering team that are experts in all these capabilities are looking for NXP as a partner to enable them in what they need, Schneider Electric being a great example. That is another aspect that's really driving it.
The breadth of the portfolio, again, no one has portfolio with this dynamic range of performance and feature sets that allows customers to tune, 'cause they don't have the perfect view of their future. Having this platform also allows them to pivot when they see different needs and can quickly do it on an NXP-based platform.
Take one last question here for this session. Vivek.
Thanks. One more question on growth. You know, if I go back to the last analyst day, there was, I believe, about an 8%-11% CAGR that was suggested for this segment.
Mm-hmm.
It did actually grow 10%, but you know, when I include you know, the M&A that was done during that timeframe. Can you talk about what the organic growth was in the last few years? More importantly, you know, how did those three years evolve relative to you know, what was assumed before? Because now when I look at the 9%-14%, it's again a very strong growth rate. I guess part of you know, following up on other questions that were asked, you know, what has history shown us in terms of assumptions and you know, what does that tell us about what the next three years can be?
Sure. The way I would answer that is you referenced the 8% -11%. In the teach-in earlier this year, I brought that up 100 basis points, and that was linked to the strength of our connectivity portfolio, combined with the processing and attach in these solutions that I described. That brought us up to 9%-12%. Then, the 9%-14% is the added success on the system solutions, this more intimate customer engagements. We've been able to invest the R&D dollars into richer portfolio and enhanced portfolio. We've just announced the i.MX 9. We just announced the next HMI solution that's growing at that 35% CAGR in the market as a reference.
The richness of the product offerings through our investments, combined with the connectivity, and the engagements are just creating a multiplier effect. Okay.
Great. Thanks, Ron.
Okay. Thank you.
Let's move along. Rafael, just want to hand off the clicker.
Thank you, Jeff. Let me go here. Early, Kurt talked about UWB as one of the growth areas for mobile. I'm gonna be talking to you about UWB, but not just in the context of mobile, but in the context of all the section and everything we're doing to grow the ecosystem. Just to level set, since UWB is somewhat of a nascent technology, UWB is a wireless technology that is used for sensing. The reason it's being adopted in the market, both in IoT, auto, and mobile, is because it solves a difficult problem, which is knowing the relative location of an object relative to you or to another device with enough level of accuracy to matter, to make a particular use case happen, to take actions, to make decisions, to react.
UWB delivers awareness of where a trusted object is relative to you with such level of accuracy that it's gonna be the catalyst for new use cases where objects anticipate our actions. I think the way I think about UWB is UWB is the closest thing that to giving objects intuition. Knowing accurately a tri-dimensional position will transform what devices can do. It will create inflection points in the value proposition of many of the devices that we use today. There's some technologies today in wireless that can be used to calculate distance, right? But none of them, and I really mean none of them, with the technical attributes and the value proposition that UWB provides. I mentioned already accuracy, right, as being more important. UWB is very accurate, both in angle and distance calculations. Look at 10 cm.
Those 10 cm, that's really what unlocks these new use cases. Not only has unmatched accuracy, it does it under very difficult environments. Think of rooms, think of multiple walls, reflections, RF blockers. UWB remains resilient. It remains accurate under those conditions. It's got long range. I mean, 100 m is more than enough for most use cases. Security is a very strong attribute. UWB was designed from the beginning with security in mind. Unlike other technologies, it has RF security. The preamble, which is a sequence of bits used to synchronize the transmission packet, that is derived from a unique key every ranging measurement. Essentially, every ranging measurement is encrypted with a different key. Low latency, that also unlocks a new type of category of use cases.
Think of gaming, right, where latency is critical. Think of public transportation and subway turnstiles where, you know, the transaction speed is also critical. These technical attributes, this is what makes UWB very good at what they, at what it does. Now, at the early stages of every technology adoption, there's always a killer app that propels the adoption. For UWB, it is secure car access. This is what triggered the adoption of UWB in mobile and in auto. Now, UWB is, on its own rights, a great access technology for auto, even without the phone. It solves the relay attack. Just relay attack is when someone, an aggressor, picks up the signal from the key fob while the user is away from the vehicle, relays it to the vehicle and gets access to it.
UWB prevents that because every ranging measurement is checking where the position of the fob. That cannot be manipulated because as I said, every ranging measurement is encrypted. This is a use case that really brought two ecosystems together, the auto and the mobile, and we are really proud of the role that we play on making this thing happen. Secure access is not the only use case for UWB. Just wanna give you what has been happening with UWB, just to give you a sense of the progress. The FiRa Consortium, and this FiRa stands for fine-ranging. It was a consortium that NXP co-created to develop an open UWB standard, to ensure there was interoperability among different OEMs. That membership has doubled since last year.
Now we're talking about companies like Google, Cisco, Sony, Samsung, Xiaomi. Last year, there was only two phone OEMs with UWB in the market. We estimate by the end of the year, there will be five. Five more working on adding UWB in their phones next year. Eight out of those 10 are working with NXP. Now our partners, this is NXP partners that help us develop UWB solutions across all markets are now 150, over 150. Cars. There was no cars with UWB in the market last year. Yes, there will be three this year. We're working on another 19 to make sure that we migrate their systems to UWB. All of them working with NXP. Again, the use case continue to come.
The FiRa consortium now has over 35 use cases they are working on for UWB. Indoor navigation, residential access, payments. The ecosystem is very healthy and is growing, and we at NXP are a catalyst for this growth to happen. When we started working on bringing UWB to market and bringing this spatial awareness feature to market, we did it from the beginning with our partners. From the beginning, we did it with industry leaders who had a vested interest on the co-creation of this technology. This is customers that we have in mobile and in auto, where we already have a very strong position to leverage. It's a lot of really good progress in the last two years. You know, we launched a couple models with Samsung last year.
We launched Xiaomi with a UWB phone this year. BMW announced their Digital Key Plus system, where UWB plays a central role. More things to come. Samsung launch a UWB tag. Tile announced their own UWB tag to be released soon. On top of that, now we have the support of the software ecosystem. Android announced support for UWB in Android 12, and Apple release UWB support for iOS. Again, this is yet another sign of our thesis that UWB market is indeed happening fast. Now, creating quickly an install base of UWB devices is critical. New companies are not gonna add UWB unless they really believe there will be more UWB devices out there. This is the reason why getting UWB in their phones is critical.
Phones are pervasive, and the moment UWB makes it on the phone, then that UWB technology becomes a free resource for other verticals to use. The reason we're bullish is because we think that this is indeed happening. If you look here, the chart on your right, it shows you the cumulative devices, key devices that are in the market. We see now that by the end of the year, there will be over 350 million UWB-enabled phones in the market. There will be two generation of iOS, multiple Android OEMs. Also, a lot of these mobile phone OEMs are also developing IoT products for UWB to connect with their phones. We estimate the IoT devices will be about 50 million units next year.
2023, now the number of cars with UWB in the market starts to become interesting. We estimate this to be about 50 million units. Ensuring that we have this virtual cycle of early UWB adoption, leading to more adoption, leading to more adoption, requires, right, focus and support in these three segments, talk about mobile, IoT, and auto. Because of that, we focus here at NXP to develop a UWB roadmap for these three segments. We call it NXP Trimension. Different products with different attributes going to three different markets, but all with a common platform. That way we ensure interoperability, we ensure reuse for our customers. I mean, this is one of the reasons, companies like Samsung or Xiaomi, they have peace of mind that even the early adoption UWB, their phones are going to interoperate with other verticals.
Because those verticals are most likely powered for NXP anyway. We announce Samsung and Xiaomi are using our UWB in their phones. On automotive, we're working with BMW and Hyundai. We now have tag solutions with Samsung, residential access with ASSA ABLOY and Tile. We talked about how important it is to get UWB in these phones. You should not think about UWB as yet another connectivity technology in mobile. The way you should approach it is, it's a secure sensing technology that has the capability of enhancing secure transaction applications. In that sense, UWB is an extension of the NFC mobile wallet. Clear examples are the access use cases that we talked about car door, but moving into payment, moving into transit, all of them require a certified security solution in the phone, not just UWB.
We've been a leader in mobile wallet for years. We have a unique end-to-end offering from hardware, software services. We have very strong ecosystem partnerships. We have a deep expertise on the application, and that is what allow us to quickly develop a UWB solution that it was based on the mobile wallet. That's what led for us to have four OEMs design wins with our UWB and four more working with us to develop UWB phones next year. Automotive. We're seeing a lot of traction in auto as well, and this is pretty exciting. NXP is the leader in car access, so we are the natural go-to partner for anything car access. We are working essentially with all our partners to migrate their existing car access solutions to UWB.
The way you got to think about UWB in this system is it is part of a much bigger system that includes BLE, NFC, and the door handles security. NXP is the only provider and the only vendor out there with all the technologies, connectivity, security, localization algorithms. We have won over $700 million in design wins with 16 OEMs. The other six, they're working with us to develop proof of concepts and develop their system with UWB. In IoT, it's early days of IoT, and the beginning of IoT. Most of the use cases were very similar to mobile: access, trackers, indoor navigation. Now what we're seeing lately is a lot of innovation, a lot of creativity in our customers.
We're starting to see things like our customers using presence detection to be able to customize the behavior of the device. We've seen customers now using UWB to do audio tuning depending on the physical placement of wireless speakers or robots finding their way back to their charging station. We're starting to see a lot of creativity and kind of put some big smile on my face because quite amazing what our customers are thinking of doing with this technology. IoT requires flexibility, requires options for all these customers, and we're providing them. We have partners, module partners, design houses, reference designs, turnkeys, SDKs. A lot of these customers in UWB for IoT, for us, are already customers of NXP.
Either they use our microcontrollers or they use our connectivity, so usually the engagements are relatively smooth. We have right now, obviously, IoT, big, big fragmented market. We're working with over 100 companies today to develop UWB solutions for IoT. There's gonna be a lot of innovation coming out of our customers using UWB in this segment. UWB is a growing market for the reasons I stated. There's a huge value proposition that technology gives. We estimate the market to grow over $1 billion by 2024. Yes, a big part of this segment is mobile, but IoT is starting to become significant by 2024, you know, just shy of one-third of it. We have a strong position in the market. We have...
That has given us early access to our customers and deep understanding of the application and an ecosystem, and that allows us to define not only today's solutions, but tomorrow's. We expect to outgrow the market by 2x and get to a revenue of $400 million. NXP will continue to play a critical role and a catalyst to continue to grow this market. We expect to continue to innovate for our customers, and we're very proud of the role that we play in growing this segment. Thank you.
Take some questions now if we have them. I'll stand behind the speaker now. The light's in my eyes. Who would like to lead us off on some questions here on the ultra-wideband topic? Don't be bashful. Chris, down in front.
Thank you. Chris Rolland, Susquehanna. Perhaps you can talk about UWB and its eventual use for automotive keyless entry, and kind of the size of your keyless entry market today and whether UWB would cannibalize some of this market or enhance some of this market.
Actually, I have the right guy to answer the keyless entry segment in terms of cannibalization, or which is basically to tell you no, but like.
Thank you for your question. Well, as outlined by Rafael, we are truly the leader in the secure car access business, and we are coming back to the relative market share model. We are close to three times larger than the next competitor in the market. We are not disclosing the size of the business, but to give you an idea, the secure car access business will not be, let's say, taken away by the ultra-wideband mobile access business because it will be a complementary business. All of the use cases we do see right now is that every car still will come with a key, so it will have a full-blown passive keyless entry solution.
Then on top of it, there is the ultra-wideband car access, mobile car access being enabled, which will be really an add-on to the existing bill of material as a complementary feature going forward. We don't see that in the near future, and because this is also a very long-lasting business, we don't see in the near future that the existing classic or passive keyless entry business will be gone away.
Gary, in the back.
Might as well stay here.
Yeah, thanks for taking my question. I guess one of the more notable smartphone OEMs in Cupertino, you know, has its own ultra-wideband chip. So my question to you is, how do you convince Samsung, Xiaomi and others that they don't need to develop their own? Related to that, how do the hooks of NFC, you know, prevent that from happening?
We didn't have to convince them very hard because the know-how and the need they had to actually connect the dots and create the ecosystem. Remember, the first use case with the auto has created high barriers of entry. I think it was more of a partnership to develop the use case more than a conversation about whether they should be doing UWB or not. The relationship with UWB as the second radio for the mobile wallet is very strong, okay?
There will be use cases where you use your mobile phone to enter a UWB-enabled car, but there'll be a point in time where maybe your phone is out of battery, and now you use your NFC phone as completely a low battery or battery less operation to open the car. You can see now that NFC is more of a backup radio for UWB in many of these access options. There's a very, very strong relationship between UWB radio and the mobile wallet.
Sure. John?
Rafael, maybe I can add on to Gary's question. I mean, to the extent that it is an extension between UWB and NFC, is there a single sort of chip solution in your future around these two technologies? If you kind of drive towards that, might that be enough of a value proposition to expand your footprint in the mobile phone space for UWB?
I mean, I think there's history of that happening, right? It happened with Bluetooth and Wi-Fi became combos because once the attach rates and the use case were so synergistic, there was no reason to it. Right now we're at a phase in UWB where the attach rates are not as high as the mobile wallet, so it doesn't quite make sense. Yeah, I see a future where that would make perfect sense, and that would be the logical step to take.
Anyone else? Okay. Well, if we don't have any other questions just yet, then we can move on.
Thank you.
Okay, great. Thank you very much, Rafael. Thanks, Jens. I'd like to introduce Paul Hart. Clicker.
Good morning, everyone. Now, I'd like to change gears and talk to you about NXP's radio power opportunity in the 5G communications infrastructure market. To get started, I wanna kinda step back to our investor teach-in, our 5G teach-in last year. At that time, I said that we felt that we were at the early stages of an extended super cycle for 5G deployment globally. We were seeing spot, well, actually significant deployments in a few geographies, South Korea and in China in particular. We were also seeing enhanced mobile broadband deployments at millimeter wave frequencies here in the U.S., led primarily by Verizon. Today, one year later, we can safely say we've passed the early stages, and we're now in the growth phase of the 5G deployment super cycle.
What I mean by that is, this is no longer a geographical deployment only. We're not talking about certain regions. We're now talking about all regions across the globe adopting 5G in a more aggressive manner. The primary driver today is still around enhanced mobile broadband. It's still about more data, faster data to the end user. That's the value proposition that operators are investing in. They're investing in more data and a more efficient means of delivering high data to their customers. We're also starting to see the emergence of new business cases with 5G. First in the industrial segment, but growing. Growing not necessarily at the rate that was forecast three, four years ago, but we're starting to see it now and seeing it develop.
That combined with the fact that 5G is really a multifaceted network solution gives us the confidence to say that this really will be an extended deployment cycle, dissimilar to what we saw in the 3G and 4G eras, but something that really continues to grow really throughout this decade. There are some impediments. Cost and power consumption of these systems are both high, higher than previous systems. These are areas that we as the enablers of 5G infrastructure need to keep our focus on. We need to keep driving, investing to drive down cost, investing to improve power efficiency, and investing to make these systems easier for our customers to deploy. At NXP, we're focused primarily in two areas. One, on the radio power devices, which is the theme of today's presentation.
We also have a family of edge programmable modems which layer into 5G as well. All of these investments really focus on the three fundamental pillars of what we see as being key to bringing 5G to a global market. The first are coverage and capacity. Coverage, meaning establishing a 5G signal everywhere you go. I think that's first and foremost in operators' thinking when they think about investing in 5G. Second is capacity, making sure there's enough data transport available so that users feel the difference when they experience 5G versus previous 4G systems. The fundamental differentiator of 5G versus 4G and previous systems is truly on the latency front. This is the real-time interoperability of the network and your engagement with the ecosystem around you.
It isn't just downloading a large data file to your phone, but it's about how you interact in a dynamic way with the world around you. I mention it here because latency is fundamentally an edge compute-enabled feature. Without decisions being made at the edge, compute happening at the edge, 5G can't deliver on the latency promises. Ron touched on this heavily, and I just wanted to follow up on that because it is a key point to the value proposition of 5G. At NXP, in the radio power space, we're focused on coverage and capacity. From an RF perspective, there's no delay in our signal, so it's not a latency issue.
The opportunity for us is about getting more 5G into the systems and making it easier for our customers to create systems that are cost-effective for the operators. When I think about 5G, I think about a complementary network ecosystem. We hear people talk about 5G in different ways. They talk about 5G deployments in mid-band. You hear operators like T-Mobile talk about 2.6 GHz deployments, but also going down in frequency to 700 MHz for a big coverage solution. You hear operators like Verizon talk about millimeter wave deployments and delivering 2 Gb per second data to the users in dense urban areas like Midtown Manhattan. We also hear about C-band deployments and mid-band spectrum as being the catalyst for really ubiquitous 5G connection globally. At NXP, we're focused on all those aspects.
We're focused on delivering macro solutions for coverage. We're focused on mid-band 5G mm wave solutions, and we're focused on 5G massive MIMO solutions, and we're focused on 5G mm wave solutions as well. We're also focused on the indoor handover of those data-driven connections by driving Wi-Fi 6 and Wi-Fi 7 solutions across the board. For us, it doesn't actually matter which format of 5G a specific operator chooses to invest in, but our value proposition is fundamental. We're really agnostic whether they turn around and talk about mid-band deployments or coverage deployments or millimeter wave, we're invested to grow all those segments together. That means coverage in all frequencies, power levels and architectures, which of course means complexity. When we look at the spectral allocations for 5G's network systems, they really have exploded.
They've exploded dramatically over the last few years. If we go back in time to a simpler era of 2G to 3G, fundamentally, the shift in a radio technology was going from a single carrier radio, single channel radio to a multi-channel radio. We increased bandwidth a little bit, improved the modulation format to allow greater data rates in a given frequency allocation. But from an RF content perspective, which is the theme of my presentation and what the business is invested to deliver, it was pretty stable. Right? A little bit of power increase moving from 2G to 3G. As we move from 3G to 4G, the major breakthrough was on the modulation format. WCDMA moving to LTE, just changing signal modulation schemes allowed far more data to be delivered to the user.
Expanding the signal bandwidth multiplied that data delivery mechanism further, and we saw a great pickup, I think, in mobile data as a result. Still, from an RF perspective, we aren't fundamentally talking about significantly different architectures versus this 3G era. Slight increases in diversity of the systems, a few more antennas delivered per radio unit, but fundamentally the same, and we added a few frequency bands. Here in the 5G era, what we're seeing is an explosion of complexity. New frequencies that haven't been touched, brand-new deployments, new architectures. There's no Moore's Law in RF, which means if you wanna deliver more antenna systems and more content and more data to the end user, you have to add more content. We're seeing that explosion now today.
Right now, mid-band spectrum looks to be the global workhorse. 3.5 GHz is really a key spectral area across the globe. We're seeing millimeter wave deployments happen in the U.S. and expect those to come in years ahead, and we're also expecting frequencies to start spanning between 6 GHz and 15 GHz in the next few years. All of this is enabled by custom technologies. Technologies spanning silicon LDMOS, which Kurt mentioned, gallium nitride, which we've invested heavily to bring to market in the last several years, and proprietary silicon germanium, all of which have been fine-tuned and tailored for the communication infrastructure market. These aren't one-size-fits-all technologies, but really spot-optimized solutions for this market segment. That sets us up really to be able to deliver on any of these deployment mechanisms.
Whether we talk about, you know, 700 MHz, we talk about 39 GHz, or even we start thinking about 6G and 100 GHz applications. NXP's technologies span that spectrum of frequencies and enable us to create competitive offerings across the board. I want to show you a few, just a few of the systems here. 5G macro systems are kind of the workhorse of early days of 5G. Fundamentally, they're very similar to what might have been seen in a 4G radio. High power per antenna delivered by high-power technologies like 28-, 48 V silicon LDMOS are now gallium nitride. The migration from 4G to 5G has really been on the technology front, enabling smaller size, enabling higher power density, and enabling improved efficiency.
The new market segments are around the active antenna systems in the 5G massive MIMO domain. This is where the antenna diversity has increased from two to four transmitters, up to 32-64 transmitting elements per radio. Here we see a completely different dynamic unfolding. The need for massive integration, the need for making parts pin-to-pin compatible, the need for driving consistency in product offerings across frequencies and powers. That's another area that we've invested heavily in. At the end of the spectrum, in millimeter wave domain, now we're looking at 256, some cases up to 1024 transmitting elements per radio. Huge integration complexity is required in order to make these systems realizable, and we're happy to say that we have these systems also deployed in the market today.
What do you think about coverage solutions or capacity solutions? What we have already developed and what's deploying in our customer base spans all ranges, and it touches every addressable customer that NXP has access to today. Whether you talk about traditional OEMs focused on traditional operators, or you think more about O-RAN enabled equipment vendors, we're engaged with and working with and designed into 100% of them today. I mentioned there's a challenge to rapid 5G deployment. There's two challenges, in fact, and the first of which is size. These 5G radios and we'll focus on the mid-band radios today because that's the main deployment vehicle for 5G in the next couple of years. These radios are fundamentally big. We're talking 25 L a few years ago with aggressive roadmaps to reduce size.
At NXP, we've been investing heavily to enable that size reduction. We've been investing in integration, non-monolithic integration capabilities at very high power, which is not an easy problem to solve. If we go back to 2018, we had a largely discrete portfolio of products that were going into 5G radios. By last year, we transitioned the majority of those products into integrated modules, including all the respective matching, combining multiple technologies to create performance-optimized solutions that are cost-optimized as well for our customers. Moving from yesterday to 2020 into today, what we're really enabling tomorrow is the further acceleration of this integration roadmap and bringing in advanced thermal management techniques so that we can scale up the power in these small form factor solutions and continue reducing radio size.
The developments we've enabled over the last several years are resulting in up to 40% smaller radio boards, which means the volume of these radios that go up the tower can dramatically reduce. This integration is enabled through a broad range of technologies, most of them in-house developed, some of them available in the general market, but all of them tailored for these applications. Whether we talk about gallium nitride, we talk about silicon germanium, we talk about SOI, all those pieces have to fit together in a uniquely optimized solution in order to continue reducing system cost for our customers. Advanced packaging, high power packaging techniques are all fundamental tenets also of this value proposition, as well as wideband and multiband IP.
If you think about enabling wider frequency ranges to be created through a digital radio system, you can think about fewer radios being required on the tower. Therefore, there's a cost reduction enabler for our customers. What's nice is the technology we've enabled, the technology features and the integration capabilities span again the full range of 5G opportunities and Wi-Fi 6 and Wi-Fi 7, which gives us confidence that we really have the right value proposition to grow in this market. The second tenet to reducing cost for the operators is system efficiency. We fundamentally have to make these boxes smaller, and the easiest way to do that is to reduce the power consumption because the greatest weight in these 5G radios comes from the density required to dissipate heat.
We've been investing to increase efficiency, and over the last three years, we've increased system efficient radio or PA efficiency 15 points, which is an unprecedented achievement in the last 20 years of this market. Massive efficiency increases coming through the means of high efficiency gallium nitride, advanced thermal management, improved system design techniques, and really some new architecture concepts that we've brought forward. Going forward, though, we don't see that to be the end of the journey. We see further efficiency enhancements enabled for us, focusing on areas like enabling low power compute, enabling dynamic and optimized behavior over the real-time life operation of the radio. All of that will continue to drive the PA efficiency up, which means the radio size can continue to reduce.
Focusing on both area and efficiency for weight, we've enabled massive reductions in radio system volumes and mass over just the last couple of years. When we look at that through the operator lens, this to me, this graph stands out. Today, what's deploying for a mid-band 32 T or 64 T active antenna radio is about 20 kg and 20 L in terms of radio size. Just two years ago, we were 25+ kg and 25+ L. What's going to deploy next year across the board is about 20% smaller. 16 L, 16 kg is the target. Both of those.
That drive from 20 kg to 16 kg has been enabled through a lot of innovation on our part, both on the discrete front with new technology introductions and also on the integration front with advanced design features. We found ourselves able to create something that's fundamentally lower cost for the operator. Industry leaders, however, are already signposting 12-kg systems that are going to hit the market in the next one to two years. That's a 40% reduction from where we stand today, and it's something that is really only enabled now by this high-power integration capability and this advanced thermal management that we've put together to create the most advanced products for this market space. We see this as a huge differentiation for NXP and really a huge enabler for 5G on a global basis as well.
From an operator's perspective, smaller size results in simplified installation costs. You're no longer rolling cranes out to the towers, but you can imagine a single installer carrying one to two units up a ladder at a single time. Reduce tower loading, which means you can put more equipment on the tower and reduce, again, site rental costs and improve tower utilization. Then on the OpEx front, lower rental costs, size equals cost on these towers to the operators, but also lower energy consumption, I mean, which leads to lower OpEx. Again, further accelerates the 5G mid-band deployment and the opportunity that NXP sees. When we put the pieces together, we see a market that's growing. We see the market is growing at a rapid pace. We see the 5G radio power market increasing at a 14% CAGR.
What's fundamental to this calculation is the realization that this market is growing not because more base stations are being deployed, but because the same number of base stations will be deployed in a more content-rich manner, RF content-rich manner than in the past. For the last 20 years, 1.5 million base stations per year on average have deployed. Our model in the future assumes the exact same scenario, but it assumes that more and more capacity solutions will be applied, deployed by year versus coverage, that the more content-rich systems which enable higher data rates to the customers and a better user experience will be deployed globally, and that more 5G networks will turn on across the globe. NXP approaches this market from a strong leadership position today. We currently enjoy an RMS of 1.2 in this space.
From that high level, we expect to continue outgrowing the market at a rate of 1.1x, meaning our revenue CAGR will be at 15% for the next three years, and it's driven on technology leadership, our system expertise, and last, but really not least, our manufacturing scale. These are complicated products that we're creating for a pretty specialized market segment, and we've invested to create the scale to bring those products to market at the volumes that the industry needs. Final summary, this CAGR that we describe both from a market and a revenue perspective is driven by more radios per base station, more RF power content per radio, and an increased addressable opportunity for our share of the board through the integration initiatives we've been driving.
That's why we're confident that the next few years in the 5G comms infrastructure space will be a growing area for NXP. Thank you.
We're almost near the end. Let's open it up for the last few minutes of Q&A with Paul. Anyone have any questions about the RF power business at this point? Will, down here in front. Oh, the slides are back.
Thanks. This may be a little bit off the run.
Some years ago, there was an idea that perhaps RF power components would be adopted in non-comms end markets, in particular in industrial.
Mm-hmm.
For heating and things like that. Has that not taken off or has it taken off but it's smaller? Or what, maybe any update in that aspect of this technology would be helpful. Thank you.
It's a good question. We have seen some RF power products growing in the industrial segment, growing through applications like industrial heating, drying, et cetera. There was talk about deployments in the cooking space. I'll say that those business opportunities have been slower to develop than were previously forecast. We've seen the majority of that happening in industrial cooking applications today or restaurant environments, but nothing yet that has migrated into residential settings.
Lane. Right there in the corner.
Hiding behind the pole. Wondering if you could just talk a little bit about millimeter wave. You obviously have a leading position in sub-6 GHz, just your competitive position in millimeter wave. In terms of the forecast, just any perspective on your outlook for millimeter wave deployments through 2024.
Yeah. To be fair, we saw an initial surge in millimeter wave a couple of years ago, driven here in the U.S. However, today, what we're seeing globally is more focused on the mid-band spectrum for coverage reasons than previously available. A lot of that came through new auctions across the globe. We see millimeter wave continuing to be a strong element of 5G in the future, but it will probably start layering in in the next two to three years versus broad mid-band deployments, which are happening in the next one to two years. Our position in millimeter wave is quite good. We're designed in at several key OEMs, and we're seeing constant steady state volumes, I would say, being deployed.
We're not yet at the stage where the mid-band spectrum is built out and then the millimeter wave is needed to complement that and deliver massive amounts of data, yeah, beyond what mid-band can produce. I see it as really an extended development, part of the extended development cycle of 5G.
Vivek, right here.
Thank you. First, a clarification. I know you only discussed the RF part, but the rest of the comms infrastructure business is the assumption that it kind of just flattens out the next few years, because that's what the CAGRs were. What exactly is in there, and why would it just stay flat? That's part one of the question. Part two is the GaN market more or less or equally competitive versus the LDMOS market as you see it?
Vivek, why don't I take the first part of that question?
Sure.
The other product portfolios inside of comm infrastructure are our legacy digital networking business, kind of multi-core PowerPC products, and then also our embedded secure card business, like our ePassports, access business, things like that. You're correct. Our numbers do assume they stay roughly flattish.
The technology question is also interesting. We have products that span both, you know, gallium nitride and LDMOS, and we see strong growth in GaN today, particularly because the frequencies of 5G deployments continue to go up, right? We see GaN growth in the mid-band spectrum. We see GaN growth in 3.5 GHz. All of the products that we're creating in those frequency ranges are GaN based. Depending on which operators choose to deploy and what spectrum they use, you know, what we see going forward in terms of technology mix is kind of open, right? We have solutions that deliver it all, and like I said, depending on which operators deploy, we'll ship what they sell. GaN is definitely a strong growing market for us.
Is it more competitive?
Well, at higher frequencies, it enables higher efficiency. In higher frequencies, the difference between what can be achieved with silicon and GaN is. It grows. Lower frequencies are pretty similar and silicon optimized is more mature, more reliable, and more cost optimized. Again, it depends on which operator is doing what, when, and how.
Any other questions? Harlan, right here on the right or left.
Morning. Harlan Sur, JP Morgan. We keep hearing more and more about the virtualization of the radio access network, so O-RAN, vRAN, greenfield buildouts, potential brownfield buildouts, but also private networks.
Right.
Other than the potential just SAM expansion because of the adoption of O-RAN and vRAN, is there anything that NXP, your team is doing to differentiate your products to take advantage of this trend?
Absolutely. If you think about the O-RAN players, these aren't your traditional OEMs. They don't have 100s or 1,000s of engineers capable of creating high power RF systems. What we've done is really focus on ease of use. The integration activities we've undertaken make it fundamentally straightforward for new entrants to create these 5G radios, at least from an RF design perspective. Now, I think that sets us up very nicely in the O-RAN space, and it actually enables new entrants to participate, where previously it was very difficult.
Anyone else? No. Well, with that, Paul, thank you very much, and it's my real pleasure to introduce Bill Betz, our new CFO. I've worked with Bill for as long as he's been at NXP, and it's been a real pleasure those years.
Thank you, Jeff. Again, I'm Bill Betz, and I am the NXP CFO, and I'm just thrilled to be here in New York City. I grew up right across the river in New Jersey, and more importantly, you know, as a team, we haven't seen each other over two years. Thank you for you all being our supporters and our owners coming face-to-face. This means a lot to us. Thank you for that. With that, let's get right into the NXP financial value proposition. I'll try to keep it very simple, very easy to follow. Here's today's agenda that I plan to go through. We're gonna start off with the NXP investment thesis, and Kurt showed it, that slide in the beginning.
I'm gonna try to summarize our growth areas of secular growth areas specific to NXP that our team covered for the last four hours. I'll then spend a couple of moments on recapping our financial progress since 2018. I will share, which I know you've all been patiently waiting for, is the NXP model for the next three years and our commitments. Kurt hinted, I'll also provide a little bit more color about 2022, and then we'll wrap it up with our capital return policy. Again, our NXP investment thesis is very simple. It's very easy to follow. First, we're a market leader driving strong, profitable growth in everything we do. Second, we have a very disciplined and proven financial model that is resilient to the bottom line.
Third, we have a reliable capital return policy to return all excess free cash flow back to our owners. Let's talk about this first principle. It's about driving strong growth in the markets we lead into. Behind me, what you can see is a chart showing what we talked about today about the accelerated growth drivers. If we just look at that chart in 2018, that portion of the business that we talked about for the last four hours was $1.6 billion, 17% of our business. That doubles by the end of this year in 2021 to $3 billion, representing 28% of our total revenue. What you heard again from my colleagues, we're gonna double that from $3 billion-$6 billion in 2024. Let me recap what those growth drivers are.
Torsten talked about the leadership position in the radar system business that we have. Today, that is $600 million growing to $1.1 billion in 2024. That has a compound annual growth rate between 20% and 25%. If we talk about the auto processing, and Henri did a great job talking about the S32 platform, the future growth of this franchise here in this business. That business, Henri, is about $1.7 billion today, going to $2.3 billion in 2024. That has a compound annual growth rate of between 10% and 12% over the horizon. Jens talked about electrification, battery management systems, e-motor control, the inverter part of that business. It's early.
It's at $200 million today, and we are confident that business will be over $500 million in 2024. That has a compound annual growth rate of over 30% during that horizon. Ron shared with us the connected edge, the Wi-Fi attach, the analog attach. He has over 25,000 customers. He focuses on the smart home, the smart factory, the smart devices. His business is $2.4 billion today, and that business is going to over $3.3 billion in 2024, representing a compound annual growth rate of 9%-14%. Rafael talked about ultra-wideband. He talked about the ecosystem focused on the mobile phone, the car access, the Internet of Things. That is super early.
This is the early innings, just like the mobile wallet. That business today is $80 million in 2021, and it's gonna grow over $400 million to 2024. That represents a compound annual growth rate of over 80% during that horizon. Then lastly, what you just heard from Paul, he shared about the RF 5G infrastructure business, where today it's $500 million, and we are confident that'll get us to $800 million in 2024. With a compound annual growth rate of 15%. Now, also on this chart, which we did not talk about at all today, is the rest of the portfolio. This portfolio has relatively high market share, high barriers to entry. Yes, some of it goes up, has higher growth rates than what we're showing here, some a little bit lower.
On average, if you just assume that has a 5% compound annual growth rate over that horizon, you combine both those, you get to 8%-12% at the total NXP level, and it's being driven by those accelerated growth drivers. Now, this is a slide that you're all used to seeing. We share this slide every quarter, every year. This is how we report our revenue. It's important to note these are our growth franchises. When we compare our growth franchises to what we said three years ago, they are all increasing from a growth standpoint. Just think about it. You take auto, you take industrial IoT, you combine them, that's 72% of the total business. We expect that part of the business to grow 9%-14%, three-year CAGR.
Now, if we just look back, auto, what we said three years ago, that business was gonna grow 7%-10%, it grew 7%. If we look at industrial IoT, we said that business would be 8%-12%, it grew 10%. Now, if we look at the mobile business, the mobile business represents 13% of our total revenues. If I look back three years ago, we said that would grow 4%-6%, it grew 6%. We're now raising that based on what Rafael shared with you about UWB, the ecosystem, along with the attach rates continue in the mobile wallet, and that's now a growth rate of 8%-10%. Lastly, in the common infrastructure business, three years ago, we said 0%-2%, we were basically 0%.
Now, with the investments we made in the RP business, in Paul's business, we expect the total common infrastructure business to be lifted and grow between 2% and 6%. Again, overall, this drives an 8%-12% total NXP CAGR. Again, three years ago, Kurt mentioned this, where basically it was, I think from an actual standpoint, by the end of the year, we would have grown 5.4%. We're basically doubling it from where we were three years ago, and that's how confident we are with these accelerated growth drivers. Now quickly, I'm not gonna go spend any time on this slide. You're the owners of this company know these numbers inside and out. Kurt talked about them in the beginning. Without going line by line, we are proud that we have and demonstrated a strong financial proven discipline model.
I've been part of that model for the last eight years. I'm very proud of it. Good times and bad times, we know how to. We have our levers. We know what to go deliver to. I'd also like to thank our NXP employees to get us where we are today, as well as you as the owners and the supporters of NXP. Thank you for that. Now let's go to the three-year model for the next three years. Okay, talked about growth enough, 8%-12% for the last four hours. If we look at gross margin, our new gross margin range is 55%-58%. If you recall, three years ago, the high end was 57%, now it's 58%. That 58% is gonna be driven by our new product introductions over the next several years. It won't happen tomorrow.
It'll take some time to get there. OpEx, we've been talking about this for the last year. We feel very comfortable with 23% OpEx, 16% R&D, 7% SG&A. Now, we talked about as we grow, we probably, and we will have leverage fall through on the G&A side. For simplicity, going through the horizon, 7% is a good way to model the number. Now, let me talk about R&D. R&D is the lifeblood of this company. It's what drives the innovation. What we spend today really doesn't impact our revenue for the next three years. It impacts it post-2024. Now we take that all together, that brings an operating margin of 32%-36%. Again, this is above what we said three years ago of 31%-34%. A couple housecleaning items on the slide.
Our cash tax rate at the moment, this is on current legislation, I think, I hope this is the worst case, 15% for 2022 and 18% for the next two years. Tax is very complex. As the governments around the world decide to figure out what is the final and what they put in law, we will follow that. We pay our fair share of taxes. Paying taxes is not really a bad thing 'cause that means we're making money. We'll get back to you on a quarterly basis. We hope to refine that number as we go forward. Non-controlling interest, as you can see, just to remind everyone, this is related to our joint venture with TSMC. We have a joint venture called SSMC in Singapore, and we consolidate the full results to NXP.
Net CapEx, as you can see, this is 6%-8% over the cycle. This is higher than what we said three years ago. We said 5%-6%, and in the last year we said 7%. This is driven by the fact for us to grow at these higher levels of revenue, as you all know, 75% of our supply for back-end and tests comes internally. This is where we're spending the money. Stock-based comp, you can see the numbers there. Then let me talk about the working capital metrics. We're very proud of these. I'm proud of these. I've been involved weekly basis, making sure these are best in class. DSO of 30 days, DPO of 75 days, and inventory at 95 days.
We wish we can have that inventory today at 95 days, but we don't see it, as Kurt mentioned earlier in his one of the Q&A questions. With that, our target is to drive free cash flow as a percentage of sales of 25%. That's what we wanna drive to. We think that's a healthy level after we invest in the business and deliver to these set of numbers. Now let me share, maybe provide a little bit more color about 2022. For modeling thoughts, revenue 2022 forecast near the high end of our model. That's what we're expecting. We expect our non-GAAP gross margin and OpEx in line to our long-term mid ranges. Again, other items, cash tax, we'll get back to you. We think we can do better. We'll keep you up to date on it.
Now, CapEx will be slightly higher than our 6%-8% by a couple percentage points. Non-GAAP free cash flow as a percentage of total sales will be below the target that we wanna drive through in the long term. We believe this is temporarily, and in 2023 and 2024 we'll get it back where it needs to be based on these targets. Again, this feeds nicely. I'll talk about capital return. There's no change in our policy. As mentioned, we have a very consistent with our capital allocation strategy for the past several years. Therefore, we decided to add a simple graph that shows four time periods, 2015, 2018, 2021, and 2024. These blue bars behind me represent the cash that we return to our owners.
The percentage above the bars represent the cum cash that we return over our cum free cash flow percentage. What you can see in 2021, since 2015, we have returned over $14 billion to our owners. As we look ahead with the three-year financial commitments that we just provided, we expect to return another $10 billion-$13 billion to our owners, and this is roughly at 100% of our free cash flow. Then finally, we're very disciplined with our capital allocation strategy, and you can see our leverage ratio. That's very important to us. At times when it's above 2x, we don't. When it's below 2x, we provide that cash back to you all. With that, let me close with the NXP investment thesis slide, and then we'll open up for Q&A.
Again, as mentioned several times, our investment thesis is very simple. We're gonna drive that strong growth above market, 8%-12%. We're confident in it. I hope you are after four hours of listening in from the team. We're gonna continue to execute to our financial discipline model and drive that operating margin to be between 32% and 36%. Lastly, as I mentioned again, we're gonna return all excess free cash flow back to our owners with a range of $10 billion-$13 billion over the next three years. We're very confident we can deliver these set of financial commitments, and I'd like to thank you all for owning NXP and supporting us. We are NXP, and with that, maybe I'll invite Kurt to come up and the whole MT to take questions about today's presentation.
All right, everyone, we're, we'll undertake some questions here for the next 15 minutes or so. Please feel free. Will, since you had your hand in the air first.
Okay.
It's going.
Thanks. Nice. Very good messages around growth. Certainly margins look good too, but I want to focus on the growth for a minute. Not only good numbers, but very strong messages around justification of that growth and new products, more so than what we saw last time. When we compare NXP's history, let's say the last couple analyst days relative to what we're seeing today, and we see these new products driving this growth. Is this just better focused R&D, better market conditions, or are there other things contributing to this, perhaps historical M&A where you maybe one plus one delivered more than two. And then I do have a follow-up, if I can.
Yeah. Thanks, Will. Very clearly, this is a result of the tuning of the portfolio into this edge compute opportunity, which I described in the beginning. It really took a number of years. First step, Freescale, NXP. I mean, you've seen all the way today the combination of processing with surrounding components. We couldn't have done that without the combination of Freescale, NXP. Then the addition of connectivity was kind of the missing piece. Now, having all of these pieces is not good enough. Now we have to enable customers, we have to get the tool chain ready, et cetera. So that took a couple of years. Yes, now it is a matter of the availability of an unrivaled portfolio, which puts us in that position to be very specific and very confident about the accelerated growth. That's really the point.
The follow-up is there was in the capital allocation discussion no mention of M&A. Yet it's been, you know, really quite useful from a growth perspective. Is there anything else to consider relative to that trajectory for capital allocation? Thank you.
Well, we definitely continue, and you heard it a little bit on the sidelines through a couple of the comments of my colleagues earlier today to do tuck-in acquisitions because this is a dynamic world. I mean, the completeness of the portfolio continues to need little additions. I would especially name artificial intelligence, algorithms, low power compute, and in general, the software capability in the company is something which we need to continue to strengthen. There was actually a question earlier of monetizing software. It is an integral part of our value proposition. On the tuck-in side, definitely we continue in favor of that strength and differentiation of the portfolio. On anything else, Will, we are not here to speculate about that today.
Great. Chris, right over here.
Thank you. A question about your capacity plans and, you know, how that's affecting you folks in the industry now. Can you say that going forward the capacity for this plan is now available? Can you talk about the commitments that you've made and you've had to make in order to do that? What your suppliers are talking about to us is that it, you know, now to get capacity, it requires firm commitments. If you're making those commitments, you know, structurally, how is that different than what it is in the past? I guess one of the things that, you know, is a concern is when the next downturn happens, does that have more of effect on margins because you've had to make volume commitments?
Capacity availability currently is definitely a key revenue driver because as I said on earnings, demand is outstripping supply capability. That has been the case for the past couple of quarters. It continues to be in Q4, and we actually expect it to continue through a good part of next year, which in itself is very healthy, by the way. I mean, that's very positive. On the capacity side, Bill talked about it earlier, the back end test and assembly, 75%, 80% we do in-house, and some of the lift in CapEx is actually to enable what you were asking for. It's gradual, and I think we just tune it to the wafers becoming available. I mean, that's something which is very nicely under control.
On the wafer side, you know our hybrid manufacturing strategy, just short of 50% is in-house manufacturing, which we are optimizing. Most of the future growth is coming from foundry partners. That's, I guess, where your question is coming to. The nature is changing towards much deeper partnerships, longer-term partnerships, so alignment with foundry partners for multiple year deals. We published this and again, in earnings, we discussed about it. We currently have like $4.4 billion , I think, standing there as commitments, which is not just foundries. I mean, that's overall, but it includes, of course, also wafers. I'm not worried about that at all. The $4.4 billion Are for multiple years. I mean, that's not just a number for next year.
Secondly, we are balancing this very nicely with the same commitment from our customers, actually higher, which is the concept of NCNR orders. A large part of our customer base is actually rushing all over us offering NCNR orders. It's not like we have to ask them for it, but actually they want to offer it like orders which are non-cancellable, non-returnable for the full year of next year, by the way, volume and price being fixed over each quarter. That puts us in a good balance. No, I'm absolutely not concerned about a downturn which would eventually have too much capacity and too little demand. I would rather say we work very hard, and that continues on a daily basis to make sure we get more capacity available for the enormously strong demand.
I hope you understood from today's presentation, a lot of the demand is secular in nature. It isn't cyclical, it is actually secular. From that perspective, I think the bigger concern, if you will, is how can we make that capacity available to match the demand? It's not about over-committing on a supply side.
Okay. Vivek, right here.
Thank you. Actually, one question on gross margins and second on the operating margin.
Leverage in the model. On gross margins, you know, one of your very microcontroller exposed peers presented recently, so they had a lower growth rate than yours, but they targeted gross margins that are 10 points higher. My question is, other than mix, why are you in this gross margin range when many other microcontroller peers, right, granted different mix, are in such a different state of gross margin? What is it about your mix or customers or anything else that limits you to this gross margin range? Secondly, on the EBIT side, if I take the midpoint of your 32%-36%, you're already there in Q4, right, or in that neighborhood.
Given all the investments you're making and all the growth you're seeing in these new products and software and other areas and the market leadership, why isn't there more operating leverage in the model?
Well, we are here, Vivek, to give you a plan which we can back up with confidence. That never means there is not a likelihood we can outperform the plan. That doesn't say this is a conservative plan, but it says we are committed to the model which we are providing. Given that commitment which we want to spit out, we are here with something which we feel is absolutely feasible and where we want to go. Now on the margin question, I guess I know which competitor you're speaking about. The portfolios are absolutely not comparable. The difference was in your question, is about growth. We are here with a portfolio which is geared for long-term growth. That is very different to a long-tail business, which is more commodity-like on, I wouldn't say older, but more marginal kind of applications.
We really believe that long-term growth of NXP is the biggest value enabler, which we can do for our owners. That's actually where the focus in the portfolio is, and there we feel that the margin guidance, which you just gave, is about right. It's about meaningful. Now, Bill, however, said the lift in the high end of the margin, the 58%, so the 57% now going to 58%, is indeed a result of NPI. We are fully utilized today. Where we move the margin gradually up is new product introductions. That doesn't happen overnight. All the businesses you saw today, they all have very clear targets that new products coming out will beat the old ones from a gross margin perspective. That's our mechanism to grow gross margins. It's not about utilization anymore.
Operating leverage.
Bill, I think you said it on operating leverage. The R&D 16% are the lifeline of what I just said. I mean, there is no growth and there is no leadership in these applications if you don't sufficiently invest. Now, we do it in a very focused manner, but we believe in the mix of businesses we have between digital, analog, and RF, 16% is about the right number. Bill carefully indicated, especially on the SG&A side, there might be some operating leverage. We give a number for three years, so we want to be confident and safe such that the 23% seem to make sense.
Chris, down here in front. Thanks, Jason.
Thanks so much, Chris Rolland, Susquehanna. Embedded in your model, I wanted to know the expectations for pricing, where you think they can go from here. Movements to longer term agreements, which is, you know, going on, I think in auto specifically. Yeah, I guess, those two things and then also perhaps talk about this movement in auto from tier ones to maybe moving much more direct and how this plays into pricing and LTAs.
Yeah. First of all, on the pricing, let's not be confused. The growth which we just projected, the 8%-12%, again, which is actually doubling our former performance, that growth is not dependent on pricing. I mean, most of the pricing adjustment is behind us. It's done with the 2021 number. Now, the numbers we gave today are 2021 to 2024, so pricing is kind of history on this. On the direct in auto and in auto in general on the growth, because I got some noise earlier in a few discussions here on the floor, we don't need the SAR for our auto growth. It is really important to understand what we talked about, and I think Jens and Torsten and Henri were very clear, this is about content growth, and it's about share gains.
We don't have to lean on the SAR, and we are honestly speaking in this environment, not in a position to project the SAR. I mean, there is all sorts of numbers out there. We've walked away from this. Our 9%-14% is our performance from a content and share gain perspective. The SAR is somewhere underneath, but that's not moving the needle. On the more direct, this is much more about partnerships and strategic alignment. So we are in much, much more direct contact with the OEMs. Both when it's about transparency for the supply chain to understand what's the future requirements from a supply perspective for specific new models which are coming up, et cetera. Enormously important and part of this change, which I think, Ross was asking earlier before in the learning in the industry.
That increasingly moves into innovation partnerships because having that partnership with the OEMs in automotive has now opened the door to do much more together when it goes into innovation, which is future business growth. That does not mean a direct business relationship from a ship to perspective between us and the OEMs. I mean, we still have actually distributors, CEMs, ODMs, tier one customers, they are all between in all sorts of configurations. The aspiration is not to enter into a direct business relationship from a ship to perspective with OEMs, but significantly increase the intimacy and the partnership in innovation, in business creation, which I think is a fabulous positive outcome out of the current situation. I mean, it's a step function. It's not like a little bit more than in the past.
I would call this a huge step function, which I think for this massive auto business in this company. Let's not forget, it's half of NXP is automotive. It's a massive move in partnership, customer loyalty with where the business really comes from. Because I mean, those people who are building the cars, they need the semiconductors ultimately. We consider this as a very, very positive move.
We have time for one more question before we go to closing remarks. Ross?
Probably perfectly suited to be the last question. You talked a lot about the sexy growth areas of the business, but you also have your core business, your high RMS, and I think you talked about that growing 5% CAGR. If you look over the last couple of years since your last analyst meeting, it's been closer to flat. Could you just walk us through any assumptions you have in that? Is it as simple as any third party say the market's gonna grow 5%, and so you do too? Or will that market benefit and change to a more pricing-friendly market? There has to be some dynamics to go from the flat to the up 5%. Can you just walk us through your thought process in that roughly $9 billion target?
I mean, it's a mix of many things, right? What is important is that in all of these areas, we have typically very large market shares. We have a significant exposure. Now, the bigger part obviously is in automotive and industrial since this is three-quarters of the company. I would just say that part of the market we are exposed to has accelerated anyhow going forward. Ross, it's just a function of this market going faster, and we are associated with it. Given our large share position, we have a significant exposure to this and benefit from it in a more significant way. On the pricing side, I just gave the answer here. Consider the pricing more behind us than ahead of us.
This growth, and that applies then to everything, so to the full 8%-12%, is actually relatively clean off pricing, which makes it, I think, very strong and very, very clean. Jeff, I think from a timing perspective, we-
We're spot on. You wanna make any last comments?
We got to the end. First of all, I wanna thank all of you actually for being here physically, live and in the room because it makes such a big difference to me. I know it makes a big difference to the team. It's so much better than talking to a screen. Secondly, I really hope that the day was insightful, that we could help you understand much better why we have that strong confidence of doubling our profitable growth of NXP for the period to come. I mean, this is a needle-moving event for us. That's why I was very happy about the question, what is it? I think, Will, you were asking for what makes it different to the history?
The difference is that the portfolio which we have been carefully architecting, the customer relations which we have been building, is now playing out. Many of these things, and Bill you said it, have been in the works for a longer period of time. Semiconductors business is not from today to tomorrow. We are building ecosystems that takes time, but once we have them, it is very sticky. That is what NXP stands for, a high commitment to significant above market growth for the time to come, with robust profitability, and I believe absolutely superior return of free excess cash to our owners. I wanna thank you for your time today, for your interest and confidence in NXP. I wanna thank the team for doing a fantastic job representing 30,000 employees of NXP. I think you did very well.
We are all proud of our teams, and I look forward to actually go and see you hopefully more in person going forward. Thank you very much. Thank you.