Ladies and gentlemen, please welcome Jeff Palmer, Senior Vice President, Investor Relations, NXP.
Good morning, everyone, and thank you very much for joining our 2024 NXP Investor Day. It's great to see you all here with us in Boston. A beautiful day, even though the shades are closed. You know, this is our fifth Investor Day we've held, and I think you're going to enjoy the program today. A couple of housekeeping items, as you know, I love to go through before we start any event.
This event is being simulcast on our website. The slides that you'll see today will be available for download within 24 hours, along with a replay of the video. As you know, we will be speaking to some non-GAAP measures today. A full reconciliation of those non-GAAP measures will be in the back of the deck, as well as other information on our website as you need it.
In terms of the program today, I think we've got an exciting program. We're going to start off with Kurt Sievers, who's our CEO and President. I think most of you know Kurt. He's been with the company since the very founding, as when we were carved out of Philips in 2006, and as part of the strategy development. He's going to give you an overview of our strategy that's going to help drive growth. Our theme today is bringing intelligence systems to the edge.
We're going to then move on to the industrial and IoT reportable segment. Rafael Sotomayor will take that for us. Most of you may know Rafael. He came to NXP about a decade ago. Prior to that, he was at Marvell, running, excuse me, Broadcom, was running the wireless business there. That was a Freudian slip. He'll give you an overview of our industrial IoT business.
Then Jens Hinrichsen will get up and speak to you about our automotive business. As you know, between Rafael and Jens, that represents about 75% of our total revenue. This is our strategic focus as a company. Jens has been with the company since its founding. He's held a number of roles, starting out with Advanced Analog, taking care of all of our analog products, a lot of our automotive products, and now he runs all of our automotive products.
So, very important to the company. You may wonder where some of our other team members are. We have most of our MT here today, but two people can't be with us today. One is Henri Ardevol. Henri has moved into a strategy role for the company. Then Torsten Lehmann is on medical leave right at the moment.
Then we're going to do a short Q&A after Jens' presentation. Then we'll do a break, come back, and we'll hear from Andy, who will talk about our hybrid manufacturing strategy, which we think is a unique differentiator in our ability to drive growth for the company. And then lastly, Bill will come on up and give you a sense of the financial model that I'm sure you're all patiently waiting for for the next number of years. And with that, I'm going to pass it over to Kurt.
Yeah, thanks, Jeff, very much. And good morning, everybody. Good to see a full house this morning. I remember vividly our last Investor Day in New York in November 2021. It was a very special one because it was actually literally the first week the travel ban was lifted, and it was like a personal reunion with a lot of friends. I'm glad that through the last three years, we had much more frequent opportunity to meet and see each other and discuss the future of the business. Now, today, as Jeff just laid it out, I will speak about our strategy to bring intelligent systems to the edge. I will also provide the updates to our long-term financial model.
And maybe most importantly, I will introduce then after me, as you just said, Jeff, Rafael and Jens and Andy, who will give you much more detailed and much more underpinning insight into the future growth of the NXP business. I know that you speak too much to Bill and I. In that sense, we really tried to make sure today you get a much broader exposure to our management team, who will provide you a lot of good detail.
Now, above and beyond all of that detail, there are three plus one things I really want you to take away from today. And those three plus one are the following. NXP is going to grow high single digit organically through the next couple of years. NXP is going to expand its gross margins, non-GAAP gross margins, above 60% over the next couple of years.
Thirdly, NXP is going to continue to get all of our excess free cash flow back to our owners, 100%. And if you put the three elements together, we can all easily calculate that NXP is going to double its earnings per share by 2030 plus. And the next four hours will be dedicated to underpin and give evidence on how we're going to do those three plus one outcomes. Talking about future in our business clearly needs understanding of some of the strategic moves of the past, which is why I want to get started here with taking a look back into the history of NXP starting from our IPO in 2010.
And I will speak about this with a lot of passion and conviction because I've been a member of the management team of this company since that very beginning with the IPO back in 2010, and I became president in 2018. So I have a lot of heartburn about that journey which we are speaking about here. And that journey, I would cluster in hindsight into two phases.
The first phase was about refining the strategy, which was about making sure that we deploy our R&D into those areas which would provide the most sustainable growth going forward. We were divesting assets. We were acquiring assets. And the most notable one was clearly the acquisition of Freescale in 2016. That is very important because the legacy NXP before that time had a lot of strength in analog, RF, and cybersecurity.
What we got on board with Freescale was leadership in embedded processing, in functional safety, and that together with the NXP legacy strength formed actually the foundation for what is our future strategy in bringing intelligence systems to the edge, not components, but systems. Over that entire period, we have about tripled our revenue, and we have doubled the operating margin performance of the company. We are OK with that performance. We clearly want to do better than this going forward. Now, looking more closely at the report card against what we gave you as targets three years ago in New York, I would personally say the environment has been more turbulent than I was hoping for.
When I think about the supply crisis, when I think about the ever-increasing trade tensions, when I think about the ins and outs of COVID-19, then things certainly have been a little bit more troublesome than we would have been hoping for, and in that context, we are kind of OK with the mid-single digit growth which we have achieved through that period.
What I want to clearly state, we missed the mark. We gave a target of 8%-12%. We achieved mid-single digit. Where we've done quite well is on the non-GAAP gross margin front. We managed to operate at the high end of our target, which was 55%-58%, which over the period gave us now a 200 basis points expansion of gross margins from 56%-58% this year, and finally, earnings have increased by about 20% from 21%-24%.
So, an OK result in a pretty turbulent period. What I think matters a lot to the future of the company is trust into and from our investors and investment into the future of the business. So, it's nice to quote here that we have $11 billion investment done cumulatively over that period into R&D and CapEx, which I would really call what manifests the future investment into the company's growth. At the same time, we have returned about $11 billion to our owners over that same period.
Now, going even a little bit closer into the scorecard, in this last analyst day, we had introduced the concept of differentiating between the growth of the core of the company, which is all of those high relative market share franchises, which typically grow more with market since we have such a high share that we cannot outgrow the market, versus what we call the accelerated growth drivers, which are really new application fields where we are growing far above market. And if you look at those two, then the core has grown at 2%. Those accelerated growth drivers have grown at 11%. Where it really worked out to our satisfaction is automotive, where we've shown a 9% growth over the past three years.
You see underneath at the bottom of the slides the individual growth drivers which were contributing positively to this, like our S32 processing for the software-defined vehicle electrification and even radar. I will speak about radar later. It looks a little bit below target, but there are reasons for that which have to do with inventory digestion. Where we are not satisfied is the performance in the industrial business over that three-year period.
Clearly missed the mark. But I want to remind all of us that we've done something different through that period, which is a great learning from past failures of NXP. We have not stuffed the channel. Our strategy of soft landing is asking us to a lot of discipline to not overship in the channel. Most of that industrial business goes through the channel.
So, some of that underperformance stems actually from very disciplined channel behavior, which should be a strong starting point for the recovery when and if it comes in the future. Mobile growing roughly with market over that period, which is very understandable when you know that most of our mobile business is the secure mobile wallet, where we have a very high relative market share. So, we by definition grow with market. Now, we also plot here a five-year benchmark. And I do that because the business is so cyclical that sometimes if you just take three years, you take the wrong benchmark.
Five years we felt like 2019 to 2024 here is a better benchmark for how we've done through the cycle, where you see the company has grown 7%, automotive 11%, and industrial 7%, which is much more back to the norm which we would actually also expect going forward. But I come to the new targets in a minute. With that, I move our minds to the future, to the strategy, to the growth, to the new model.
And I really zoom out far. This is the total semiconductor market. According to McKinsey, a $1.3 trillion opportunity by 2030. And if you look over the past years, growth has been driven by long waves of mega applications. Between 2000 and 2010, compute was the name of the game. Intel was the winner. Between 2010 and 2020, that was the age of mobile, of smartphones, of tablets.
Companies like Qualcomm and Broadcom really emerged strongly through that period. Since then, it is for sure cloud computing, AI, generative AI in the cloud with NVIDIA, the absolute rising star. We think from here the next big wave is intelligence systems at the edge, which will actually benefit from the artificial intelligence, from all the cloud computing, all of these edge applications, be it in the home, be it building automation, be it factory automation, be it the car as the ultimate edge device. They're all connected to the cloud and will benefit from that success the cloud is currently scoring. Now, from a consumer perspective, if you think about what technology is doing to us, the way I would describe that history is a little different.
I would say in this computing age, our environment, which we experience as human beings in our households, at work, et cetera, was pretty analog. From 2010 to 2020, it was more this on-demand environment. You could electronically order all sorts of things. But what comes next and where NXP is going to be strong at is what we call anticipate and automate. The idea is technology will know what you need next. And then technology is going to automate the delivery of that.
And that is about all the help we are getting from robotic lawnmowers, from eventually autonomous cars. All of that is about anticipating what is next and then delivering it in an automated fashion. And these are all these edge applications which I will explain how NXP is going to exploit their opportunity of growth.
And they go about automotive, factory automation, building automation, smart home, and health care. My colleagues, Jens and Rafael, in the next one and a half hours, will give you all the ideas, all the customer traction, all the product roadmaps, how we are engaging across those markets. And they all follow the same scheme. It is about automation of tasks. It is about trust which these applications have to provide. And trust comes with cybersecurity and functional safety.
And this is about efficiency increases. Rafael is going to speak about the horrible numbers of today's factory downtimes across the world, which is an enormous waste of value, which his edge applications in industrial will significantly reduce. So, the output of factories will get much better, just as one example. So, putting this in numbers, this is the McKinsey forecast of global semiconductors from 2021 to 2030.
It's a long period of time. What really sticks out is that automotive and industrial, which you can easily hear me speaking about a lot here because that's the future focus of NXP, those two, other than generative AI, are the fastest growing subsectors of the total semiconductor market, with 12% in automotive and 10% opportunity in industrial.
Mind you that those two together, with $170 billion, represent about one quarter of the future growth opportunity of the semiconductor market. It's interesting to note that this is just as large as two well-known markets together, which are wireless and computing. In the past, wireless and computing were the Holy Grail of semiconductors. Now, automotive and industrial offers the same growth opportunity over the next years as wireless and computing together. Now, what are the underlying assumptions for that growth?
And Jens will go in much more detail in automotive. The highlights are significant penetration growth of software-defined vehicles, which is about the digitization of the electronic architecture in the car, significant continued growth of the penetration of assisted safety systems, ADAS, eventually autonomy. I want to highlight here we don't need autonomous cars to hit these numbers. Eventually, they will still come. But also continued penetration of XEVs.
I know there has been a bit of a debate more recently. Is it slowing? Is it moderating? Is it more hybrids? We still believe, and all the market research tells us that by 2030, about 70% plus of the global car production is going to be electric. And by the way, this is going to be led by China. Similar in the industrial space, we see a quadrupling of the market for robotics. Gigafactories growing.
All of that is crying for edge efficiency, which Rafael is going to lay out how we're going to serve that. Now, where this comes together for NXP is the technical requirements for all of these applications. Those requirements are universal across automotive and industrial. It's not different between the two. We need functional safety and cybersecurity. All applications are connected.
They have to be cybersecured. Most applications become robotic-like applications, full automation. They need to be functionally safe. AI and ML is coming down much more to the edge. We have it today in our radar solutions. Rafael will explain how it comes to the industrial edge and how we are going to embrace AI at the edge, dedicated AI, ultra low power, and maybe most importantly, real-time performance.
When people often ask us, how are we differentiating against Qualcomm or maybe NVIDIA trying to get to the edge? It is the real-time performance of our processing solutions, which is very different to media processing, but you need real-time control in machines which have to operate safely at all times. I think, Jens, you're going to speak about this in more detail in a minute.
Now, next to specific technology requirements, there is a whole list of industry requirements which are a result of how the industry has transformed. The first one is a very positive one for NXP. It's the ecosystem shifts. Since the supply crisis, we have moved so much closer to OEM customers. But that also means we come into a position that we have the obligation, I would say we have the opportunity to resolve the complexity they are faced with.
Many of these large industrial automotive customers come from a mechanical background. They are struggling with the enormous complexity of electronic hardware and software systems. This is where we can cut in. Software. Most edge systems are software-defined. And finally, and Andy, you're going to speak about this, the geopolitical turmoil is firmly demanding more resilient supply chains, geographically more resilient, and a certain level of sustainability, which we are anchoring in our hybrid manufacturing strategy, Andy, which you will lay out in a minute.
So here we go. NXP is in a fantastic position, and we've worked hard over 10, 15 years to build that position to respond positively to those trends. It is, A, our complete portfolio, which is a result of a lot of homegrown, but also of acquisitions and tactical acquisitions, which we will continue to do. Deep core competence.
I mentioned cybersecurity a number of times. Think about the franchise for electronic ID, for mobile payment. This is where we've grown our security footing. Now we moved this into automotive and industrial. Functional safety. Freescale, which we acquired in 2016, has been forever a leader in airbag and ABS systems, which is all about early functional safety designs, which we are now porting into the software-defined vehicle. Resilience, hybrid manufacturing.
I get, Andy, you get there, and I think maybe most importantly, the simplification for our customers, which is system solutions. We are working with companies which cannot cope with the complexity of the requirements they are faced with, and we have to scale this over long-tail small customers. Rafael, you will explain how we translate from lead customer learnings into the long-tail channel-served customer base, which, by the way, is margin-wise very aggressive. I come back to that.
So, in summary, we sit and have built and continue to evolve a very strong and complete portfolio for the intelligent edge, which is compounded by trust, safety, and security, and the capability for system solutions. We are deploying that into two main market segments, which together comprise about 75% of our revenues, which is automotive and industrial and IoT.
I mentioned the subsegments before. And really important to understand, the underlying IP, the underlying deep core competence in the company is universal across the two of them. This is not like two separate companies inside NXP. There is one strong core which is being deployed into two end markets. Now, the opportunity is always larger than what we can afford. And you all want from us profit growth. So, Bill's and my job at all time is to focus our investments.
We have very successfully applied a relatively simple but strikingly strong measure over the past years, which helped actually to grow our gross margins, which is a matrix of: is the opportunity which is at hand above market growth, and secondly, do we have a path to achieve relative market share leadership? Relative market share leadership allows us to out-innovate competitors because we spend more dollars on the same application than any competitor can, which yields sustainably higher gross margin.
We have absolute evidence that where we have that RMS leadership, we are leading in gross margin performance. That stays. What I show here, we've done for many years, and we absolutely continue to do this, but we are augmenting it by one element, and that one element is we have to invest into what forms then a system solution. We cannot only look at product leadership.
For our lead at the intelligent edge with OEM customers, we need system leadership, and that's a third angle which we are adding to this matrix in order to focus and funnel our investments. This is a very, very living tool, so Bill and I and the whole management team of NXP, we run through the whole portfolio of the company every quarter, and if something falls off the grid, they get a second chance for another quarter, and if not, we stop, and then we invest into something else, so it's a pretty rigid process, but it has proven to drive gross margin performance. Now, what it all comes down to is the new growth algorithm for the company. I said in my intro, high single-digit organic growth, or translated into 6%-10% for the next couple of years.
We continue to apply the differentiation between core business and accelerated growth drivers. There is a small reclassification I want to draw your eye to, which is two of the former accelerated growth drivers from the past three years, namely applications processors for e-cockpit and automotive and the RF transistors for mobile base station networks. Those two are moving to the core. So, they are not lost for the company, but we just moved them from growth to core, which is why you find there a little reclassification. If you take the new classification, 40%, almost 40% of the business by 2027 will fall into growth, which is a good thing. So, a large part of the company falls into the bracket of what it's supposed to grow between 15%-25%. The core, we think, is going to grow at about 3% in the years to come.
Now, diving a little bit deeper into this, this is the new scorecard. So, every year we're going to update these numbers, and we're going to let you know how we are scoring against those new growth targets. Another element here is new, which is the assumptions on the right-hand side. Over the past years, we often got the question from you, what are your boundary conditions? What are you actually assuming in this turbulent environment, which is a bit of a, to make it a bit more firm, under what conditions you can achieve that growth? We listed the key, from our perspective, the key conditions here, which is a low single-digit SAAR, which is a low single-digit ASP erosion per year, which is kind of back to what we had pre-COVID.
Very important to again emphasize, this is not falling back down to the pricing levels which we had pre-COVID, but it is going back on a curve, which is a low single-digit decline per year, which we can easily overcompensate by operational efficiencies. I come to the margin in a second. Content per vehicle, and Jens, you're going to go into this in more detail, mid to high single-digit. We assume over the longer period, the GDP globally of more than 2.5% and a manufacturing PMI above 50. Those two are clearly not in place today. So, those two will have to improve going forward. We assume the geopolitical stress on our supply chains will continue. So, the regionalization of manufacturing will hold on.
We assume over time a normalized supply chain relative to tier one automotive inventories, which also means that we will go back to our target channel inventory, which is the famous two and a half months, or we translate it now in weeks. All of our peers speak about weeks. We are providing 11 weeks just as a better benchmark to competitors, which again is a significant difference of how NXP has operated through the past years versus peers. We kept it very, very low, which gives us some tailwind actually into the future. Now, value creation comes from more than revenue growth. We just know we have a higher margin entitlement than what we have achieved so far, while we've had a good trajectory. We had an increase of 200 basis points over the last three years.
And the new model which we are laying out today for the next three years is 57%-63%. So, a midpoint of 60, which is facilitated by continued operational efficiency, higher utilization. We told you in earnings just two days ago that we are in the low 70s when it comes to utilization today. Fall through from revenue growth. And there is another strategic element which layers in here. We have worked hard over the past years to expand the customer reach in the long tail through the channel. So, those many, many mass market customers which we are touching through distribution, we want more of them. And we are working very intimately with our distribution partners to grow that number of customers.
The revenue from those mass market distribution customers is accretive from a margin perspective to the whole company, which is helping here for the next years to further grow the margin. As much as our innovation work, I talked about our focusing of R&D funds on high RMS products. That comes down in the NPIs, the new product introductions being typically accretive to the existing business.
So, with new products coming in over time, the mix is actually mixing up to better margins. Now, while the semi-business is often surprisingly short-term, a lot of the strategic actions we have to take are very long-term. So, I offer you here a bit more perspective about gross margin trajectory because a lot of the actions which we take today are playing out in gross margin benefits in 2027 and beyond only, not in the next three years.
I think this summer we talked about the VSMC joint venture with Vanguard, which we put in place in Singapore, which is giving us access to a 300 millimeter factory at cost. We said this would add about 200 basis points of additional margin once it runs full out, 200 basis points for the whole company margin. Now, there is more than this. With that factory coming online, we can start to consolidate the existing eight-inch facilities of NXP. We will further bring down the fixed cost basis of the company because we are adding space in 300 millimeter manufacturing capability. Of course, we continue with fab throughput. We continue with mass market and NPI.
And that together, we assume will be about 400 basis points additional gross margin expansion, non-GAAP gross margin expansion above the range which we gave you with 57%-63% for the next three years. And that gets me back to the three plus one points I mentioned in the beginning of my presentation this morning. We look at a high single-digit organic growth over the next three years, 6%-10%, midpoint of 8%, driven by two main elements, the S32 processing platform for the software-defined vehicle and the intelligent edge systems in industrial. We will drive gross margins above 60%, non-GAAP gross margins above 60%, which is a continuing journey. If you go back eight years, and I think, Bill, you're going to show this, we've ever increased, and we see there is more headroom we can go for.
We continue our strict policy of 100% of excess free cash flow to you, to our owners. That, in summary, should double our non-GAAP EPS by 2030 plus. With that, I thank you for the early morning attention here in Boston this morning. I hand over to first Rafael, who's going to shed much more light in how that growth algorithm of 8%-12% is going to work out in the industrial space. Rafael.
Thank you.
Here you go.
I think I have to use this microphone.
Thank you.
Thank you, Kurt. Good morning. So, I'm Rafael, and I will be discussing our strategic vision for industrial. And before I start, I'll take a page out of Kurt's playbook, and I'll tell you that in this presentation, I'm also going to leave you with three points. And the first point that I want you to take away is industrial is going to a transformation. It requires, it demands innovation. It demands new products. And while you may not think of NXP as an industrial player, this shift towards innovation is creating a tremendous opportunity for us. The market needs align very well with our technology strengths. It kind of makes it fall right in the sweet spot of what we do best.
Second point I'll make is Kurt talked about how edge systems are becoming complex and how solving that complexity is the path to leadership, well, in industrial, we really embrace this concept of solving complexity for leadership by developing comprehensive solutions with our key customers, and three, I'll tell you, is we take these solutions to the broad mass market, and this last step is crucial since it's a significant portion of the market, and we do with a strategic sales plan that is aligned to really target and reach this fragmented market.
Now, what is driving this transformation? If just in manufacturing, this is a manufacturing number. Manufacturing represents about 16% of the global GDP. So, there's not a coincidence. The governments, the industry, they're trying to really attack the issue and enable, augment efficiencies, reduce downtime, increase resilience.
But if you look at the numbers here, we have a long way to go. I want you to focus on this number here, $1.5 trillion in unplanned downtime. That's a number, that's an estimate that came from Siemens of a wasted, this is for the global 500 industrial companies. They have $1.5 trillion of losses due to unplanned downtime. This is 10% of their revenue. So, a long way to go. Clearly, a market that is primed for improvement. So, what's driving the transformation?
Economics. So, the mega trends at its essence, these mega trends are based on the need for efficiency and the need for sustainability. And the technical drivers driving these mega trends are all the technologies that drive autonomy, energy efficiency, like low power, electrification, and resilience, security, safety. And these mega trends are best realized, these drivers are best realized by systems at the edge.
Edge systems reduce dependency on the cloud, lowering power consumption, strengthening data protection. They are most of the time autonomous, real-time, like they handle data and do decisions locally. Modern systems now are software upgradable, so you're always secure with the latest countermeasures and updated with the latest features. These systems now are having longer, they live longer.
They adapt to evolving needs through time, so it's a very interesting, fascinating time for industrial. This need for innovation is disruptive. This need for new solutions is driving an upgrade cycle, so in this environment, incumbency is not a given, and for us, this presents a really interesting opportunity, a very desirable opportunity. I'm going to shift now to why us, what makes us so uniquely positioned to take advantage of this opportunity, so if you look at, this is our game plan. This is simple.
It's a simple game plan, but let's not overlook the complexity of execution. As a matter of fact, we think the execution complexity is part of our competitive moat, so first, we develop great products. It always starts with great products, then we partner with customers and we develop, co-develop system solutions, and when these system solutions are mature and these products are mature, we take it to the broad market, which you can see from the chart. It represents the largest opportunity of the segment. Simple game plan.
Tough to execute. Products, so even though it takes more, even takes more than great products to be successful in industrial, it does take great products. Differentiated products. Products that are designed fit for purpose, not repurposed from other markets like some of our competition, fit for purpose. Industrial applications are unique, very diverse.
It requires great insight into the use case to develop the right specifications, the right products. Before I dive into our products, which we're quite excited about, I want to revisit once more this premise because it will come through in the rest of my presentation and the premises of solving and simplifying complexity because it is at the core of what we do. The reason I want to revisit is because it's critical to what we do in our strategic plan.
We know that what our customers are going to be doing tomorrow is very different than what they're doing today. This leap into tomorrow is not easy. The technology needed for tomorrow's solutions are challenging. Sometimes they're even out of reach from our customers' technical capabilities. This is where we come in. This is our job.
This, our role is to make sure they're always within reach, that we not only pioneer solutions and innovative solutions, but we make it easy to use, it's easy to deploy. Our goal, and we aim to make this complexity an asset, an asset to drive growth and to drive leadership, and the reason for that is important because the technologies needed to make industrial products are many, are large.
They continue to grow. We need to continue to invest in new IP. We continue to invest in new hardware and software platforms, and so we need to make sure that when we release them, they all come together relatively easy for the customers to use with very low investment and low effort on their part. Now, the foundation of our industrial strategy is our processors.
Our processors is the most strategic decisions that our customers make when they select a silicon vendor. This is where their investment goes. This is where their IP goes. Most of the investment, and this is a very strategic step in their selection process, processors, and we have a broad range in our portfolio. We have three main categories, MCUs, crossover MCUs, and application processors.
Having this breadth is a significant part of the value proposition that we offer to our customers, the ability for our customers to leverage the software investment from small low-power devices all the way to high-performance application processors. We offer an entire breadth of processing capabilities and performances. Our customers can come to us for any type of MCU or any type of solution that requires MCUs and application processors. We become one-stop shop for them.
You will be hard-pressed to find another industrial player with this portfolio. Yes, some of our competition have created some positions in MCU, but they have significant gaps in the high-performance segments. They have significant gaps in connectivity. They have significant gaps in security. Some of our competition comes from the analog world trying to enter the digital space and the processing space. That's hard. What you see here, this is difficult to replicate. It takes generations.
So, starting with our MCUs, this is just to be specific, these are 32-bit ARM-based microcontrollers. The MCX is our next generation microcontrollers. We released them a few years ago. We released the family, we released the product line with six families, each one of them targeting a different part of the segment of segments of the market. NXP MCUs are known for innovation in the market.
Let me kind of explain to you a few breakthroughs that we introduce first in the market in MCUs. If you look at the bottom here, the MCX N, this is our high-performance product. This was the first MCU in the market with an integrated AI accelerator with the performance and the security needed to run AI applications in an MCU. If you look at the top level here, the MCX A, this is our workhorse. This has got the Goldilocks balance of performance and cost.
This is the first MCU with a 10BASE-T1S integrated for Ethernet applications over a single pair of twisted copper wire. This is to actually take advantage of legacy environments, simplify cabling, and again, differentiated features for MCUs. The MCX E is a category of products that targets functional safety. What does that mean? That means it ensures reliable operations under faults. What faults are?
Could be memory issues, logic issues, timing issues. Sometimes you have blips and electrical issues that cause timing. All that, these products are very resilient to noise with high EMC standards. They're used for mission-critical applications or applications that control motion. What I'll leave you with here is each one of these segments has a unique value proposition and an innovation angle, and at the same time, we maintain software consistency.
We have tremendous respect for the investment of software for our customers. Now, just migrating now to following the continuum and performance from MCUs, I want to jump into crossovers. Crossovers, we invented crossovers about seven years ago. At the time, there was this performance need and gap that MCUs couldn't fill, and the only way to do it was to jump to an application processor.
But the moment you did that, developers lost the real-time characteristics on an MCU. The BOM cost went up significantly. And most likely, you had to invest more in software. Sometimes this move was not viable. So, enter crossover. That gives you a significant performance boost with an MCU platform. NXP is the benchmark in this category of products. And since we entered the market, we created the market, we continued to kind of invest on it and deliver many, many, many firsts.
We were the first with a dual core running a one gigahertz with integrated graphics. If you look here on the top, this is a new product, the RT1180. This is tailored for factory automation. It's got integrated TSN switches, enhanced security. Then we created the first tri-band radio, an MCU, monolithic integration of tri-band radio with an MCU.
And then here at the bottom, this is literally the latest. This is hot off the press, the RT700. At risk of sounding pretentious, and I'm going to say it anyway, this product is a bit of a technical marvel because it attempts to solve a paradox. It's architected for ultra-low-power applications that need high performance. So, this device is dual-core. It's got dual DSPs. It's got an accelerator for AI.
It's got graphics. It's got security. It's got an independent system to manage all the I/O in extreme low power. There's nothing like this in the market with this level of performance and a power profile. And this is actually customers waiting for it. They're actually going to go to production with this device in rapid time. You will see products next year on it. And they span everything from scanners to handheld devices to wearables.
Quite a unique product. Now, we walked through that continuum of MCU to crossovers. Now we're into the performance of application processors. And very similar to crossovers. In application processors for industrial, NXP is the benchmark. i.MX is a franchise. And the i.MX 9, this is our latest family and our latest generation application processors. We entered the market with three families, all targeting different ranges of cost, performance, features. The i.MX 93, it's entry-level Linux box.
The i.MX 94 targets industrial applications heavy with networking. And the i.MX 95, it focuses on graphics, HMI, multimedia. Now, what makes these processors different is that they're uniquely designed for industrial. They're not generic application processors. They have features to uniquely enable applications and features that are needed for industrial. So, let me give you some examples. For instance, the heterogeneous compute.
These application processors not only run Linux or Android, but they have these dedicated MCUs or subsystems, autonomous MCUs integrated into SoC with their own peripherals, their own operating system, their own timers, their own I/O, and these devices are able to run real-time applications like motor control or sensor fusion with the same level of precision that a discrete MCU would do. These products have industrial networking.
They have TSN switches, and we also support different industrial protocols like Profinet or EtherCAT, among others. They have integrated AI accelerators. They have real-time safety islands. These provide a fail-safe mode that is independent of the host, meaning no customer application can interfere with it, and this allows for a fail-safe mode operation. Think of robotic arm, the malfunctions. The ability to return this arm to a rest-safe state is done by this feature. These application processors are unique.
This is the reason why the i.MX has become the standard in industrial application processors. Now, when I speak of, and I think you sense, when I speak of the value proposition of the different products that we have, MCUs, crossovers, or the application processors, you sense that most of the time the differentiation comes from differentiated IP. It's anchored on differentiated IP. We chose these 12 technology pillars as a source of differentiation for us.
We believe that we can add considerable value by investing in these IPs, by sizing them properly depending on the product they're going to go to, by controlling the time to market, by controlling the requirements that they deliver. Each one of these IPs represents a dedicated investment and focus for NXP. That investment is what makes us different.
So, I'm going to go through two examples really quick just to kind of illustrate why this is a source of differentiation for us. And I'm going to start with something that we're known for, which is security. Security, uncompromised security, is fundamental for edge systems. More and more of these systems are connected or interconnected, and they represent a large attack surface for bad actors. So, a security breach could compromise a system with massive ramifications. So, we invest heavily on security.
We probably, in the industrial space, we're probably the largest players on security. Again, Kurt mentioned it. The security IP, the know-how, comes from the hundreds and hundreds of millions of products that we ship in the market, like mobile wallet, passport, credit cards. Those products require high levels of security. So, we take that know-how and that differentiated IP, and we apply it to edge devices.
So, we invest quite a lot on IP. We are pioneers on post-quantum cryptography. Quantum computing has the potential of rendering current crypto useless. So, PQC standards will prevent that. By the way, industrial customers want that in their products today. Those products are going to be in the market for 10, 15 years in the field. So, they want to be protected.
Now, having great IP and security is not enough. You have to design your chips and your SOCs with security in mind. And you also need to make sure that the customers implement security properly, which has been a bit of a pain in the past. So, we took a page of our secure elements, which is another product that we have, and we created an independent subsystem that gets integrated in all of our chips.
And this independent system handles all the security aspects of the chip and makes it easy for the customers to use security properly. So, this is definitely a source of differentiation. Customers in the past took security as an afterthought, and that is not the case anymore. We have the strongest offering in security. And the reason it's also relevant is because we see a trend of regulation coming.
Governments are realizing the potential impact of cyber attacks, and they're creating higher regulations. Thinking of, for instance, in Europe, the Cyber Resilience Act is requiring compliant devices by 2028. And we see similar efforts in the rest of the world. So, all these regulations are going to raise the bar for security. And the good thing is with NXP, customers are going to be covered.
Now, the second one, this is a new area of focus for us and a new area of investment, which is AI. And what I mean here by AI is the ability to run AI models at the edge. This is a new investment focus for us. If you look at AI, AI models are usually created in the cloud, and they're inferred in the cloud. But that inference is now moving to the edge.
And when you move the inference to the edge, the data stays at the edge. And then you have benefits, obviously, with privacy, with security, with robustness of operations. You don't have to depend on a reliable connection to the cloud. You have benefits of latency and real-time. So, our investment with respect to AI at the edge is in three areas. One is we invest in accelerators.
These are very efficient and scalable accelerators, so we can put the different sizes and performance in the different category of products. We invest in software, especially on the SDK. We call our AI SDK eIQ. This is a software that enables the creation of models and the optimization of models, and the last one is pre-optimized models. There are hundreds of open-source models today available for voice, for audio, for text, for video, for text-to-speech, for speech, you name it.
So, you have to actually optimize and pre-tune those models to run in your target hardware. Now, that last step is critical for, and it's what makes edge AI different. When you grab these models that were created in an unconstrained environment, remember, now the edge is a constrained environment. They're constrained with memory. By the way, AI is very, very hungry with memory.
These edge devices are constrained with memory, constrained with power, constrained with computing capabilities. You need to tune these models. You need to quantize these models, and when you quantize these models, you lose precision, and the whole concept here is to do it and lose precision, but not lose accuracy. Because the moment you lose accuracy, you lose the use cases, so these areas are critical, and that's what makes running AI at the edge different than running on the cloud.
It's a delicate balance, so now, why do we want to own our own destiny? This chart shows you there's two horizontal axes here. You have one on the top with performance. This is on teraflops, and you have one in power, which is watts, and you have the different edge AI use cases that, at least today, we get requests from our customers.
You can see that the performance requirements, both in power and AI performance, are quite different. You look at this and you think, wow, this looks very similar now. The decisions we make for processing for edge devices. AI at the edge also needs to be fit for purpose. This is a reason why controlling your destiny by controlling the performance, the size, and the timing of this IP is critical.
We intend to contribute to the adoption of AI at the edge by making AI prevalent in our products, prevalent completely in all our i.MX products and application processors and a significant portion of our MCUs. In phase one, we focus on developing a foundation of products, differentiated products with differentiated IP. Now, I move to phase two, and this is where we tackle complexity and innovation at a system level.
We engage with our customers to develop system solutions and by doing so, we not only guide them through technology selection, we ensure all these pieces and components come together to deliver the right KPIs, the right metrics, and the right functionality for them so, when you go from discussing products to discussing how to solve challenges at a system level, you elevate the discussion.
You elevate the conversation and now it's all about how to really seize the opportunity, how to actually we go from unlocking the transformational potential of your technology and the moment you do that, you become a trusted advisor to your customer. You create tighter bonds and so, in order to kind of bring that concept to life, I'm going to now give you a testimonial of the CEO of Johnson Controls on how he perceives their relationship with NXP.
We have a historic opportunity to make our customers' buildings smarter, more efficient, and more sustainable. Our customers are on a digital transformation journey using sensors, edge processing, and communication technologies that are deployed throughout their buildings. This has yielded many insights into how the world's buildings can be managed much more efficiently. But to fully benefit from this digital transformation, we need to increase the automation in our buildings to take action in real time based on this new data.
In our evaluation process, we found platforming is the best way to enable innovation in our products. Platforming reduces the development cycle and allows easy upgrades to the latest technology. We looked for partners with scalable hardware and similar software architectures that could scale across multiple platforms. NXP was one of a few companies who could support this platforming concept and serve the industrial market.
NXP and Johnson Controls recognized that digital transformation requires more than just assembling point products. We needed to transform our way of thinking and work together to develop a collaborative mindset. By building around platforms of silicon product families, Johnson Controls can create a variety of products with lower R&D costs using common architectures, software libraries, and capabilities while still providing scalability across performance and price. NXP provides scalable hardware and architecture, which allows us to quickly develop products for different markets.
Their manufacturing footprint across multiple geographies is crucial to both Johnson Controls and our customers to avoid geopolitical risks and ensures a more secure and resilient supply chain. As our relationship with NXP has grown, their value to us has expanded beyond technology to a collaborative way of working. Together, we tackle market challenges and find opportunities to create value.
Examples include our edge products, which use NXP processes and advanced security features to secure our customers' buildings. We look forward to continuing our collaboration with NXP as Johnson Controls leads the journey towards smarter, safer, and more sustainable buildings. So just before I jump into this slide, and I appreciate George's graciousness to actually kind of give us this small video here, I want to just kind of highlight how these relationships with our key customers take time.
This is not a one-to-one relationship, but it's actually many engineers at Johnson Controls talking to many engineers at NXP, co-creating, developing trust and collaboration attitude towards creating innovation and developing system solutions. So these solutions, these relationships that help us create solutions are very valuable. Now, this slide here shows you our focus markets.
Our focus markets, and this is where we develop products and we develop solutions in these target markets: industrial, from factory automation, home building, power and energy, and healthcare. Why these segments? These are the segments driving the megatrends. They're driving the transformation in industrial. It's not a surprise they're also the fastest growing part of the segment, growing at almost twice the rest of the segments in industrial. Now, I'm going to give you a few examples now.
I'm running out of time, maybe one. I'll give you an example of our reference design. Remember, the reason we make these designs, the system solutions with our customers: one is to understand a bit more of where the market is going, the technical needs that we need to put into our roadmaps, and also to enhance our relationship with our customers.
This, and I'll get to it, this is an example of a reference design we did for factory automation with a leading European customer. And this was a very powerful example because not only did it allow us to really understand the system, but actually created also and helped us design, in this case, our crossover MCU here, the 1180, which I already mentioned before, and our analog front end. They had a say into what those features of those products needed.
Now, before I get into the weeds here of what this reference design and this system solution attempts to solve, let me kind of step back and maybe create a picture of the problem statement here. In factories, operations rely on synchronized systems reacting constantly to data coming from sensors from all corners of the factory.
Data needs to be acquired, and they need to be delivered to probably hundreds of PLCs, and these PLCs analyze the data, and they take action, and they send commands to actuators, which in turn move motors, conveyor belts, robotic arms, you name it. Now, all these tasks happen at precise cycles from a few hundred microseconds to a few milliseconds, so synchronization is required, then at the same time data is analyzed, and they monitor trends.
They try to detect anomalies, so this reference design is for a remote IO, so remote IOs play a critical role in this system. They gather the data from the sensors. They collect the data from the sensors. They connect to the actuators and give them commands. It communicates to the PLCs. It just basically is at the center of maintaining data synchronization. That's where remote IO is.
So what do you need in a system like that? Well, you need really good networking. You need deterministic, low-latency, encrypted data communications. You need a programmable high-resolution AFE. Why? Because this AFE collects the data from the sensors, but it needs to be programmable because there are different sensors in the factory. Sometimes you're talking to one type of sensors. You're talking to a different type of sensors, but you do it with one remote IO. And at the same time, you need robust security. All these connections are vital and mission-critical for the operations of the factory. Now, when we did this with this lead customer in Europe, it delivered incredible insight on what real market needs are. But it gave us one more insight, and that's more of an internal insight.
And I'll let you know because we consider this a value proposition for us, is that the products stop being product roadmaps. The roadmaps and the way we think about our products is different now. They're interconnected. And now you start thinking about not no longer about isolated product roadmaps, but about systems, about platforms. And all of a sudden, your individual product roadmaps become integral roadmaps. And the only way you can do this is if you have all the components to create the system and if you have the insight of where the system is going. So we see this as a strength for us in the future. At the same time, we have similar reference designs in different other areas. We have the reference designs for vision that goes into different markets. Power and energy solutions is a big one as well for us.
This is the same technology that we use in electrification and auto here, and Jens is going to explain to it. But this, again, with the same purpose that is to get insight, to get closer to our customers, and to stay ahead of future requirements in industrial, which I remind you is a diverse segment with all sorts of unique requirements, so understanding the system is critical to stay ahead, so that gets me to my last step, the final step, which is critical. It's a critical step to our path to leadership, which reaching this segment is fundamental, so we not only want to be trusted advisors to our key customers, we want to be trusted advisors to thousands of customers, so how do we do that, well, we need to do things different. We have to act and think differently.
So we have to take this concept of collaboration and solving complexity to a new level. We need to put these concepts on steroids. And so we amplify. We do it through our global sales distributor network, e-tellers, design houses. We invest in a strong digital presence. See, friction in this segment, friction is the enemy, the main enemy. So anything that could generate a customer objection should be eliminated. And that is usually around access to collaterals, ease of use of your material, training, support. So one, we need strong collaterals. Technically, we need to make sure that these collaterals are accessible, applicable for this type of customers. And so we develop a very robust and scalable hardware platform that allows the customers to develop software code on a target hardware with minimal upfront investment. We have also an application code hub.
This is an online repository with hundreds of software samples that customers can use at their disposal. They go search, look for their solutions, and this is free. I mean, this is free for NXP. Our goal is to make sure that we jump-start their development. We also need a vibrant channel to be able to disseminate and customize our collaterals to train and support the customer in every phase of their development, so we have a very well-trained global sales distributor network. We have strong partnerships with e-tailers. We have a very robust partnership, or I would say a very robust ecosystem of embedded board solutions. EBS partners are very important for us, so let me give you an example of when things go well, what happens. We developed this design with one of our EBS partners. They grabbed our application processors.
We put some functionality for touch sensing, graphics, and voice capability, and rapidly, this particular board becomes a market leader in coffee makers. You may think, "Okay, it's not a big deal. It's not a chance it doesn't change the world," but niches like that are everywhere in the market. We ship about six million units in the healthcare business with three EBS, so we have to think different about this market. A grain of sand doesn't weigh much, but a bucket of grain of sand weighs a lot, and so every MCU that you sell, you create an annuity, a small annuity. The game here is to create layers and layers of annuity so it becomes material. There's no small customer in this market. The key is here: the ability to reach that customer in a cost-effective way.
The investment that we have in empowering this channel for demand creation is critical. So unlike our competition, which decided to do this on their own, we don't consider this just a distribution network. This is a critical part of our strategy to create leadership and growth in this segment. And for us, it requires a multi-prong approach. And NXP is very good at executing this game. So before I go to the last slide, one more commercial and video from one of our key partners. This is a key partner from us, which is Honeywell. And they're going to give us their perspective on why they partnered with NXP. With a long share of history, Honeywell relies on NXP as an important partner.
So as Honeywell focuses its business on the three compelling megatrends of automation, the future of aviation, and the energy transitions, partners like NXP enhance our progress and create value for our two companies and our customers. By designing the system in silicon in tandem with co-creation with our customers, Honeywell and NXP are able to provide better solutions for our building automation customers. What do these customers want? Safer, more sustainable buildings with easy-to-implement solutions and more autonomous controls that help boost operational efficiency.
Together with NXP, we give them exactly that. A perfect example of this is our optimizer building automation controller, which utilizes machine learning to centrally coordinate a building's overall efficiency, comfort, and sustainability efforts. We've engaged with NXP closely because we saw early on that few partners possessed the necessary ingredients of a scalable hardware edge platform offering a seamless software architecture and toolchain.
NXP also placed a strategic value on the relationship with Honeywell. This led to regular direct interactions and workshops across our technology and business organizations, allowing us to provide critical inputs for NXP engineering as they construct their next-generation products, software, and solutions. We look forward to many more years of working with NXP and are thrilled to showcase the results at this year's CES exposition, where the value we're creating for customers will be prominently displayed. So that takes me to my last slide. And I'm going to tell you what I told you, which is the industrial market is changing. It's transforming. It requires innovation. That need for innovation is creating a tremendous opportunity for us since we're focused on our products and our technology kind of fit very well in the market needs.
In order to attack complexity, we're going after system solutions that create a tighter bond and a greater bond with our customers and a greater insight into where the market is going. And then we take these solutions to the mass market with a sales strategy that is tailored to address that fragmented set of customers. And with that, our plan is to grow to 8%-12% in the next three years. And with that, I will pass the baton now to Jens, my peer, who's going to have a conversation with you about automotive.
All right. First, a very good morning from my end. My name is Jens Hinrichsen, and I'm in charge of the NXP automotive business. In the next 40 minutes, I'd like to give you some insights into the automotive market dynamics and an update on the NXP automotive strategy, as well as an update on the success of our automotive accelerated growth drivers. Actually, building on what Kurt said, NXP focuses on bringing intelligent solutions to the edge.
The car is the ultimate edge device. Not only that, the car is an edge device in the larger context of all the connected things. Now, if you look at the car electronics outline to the very left, with all the compute, with the wiring, the central processing, and the end nodes, the car is a very complex system. At the end, the car is the ultimate system of systems. NXP is very well tailored to support this very complicated application.
So we have many years of automotive expertise, and around about 56% of our revenues are going into this vertical. So we really know what that means. And explicitly, like also Kurt said, the ultra-low power capabilities, our security know-how, and especially the functional safety and the real-time processing capabilities support this application very nicely.
So accordingly, automotive is an important growth driver for NXP. And we are leveraging very clearly the megatrends in the industry, such as the digitization of the vehicle, which we call the software-defined vehicle, the ADAS support, as well as the electrification. And in the next slides, I'd like to give you a bit of an update on the market, as well as on how NXP is strategically positioning this one and leveraging the full potential of the opportunity.
So if you look at the market, and if you look to the very left, you can see that in the strategic horizon, the light vehicle production is just very moderately growing. But you can also clearly see that context and complexity is accelerating the need for automotive semiconductors. And the numbers in the bars are always in million units, the light vehicles. And in light blue, I have outlined for you there also a little bit on what the contribution from China is around this.
And you can see that the digitized vehicle of the future, so the software-defined vehicles, are growing with a 48% CAGR on the strategic horizon. The ADAS-equipped vehicles with around about 8%, and the electrification continues with strong growth around 20% CAGR. This all drives the need for automotive semiconductors. That's a fantastic opportunity the market is providing.
I have a few more information also for you outlined here down below. And let's focus on the middle there. You can see that the lines of code and the data in the car in the next years are literally exploding. This requires a lot of semiconductors because you need software, and software always requires semiconductors to run on. I also have outlined here to the right an important element, which is the power consumption of an average ECU in the car. And you can see that the power consumption is significantly increasing. So managing the power in the car, managing the energy efficiency of the car, and managing all the power of all the electronics becomes a very important element going forward, especially if the AI moves into the edge, driving the power consumption up.
So this also requires very intelligent semiconductors and gives another fantastic opportunity for the semiconductor industry. And then to the very left, a very important element, and I will come to this a few times in my presentation, the time-to-market requirements are significantly reducing. And especially emerging car makers, they can spin already these days, and especially Chinese, they can spin a vehicle platform in 12-24 months. So the car makers need help in leveraging engineering efficiencies. They need help to reduce their development cycles. Also, that one requires semiconductors, and that's a fantastic opportunity for us. Looking forward, you can also really see that the industry is shifting, right? And we see a completely new way of thinking about the future mobility.
And the consumers, they look at the car more as a consumer device, and they want seamless user experience, and they are increasingly appreciating ADAS functionality for safety and convenience reasons. And they constantly want to have the car up to date, and they increasingly want to upgrade the car, actually even personalize the car over the lifetime to make it a real personal device like you are maybe used to in other applications like your phone, etc. And accordingly, the car makers, they have to change the way of making cars, right? So they have to digitize the car to enable software-defined and intelligent features, and they have to handle this one across all their brands, all their models, and their platforms. Literally, they have to completely reinvent the car as scalable, connected, electric.
Everything I have said so far is clearly driving more features, more functionality, more software, and that drives complexity. Our customers and the car makers, they need help in managing complexity, very similar to what Rafael has said also in the industrial part. If you look to the left, you can see how the future car is having way more software than the current cars. Just simply any additional feature, any additional function is adding an ECU and requires more software. Then also over-the-air updates and then the increasing involvement of ADAS and IVI is driving software. Finally, really data-driven use cases and service-oriented architectures that really the element of software in the car explode. This is driving complexity.
And the vehicle architectures of today, which you see in this fantastic car there, which is very hardware-defined, this is not able to handle this complexity. And this is what I call the legacy complexity. And I have here a quote from a car maker. It's actually a European car maker making a statement here that they say if they want to introduce one new feature into their existing architectures, they have to touch more than 50 software modules and more than 10 ECUs to be modified. And this is just simply if you want to add maybe a seat functionality in the car that requires an ECU that's this additional hardware and they run software on top of it. And then this ECU needs to be connected with all the rest, and the interoperability is really a nightmare. That's a true challenge.
So the existing vehicle architectures have many components from many vendors across multiple tier ones with the individual software, and that is driving complexity. And that complexity hampers clearly innovation, time to market, and total cost of ownership. So the car makers, they need something different. They need a more efficient, more software-defined vehicle architecture. And this is basically what we call the digitization of the vehicle.
This is where the software-defined vehicle comes in. And based on a very scalable and super powerful hardware platform, most of the functionalities and features and actually experience is defined in software. And what the car makers really like to do is they like to abstract the software from the hardware. So whenever they want to introduce a new feature, like I have explained on the slide before, they don't want to modify the hardware. They just put it into software.
And therefore, also, the car needs to be updatable and upgradable over lifetime. They want basically that there's a perceived value increase over lifetime because you start to personalize the function of your cars, and maybe you only activate features and functions you personally want at a certain point of time. And that is only possible with a digital vehicle. And also, and I mentioned this before, the car makers need help in reducing their development cycles.
So they want to have a reusable hardware platform, and they want to have a reusable software platform that helps them to achieve the engineering efficiencies they require to be competitive in the time of market. You have seen probably slides before. I know Kurt has presented those same slides a few times. There are different ways towards this future vehicle.
You can basically think that more emerging car makers have a very revolutionary approach, and the more, let's say, legacy car makers or established OEMs, they have a more evolutionary approach. The impression might be that maybe the, let's say, revolutionary approach seems to be the easier one, but actually, it is also not easy. These car makers, they have to get familiar with the automotive technologies. They have to get familiar with the automotive ecosystem.
Security functionality and quality performance is really partly new to them. Often, in their development of their vehicle architectures, they are doing this without tier ones, and they lack the system integration capabilities. Also, not an easy path. The legacy car makers, well, they can use all the experience and all the know-how of the past, right? They are familiar with the automotive ecosystem and with the technologies.
But more digital technologies, data, software, that's also pretty new to them. And then also, especially also non-typical automotive technologies are also not familiar to them. And the biggest challenge is they have to manage this across all their models, all their brands, and all their platforms. So not easy. The tough path is, but very clearly, any of the customers need help towards the software-defined vehicle.
Let me zoom in a little bit deeper, and let's have a look on this one here because the fundamental element to drive the digitization in the car is the vehicle architecture. And these are a lot of bubbles and boxes and lines, but let me try to explain this. On the left, you see the current domain-based architecture, which is typical, popular, which is ramping right now, and it will also still be around for a while.
However, then, if you move to more a digitized vehicle architecture, you find different ways towards it. And very clearly, and that's super important to understand, every car maker has its own interpretation. Every car maker has its own path towards that. And it is also not a static environment. It's evolving. It is really changing.
And sometimes they make step functions and start to the left and go more to the right. It's a very dynamic, and there are constant subsets of these architectures in reality. So which is basically also an environment we really like. But for the sake of simplification, NXP is basically clustering the future architectures basically in three groups. And you can find on the very left the distributed zonal. In the middle, we call it the vehicle compute and zonal. On the very right, the consolidated compute power.
Let's start with the very right because this is, from our perspective, the most digital vehicle architecture we can see and we can think of and car makers are driving towards. Here you see that there is a very strong vehicle compute engine in the middle, a high-performing compute engine for the vehicle compute next to the IVI and the ADAS.
And this vehicle compute engine is basically running all the core vehicle functions in that central compute so that all the remaining zones become very, I call it dumb. They are just simply I/O aggregators, or maybe they are basically the connection between the Ethernet network, which is the more the wider lines towards the end nodes, which is based on CAN and LIN. That is the most digitized architecture.
It has got a very, very streamlined wiring, and this is probably also from a total cost of ownership, the most optimized setup. The one in the middle is a bit of a hybrid. It also has got that vehicle compute in the middle, but maybe the app integration of the vehicle core function hasn't happened to the full extent like to the other extreme, and there's a lot of compute power still in the remaining domains or in the zones, and so also, the zones and the domains require high compute engines to make that all work. It's more a compromise in the middle. The one to the left is probably the most fastest and easiest to implement one. That's maybe the most pragmatic first-step approach you can see.
Here, you don't have a central vehicle compute other than a gateway, and all the compute power is still remaining in the zones. This is the least digitized from our perspective and is also probably having the least optimized wiring and network in this particular field. But it is the fastest to implement, and therefore, we see emerging car makers and one or two Chinese car makers jumping on the very left here to the distributed zonal because it's easy to implement and do a fast step. The one in the middle, most of the Chinese are on that one, and also most of the high-volume legacy makers are in that middle platform to what we see. The very extreme to the right, the most digitized one is basically approached by the more high-end car makers in the industry. But everything exists and even subsets are existing.
Later, I will show you how NXP is supporting any of these vehicle architectures. But before I show that, let's wrap up a little bit the dynamics in the market and the changing customer needs, and you see them here to the left. First and foremost, I'd like to highlight this is still a high-tech industry, and therefore, our customers, and regardless of where they are, they want innovations. They are differentiating via innovation, and so that is for us still the most important thing. The innovation leadership is essential. However, then more importantly, in the industry, you can see increasingly more that R&D efficiency is requested. They like the reusability and the scalability and the reuse of the hardware and software to basically achieve very fast time to market and optimizing their total cost of ownership. These are changing dynamics in the industry.
And the automotive strategy of NXP is tailored towards supporting this in a very nice way. And it is based on three pillars, and I will go through this in more detail. It is, on the one hand, the innovation, so we continuously driving innovation leadership. It is the system play, a bit similar to what Rafael has shown in the industrial part, and then also to the very right, similar like Rafael has said, it's collaboration. We need to be thought an innovation leader in this ecosystem of automotive and drive collaboration forward. In the next pages, I'd like to give you a little bit more detailed insights, and I'd like to start with the innovation part. And if you look at innovation, the strength NXP is having is explicitly the deep application expertise and insight.
We nicely combine this with our broad technology we have available, and especially the combination, and Kurt referred to this one, especially the combination of the strong processing platform with all the analog we are having and the security and connectivity makes us to achieve really true innovation leadership positions in the market. I think you are very familiar with the positions I have outlined here. I will not go through this one in detail,
but really strong proof points of our innovation capabilities. Important is also that we are not staying there, right? We don't remain there. We drive this forward. We were the first ones offering a 28-nanometer one-chip radar solutions. We are way ahead of our competition rolling out our ultra-wideband-based mobile phone car access solutions, really leading in the industry in this field.
Recently announced, we were the first ones announcing an automotive 5-nanometer processor for the vehicle core based on ASIL-D functionalities. We continuously drive innovation because this is the essence of the high-tech industry, but never forgetting our strong heritage of automotive quality, functional safety, and robustness. I have for you, and not a video, but I have for you a quote here from Hyundai, which is, from my point of view, very powerful. They are a very innovative car maker, and you can please read through this. They awarded us as one of the first semiconductor vendors in 2022 as the supplier of the year, explicitly appreciating our game-changer mentality and our innovation spirit. This is a real innovation partner for us, and we are very grateful for this powerful quote.
Let's move to the next part, and that requires a little bit more time because we need to dive into this one because this is really essential and important. It's a real game changer from my perspective. That system solution, that's a key element of our strategy. So let's resume. We have clearly seen the car makers, our customers, they do need help because they are facing a dilemma.
They have to drive innovation. They have to improve the time to market. They have to optimize their total cost of ownership in an environment of significantly increasing complexity. So they need help, and the way to overcome this dilemma is for us that we have strategically decided, as outlined by Kurt, that we step up beyond a normal traditional semiconductor component vendor.
We are offering system solutions, and we are leveraging explicitly our strengths, and not many in the industry do have that, that we have this very powerful processing platform, can connect this with all the analog and the power, and link it all together with our wired and wireless connectivity, so we basically engineer system solution, and what I mean with this is we are not just simply putting the components together and hope it sticks, no, right from the beginning, we really develop the products in a way that we are optimizing them on system level, so we really make them as complements, and the combination of the sum or the value of the system is higher than just simply the sum of the components in itself. It is creating a real system additional value.
And then on top of it, we basically support it with pre-engineered software to offer truly solutions for our customers. And that helps them to have a true engineering jump start. That helps them to really optimize their performance on system level and managing their interoperability on system level. And that helps them to manage total cost of ownership and not only maybe improved component cost. It's really a system view, and that's a real differentiator NXP is using. So let's dive into this one a little bit. So let me start first to tell you on what we are doing in this system play. So where do we focus on? What is it what we are doing? And NXP is focusing on the vehicle core. So what is the vehicle core?
The vehicle core are basically all the functions of the car that make the car operate and make the car move. So really, the things like the vehicle dynamics, the vehicle propulsion, the gateway, the connectivity, body, and comfort, these are the core vehicle functions that really make the car operate and move. And this is where we focus on, and I tell you basically why we do this. But in the system play for these vehicle core functions, we are offering the complete processing platform, and this is for the central, for the domain, for the zone, and for the end nodes. We have them basically all supported with our networking and power management solutions. On subsystem level, then in this vehicle core architectures, from a safety perspective, we continuously driving our radar solutions forward. Radar fusion and safety systems are really in our focus.
In the electrification field, the battery management and the inverters we presented this a few times in the analyst days are very successful, and we drive them forward. From a user experience point of view, car access, radio functions, and also our e-cockpit solutions are on our radar screen and focus for our system solutions.
The NXP technology fits very well to support the vehicle core because basically the lifetime requirements, functional safety, and robustness requirements of these true automotive SOCs for the vehicle core differ a lot from the more consumer-oriented SOCs for the IVI and for the ADAS. The most important thing, and this is also why we really believe, and you have seen this in the architectures, that the vehicle core remains to be a dedicated standalone domain in itself, is the vehicle core has to be security hard protected.
It cannot be that you can hack into the vehicle core functions. If you hack into the vehicle core function, the damage in the car can be severe. So it needs to be security hard protected. That's really essential, and we know very well on how this is to be done. But then also think about the vehicle core functions. This is an environment of zero latencies.
So that means if a radar system is maybe identifying any object on the road, the car has to brake immediately and any millisecond counts. So this is not an environment of latency. So that basically means our hardware-enforced isolation capabilities and then enabling the real-time processing is really allowing us to run all the vehicle core functions in parallel without jeopardizing at any point of time the performance of this core vehicle. So there is no latency allowed in this vehicle core.
As there is no latency allowed, there's also no lack of functional safety allowed. This is a very super safety-critical environment, and you cannot basically back down your processing because it is running on certain applications for our data on the application processor. It needs to be instantly ensuring functional safety. It's like in an airplane. You cannot jeopardize functional safety at any point of time. Again, also here, this is only possible because of our real-time processing capabilities.
Of course, no need to mention anymore, clearly, this is true automotive quality and robustness requirements. I told you we focus on the vehicle core, but now how do we do this? How do we do this system play now? This is where the CoreRIDE comes into the game, and this is what we have recently announced. You might have seen it.
The CoreRIDE platform is basically a platform for our customers to basically develop and produce vehicle architectures. Here we do exactly what I've said in the intro slide of the system play. We are combining here a hardware platform of our compute engine, the networking, and the system power management, and then we complement it with software, which is basically coming from our CoreRIDE partners, which I have outlined here below.
To the right, the CoreRIDE partners are building with us a joint development platform. We take our hardware. You see that they're in green. We have our drivers on top of it, but then the OS, the hypervisors and the middleware comes from our CoreRIDE partners, and we do the system integration all together for the car makers.
That basically allows them, the car makers, to focus on the application software because that is, from a consumer point of view, what really differentiates the car performance is the application, and then the car makers can really focus on what really matters while we wish the CoreRIDE partner do all the middleware integration and providing them a very powerful hardware and software platform.
So what we literally do is we are offering system-verified production-ready reference designs on hardware with pre-integrated software, which gives now a really, truly our car makers an easy path to manage complexity and an accelerated path to drive the digitization in the vehicle architecture. The CoreRIDE is a key element of driving this forward, and I'd like to come back now to the vehicle architectures I have introduced already earlier on.
Here you can see now how we are able to support all of the vehicle architectures regardless of what kind of construction you can look at. Here to the left again, the CoreRIDE platform a little bit more in detail. Again, the building blocks, the hardware and the software you can find here to the left. It is all tailored around a bit similar to what Rafael also explained in the industrial field.
It is tailored around our very powerful and scalable processing platform, and it starts with the very high-performance S32N leveraging the 5-nanometer technology all the way down to a simple legacy node 90-nanometer technology microcontroller to support any of the processors you can use in the platform on the right. Then you can see it comes with the network.
So this is Ethernet, but also CAN/LIN, and we are clearly leader in this field. And every processor has got a functional safety SPC and a PMIC to be powered, complemented by the software of our CoreRIDE partners. And now hopefully the color code comes out. You can see that a little bit. The more darker blue ones are nicely tailored to support the vehicle compute and the central functionalities.
They are really driving the digitization of the vehicle architecture. The lighter blue ones are nicely tailored to support the zones, and the gray ones are the smart actuators and the end nodes. We have it all. It's like Lego blocks. You play, you get it. And regarding what kind of topology the car makers are using, even subset out of it, we can support very nicely, very efficient. That's the path to digitize the vehicle architecture.
I’d like to keep your focus once more on the darker blue ones, especially on the S32N platform, because I come to this one. That is really, truly then enabling the digitization of the vehicle architecture. Because remember, the car makers want to abstract the software from the hardware, define the functionality in software, and run them virtually all centrally on a compute engine. So you need a really powerful compute engine to do this digitization, and that’s where our vehicle processors come into the game, and that’s our real differentiator in this system play. I come to this in a minute. But let’s have a look on a quote from Leapmotor.
So this is a leading car maker in China, and this is a nice EV with a range extender, and they are using our S32 platform, and they leverage the platform exactly in the way I have said, and you can read it there. They are basically using it to centralize the vehicle core function and reduce the wiring. So they are really totally optimizing the car and digitizing the car based on the platform exactly in the way I have outlined to you. So very nice, and also a big thanks to Leapmotor for providing us this quote. Before I wrap up now the system solution part, I just want to give you a very brief overview that there is more than just the vehicle architecture. This is the element here to the very left, and you have seen on how we do this.
We continuously drive also the energy network on a solution level because you have seen on how the power consumption and the ECUs are increasing. Every watt counts, energy efficiency is very important. Electrification, you are very familiar with this one. The BMS and the inverter solutions are really key in the infotainment field. We continuously driving our system solutions and also on the radar side. I'd like to take now the time a bit similar to Rafael and dive a little bit into some system solution examples how we really, truly bring this to reality. One is this vehicle compute engine I have set. The other one is the battery management, and the other one is the radar. These are also very much the accelerated growth drivers of the automotive strategy, and therefore they are very interesting.
And here again, I have clearly said that the car makers are starting to digitize the vehicle and that there will be a central vehicle compute engine, which is there in the middle of the future architecture. And based on today's domain-based architecture, you see on how now the vehicle core functions are all resulting into the vehicle compute engine. And the element, the solution on driving this one is the S32N platform. It's actually not only one device.
It is a roadmap, and there are more devices to come. We have only one device so far announced. That's the S32N55. That's the first 5-nanometer automotive processor with functional safety SOD capabilities, but there is more to come. And this S32N platform is basically a platform to enable scalable cross-domain super integration. That's that thing that is digitizing at the end the vehicle core functions and therefore the vehicle architectures.
Especially with its inbuilt hardware isolation and virtualization technology, the car makers are then able to integrate dozens of ECU into this engine without jeopardizing any of the security and functional safety performance of these vehicle core functions. You can imagine that thing comes, I believe, with a 16 split-lock Arm Cortex-R52 cores. This is such a powerful real-time engine, so you can really abstract in this case the function from the hardware and put it in software and virtually run this in the central computer without jeopardizing any of the performance or functional safety at any point of time. There is more to come, and you see that the S32N55 on the breadth of all the core vehicle functions we are focusing on is largely focusing on the real-time essential functionalities first.
There are more devices to come which will do next to the real-time processing on top of it, also application processing, even more powerful engines so that you also can then collect the data centrally and run the data centrally and then also enable ADAS Fusion safety features as well as OEM applications as well as services. That is also a real enabler for AI functionality and use cases even in the vehicle core. This is what we are planning to do. This is really driving the digitization. Radar, I think you are very familiar with this one. We are the leading semiconductor vendor in the field of radar, and we are leveraging here our scalable radar solutions all around the radar sensors, the radar fusion, as well as the radar processing, and we can offer any solution the customer wants for their car.
I have outlined the car here down below, and you can see that they're one chip, very highly integrated, very efficient edge radar or corner radar solutions are nicely tailored for the edge radar solutions. Typically, four of those go into the car, and then the imaging radar. Typically one system goes into the car, sits typically in the front and provides high-resolution imaging radar capabilities.
But also I'd like to highlight our distributed radar capabilities, which is basically now having multiple smaller radar sensors typically in the front, and they are pretty small, and the radar sensors talk to each other very nicely, and they create a gigantic virtual antenna, and they can improve the resolution even by another 50% in creating really high-resolution also imaging radar quality in the car.
This is also a very nice set on leveraging this, and then the radar information will be combined in a radar bridge and/or in a radar fusion engine. Very powerful. You see the customers can do plug and play, and for every radar node you see here on top, everything comes from NXP, the processor, the connectivity, the power management very nicely as a system solution.
Last but not least, our battery management solution is very successful. I have presented this a couple of times in the analyst days. There are also two here in blue. Once more, everything that comes from NXP are very nicely the components coming together, and very clearly, battery management is more than just the battery cell controllers. It's more than just the analog front-end devices, which are part of the cell supervision circuits.
There are way more components around the processor, the connectivity, and other devices which are really critical to make the applications, and typically, these are multiple boxes in a battery management system, so there is, on the one hand, multiple supercell cell supervision circuits. There is the battery junction box, but also the battery monitoring unit, and for everything, we are providing also here system-verified production-ready reference designs like you see here with the board, and then the car makers can really build the full electronics for the battery, which is shown there to the left, to the right, and this is clearly very appreciated by emerging customers, extremely successful in China because of the capabilities to adopt very, very fast our solution, which are nicely pre-engineered.
All this hopefully gives you an idea that DRs are really true system plays, and we bring this nicely to reality, to life in our automotive play. Before I come to an end, let me conclude our strategic part with the last remaining pillar. That is the collaboration. And you see that the industry is in transformation and that the car makers really, truly need help.
They need help in this dilemma to drive innovation while they have to also reduce the time to market and to reduce the total cost of ownership, and this in an environment of increasing complexity. So the way of working has to change a bit, and it is changing. And the very traditional, more sequential supply chain you can see here to the left, or engagement model you can see here to the left, is evolving over to a more modern ecosystem collaboration.
And NXP is right in the middle of this one, being a real true thought and innovation leader in developing this ecosystem. We are very, very strongly engaging with the car makers driving innovation forward, but never to be underestimated and forgotten. The tier ones are still very important. They also drive innovation, but on a different level, and they are needed to do the system integration and do the model making. But also ODMs, IDHs, EMSs come into play in parallel to the tier ones.
So we are engaging with all of them, fostering this very nice ecosystem. We need to bring the partners together to do joint investments, bring the partners together to drive standardization. That's what we do. And you see here a couple of examples. Internally, for instance, here on NXP, I show you how we are providing.
This is based in China, where we are providing our customer an enablement center for our battery management solutions. If they don't know how to bring this solution into the car, we can really help them. That's what we do. The CoreRIDE partners are there. The SDVerse, which is basically a marketplace for software in the digitization of the software-defined vehicle, and then you see a couple of OEM and tier one engagements. I'd like to highlight a little bit Foxconn because this is really essential.
You might have seen a couple of announcements here. We have a joint lab with Foxconn to accelerate the development of the software-defined vehicle, and I also have a little video for you where the chairman and the CEO of Foxconn reflecting a little bit on what I have said. Please. Here at Foxconn, we see big in our growing electromobility.
Automotive portfolio leverages our excellence in ICT and our strength in time to market and time to cost. Our long-standing partnership with NXP excites us because NXP also sees big. Collaboration is critical to success during transformative times. NXP is a key partner for us. They provide the technology and system expertise to enable Foxconn's EV developments. NXP is a leader and the ideal partner in uniting the automotive ecosystem.
By leveraging NXP's broad portfolio and working with them on their CoreRIDE platform with pre-integrated software and hardware, we are able to accelerate software-defined vehicle development significantly. Together, we revolutionize towards cleaner, greener, and smarter. All right. We are also very thankful for this video. And with this, I come to the last slide, and I'd like to wrap up the automotive session.
You can see also here the growth model going forward based on the key assumptions to the left. You see that we have nicely been able to grow from $5.5 billion-$7.1 billion in the last strategic horizon and look forward to grow close to the $10 billion going forward with a growth of 8-12, significantly supported by our accelerated growth drivers, which I have presented to you. If there are three things, like Kurt also summarized it, I have for you to take away, it is very clearly the industry is transforming, driven by the digitization of the vehicle, by the electrification of the vehicle, and by ADAS. The second thing is our customers need help. They need help to drive innovation, faster time to market, and improve total cost of ownership in an environment of complexity.
And there, the NXP strategic play, especially the system solution play based on the CoreRIDE platform, is nicely tailored to enable our customers to achieve that. And the last and very important item is also we clearly have a leading position in the processing architecture of the digital vehicle of the future. With this, I'd like to conclude and give it over to Jeff, I think, for Q&A.
Thank you. Great. So I'd like to call Kurt and Rafael up to the stage, and we'll hold a Q&A for the next 20 minutes, 30 minutes or so. Thank you. What I'd like to ask of you when the microphone comes to you first, say your name, the firm you're with, and what we're going to do is fast round robin so everybody can get an opportunity to just one question, please, to start with. All right. Mr. Seymore. Ross.
I know who I am. I just need to like it. He at least listens. Nicely done. Sorry about that, Jeff. It's okay. I am Ross Seymore from Deutsche Bank. So, a question on the auto market in general for either Kurt or Jens. The adoption of the software-defined vehicle, all the technology seems quite obvious. China's leading the way, but I guess a two-part question.
First, in China, how's NXP positioned relative to the desire of the country to have more in-sourced local production? And second, in the Western companies, how are you seeing them respond to it? Can they keep pace economically, or is the pressure from the Chinese vendors allowing them to slow down if they put tariffs into place or accelerate to fight against the Chinese? How are you seeing the kind of Eastern versus Western adoption curves?
Yeah, Ross, let me take a stab at this because it connects actually to something we will have after the break. So the local competitiveness of NXP, I hope you heard from what Jens said, is actually very good. I mean, we are leading with battery management systems. We are leading with processing solutions.
And in spite of everything you might hear left and right, we are not faced in those sectors with any serious local competition yet. I say yet because obviously we stay paranoid about it, but we don't have that as of today. The one requirement which comes up very hard in China to stay competitive is local manufacturing capability. Chinese OEMs, Chinese customers want us to give evidence that we have a plan that we can produce locally, and that's mainly about front-end production, what they want to purchase from us.
And Andy is going to speak after the break what our strategy is to follow that requirement. What I can take ahead already is that I think, given our hybrid manufacturing strategy, given more outsourcing, leaning more on foundries, we have a significant differentiator here versus several of our peers which are more internal manufacturing oriented, and they cannot follow that requirement the way we can. So that is a clear advantage.
Now, from a global perspective, Ross, I think the numbers speak for themselves. Western car companies are losing out currently broadly against Chinese because all of the transformation which Jens was speaking about, at the moment at least, is mastered better and faster by the Chinese than by the Western companies. I will not comment if tariff policies, etc., will be a lasting weapon against it or not. That's a political question.
What we see on the ground is that the Chinese companies are actually ahead, and I would say Asia at large. Jens, you spoke about Hyundai, which we see as a mass volume company also being pretty successful from a global perspective. Now, our position in China is at least as strong as it is in Europe and the U.S. So at the moment, we are actually benefiting from the fact that China is running ahead.
I say that, however, with the necessary paranoia because clearly China is a very turbulent place. So we have to continue to work hard to stay competitive in China. Today we are, but I don't want to say this is a given that this stays always like this. We are working hard on it. There are two angles. The one angle is what Andy is going to explain, which is local manufacturing.
And the other one is actually very straightforward. It's simply competitiveness on the product and system level. Everything you said, Jens, is an offering to Chinese customers where we are a step ahead to local competitors in China. So it's not rocket science, but this is really where it comes down to. So we are optimistic about keeping that advantage which we have, but clearly we have to work hard.
And I will also reveal here, we have started to dedicate and deploy R&D funds more to China, more to Chinese customers' requirements than we've done in the past. In the past, Western OEMs were our lead customers, and we would then sell the resulting products to Chinese customers. Given the advancements of the Chinese OEMs in SDV and electrification, we start to use them and leverage them as lead customers and eventually sell them to the West.
So that is actually falling upside down. And it is culturally, and that's what we are working on hard across the company. Industrial is doing exactly the same thing, by the way. It is, of course, for us a change, and we are very aware of this. We have to give more preference to the specific requirements from China than we used to do in the past, and we do so. Great.
Tammy, why don't you come back to Chris Caso right there?
Thank you. Chris Caso from Wolfe Research. I guess a question on some of the growth projections that you provided and specifically kind of what are the starting points for that. And going back to the last analyst day when the growth projection provided different market conditions, things are pretty strong. So coming off a strong base, harder to grow off a strong base. Now, you're still in a situation where your customers are burning their inventory now, so interested in how you accounted for that going forward?
Yes, Chris, and that accounts for the whole business, so I will take that question. We have indeed moderated our growth ambition. We had 8-12 last time. We come down to 6-10. This is largely indeed attributed to the current weakness, because it is obvious that calendar year 2025 is likely a touch below model, so if over the three-year horizon we want to achieve the ambition, we just have to account for that first year of the three years, which is why we had to bring down the overall ambition, so that current weakness, which we talked in earnings two days ago, obviously reaches at least to some extent into next year. As we said, we don't know how long, but it does reach into it. In order to be reasonable about this starting point, we have moderated the total from the 8-12 to 6-10.
Right. Just as you were looking at this over the years, it's a derivative question, so not a new question. You've talked about looking back a couple of years and saying at the current level you're at, it reflected no industry growth. That was why you were confident that you were still undershipping demand. Is that the case, and is that baked into your assumptions as well?
Yes. That goes to this whole soft landing strategy. As of today, and I've said that holds for the past six quarters, we are undershipping end demand. Absolutely.
When that returns to shipping to end demand is a function of when is the macro recovering and when is the inventory burned away, which is still sitting at direct customers. I say it again, I've said it a hundred times. We are very disciplined in the channel. So we are not having that burn inventory issue in the channel where we are at eight weeks at the moment. We have that challenge with direct customers to a certain extent. And given the negative mood in the macro currently, they even further reduce their inventory targets, which means we bring it even further down. Now, you know the trick, Chris. Following that, it's going to come back even harder. I mean, we've seen this a number of times before. But yes, we are undershipping.
I think we started undershipping earlier than our peers, which is why we have in 2022 undergrown our peers and which is why we are now overgrowing our peers. I mean, this year with the guidance we gave on Tuesday, we're going to be down like 5% on a year-on-year basis. All of our typical peers, which you have in your benchmarks, are much more declining this year.
And that is a function of us tackling that problem earlier, never shipping too much. And for me, that also means we come out better than the others. That's the consequence. I mean, that's at least our thesis.
Hey, Tammy, how about here, Mark down in front on the right side?
Great. Mark Klopotzis, Evercore ISI. Thank you very much for the very informative presentations. Really appreciate it. Kurt, a question for you, I guess.
Obviously, the system solution is an important message that you all are delivering today. I'm wondering how far up the value chain do you think NXP can get? Can you get to the point where ultimately you're like the de facto standard, like a Wintel where you get your automatically 80% designed in? I imagine the software element is an important part of that. Can you share with us what your software spend has been five years ago today? Where do you think it
has to get? Yeah, the software spent, and Rafael and Jens might detail this, has surpassed the 50% of our R&D bill in our processor businesses. It was less than 50%. It is now easily above 50%, and it keeps going in that direction. We are also often asked about tuck-in acquisitions and where is it that we see gaps.
Clearly, the software side is an area where we are strengthening the muscle. So that ratio will even tilt more into software. That doesn't mean we neglect hardware. Hardware is an important platform element for the scalability which my two guys talked about, but the software becomes very prominent. Secondly, we don't want to be a tier-one supplier. I want to be very clear. It is not our ambition, absolutely not our ambition, to become a direct supplier to OEM customers, especially not in the automotive world. There are a lot of reasons why we don't think that's our natural place. We also think they continue to have a certain place in there. But for the digitization of the car, for the digitization of these industrial systems, we are becoming a very, very, very important innovation partner for the OEMs because we oversee the whole thing.
These tier-one customers, they only have one system or one application. The whole processing platform, which is across different applications, which Jens talked about, none of the tier-ones has that breadth. We have it, which is why we are a very much wanted collaboration partner. I will not repeat the name you had for who we could become, but yes, it does become more of a winner-takes-it-all game where we used to win in the past many, many, many, many deals and design wins with tier-ones.
We make now much larger deals on OEM level, which are then deployed back into the tier-ones. So the whole mechanism of design wins is actually changing. And eventually, I think it moves much more into a winner-takes-it-all mechanism. So the answer is yes.
Thank you.
C.J. Muse, right here. C.J. Muse with Cantor Fitzgerald.
Kurt, a question for you, a broader question on the competitive landscape at the edge for both industrial and automotive. You have digital players attacking. You've got microcontroller competitors and others attacking from the other direction. Curious what you think gives you the biggest advantage. Is it the system? Is it software? Is it the breadth of your entire offering? And is it kind of the first-mover advantage that drives the stickiness? Would love to hear kind of your rank order of what's driving that competitive position for you all.
Now, let me first speak about industrial, and Rafael is going to jump in there because you made the argument, Rafael. I think when you guys think about industrial, you think about these large analog companies. But the core of the industrial market moves to intelligent systems at the end, which is about processes.
That doesn't mean there are no analog products, and we are also happy to sell our analog products, but the core is scalable platforms and software. The breadth which we have in processing for industrial, and Rafael, you went through this, from relatively simple low-cost microcontrollers all the way to very, very complex application processors, we have no competitors which has that, CJ.
I mean, just to be very clear, there might be high-end competitors trying to come into industrial with the very high end, but then they don't have the scalable platform. They just have that high-end price point. We have a number of microcontroller competitors in the industrial field, but they don't have application processors. I will not give you the names, but you know it. I mean, they cannot scale up to what Rafael was explaining.
When you heard the CEO of Johnson Controls speak, everything was about platform and scalability because they cannot afford to write software for each of their applications separately. They want one reusable block of software which needs them to scale across performance levels. That breadth is where the big advantage of NXP is. Rafael, I want you to give a bit more color just to say upfront. It's exactly the same in automotive.
It's the same thing, but here it is compounded by real-time performance and functional safety performance. This 5-nanometer device which you spoke about, Jens, there is nobody else in the industry which has a fully ASIL-D-compliant 5-nanometer processor in automotive. Just not available. It is only NXP, Rafael.
Yeah, and the way to think about industrial is a different paradigm, right? The paradigm has completely changed. And people are trying to, you can see your competitors trying to use the playbook in the past, the playbook of the PC or the playbook of the phone, which is I develop a turnkey, I complete the BOM, and I complete a turnkey software, and I own the market. And then you come into industrial and you realize there are thousands of applications, hundreds, not thousands. I'm sorry. I'm sorry. It's not hundreds, but thousands of applications and thousands of customers.
And then you start kind of saying, "Well, how do I get around this massive fragmentation of requirements and customers and trends?" And so our strength here, right, and this is where one is our focus. And when I talked about system solutions, one thing that maybe didn't come across very well is about this transversal functionality that you can apply and dominate.
You create platforms that deliver certain functions for AI, and then you apply it to the different needs by having a very scalable and focused platform. And so we're modifying our playbook as these edge system solutions start taking hold. But very clearly, the playbooks of the past don't apply to the industrial segment. And again, I mean, this is part of our opportunity.
Great. Vivek, right down here.
We have a. Thank you. Thank you, Jeff. Thanks. Vivek Arya from Bank of America Securities. So Kurt, if I look at the last three years, your automotive business, so I'm kind of trying to separate the cycle out of this, last three years, your automotive business grew roughly 9% CAGR. Units grew 5%. Pricing, I imagine, grew a few percent also. So it doesn't really seem content played a part in those three years.
If I look at the next three years, units are not growing. Pricing is declining. So is it all content that's going to drive this growth? What am I missing in this looking at the past three years versus the next three years? Thank you.
Yes. Yeah, maybe I build on this one until you can chime in. First of all, I prefer in automotive to look at a longer horizon, right? Three years is a little bit difficult. If I go either back, maybe 2018 to 2023, or maybe going forward looking 2024 to 2030, I can clearly see that we are strategically growing with a double-digit growth in automotive. At the same time, if you look at the SAAR, maybe especially if you go back from 2018 to 2023, the SAAR is actually even declining.
So there is a significant growth above the SAR. And this is largely driven by content, right? And that is basically 50% that there are more semiconductors in the car roundabout, and it is 50% that there is a higher value, a higher ASP in the car. And the ASP is largely driven the mix because the value of the semiconductors is significantly greater. If I look forward, it's nearly the same math. Maybe even if I have a little bit of low single-digit growth in the SAR, there is still close to double-digit growth above the SAR going forward. And that is typically 50/50 more content and higher value. And Vivek, when Jens says higher value, that is not the dump price increase. It is that per unit we charge higher price because these units are getting bigger and more powerful. Jens showed the integrated radar solutions.
We used to ship discrete radars in the past. Now it is one chip which does the whole edge radar. Jens talked about the S32N, and you were very humble, Jens, but you didn't hint to the slide, but I know you guys all took a little note. That is a processor which sells between $50 and $130 per piece while we are selling microcontrollers in low single-digit dollar numbers. So that value increase which Jens is referring to is not the price increase, which also happened over the past years, but it is actually the value because we have higher integrated and higher performant products. It's a very important mechanism which, if you allow me, you forget in your algorithm.
Great. Gary, I see you in the back.
Thank you. Thank you, Jeff. Kurt, I wanted to ask something a little bit less obvious. I wanted to ask about the 25% of the revenue that's not necessarily highlighted at this event. I noticed that in your long-term projections looking forward, there's a big tick down in growth expectations on the mobile side, 900 basis points at the midpoint, and as well on the comm side. So in mobile, is that market-related tick down? Is it more of a strategic focus? And then as well, maybe the offsetting factors to consider in comm as well.
Yeah. Thanks, Gary. I completely acknowledge your point. Let me be specific. We are guiding our mobile business 0%-4% over the next three years, which at the midpoint is about market growth. We think the mobile market might grow at probably 2% CAGR.
Given our position, and you know it, which is this enormous leadership in the secure mobile wallet, which is the main part of our mobile business, we think it's just fair to assume we grow with market. So that's the simple secret behind the source. Comms Infra is more complicated, and I'm glad you asked because it's always a bit of a, "Yeah, what is it?" So Comms Infra is guided at 0%, so as being flat over the next three years. Underneath, there are three pieces in the Comms Infra revenue segment. One is our former, I call it former Freescale legacy digital networking business, which is a business which we actually haven't invested in since we acquired Freescale. It is starting to go end of life. So over the next three years, just assume and model this business to come down. That's a negative contributor.
Then we have the radio power transistors. You've all seen it in the past years. It's always a lumpy business. I say very clearly, it's not my favorite business because it's very hard to forecast. So we don't want to give too much merit to that business. It could turn out better, but I think it's reasonable to stay conservative. The third element in this segment is our secure card and RFID business, which is growing. So that is the element under the cover of the total, which actually contributes pretty nicely and positively to growth. If you put the three together, it is reasonable to assume a flat line. But again, those three are completely different. They also serve different dynamics in different segments. The one which is, say, strategic for us relative to investment and growth expectations is RFID and secure cards.
Great. Will, you're right here. Thanks. Will Stein from Truist Securities. Kurt, one of the sort of long-standing, very positive views of this sort of diversified analog and microcontroller category is that the design wins, especially in automotive, were always perceived to be very long duration. Share shifts were moved at a glacial pace. And I think one of the points you're making today is that innovation is accelerating, and as the Chinese become more effective, they're increasing that pace, which means there's some opportunities, but some risks also for everyone, right? How does that influence your thoughts about strategy and also value and your investment in R&D? Because it's less of an annuity. Well,
Will, no risk, no fun. Let me confirm what you say. Jens clearly said that these new players, and mostly they come from China currently, cut down the development time for a platform, I think you said, to one to two years, where they used to be four to five years in the Western world. We consider this an opportunity because that means that our newer products, and that goes back to the value argument I just had with Vivek, our newer products, which on average have a higher value per unit, can go to market faster. And that in itself offers a growth opportunity. Now, of course, we don't ignore the risk that, of course, faster somebody else could come from the side and take the business away. Yeah, but I mean, that's what we do. This is our job to go compete.
We actually like the fact that the dynamic is going to a faster pace because it offers us a faster path to growth. Great. With that, we're going to hold it here. We're going to take about a 10-minute break. Don't worry. We're going to have another Q&A after Bill finishes up. So we'll get to everyone, all right? We'll see you in about 10 minutes.
Thank you very much. Thank you. Thank you.
Our program will be back on in about two minutes. Our program will begin in one minute. Please take your seats. We will be beginning in a moment. Hello. Could we take our seats, please? We'd like to get started again. Good job. Great. Thank you very much. We'll have plenty of time to socialize later and ask more questions. Now, I'd like to introduce Andy Micallef, who is our Head of Operations. Andy?
Good morning, everyone. Welcome to the manufacturing section of our presentation today. I'm Andy Micallef, and I lead NXP Operations and Manufacturing. I'm excited to update you on our hybrid manufacturing execution strategy, which enables NXP growth, resiliency, flexibility, and cost. In order to deliver the 6%-10% growth that NXP revenue looks at going forward, operations needs to deliver 50% more wafers than we do today, servicing our 25,000 end customers with over 10,000 products in our portfolio today. We expect these to grow going forward. At the same time, the technology shift is mixing. This is a good thing for NXP. More and more customers are adopting our new products and platforms. And to address the Q&A session last time, we see product life cycles shortening. Jens mentioned this. And so it's accelerating.
Historically, a lot of our products were produced internally at 130-nanometer and older technologies, and moving forward into 2030, that's going to accelerate to more advanced technologies, and again, I think it's a good thing. It's something we have to manage. It's something that's difficult in transition, so we'll use our hybrid manufacturing strategy with really four principles in mind. We developed the strategy over the last few years when the supply crisis, we were fighting all of that, and so control of supply and geographic resilience are at the forefront of our strategy. We need access to competitive cost to expand our gross margins, and we'll do that through the 200-millimeter fab consolidation, as well as migrating existing products to 300-millimeter platforms, and we'll do all this with a low environmental impact.
Our hybrid manufacturing strategy has several great points for the company, has several benefits, primarily controlled access to supply, minimizing structural cost, which expands gross margin. What we've seen, which is driven by the technology shift of older technologies to newer ones, that acceleration also is accelerating less need for internal capacity utilization within NXP. When I joined NXP in 2021, it was about 50% internal and 50% external. Going forward, we can see it migrating to 2030 at 20% internal and 80% external. NXP has a very long history of technology deployment and manufacturing. We have four internal fabs today, three in the United States, one in Europe. We also operate a joint venture with TSMC called SSMC in Singapore, where NXP is the majority shareholder.
Now, we also have a global footprint of partners around the world, foundries, in order to cover all of our other needs outside of these internal manufacturing sites. So controlling the supply going forward is going to have to be actively managed through this migration forward. Okay? At the same time, we want to manage the control. So we'll be investing in joint ventures in Europe and Singapore. And through this transition, we actually improve our control by four to five percentage points. Today, we have 40% control. At the end of this transition, by 2030, we'll have 44%-45% control of our supply. So again, huge manufacturing footprint, opportunity to manage and expand gross margins through the consolidation of our 200-millimeter factories into 300-millimeter factories, primarily through the joint venture in Singapore. Assembly and test is an area of strength for NXP.
We have a huge advantage for our customers, as well as a cost advantage through our internal manufacturing and assembly and test. I didn't believe this when I joined, so I tested it, and I know that our cost is at par with any OSAT you can consider taking out their gross margin. We do not have to be 100% utilized or 80% utilized in order to have competitive cost, and we've got the scale and the team, and it was already in place, so I inherited this excellent area of supply, and we're also geographically dispersed. We have a factory in Kuala Lumpur, Malaysia. We have a factory in Bangkok, Thailand, a factory in Kaohsiung, Taiwan, and a factory in Tianjin, China, so this gives us internal geographic diversity going forward.
Now, during the supply crisis, we exercised a lot of dual sourcing between these sites and both inside and outside. So we improved our flexibility during the supply crisis because we had bottlenecks we didn't expect. With this, we expect to maintain our 80% inside and 20% outside throughout the horizon because this gives us great supply control and low-cost delivery to customers that, even though it's Asia-centric, much like our competition, much like the industry, this gives us good geographic resilience between the different sites that we own, as well as we use other OSATs to complement that for package types and test platforms that we don't have in-house. Our smart hybrid manufacturing plan is in motion. I introduced the word smart because we will introduce a lot of automation in these factories. Certainly, 300 mm factories are fully automated.
And any assembly and test sites that we use, we use increased automation and all the AI tools and the yield enhancement tools that are available. They will be implemented and also rolled out to the rest of our internal factories. Huge umbrella of geographic resilience. Over the last two years, we've made multiple announcements that improve our ability to grow at competitive cost with flexibility in mind, a lot of dual sourcing, and the ability to expand our gross margins going forward. So let's quickly walk through these, and then I'll go through in detail. So the European Semiconductor Manufacturing Company is in Dresden, Germany. Vision Power Semiconductor Manufacturing Company is in Singapore. We're announcing today a second factory in Malaysia, which we call ATKL2. So that's another internal factory to support our growth of 6%-10% and maintain our 80% internal. We need another factory.
We think that Malaysia is a great site to do that. Then we'll talk about our China supply chain. Already a question earlier today about that. We've got a great plan that we're sharing with our customers. We've already started implementing it, and I'm happy to share that too. Again, minimizing fixed cost, migrating to more of a variable model with control in mind, geographic resilience, and flexibility. Now let's go through each one of these. ESMC, we'll start with TSMC first. TSMC has been a 30-plus-year partner of NXP, going back to our 1999 joint venture SSMC in Singapore, where NXP is actually the majority shareholder. TSMC is the world's most advanced pure-play foundry and the most capacity in place of anyone else.
We are participating in their globalization strategy through ESMC, along with Bosch and Infineon, each of the three of us having a 10% equity stake. This is their globalization strategy. It's an alternative source to Taiwan. TSMC sees 28 and 16 nanometer as a huge growth node for the industry. They've announced that they're going to add new factories in Kaohsiung, Taiwan, in Japan, and in Europe at 28 and 16 nanometer. Now, today, NXP's sweet spot, our largest volume node, is 40 nanometer, and it makes logical sense that as we move forward along Moore's Law, the next logical node for us to grow into as our highest volume node would be 28 and 16. Time will tell if that's true, but we believe that this is a great project for us to participate in, to give us controlled supply.
This will have 40,000 wafers volume capacity by the end of 2027. And this is a $500 million investment by NXP and supports $2 billion of incremental NXP revenue going forward. So to summarize, this is controlled capacity with our 10% equity stake, gives us geographic resilience for customers who want European supply. Okay? So European customers are very happy about this. American customers are very happy about this. And so we're really excited to get moving on this project, which is already under construction. VSMC is a joint venture with Vanguard. Vanguard International Semiconductor is a publicly traded Taiwan company. We have a long history relationship with them. They're a specialty analog mixed signal, high-power boutique firm. They're owned by TSMC. A lot of the executives at Vanguard are ex-TSMC executives. So we're really excited about this project. We'll have a groundbreaking in early December.
We announced it earlier this year. NXP is a 40% shareholder here. The most exciting thing about this is it gives us that migration to these legacy technologies for which there is huge demand. We expect this to be very successful from the day we open and have our first outs in the second half of 2027. Many of our competitors, even some of our customers, we're already working with Vanguard in order to secure their orders because TSMC has said they're not going to invest in any technologies older than 28-nanometer, and this is an area which is geographically resilient in Singapore where we expect it to be full from day one, and we're really excited to participate in this project, so there's going to be three initial nodes: 40-nanometer, 90-nanometer, and 130-nanometer. This will be built in the same blueprint as TSMC Fab 15.
It'll have a phase one and a phase two. It'll eventually be a gigafab. Phase one will have peak wafers at 55,000 wafers starting in 2029, but first outs will be in the second half of 2027. So NXP contributed $2.8 billion into this joint venture for phase one. And this enables $4 billion of incremental revenue to NXP. Now, this will be the primary landing spot for our 200-millimeter consolidation. So we'll gain the efficiency of moving from 200 to 300. And also, as Kurt mentioned earlier, this is access to cost. Okay? ESMC will be market wafers. These will be cost plus administrative fee. So to summarize this one, huge control of capacity, access to cost, and geographic resilience in Singapore. Area of strength for NXP is assembly and test. We're building a second factory in Kuala Lumpur, Malaysia.
It's going to be the same industrial park as our existing factory. We bought an existing building from Wistron. And so we don't have to build a greenfield factory. So this is something that we just have to modify and move forward. We'll control this expansion. So the reason we didn't make a big splash announcement when we started this project was because it's being managed within our existing capital structure that Bill's going to talk about in the next presentation. So this supports our growth, six%-10%. We need another factory, great geographic location in Malaysia. This will ramp to six million units a week. We'll control this with demand. We're not going to accelerate. We are doing product qualifications there today and putting some assets in there to build out some flexibility and geographic resiliency for products that are currently made in either Taiwan or Tianjin.
This is $1 billion of capital that we're going to slow roll and contain over time. And it supports $3 billion of revenue, not incremental. So this can handle wafers from any fab or foundry. And so it's not incremental, $3 billion, as compared to the two and the four that I talked about with the other projects. So this provides us great flexibility because of the opportunity to dual source, cost-competitive site, and geographic resilience. China. China is a great market for NXP. 35% of our revenue today is in China. Half of that, or 18%, stays and goes to end-China customers. So we expect that to grow. And we're investing in China. However, we're going to partner with local foundries in order to redesign or port current designs or develop new designs in order to serve that market in close collaboration with our customers.
So just listing a few of the partners that we're investigating with: TSMC Nanjing for 16, 20 nanometer, SMIC. We already do business with SMIC at our largest technology node. I already shared with you 40 nanometer. That's a source for 40 nanometer. And we're looking at Grace for analog products. So this will require us to redesign and target specific technologies in a China-for-China supply chain. From an assembly and test perspective, NXP Tianjin is our largest site for assembly and test. Very cost-effective. So we have a huge benefit there with an existing team in manufacturing. Then we've got a design center in Shanghai. So this sort of works for NXP. For the remaining, we'll use the existing OSATs, JCET primarily, TFME, ATX. That's the companies that we do business today for the 80/20 mix. And we'll use that going forward.
So this gives us great flexibility, the ability to dual source between China and other sites, and to introduce new technologies into China to make sure that we're cost competitive for that market. We're committed to a sustainable future. This is an important part of our manufacturing strategy. We've published our goal for carbon neutrality by 2035. We've remained on target to our annual goals. You can see the 2023 achievement there for water recycling, renewable energy, and carbon emissions. We'll continue to fund this as part of our capital structure that Bill outlined in the next slide. And we remain committed to our 2027 goals for all these things. So continued importance and investment in renewable energy and a sustainable future for NXP. So to summarize, our hybrid manufacturing creates value for our customers and creates expanded gross margins for NXP through controlled access at non-structural cost.
We need to grow our capacity to enable the 6%-10% growth. Lots of attention, as you could tell, not only around control, but geographic resilience for our different customers, regional supply chains that require that, geographically neutral locations for those customers who desire it. Focus on JV partnerships rather than a lot of investment in really heavy fabs and high risk for technology as life cycles move shorter and shorter. We'll be consolidating our 200-millimeter factories into 300-millimeter factories. Both a benefit of the consolidation. When you announce a consolidation, the factory immediately goes to full utilization as you work with customers on last-time buys. Also a benefit going to 200- to 300-millimeter, you get higher yields through edge. So there's also an economic value there. Continued investment in our strength in assembly and test. We'll do all of this with minimal impact to the environment. Thank you. I think that's the end. All right.
So I'll pass it on to Bill Betz, our CFO.
All right. Good day, everyone, and thank you for joining our investor day here, especially our owners and supporters. I plan to obviously cover the updated financial model in a little bit more detail than Kurt shared earlier, as well as refresh our capital return strategy to you all because there is no change, and we're very disciplined here going forward, but before doing so, I think it's always good to zoom out and reflect on the value we have created so far. So real quick, one slide on it, and I chose the same period that Kurt chose earlier of a five-year history because I believe this represents the portfolio and the markets we serve, along with the many macroeconomic factors outside of our control.
And now, despite again you heard living through a global pandemic, experiencing a global supply crisis, increasing global tensions, and several semi-cycles over the last five years, despite all that, we still grew our revenue at 7.3% compound annual growth rate over that five-year period. Gross margins, we're very proud of this. Over the last five years, representing the quality of our portfolio, grew by 460 basis points over the same period. Furthermore, we continue to return 100% of our free cash flow back to you all as our owners in this room and many that are online listening into today's event. And the size of that is $12.3 billion over five years. And again, representing just over 100% of total free cash flow. And then finally, I think we're delivering a respectable non-GAAP earnings per share growth rate of 11.5%. But this is history.
We can only learn from the past to improve the future. With that, let's transition into the future and recap our revenue growth rates that you heard from Rafael and Jens in great detail of where we plan to grow and how we will grow. On the left, we have our revenues for 2021, our 2024 guidance, and our 2027 midpoint of our CAGR for the next three years. The light blue is our accelerated growth drivers, and the dark blue is our core businesses. Then on the right is the total NXP growth, core and business, what we plan to grow at. Then, more importantly, our end market segments that we plan to grow over the next three years. Again, 75% of our business is auto plus industrial IoT markets.
We plan to grow this between 8%-12% CAGR over the next three years, representing over 80% of our revenue in 2027. Mobile, Curt answered this during the Q&A. We're just assuming we're a niche player. We'll just grow along with market. And then, of course, our common infra, again, Curt shared earlier, we have two effects going on. Our digital networking business declines, but that is being offset by the strength in our secure card business, including RFID tagging. So summing this all up, we believe this is a good growth, a reliable, reasonable growth for the next three years, especially where we are in the cycle. And we do expect to grow 6%-10% CAGR over the next three years. So let's recap on our non-GAAP gross margin journey.
Over the last three years, we have expanded our gross margins by 200 basis points, and we remained at the high end of our model over the last two years. And again, as I mentioned many, many times over the last three years, 58% is not our final destination. As we look ahead over the next three years, we are confident to continue to expand our gross margins by the many levers that we talk about in our earnings calls, and they're labeled here, very simple. And then finally, you get a glimpse, and Andy shared some of this as part of our hybrid manufacturing strategy. You heard this deep in our accelerated company-specific growth drivers, growing at 20%, which have better margin than our existing portfolio.
So we believe after 2027, as we continue to grow and focus on innovation in our R&D portfolio, our go-to-market mass market strategy of focusing on that long tail that Rafael explained, huge opportunities there, as well as the improvement of our cost structure. And this is important. VSMC, as shared over the summer, will boost us up by 200 basis points alone. So half of that 400 comes from our VSMC investments that I'm going to talk about and lay out the timing of those cash that we're investing in. And again, more importantly, this reduces our total fixed cost structure as a company. When I joined 10 years ago, actually, it's almost 12 now, our fixed cost structure was 70%. And now, today, we're at 30%. And in 2027, we'll be at 25%.
And in 2030, we'll be below 20% as we continue to work on making NXP more variable, more focus to access at competitive costs, as well as supplying in order to control our supply longer term, bringing that geographic resilience to us. So with that, let me move to the financial P&L, the scorecard that many of you are waiting for impatiently before lunch. So I do appreciate that. Again, quickly sharing our financial metrics. On the left, we have our targets in 2021. We have our 2024, where we're estimated to be. And again, revenue is a bit disappointed because of the macro, and we talked about that. Gross margins, I think Jeff will give me a silver star for hitting 58%. Not yet that gold star. We got work to do. Operating expenses, you know, this is timing. Some periods we're below, some periods we're above.
But we know in order to foster that organic growth, we do have to spend 16% in R&D. You heard it earlier. We're investing in more software. And so as we invest in that R&D, we're going to produce more quality products that increase our gross margins. But it takes time. And that's called our new product introductions over time. And again, I think we have leverage in our SG&A model, and we expect that to fall through as well at different times over the next three years. And again, putting this all together, again, through the help of gross margin, we're able to be above our midpoint of operating margin at 34.6%. Below the line, again, for tax, we did what we say we do. Non-controlling interest of the joint venture that we do consolidate. This is the TSMC portion of those profits.
And again, we just tightened up that range. That will be always a little lumpy within the quarters. And again, we came toward the low end of that over the last three years. And then finally, we have this new line that we have to add, and we want to add because we're investing in our VSMC and ESMC joint ventures. And so, as you heard Andy earlier, we just broke ground. And so we're starting to see and feel the small losses that will occur for our share, our equity share, which will be part of our non-GAAP earnings as we move forward. And now in blue, here is our new model. We hammered away the 6-10 already. I talked about the 57%-63% gross margin model as we move ahead. No change to our operating expense plan, 16% and 7%.
And that leads to an operating margin of 34%-60%, with 37% being in the middle. Now, again, below that line for modeling purposes, for 2025, I ask you to use 17% for our non-GAAP effective tax rate. And after 2025, please use 18%. And again, really not much change on the non-controlling interest from the JV of SSMC. But I do want to highlight, our portion in the construction of these factories will impact our P&L in the short term, but they will become a profit sometime in 2028, 2029. So the next three years, we will be recording roughly a $200 million loss because of those equity investments that we're making inside ESMC and VSMC over that time. And obviously, in the quarterly earnings reports, we will provide those numbers and best estimates to you as we move forward.
Overall, we believe this P&L is attractive and rewarding to you all, and it's very achievable as well, and so we're confident that we can do that. Now, the next slide, I want to do the same about our allocation, capital allocation strategy, and the source and uses of our cash. Again, very simple, similar to the P&L metrics. We want to break out, starting with our non-GAAP free cash flow. Clearly, here over the last three years, we didn't achieve the 25%. And you can see mainly the reasons is in our working capital. Again, in a supply crisis, unfortunately, we had to pay our vendors a bit earlier than we would like to make sure we can get that supply. And so you see an impact in the DPO metric on why we didn't achieve 75 days.
You also know very well on our inventory, it's above our long-term target, and we're holding over three weeks in our finished goods at the moment, 21 days of it, and we expect to replenish that channel sometime in 2025, allowing us to improve our free cash flow, and then this new item that you can see, it's called capacity access fee, and what this is, it's part of its core VSMC. So we have an equity position for that 40%. But guess what? We need more than that 40% the first three years. So we're buying capacity above our 40% equity position so that we can ramp and secure our customers' revenues that we will likely generate, as well as a competitive cost structure similar to a cost plus as Andy shared earlier.
If we keep on going down, CapEx of revenue, again, we said 6%-8% over that period of the last three years. We delivered slightly better as we controlled our CapEx as the revenue fluctuated throughout the last three years. Stock-based compensation, three years ago, we set out a set of targets. We are on the high side of this. We're going to tweak this a bit lower as we go forward. Equity investments, a total of $500 million for ESMC and $1.8 billion for VSMC. We've already spent up to the end of 2024, $343 million. There is no change with our capital return policy. We returned $6.7 billion back to you all in the form of buybacks of $3.9 billion and dividends of $2.7 billion. Now, you're all going to ask me, why is this not 100%?
Very simple. Back in March of 2024, we took $1 billion that we generated through our cash flow of operations and retired debt and reduced our gross debt for the company. So if you added that back in, you'd be very close to 100%. Now, going forward, looking ahead in blue again, we feel very confident we can get this free cash flow metric back above 25%. From a working capital standpoint of the metrics, there are no changes except for our inventory DIO metric. We are increasing this from 95 days to 110 days, about two weeks. And the reason why we're doing this is we're learning a lot. And as we learn from the past, we want to be able to service our customers, especially those short-term orders that come in within the quarter. There's a lot of that happening.
We realize that holding a little bit more finished goods, we'll be able to ensure we get our fair share of the market and be able to continue to improve our market share throughout all our businesses as well. And then again, the capacity access fee, we have $800 million left. And the way to think about that $800 million, $500 million goes out in 2025 and $300 million goes out in 2026. CapEx percentage of revenue, we're bringing this down. Obviously, we're making a lot of equity investments. This will run below 5%. And this will fund ESG and all the backend, including ATKL2 that Andy talked about. That's where this line would show up over the next three years. SBC, we're going to run this as a percentage of revenue. It gives us a little bit more flexibility so it doesn't run up on us.
We're going to run this around 3% over the next three years as well. Again, there's no change to our dividend policy. Very simple. The way to think about it, our cash flow from operations, we take 25% of that fixed right off the bat and we give it back in form of dividends. As the company grows, our dividend will grow as well. Buybacks, very, very simple rule of thumb. We base this all on our capital allocation structure that we like that in a way because you got to find that sweet spot. That sweet spot for us is between one and two times net debt ratio. We're very comfortable when our net debt is below two. We'll be buying back shares, very comfortable to buy back shares at that time.
When we're above two, that's when we pause to make sure we get back into model because we want to keep our investment grade and we want to continue to have access when needed to grow the company. And then finally, again, we plan to return 100% of our free cash flow back to you all. Now, I thought it would be important to help illustrate our return capital strategy over multiple years. And so there's a couple of dots and there's data. Let me just start at the bottom. There's our net debt leverage ratio. That's inbounds, out of bounds, the way to think about it. And then you have these blue bars, which includes both our dividends and our buybacks over those periods. And the blue bars represent the Q. And then you have this arrow, this swooshing over.
And between those two blue bars is what we return in that period. And so I mentioned we returned $6.7 billion. You can see that over the last three years. And that's where it is. And then that white dot is the cumulative returns over the cumulative free cash flow since we started this strategy back in 2018. And so again, at the end of 2024, we would have returned $19.2 billion back to our owners. And still, that's representing about 109% of our free cash flow. Now, many of you probably are trying to have folks trying to pencil in all this and what does that number mean for the next three years? What sort of capital return we plan to return back to you all? And so we did that model for you. And we plan to return anywhere between $9 and $11 billion.
By the end of 2027, over a 10-year period, NXP would have returned approximately $29 billion back to you all. Again, we believe this is quite attractive and we believe we can deliver it and we will deliver it because there's no change in our capital return policy. With that, I have one last slide that I'd like to review with you all, and then we'll open it up for Q&A. I'm going to do something that Kurt did, Rafael did, Jens did in their presentations. You heard a lot from us, from me, a lot of numbers, a lot of content. There are three things I want you to remember today. First, we plan to grow the company over the next three years from 6%-10% as a leader in bringing intelligent systems to the edge.
Second, we plan to expand our gross margins to over 60% with that new range of 57%-63%. And there's more after that, but it takes time. And that's through the innovation, improving the quality of the portfolio that Jens and Rafael shared in many, many details and examples of what we're doing, our go-to-market in that long mass tail, as well as the hybrid manufacturing strategy and our investments we're making as we move forward. Thirdly, no change. We will return 100% of our free cash flow back to you all in forms of dividends and buybacks. And again, that amount is approximately $10 billion over the next three years. And if you only want to remember one thing, remember this: NXP will double their non-GAAP earnings per share by 2030.
So with that, I would like to say a big thank you to you all coming out here and for sharing and supporting us over the many years being with us. And I look forward, we look forward, and super excited to bring NXP to new heights.
So thank you.
So we're going to start another Q&A session. I'm going to invite all the presenting MT up to the stage for right now. And we're going to go with Blaine Curtis right here. And if we could say your name, your broker you're with, and one question, and then we'll keep going.
Thanks. This is Blaine Curtis at Jefferies. A question for Kurt or maybe Bill. Just on the China for China strategy, obviously, there's a lot of talk about China for China. So I understand the motivation. I'm kind of just curious in terms of other multinationals matching what you're doing as well as local competition. So when you say we need to get the price point down to remain competitive, I'm just trying to understand who's that for first. And obviously, you had a great presentation to show why your products are better than most. So would this be you being conservative and maybe down the road you actually can sustain the ASPs and the cost is lower? I'm just trying to understand the moving pieces there.
Yeah. So Blaine, on the local competition, we have a paranoia about this. And I see it in two phases. Today, there is not a competitor which is in any of the strategic areas which we do competitively from a performance perspective.
There is a lot of focus from the local Chinese competitors in discrete power, which is very understandable because it combines manufacturing capability with device manufacturing. So they can do both at the same time. Since we neither do MOSFETs nor silicon carbide nor gallium nitride or any of these things, that doesn't touch us. And it takes a lot of their effort away. The other thing we see them doing is low-end microcontrollers, which there is Rafael here, which touch the very low end of what Rafael was showing that would be the MCX portfolio. We think the total revenue these Chinese competitors currently achieve is in the $3 million-$400 million range. So it is still relatively small. But there is an enormous stickiness through software play. I don't know, Bill, I think you asked earlier about these faster cycles.
The stickiness of the product goes away, but the stickiness of the platform goes up because the investment of the customers, and that holds also for China, is in software. So as long as they cannot offer a processing portfolio where software can be reused from the low end all the way to the high end, it isn't that useful at all. So I mean, again, I don't want to talk them down, but we haven't seen a lot of progress from these companies to come into where we really play, which is the mid and higher end of processing. So therefore, when I say two phases today, they just don't really compete with what we do, but we stay absolutely paranoid, and we see this on two axes. One is manufacturing.
That is why Andy showed you how we invest actually into relationships to offer Chinese-made product, which is partially a government-induced requirement plan. Cost is one side of this, but the bigger side is that it becomes a tick-box item. If we cannot quote a product which is locally manufactured, we fall off the list. Now, that was a burden to start with, but when Andy offered us a great strategy how we can deal with it, it becomes a different story for NXP because most of our peers have their factories in their Western home countries and cannot flexibly move into China like we do. So it's actually a good thing. The other one is following the lead customer requirements in China, which I think I spoke to Bill or somebody about earlier, which keeps us competitive. Why does it matter, Blaine?
I think last earnings was kind of very illustrative for this. What currently grows in automotive and industrial is China. It's not the Western world. It is the Chinese customers. And my sense is they will continue to make more advancements into the world with their not only cost structure, but also competitiveness from a performance perspective. I mean, let's not say this is just cheap. They are actually very advanced. I mean, you talked about SDVs, Jens. Chinese companies are managing this transformation very successfully currently. So for NXP, and you had the number, Andy, 35% of our revenue is in China. We think half of it is China for China approximately. And at least the way the economy currently goes, we think that will grow as a ratio of NXP.
Josh, right down here.
Thank you. Josh Buchalter from TD Cowen. I wanted to ask about the gross margin target. I guess I'm a bit surprised to see it widened and compared to the last one, and certainly not trying to dismiss that it's going higher. But coming out of a very volatile cyclical period and as you move to sort of more outsourced manufacturing, can you maybe explain why the gross margin target doubled in width? And maybe within that, should we expect more volatility and cyclicality going forward? I guess is your broad industry view in particular as China takes more of the auto and other markets? Thank you. And thanks to Jeff for the assist there.
Sure. Let me take that one. If you think about next year and what I said during earnings, obviously, we don't want to be right out of model the first year. And revenues playing a role in that.
So I would just say that's the only thing, to be honest with you, taken into the short term on the widening. But we are very confident to get toward that 60 based on all the levers we talked about, Josh. Quinn in the very back.
Hi, Quinn Bolton with Needham. I want to ask a longer sort of question on the manufacturing strategy in VSMC. You say you'll be consolidating the 200-millimeter facilities into VSMC over time. Do the 200-millimeter facilities go away entirely? Do you keep operating those for supply chain resiliency or redundancy purposes? And then as you manage that transition, are you anticipating building up significant buffer stock to manage that transition, or just how does that process work? I know it's kind of a 2027 to 2030 question, but just sort of thinking about that. Thank you.
Yeah, sure. So it'll take about 10 years to make this migration. We work closely with our customers, last-time buy forecast, product transitions to other NXP products to make sure. But again, as I mentioned in my presentation, once you start that project, the utilization immediately goes to 100%. And it can take two, three years in order to satisfy those last-time buys. So that's a utilization benefit to gross margin, right? And to answer the second part of your question, no, they would be exited completely. So we're not going to maintain some small pilot line or anything for geographic resilience. We're moving to geographic resilient locations. And I think that hybrid structure works f
or us. Great. Right there. David, please. David O'Connor.
Thank you. It's David O'Connor from BNP Paribas. I just want to go back to the systems point of view that Rafael and Jens mentioned earlier.
And maybe also you can chime in there, Bill. You all talked about more compute we need, more reference designs, more software, more accelerators. Things are getting more complex. At the same time, I'm just wondering how you're getting paid for that because you still talk about pricing down to a single digit. And then, Bill, on the gross margin, I didn't really see anything on the system side of things. Maybe it's in there on the mix, but it seems to be not a big part of that. If you just talk around how you're getting paid for that system approach. Thanks
. All right. Yeah, let me take that because this is also a real key element of our CoreRIDE platform. And basically, if you talk about the processors and you have your embedded software, typically, you don't get extra money for that.
That's part of the hardware price, right? So the enablement software, that's really typically part of the hardware. But if you go into the CoreRIDE model where we are approaching really the system-level reference designs where we have these system-verified and production-ready reference designs, we do see basically two effects. On the one hand, the entire bill of material is becoming bigger as we pull way more components into the deal, right? So the amount of products we are selling is accelerating and it is growing. Plus, we have the possibility with this approach to get also way more opportunities from the customers. We heard that a little bit. One takes it all. They want the solution from one player and they want it from the low end to the high end. So this is pulling way more hardware with it.
But then, the element of the software where we do the middleware integration with the CoreRIDE platform partners together, the business model is basically made as such as the recurring revenue from the software from the middleware, for instance, is still going to the CoreRIDE partners because we also don't want the liability, etc. They still do the recurring revenues. But we will have a margin on our hardware for the integration work we are doing in bringing this all together. So there is a bit of a financial reflection on this activity as well. That's what we call mix in NPI. So as we move up, it's more valuable product that we're selling over time. So that's where you get the mix up. So let me just kind of capture just a subtlety on the answer for industrial. There are three ways.
At the end of the day, NXP needs to get paid, right? Especially if you're adding value, you need to capture a portion of the value. There are three ways. So the first one, the way you think about it, especially in industrial, is today we have about every time we sell a processor, at the average, it doesn't matter what processor it is, we have about 25% uplift in attach. The whole thesis is that as the customers use more and more of these systems, we are going to have a multiplier in the attach from 25 to. And so we have a particular target, and the thesis is that we can actually go and double this in the next three, four years. So that's one way of saying that's one way to get paid.
The second one is, by definition, the value that you're adding. You will be able to command more gross margins, and the last one is the insight that we get with the customer and the stickiness that we create with the customer. We have created longevity, so the amount of time that we last on that socket will last, and so the thesis is that with three ways that we get paid in industrial is those three that I explained, and I want to highlight your last point, Rafael. It's really important because it touches a lot of earlier questions. The stickiness of our portfolio moves from the component to the platform, which gets the customer bought into the software, which is written on the platform, independent of the individual product.
So the stickiness expands into the total lifetime of the platform, which is a total transformation of how our business operates. It's not a gross margin answer, but it is an answer on where is the future stickiness of the NXP portfolio. It is across those processing platforms in the software which customers have invested. Great. Tammy, down here with Ross right here. He knows his name. Thanks, Jeff. Ross Seymore from Deutsche Bank. Bill, one for you on the free cash flow side of things that you wanted to have the 25% target. It seems like there's some significant differences in just how money is going out the door where CapEx in and of itself doesn't capture it as much as it used to with all the JVs, equity investments, etc. So how do we think about when the total cash will start to grow from that perspective?
Does the DIOs that you had going from, I think, 95-110, does that include some buffer inventory for the fab consolidation you're talking about? So it could go above that. Just trying to think through working capital and then the equity side to get really what cash is doing. Sure. So the working capital that I shared in 2024 of the 19%, I'd call a point is tied up in DPO. About a little over two points are tied up into inventory. And then that capacity access fee is probably another point. So next year, we won't be modeling, obviously, because we have a lot of cash going out for those two investments that we talked about, both in a combination of capacity access fee as well as the equity investments we're making in both those factories.
But after that, as it tails off, we will then improve our free cash flow as we move ahead. Related to the inventory target of 110, really, that just takes into account of ongoing operations of servicing our customers and really focused on those churn orders that we receive in the quarter. What we're realizing in this environment, especially with some bad behaviors by our customers, we're getting so many orders late, and now we're able to service them. Related to any consolidation, they probably take about two to three years, and we haven't shared exactly the timing of it. Clearly, we would love to find partners out there that are willing to take over the factory, which I'm sure we're exploring at the moment.
But more importantly, when that occurs, we will then say, "Okay, how much of that pre-build, what impact does it have on inventory?" And I would share that with you on a quarterly basis, of course. Great. Mark, right down here, Jennifer. Thank you. Mark Lipacis, Evercore ISI. Thanks again for the presentations, Bill. I think a follow-up from Ross's question here. When you talk about the free cash flow that is available to be returned to shareholders, the investments in the JVs, are you kind of considering that to be like a CapEx that would come out of the cash flow from operations before you have the free cash flow available, or is that something separate? If you could spell that out, that would be great. Yeah. So the $10 billion number I included includes all that.
So after we've paid, includes we generate to be able to deliver $10 billion by the end of the year. So those investments effectively are like CapEx leveraged? Correct. That's the way we kind of treat it, even though it shows up on a different P&L line item. Or I'm sorry, cash flow line item. Right. Okay. That's clear. Thank you. Great. Thanks. Vivek right here. Jennifer. Thank you. Vivek from Bank of America Securities. The 57%-63%, the gross margin model, is it a way to think about incremental gross margins, right? Or does it literally correspond to kind of the bottom and low end of the grid? How sensitive is that to the actual growth rate?
And if I can squeeze in kind of quick part B of that, I think, Kurt, during one of your presentations, you mentioned that for next year, you expect to be a touch below model. And I was hoping I could just push you a little bit more on how to think about 2025, right, based on that comment. I answered this, and then you go, Bill. I said a touch below model, or I meant maybe I didn't say it clear enough for revenue. Because that was associated with a revenue question over the three years.
Somebody asked, I think Chris was asking or somebody, "Why is your revenue target moderated versus the former one?" And I said, "The revenue target is moderated because we just don't know with all the uncertainty for next year, and I want to capture this in a somewhat lower CAGR for the three years." Bill will explain to you now that we will be for all the rest of the P&L totally in model. Right, Bill? Yes, that's correct. So the way to think about it, clearly, we have a step down in Q4 on revenue, and we guided a normal seasonality down into Q1. So revenue has effect. And then you also know our internal utilizations, what we're running at are in the low 70s.
So again, at some point in 2025, when we start to sequentially grow, you will then start to see the lift back to our gross margins because of the fall through of our fixed costs through the improved utilizations. It's just timing the way to think about it. And then mix and NPI and everything that Jens and Rafael talked about, that NPI of moving up, the quality of our portfolio staggers on over time. It just takes time as they ramp. It's the best way to think about gross margin. And SSMC is 200 basis points, but that's post-2027. Great. Toshiya, I think you had a hand up right there. Thanks, Jeff. It's Toshiya Hari from Goldman Sachs. I wanted to ask about your go-to-market strategy, which so far really hasn't come up too much. Distribution has been 50%-55% of revenue.
A couple of your peers are taking a fairly, well, I guess one in particular is taking an extreme route. A couple of others are no longer paying for demand creation in some instances. How are you guys thinking about the balance between direct and distribution? And in your long-term model, from a profitability standpoint, what's embedded? Thank you. So the business through the channel is key to NXP with our focus especially on the industrial market, which is characterized, as Rafael showed, by 75% long-tail customers. We have a strong focus on that. We definitely want to continue to go with distribution partners. So we don't take that step which others might have taken or not.
We need the distribution partners as well as other design partners in order to actually serve that multiplied distribution channel such that on a go-forward basis where today it's often like 50-50 between direct and distribution, it might actually touch up in favor of the business which goes through the channel, and that is a creative. I also had, I think, in what I said and in what Bill repeated, as one of the margin levers, we had mass market as a statement, and that is the expansion of mass market customers, so it's not more business with the same customers, but it's simply touching more customers, and that is adding to the margin, but of course, it also adds in the relative way to revenue in favor of the channel, but we are committed to the channel. Let me just be very clear.
We don't do anything else we couldn't do without. Great. Hey, Will, I think you had your hand up, and then we'll get to you, CJ. Sorry. Thanks. Will Stein from Truist. Can you talk about your anticipated spend on M&A? I don't think that was really discussed in any detail in the capital allocation plan. I know that on the call asked about this, and I think it was discussed that this was going to be software-focused. But any comment in terms of sizing or the potential for something that would be sort of a material contributor to revenue? Thank you. So that was about M&A. It was hard to hear. So that's the same question you asked before, Will. You will get the same answer.
I think you saw that my two friends to my right-hand side have an insatiable hunger for innovation, rightfully so, to deliver growth with our customers. And we will continue to, and we have done tuck-in acquisitions, small size. Sometimes we publish them, sometimes they fall under the threshold, and we don't. And the growing part of that, Will, is in the areas of software, which goes back to this CoreRIDE platform idea, which goes back to the software, which is pre-integrated in the solutions from Rafael. So that has been part of the strategy, and it continues to be part of the strategy. And we do not have a big part in our portfolio where we have a big need. So we don't stand here and say we have a huge gap in our overall portfolio where we absolutely need to acquire something to fill a striking gap.
On the transformational M&A, Will, you heard that song before. I personally, I am a big believer that the semiconductor industry needs more consolidation from a global level. That's a philosophical statement, and it's a philosophical statement because I continue to believe that the regulatory environment is not allowing easily deals on that level. Therefore, in our book, in our planning, are these tuck-in acquisitions which are helping the strategy you heard about today. Okay. Great. Jerry, right down here. Jennifer. Thank you. This is Jeremy from Stifel. I guess if I could, two quick questions. First one is regarding a little bit more clarification on the mass market strategy. Can you talk a little bit more about what specifics maybe that you're doing?
Is there more investments that you may need to do, more tools that you can provide to these smaller customers, maybe an e-commerce strategy, and what kind of investments this might entail? And a really quick clarification on the capital or capacity access fee. It sounds like it's not a deposit. It's a cost. How does that flow through the financial statements? Is there any impact on the gross margin? Thank you. Yeah. So on the mass market, that's very straightforward. We found out that from a competitive benchmark perspective, we should touch more customers, small customers, but we need more. We benchmarked ourselves against the number of peers, and we just found that they touch more customers through the same distribution network. So we are working with our distribution partners to expand actually our reach to a broader basis of individual customers.
That takes some investment, but it's all captured in the P&L and in the model which Bill just showed you. But it's not material. The other question I think was the capacity access fee, and I don't know, Bill or Andy. Yeah, I can take that one. Again, what we're doing with the capacity access fee because we need to ramp, we need the capacity. So we're buying more capacity above our 40% at a cost-plus model. That's different than buying it on a foundry model. And then you can do the math. It has an impact to gross margins. That's favorable. And a key thing I just want to add, Jerry, to Kurt's response. The mass market doesn't imply new products having to be developed. These are existing products taken to a longer tail of customers. Okay. With that, CJ, right down here. Thanks, Tim. Great. Thanks.
Bill wanted to clarify just one thing on free cash flow. Are you including the equity investments in that calculation, or you're excluding? I'm including. Within the CapEx line? No, not in the CapEx line. Below, I called it out as a separate line. But when we generate the free cash flow, the $10 billion, the $9-$11 billion, we included it. Okay. So your adjusted free cash flow is included? Yes. Okay. Perfect. And then the main question was really around auto. And one of the great aspects of the business is long life cycles, very sticky. You're talking about China now moving to a very fast annual type of cadence. Is that a good or bad thing for the industry? Maybe I jump on this one. Basically, I do see it as a good thing because this really helps us to drive innovation way faster.
You can clearly see that, especially in China, and this is also why I really like the engagements in China. You have seen also some of the quotes and the references I gave. They are really, really fast-paced catching up and implementing innovation way ahead of the rest of the market. On the one hand, because the innovation cycles are so short, but also because they just simply try and take the risk. What they typically do is they spin a car platform in 12-24 months, but it follows right away with a 6-9 months improvement cycle of the platform they have done. They take the risk, they go fast, they implement it, they bring the innovation to reality. It's fast time to market also for me. They work with me on driving the total cost of ownership.
They do improvements, etc., right after. That is spinning way more. It's super helpful for us. It's content-driven, significantly stronger in China. Thank you. Any other questions? Chris Caso. Yes, thanks. Chris Caso, Wolfe Research. Bill, I just wanted to clarify the equity investments because it's a fairly big number. I'm not sure it's been fairly fully dialed in. So we should be assuming it's $200 million a year, 2025, 2026. Oh, not per year, sorry. It's the cum? The cumulative. Yes, sorry. Right. Okay. So that's an important number. And then essentially that's, well, I guess it's more than offset by the reduction in CapEx on the model because you've reduced the CapEx on that. That's correct. As well as eventually getting our wafers at a cost-plus model as a JV partner versus a true foundry where there's margin stacking to be had as well.
And then eventually that turns positive, so we get a return on that investment as well of our 40%. Yep. Right. And that's assuming that the market goes well, that the JV's turning a profit. And if during a downturn, you would see underutilization from that JV fab on the non-controlling interest. Yeah, but Chris, don't forget, we fill more than 40% of that factory. And we fill it by, amongst others, not only by consolidating our other 200 millimeter factories. So we can make sure it's full. It is in our control, which has then a double positive because we have then more business running in 300 millimeter at cost while the fixed cost burden of the company on our internally owned 200 millimeter factories comes down at the same time. Understood. Thank you. Great.
I think with that, Kurt, I'm going to pass it to you for any final closing remarks that you might want to make. Yeah. I will not ask you for the three things to remember because I know you know. So I just mentioned number four, doubling of the non-GAAP EPS by 2030 plus. And since all of you have taken the time to come, you on the webcast to join, I want to just send one word of big, big thanks to the NXP team globally, to the NXP team here on the ground supporting the event, but much more so to the global NXP team to facilitate the business, which, as Bill has shown you, has delivered proper results and returns over the past years. And we want to drive even much harder and much better in the years to come.
So thanks very much to the entire global NXP team. Thanks to you. We will have a lunch now, I think just next door. And I look much forward to see you both over lunch, but also at the next conferences, etc. Thank you very much. Thank you. So that brings to a close our formal presentation. As Kurt said, we will have a lunch out and around the hallway here, and we look forward to seeing you there. Thank you very much for all your support.