We, as the stewards of Moore's Law, will be relentless in our path to innovate. Moore's Law is alive and very well.
I am just really thrilled that we have the opportunity to take this great icon of a company forward as never before. Our best days are in front of us. IDM 2.0 is our evolution.
Historic investment for Ohio, one of the largest investments in semiconductor manufacturing in American history.
Every aspect of human existence is becoming more digital. Demand for semiconductors is truly unprecedented.
Intel Foundry Services will open up the Intel fabs for the industry with leading process technology, a wide range of our own and third-party IPs. We have a clear path for the next decade of innovation to go to one nm and well beyond. We will be relentless in our path to innovate with the magic of silicon.
We have to control and manage our supply chains.
Introducing the all-new 12th Gen Intel Core processors. Intel's most significant shift in x86 architecture in more than a decade.
The Intel Xeon platform is the foundation. It is the most pervasive platform on the market, delivering everything customers need.
As gamers, we expect more. As Intel, we want to deliver more.
You know, today we have Mobileye technology being used by 29 automakers. More than 88 million cars are on the road today.
We're really on a mission to let those who build networks take control to be able to program them for themselves. We're gonna give them the hardware and the software tools to do that.
Today's developers are discovering leading-edge concepts and creating the future. We can, and we will do better.
Most trusted companies in America. I've seen the technical work and progress of our team. I have full confidence that Intel is in a very strong competitive position, and our best days are ahead.
Welcome. For those of you in the room, welcome back. For those of you joining us online, let me be the first to welcome you to Intel's Investor Meeting for 2022. My name is Tony Balow, and I have the pleasure of running the investor relations team here at Intel. I'll also be your emcee for this afternoon. Before we start, let me read a relatively brief safe harbor statement. It'll be a little shorter than that. Today's event includes forward-looking statements which are subject to risks and uncertainties. Please refer to our SEC filings at intc.com for more information on the risk factors that could cause actual results to differ materially. We will also be presenting non-GAAP financial measures during today's event. Please visit intc.com for reconciliations of these measures to GAAP measures.
For those of you who couldn't join us here in San Francisco this morning, we held a series of small breakout sessions covering several of our business units, Foundry, Mobileye, Network and Edge, as well as several of our functions like technology development, software, and manufacturing. For those of you online, that material, as well as recordings of those presentations, is posted on our investor relations website, intc.com, and you should be able to download it now. This afternoon's session will start with hearing from our CEO, Pat Gelsinger, who will cover an overview of our strategy as well as our plans for the next several years. He'll be followed by brief presentations from both our client group as well as our data center and AI group with Sandra Rivera, Michelle Johnston, and JJ Johnson.
We'll have a short break then, and for those of you online, you'll have an opportunity to step away as well. After our break, we will hear from our new CFO, David Zinsner, who will present the financial outlook for the next several years as well as a full year 2022 outlook. After that, again, we've saved a healthy amount of time for Q&A, and we'll do that in two parts again. First, we'll start with Pat and Dave on stage, and then I'll invite the rest of the executive management team up so you can ask questions of them as well. Finally, just take a moment.
I would be remiss if I didn't say thank you to everybody who has helped us get this far in the presentation, as well as thank you for all of you in the room and online who have joined us here today. With that, I'd like to welcome to the stage our CEO, Pat Gelsinger.
Thank you. Thank you, Tony. On behalf of 121,000 loyal, passionate Intel family members, it is my privilege to be representing you, representing them, and welcoming you to our investor conference today. You know, I think we have a lot more exciting stuff coming up. The morning was fabulous, but let's dive in. In 1979, as a pimply-faced 18-year-old kid, as I say, I went through puberty at Intel. I started at this great company, and I had the honor to lead the team that created the 4006 microprocessor to become the first ever CTO. I joke that I, you know, learned at the feet of Grove, Moore, Noyce, and really got to seep in the rich history of this iconic company. Here we are one year after I had the opportunity to come back and be its CEO.
30 years with the company, an 11-year CEO training mission as I took vacation from Intel, and now back for a year. I came back to the iconic company that I was born and raised in, the iconic company that defined the technology industry, but it's also so critical for our nation and the world. Today, we're gonna describe how we are rebuilding this iconic company. Since my return, we have had a torrid pace. We've added a new word to the English, and certainly the Intel dictionary, torrid. With our IDM 2.0 strategy that we laid out, you know, we began this tenure with a bang. You know, we laid out a new strategy for the company.
We followed that quickly by, you know, laying out our process technology advancements, our new factory expansions in New Mexico, Malaysia and Arizona, and last month, the Silicon Heartland. We've continued that torrid pace. We gave the most detailed and transparent roadmap update that we've ever done for our process and package technologies, a path to regain unquestioned leadership, five nodes, four years. We laid out our most dramatic set of architectural breakthroughs as we launched Alder Lake, our first hybrid core, and our new graphics product line. We laid out how we're getting the geek back with our innovation conference, this revival and renewal of our commitment to developers everywhere.
We also said, "Hey, we are the stewards of Moore's Law, and we're gonna keep that legacy alive and well for decades to come." If anything, I feel like we're a bit ahead of schedule where I thought we'd be at the one-year mark since I came back. In all humility, we still got a lot of work to do. Today we're gonna show you how we will accelerate the growth of this great company, delivering a compelling strategy and creating extraordinary shareholder value. Tremendous amount of information that we're covering today, and I want you to take away four things. First, we're rebuilding that Grovian execution, as we call it, you know, bringing back that heart of Andy Grove, the confidence, the engineering centricity, the discipline, the competitive spirit. Second, we got the right strategy. We've laid out the strategy. We know the path to the future.
Our core technology, our leadership products, how we're driving these open platforms and industry standards, how we're building capacity at scale. Third, we're harnessing these core strengths in traditional markets, but also accelerating and disrupting new markets as well, and how Intel is the only company that has the breadth, depth of software, hardware, manufacturing at scale to execute on this strategy, and how our business model has fundamental advantages in how we can invest in process technology and leverage it across all of those business models. Only Intel can do this. Finally, the next chapter of the Intel story is one of growth, and it starts today. Our plans won't come to fruition overnight, but by the end of the day, I think you're all going to agree we are well on our way.
We've repositioned this company, and we are executing on that strategy, and the future of growth begins today. There's three parts to my presentation today. First, our beliefs in how the industry's gonna evolve. Second, our strategy. Execution, how are we gonna get it done? You know, as we look at that, we've rearchitected the company into these six distinct business units so that we can compete vigorously and drive growth. It isn't just six random business units. They reinforce. They build on each other and reinforce each other, building on common capabilities. We're gonna cover each one of them today and then show you exactly how we can accomplish that double-digit growth that we described. Some of you have said, "How are they ever gonna do that?" Well, we're gonna tell you today how we're gonna do that.
Our four critical beliefs, how we're shaping the future of the technology. First, we believe we're on the front end of sustained growth in semiconductors and technologies, driven by the digitization of everything. Technology is increasingly central to every aspect of humanity, and semis are needed for everything digital. We believe this is driven by what I've called the four technology superpowers. First, 4004s everywhere, right? Ubiquitous compute. This thing called microprocessor is now permeating every aspect. Second, infinite infrastructure, from the largest clouds to high capabilities, low latency edge devices. Pervasive connectivity. Everyone and everything is becoming connected. Finally, making sense of it all with artificial intelligence.
Those superpowers are enabling us, and as you saw in Raja's presentation this morning, you know, that it's creating this opportunity, this challenge to continue to create compute opportunities where we're simulating every aspect of human existence, fully immersive, digital experience, intelligent avatars, digital twins, all of these, you know, these experiences. There's an insatiable demand for compute, and that insatiable demand for compute is gonna require general purpose compute like we do, but also graphics, immersion experience, I/O capabilities, accelerators. This demand is insatiable, and we believe that it's foundational and will change the structure of humanity and the semiconductor industry. We estimate that to be a $1 trillion market by 2030, the doubling of the semiconductor industry over the next decade. We are uniquely positioned to capitalize on that. It isn't just that it's a larger market, it also needs more powerful compute capabilities.
This is our second belief, that these technologies demand, require leading-edge technologies. By the end of the decade, 50% of that demand will be for leading-edge technologies, growing 6x the rate of the industry overall. As we've seen, the shortage is also driving older nodes to more modern volume nodes as well. People are redesigning more rapidly as a result, driving towards this leading-edge capability. There's only three companies that have the R&D, the capital capacity to do that, and only one in the West. We see this wall at 10 nanometers, the Angstrom era, EUV technology that creates the haves and the have-nots. Moore's Law is alive and well. With our RibbonFET technology, PowerVia, 3D packaging, EUV, High-NA, we're on a path from 100 billion transistors this year to 1 trillion transistors on a package by 2030.
As I like to say, until the periodic table is exhausted, we are not done. We're going to keep pushing that arc and bending physics to keep Moore's Law alive and well. We also believe that open ecosystems are more powerful than closed. They unleash innovation, democratize compute, and we have a long history as a company in this area. Personally, I helped to create USB and Wi-Fi. I love it when my granddaughter, she plugs in a USB stick, and she says, "Thank you, Papa." The things that we do enable industries. We open compute with Xeon. We've displaced mini and mainframe computers. We've created the foundations for today's cloud computers. We're seeing similar things happen in new markets. We saw the industry move from centralized to decentralized to centralized computing with cloud.
We believe that next swing to the edge is now underway, this pendulum of the computing industry, where the edge enables low latency, high bandwidth, inference, and AI at the edge. We see machine learning and deep learning moving from proprietary today to open systems in the future. We're committed to be that company that creates, that democratizes, that enables these open ecosystems. Number four, the world desperately needs more geographically resilient supply chains. Clearly, the COVID experience, what we've gone through with the semiconductor shortage, has caused this rapid, extraordinary disruption of every aspect of the economy, driving large portions of inflation today. Over the last 30 years, we saw that we went from 80% of semiconductor manufacturing in the West, in U.S. and Europe, to 80% in Asia, right?
All of a sudden, we realized that we had the drifting of our entire supply chains for something more important to humanity than where the oil reserves are now highly concentrated and not in any way resilient. My moonshot is that by the end of this decade, the US would have gone from 12%-30%, Europe from 9%-20%, that we'll have gone from 80%-20% to 50% by the end of this decade. We've seen extraordinary tailwinds to help to reestablish this geographically balanced, more resilient supply chains for the future. A bet on Intel is a hedge against geopolitical instability in the world. Government leaders are recognizing this, how vital semiconductors have become to every aspect of the economy, every aspect of national security. The world has woken up.
Both the CHIPS Act and the America COMPETES Act in the U.S., you know, are now headed to conference process to finish and enable this $52 billion in incentives to fuel the semiconductor industry. Just last week, the EU issued their European Chips Act, which is moving quickly. We see the set of incentives and strong geopolitical support as a further tailwind for Intel. You're gonna hear from David this morning about how we see these financially helping and driving and improving our CapEx and cash flows over time. With those beliefs, our strategy then capitalizes on them, believing that, one, we're gonna deliver leadership products in everything and every market that we compete in. The answer is really simple. Build the best products, and that's what we're gonna do.
Those are powered by these core strengths, supercharged by the incredible talent that we have in this company, and building on our traditional markets and businesses, but accelerated by our entry into rapid, large growing markets as well. Our business model has resilience and flexibility built into it. We have the right strategy, and we're executing today. Starting with leadership products, here's five really good ones that we got. When we launched Alder Lake last year, it was like, wow, we now have the best mobile product, the best desktop product, the best computing client that's ever been delivered. Over 140 customers in 30 countries with rave reviews. It crushes our chips as well as all of the competition. With Sapphire Rapids that begins shipping this quarter, up to 30x the performance in AI and key new features and security, a great product.
Mount Evans IPU redefining the data center network and partnering with Google and cloud providers to enable this new model of computing. Ponte Vecchio, a tour de force demonstrating our leadership in packaging and GPU capabilities, enabling the two-exaflop Aurora supercomputing. Finally, Mobileye. Ain't the car out there pretty cool looking? Yeah, I love it. Yeah, I want one of those. I want. Somebody said when we were practicing, they said, "Hey, you know, can I drive it home?" I said, "No, no, it drives you home." We're laying out this vision of the supervision strategy for the future. These products are leadership, and our roadmap that we'll describe to you today only gets better. We've recommitted ourselves to being an IDM, an integrated device manufacturer, and that starts with regaining process technology leadership.
In July, we said five nodes in four years, and a bunch of you looked and said, "Have you gone nuts? Doesn't it take two years per node?" We in confidence said, "five nodes in four years." We remain on or ahead of schedule against the timelines that we laid out. How can we catch up so quickly? Well, we've made some fundamental changes, and if you heard from Ann this morning's session, you know, we've leaped forward with our embrace of EUV. We've built deep partnerships. You know, we used to tell the equipment vendors, "Drop it on the shipping dock, and we'll take care of it from there." Now we're deeply partnering with them, and particularly, the ASML relationship is superb. We rebuilt the leadership team. We reinvested.
I whipped out my checkbook, and, you know, one time Andy Grove said to me, "I thought I gave you an unlimited budget, and you still overspent it." Well, to some degree, we've given Ann an unlimited budget, and we said, "Get us on track." We've invested heavily in the equipment and the engineering to go do that. We created a parallel tick-tock development model for our teams. We've organized into common technology teams. You know, the Intel 4-three team is separate from the Intel 20A, 18A team, so we've created this parallelism and risk reduction in our approach. You know, our process integration was stalled at 14 nm bringing it into production, but our components research team wasn't, and they kept alive and well with innovations like RibbonFET and PowerVia. Finally, we've simply become more open and more engaged.
You know, we're leveraging the expertise not only of the equipment industry but the EDA industry and, you know, moving to industry standard design tools and PDKs. Those of you who might talk to industry and equipment vendors, they will resonate. This is a new Intel, more open, engaged, and proactive than ever before. You don't just have to believe us. We're gonna give you proof points against it, and some of those have already started to emerge. Intel seven, Alder Lake, best product shipping in volume. five nodes, four years, one done. Meteor Lake, taping out the production stepping later this year, well on our way to having Intel 4 Meteor Lake. Intel 3 , you hear from us this morning, huge acceleration in that, and we've accelerated our server roadmap to embrace and deliver using Intel 3.
In 2022, we'll have the first data points on 20A and 18A. To just bring that home today, let me show off. See if I can do this here. This is sort of like the Simba moment, right? You know, boom. Intel 18A, right? Our first SRAM wafer's up here for it. You know, we're processing. We're running gate all around with our RibbonFET technology, our PowerVia technology, SRAM proof points that are on or ahead of schedule to our internal as well as our foundry customers as well. Thank you. You know, we expect these major steps forward as we go through the year with proof points that we communicate will rebuild, restore your confidence. If you weren't just a little bit excited after listening to Ann's presentation this morning, you're obviously not a geek.
Foundational to leadership and process and products is our ability to manufacture at scale. As Keyvan mentioned this morning, we are executing the largest build-out, you know, of capacity in the company's history. 24 concurrent construction projects. I joke that I have more concrete trucks working for me now than any other human on the planet. We are underway with major build-out. From the company that created Silicon Forest to help to establish the silicon in Silicon Valley, the Silicon Desert in Arizona, now the Silicon Heartland. I'll tell you, this announcement that we did in Ohio, I was expecting it to be good. I underestimated it. This idea of manufacturing heartland in the middle of our nation, I'll say the enthusiasm we've gotten from that across the nation has just been spectacular.
Coming soon, our next mega fab location in Europe, and this public-private partnership, you know, has been an extraordinary journey over the last year, and you'll hear Dave talk about some of the benefits of that as it rolls through our financials this year. You know, another core aspect of our manufacturing is sustainability. As Christy spoke about in her session, our commitment to our RISE goals, you know, that we're setting the standard for sustainable manufacturing, aggressive environmental targets, renewable energy use, water reclamation, landfill conservation, you know, advanced chemistry to minimize hazardous substances. We are the unquestioned leader in environmentally responsible manufacturing. We're good neighbors in our community, and just last week, Barron's named us the most sustainable company in the industry. You know, it's just like, wow.
You know, this is just over and over, we believe so deeply in what we do, but it's not only doing great things, it's doing the right thing as well. Given all of the massive supply issues that we've gone through, our manufacturing network and what Keyvan has done with the team has just been fabulous. We have customers now saying, "I only want products that run in Intel factories. I want you to build everything for me," because every other supply chain is so disrupted around the world. In summary, IDM gives us scale, cost, flexibility, control over our supply chain, a business advantage that no one else can deliver at scale. We also have a rich history of embracing open. We democratize compute. We create these growth engines for the industries. You know, these platforms become sustainable engines of growth.
How do we truly unlock that trillion-dollar TAM opportunity? You're gonna hear this morning from Sandra or this afternoon from Sandra, MJ, and JJ talking about that. You know, in Greg's session, he talked about how we're democratizing AI. You know, we also have this recognition that not only do we democratize and open compute, but we create open platforms. You know, and things like Wi-Fi, USB, PCI, and now CXL, we drive these standards that create the platforms. You know, we've launched the Open Edge, open software platforms, and now we're also opening our manufacturing.
We heard from Randhir about how our strategy in IFS is fundamentally about opening wide the doors of our manufacturing engine, driving the chiplet standards in the industry, opening for the first time ever the x86 architecture to competitors and customers and partners alike, changing how we engage as an open manufacturing engine for the industry as well. You know, we announced last week our billion-dollar incubator program and how we're investing in and creating that fervent, vibrant, bubbling cauldron of innovation through our incubator program as well. You know, and I'm just personally proud of Intel's role in shaping and bringing forward these foundational, fundamental technologies that shape industries, and increasingly that's built on a common end-to-end software platform that Greg described, our oneAPI approach. Our software assets are one of the underappreciated competitive advantages that we've had.
You know, we have these four decades of, you know, hardening of these software technologies and, you know, having run a software company, and sometimes I wonder why did God take me on an eight-year journey to come back to Silicon? It was learning how to become a software executive at scale. Software is like a fine wine, it just takes time to harden those components. We're creating these core platform capabilities and then driving them into the open source community efforts. We're the leading provider of Linux and Linux OS technology. You know, in fact, with the 19,000 software engineers we have, I have more software engineers working for me today than when I was running a software company. It really is quite spectacular. Software is not just one thing. Developers are everywhere.
You know, for us, software is key to exposing our hero silicon features, optimizing the software to run best on Intel, but pulling that platform value up the stack. You know, our software strategy, if you describe it in just two words, is move up. What Greg described is how we're moving up the software stack. Today, you know, about $100 million or so from software licensing subscription. In the future, Greg's agenda, a lot more. You know, the value of software, of exposing our platform, of creating and enabling an industry of developers is so critical to the future, and why I was so excited to have a leader like Greg join us here on this journey, making software a competitive differentiator for Intel on this journey.
At our innovation event in October, we laid out this three-part strategy that says we're gonna be open, we're gonna deliver choice, and we're gonna create trust in the industry. Hugely positive response. I wrote an open letter to the ecosystem in October, and it's just been overwhelmingly responded to by developers. You know, their view of us is, wow, you know, they need choice. You know, they want an ecosystem champion. They see us as a company that shapes technology as a force for good as well. They need trust in this digital era. You know, we're investing in the silicon and the platform features to create these secure platform capabilities in what we do with the edge, and the server, and the client, and the cloud, and our HPC, our auto platforms.
You know, Sandra will talk about that a little bit more when she comes up today. Since our innovation event that we had in October, the geek is back and developers have responded and say, "We are thrilled that Intel is back driving these ecosystems and software capabilities." Now, today you've had the opportunity to meet many of my leadership team here. You know, we've brought in key members from the industry to help us drive fresh perspectives with Nick, Sunil, Christy, Greg, most recently David, and we're about to announce our new commercial officer as well. You know, fresh talent coming from the industry, but also building on, you know, extraordinary and long-tenured strength inside of the company, Sandra, Anne, Michelle, Keyvan, Raja, Randhir, and Amnon. This is my team. This is the best team in the industry.
We have the right people who are experts in their domain, who are committed to the mission that we are on, and we're unleashing them with the right strategy. It's not just me, it's not just the leadership team. It's the 121,000 brilliant, dedicated individuals working together in service of our customers, the mission that we are on. You know, there's this, you know, there's this depth of passion around what Intel does. When I came back to the company, we were losing talent, you know, and many of you were writing on that, you know, the brain drain. Well, that changed, and we now have brains coming back. We've hired 17,000 technical employees in 2021, many joining from key competitors saying, "Wow, that's pretty cool.
Let me go join them, what they're working on. Many of them are returnees who are coming back and saying, "Ah, Intel is back, and I wanna be part of that." The band is coming back together and the mojo is back. You know what? Intel is known for this powerful culture of discipline, innovation, and execution. You know, and I call it the Grovian culture that we want. We recently refreshed our values. We've rolled them out across the organization. We've rebuilt our decision-making processes. We brought back OKRs, objectives and key results. Why'd we stop doing that? We invented it. Everybody else in the valley embraced it, and we stopped doing it, so we're doing it again. We've tied every person in the company, you know, and part of their financial rewards is based on their OKRs and their individual execution against them.
The culture is being rebuilt. The evidence is building. We are well underway. I've shared our beliefs, our strategy, and now to the third chapter. How we've re-architected the company for growth, how we brought these six distinct but reinforcing and powerful business units together, how they're gonna enable us to grow in the large traditional markets, but also disrupt and enter new emerging markets as well. You know, and these emerging new markets for us, these are areas that build on technologies that we've been working on and are gonna enable us to be a force. We're gonna build on our core competencies, and they're gonna give us new opportunities to grow this company, and we're gonna make it all transparent to you.
We're making these reportable segments, so you get to measure them just like I do as we're building our momentum in these growth areas against these six business units. We said we're gonna reunite and grow the company. We're gonna accelerate our revenue growth. You know, we said we're gonna get to the low double digits, and some of you said, "Huh?" Well, how are we gonna do this? Let's just walk through that and why I'm so confident for this. You know, we see this year's growth to be moderate. You know, the impact of supply chain challenges, the emerging businesses are launching but aren't yet significant. 2023 and 2024, we expect them to start to gain more momentum. We're getting our new capacity in place, and we're gonna see traditional and some growth from our emerging businesses.
By the end of the period, 2025 and 2026, 10%-12% growth rate as we're getting the full benefit of the investments, as well as the size and impact of our emerging businesses as well. These markets are large, and as we have leadership products built on leadership process with at scale manufacturing, we believe we can not only grow, but gaining market share and ASP. Let's walk through each of the six business units now to give you some view of them. We believe none of these assumptions are heroic. You can be confident I'm gonna be driving my team to well exceed the numbers that we're about to show you. Let's start with Client. Last month in our earnings call, you know, our Client business reported a record year in 2021.
You know, the PC has become more essential to everyday human life than ever before. This new normal of hybrid work and education. You know, year of 15% growth in the client business. This market is now structurally larger, and we see this 1 million units per day kind of number as a sustainable one for the future. We share this with our ecosystem partners and Microsoft that, yep, this is just a bigger, more important industry. Looking forward, we don't expect that kind of growth to continue. Market growth may be low single digits, but we see that we have opportunities to grow from other factors as well. A stronger and stronger roadmap with Tiger Lake and Alder Lake. We saw in Q4 that we're already regaining market share.
This continues with Raptor Lake, Meteor Lake, Arrow Lake, leadership products that allow us to have more ASP leadership in the marketplace. We've also invested in reviving the ecosystem. You know, we have to bring excitement back to the PC, the best experience, and with Microsoft and Lenovo and Dell and Samsung and HP rebuilding that vibrant PC ecosystem. We also see that there's key next generation use cases like the immersive experience and metaverse that are gonna keep sizzle and excitement building in the PC industry of the future. You know, we're also expanding our share of the bill of materials and the PC BOM with Wi-Fi, graphics, 5G, I/O capabilities as we grow our footprint. A larger TAM, leadership products, innovations in the platform.
As we've done the last six years, we believe we can continue to grow this business over the horizon in the low to mid-single digits. Nothing heroic. Continuing growth in our business, and MJ and JJ will describe this more in a little bit. Next up, maybe the most important for many of your eyeballs is data center and AI. This business for us is our core Xeon server with Habana and FPGA. Record unit volumes in 2021. In fact, in December, in one month, we shipped more server products than our nearest competitor did for the whole year. The scale of what we do is simply dramatic. We remain the supplier of choice for the data center and the cloud providers. The market remains robust, driven by an enterprise refresh and the continued growth of cloud, but they're also looking for power, performance, security, privacy.
Workloads continue to be deployed both on-prem as well as in the cloud. You know, today we're laying out our new server roadmap. Sapphire Rapids, you know, the leadership platform, PCIe, DDR, CXL. Initial shipments beginning this quarter. Compelling features in areas like security and dramatic, up to 30x improvements in AI. Now, as we were looking and engaging with our customers clearly said, "You know, we really want a different cadence to how you deliver products. You know, we want a yearly cadence to products, and we want our platforms to be every two years." We listened, and we've reworked our roadmap to deliver this two-year platform cadence with one-year product cadence. Into the Sapphire Rapids platform. Next year we'll deliver an upgrade called Emerald Rapids. Socket compatible, but improving performance, features, core count, new capabilities into that same platform.
Then in 2024, this is the big deal. We're doing several major things. We're first bringing out our next generation platform, a major leap in the platform capabilities, but we're also doing exactly what we just did with Alder Lake. We have P-core and E-core, a performance, you know, core, as well as an efficient core. Alder Lake, part of the reason it's such a great product is leveraging that hybrid architecture. Well, in the server, you're at a different scale, so we're creating a dedicated efficient core line and a dedicated performance core line in a common platform. As we looked at that cadence and the great health that we were seeing from Intel 3, we said we're gonna accelerate our embrace of Intel 3.
This dual track acceleration of Intel 3 gives us a very competitive product line in 2024 with Granite Rapids, Sierra Forest. This capability allows us to meet the market needs more effectively as well as deliver unquestioned leadership product performance. A strong roadmap aligned to segments and meeting our customer cadence. You know, this is about real-world workloads. You know, not a variety of synthetic benchmarks. Sometimes, you know, people give me numbers with some benchmarks that I created 30 years ago. It's like, that's a bunch of crap. That's not real-world workloads today, you know? We see these real-world workloads where Intel just shines, and this roadmap delivers against them. We see our business accelerating up to the mid-teens%. We see this 2024 as an inflection point in our competitiveness.
Right after me will be Sandra, and she'll go into this in a little bit more detail. We and our customers are excited about this new roadmap. Next up, networking and edge. This is our former IOTG business and some of the data center networking and some of the dedicated Xeon use cases in communications. You know, I specifically architected NEX to focus and capture the transformations that are happening both in the network and at the edge. I was thrilled when Nick agreed to become the leader. Somebody who had been basically architecting many of the core network transitions for decades, now coming in to lead doing this. It somewhat felt like the Andy Grove when I went to leave to go get my PhD.
He said, "You can learn in the simulator or you can stay here and fly the jet," when I started running the 4G. Nick was learning on the simulator or teaching people how to run the simulator. Now he's running this business for us. The perfect person to lead and capture two seismic shifts, the proliferation of programmable software-based infrastructure and the arrival of the intelligent edge. That's exactly what we're laying out. You know, we've seen this occur in the data center as we move from, you know, minis and mainframes to open systems. That's exactly what happened in the telco network, and they virtualized the core network. Now that's moving to the edge as well. This is where the Open RAN is now replacing fixed function platforms at the edge.
You know, open standardized platforms that are enabling software and 5G and workloads like AI at the edge. We are incredibly well positioned. We've architected our product line to capture these shifts and play to these strengths in the marketplace. We're also driving the rearchitecture of the data center network as well with our IPU product line, creating capabilities that gives superior security, infrastructure tasks moves off of the CPU onto the IPU and enabling at scale cloud environments to be more efficient and get more capacity from the CPU capabilities, you know, driving this whole shift with FlexRAN and Open RAN. The edge, as I've said, is this next architectural challenge where the pendulum of compute has been moving between centralized and decentralized.
We're uniquely positioned to capture that swing to edge-based computing with a multigenerational software-compatible platform with the standards like OpenVINO already being well accepted in the industry. We see this business having sustainable growth in the mid-teens%, growing faster than the market over the five-year horizon that we're laying out. A great unique business for us. If we take those three businesses together, what you see is that, you know, this represents about half the growth of Intel as we look to the future. You know, our position in client, position in data center, and the strength in our networking and in Edge business. We combine that with our capabilities in software and manufacturing and developers, a great solid business. We're not done there. How do we then go execute on these new growth opportunities?
I am so excited about these new areas for Intel to reach into aggressively. First up is our accelerated computing and graphics business. You know, this area has been an area that, you know, for many years has been growing, but we've never stepped into it, and today is a great opportunity. We're seeing insatiable demand from customers, hungry ecosystem and OEMs. You know, I'm now getting open letters to the CEO of Intel as the only hope to fix the gaming community in the marketplace. Well, that's pretty exciting. You know, those are exciting customers. I like getting those kind of mails. Some of them that I get aren't quite that good. Today, we are jumping in with both feet. We can disrupt this market.
You know, we're leveraging our existing IP and software, and in many respects we have like 70% of the software work done because we're the integrated graphics leader by far. You know, it's like starting the giant slalom halfway down the hill, right? Our competitors. It's just not fair. Well, business ain't fair sometimes, 'cause we have a huge set of assets in our integrated graphics, our software capabilities that we get to build upon. We're creating new capabilities in the platform called Deep Link to combine our chipsets and CPUs with our discrete products to create unique and differentiated advantages for our customers, and the response has been nothing short of extraordinary. We have 50 new mobile and desktop designs. Alchemist is shipping to customers now. You know, and we're gonna see the scale from mainstream to performance sector segments.
You know, we're in these financials, we'll ship 4 million units this year and ramping rapidly as we go through the year. We've also architected this business, as I call it, no wafer left behind, or actually Raja coined that, you know, because we're dual designing these products to be able to leverage internal and external capacity, 'cause we will take advantage of capacity capabilities everywhere we can find them, given this insatiable demand. We're also out of this business in HPC and a high-end AI. You know, in HPC, you know, I've been in the industry 40 years. You know, it's always been sort of this niche, funny segment on the side. Well, what's happening is HPC and AI are colliding, going from niche weather and nuclear modeling to mainstream applications as well.
Again, we are well positioned to capture the demand for this with our Core Xeon platforms, but also with exciting, new, compelling tour de force products like Ponte Vecchio. You know, we're also taking Xeon and extending it into the space with our HBM product line, high bandwidth memory being added to Xeon cores, a dramatic step up in performance in a software-compatible way. The developer community is fired up, tremendous momentum as they get access to Xeon HBM and Ponte Vecchio. The roadmap is robust, and Raja and I are both thrilled as we look out to the 2024 Falcon Shores, where we fully bring GPU and CPU into a single socket with coherent views of memory to redefine a computing model for the future. The roadmap is powerful. The market is exciting. Key addition to build on our business today.
We're doing even more. We're also launching our custom compute business, where we're taking these compute-intensive capabilities to create specific product lines where high-performance computing needs. The first of these is efficient blockchain, and we're entering that market where there's compelling opportunities to improve power performance and strong industry demand. We're gonna attack new segments like auto infotainment, automotive systems, immersive displays and more. In 2022, we expect this to be a billion-dollar business and growing rapidly over the horizon to be almost $10 billion by 2026. An exciting new business area for us. Next up is Mobileye, unquestioned leader in ADAS, and we're building on that into the AV, the autonomous vehicle segment. Last year, 40% year-over-year growth in this business. Our ADAS is now deployed in more than 100 million vehicles.
Those vehicles are now 13 of the 15 top automotive companies. You know, if you were in Amnon's session today, you know, the strategy that we're laying out is comprehensive for the entire ADAS to complete AV sector. We're launching fully capable hybrid systems of both vision and radar LiDAR. The car out here, if we turned off all the radar LiDAR, it's a level four car with just vision. If we turned off all the vision and just used radar LiDAR, it's a full level four car being able to drive itself. We fuse those together for increased reliability and accuracy. It also scales all the way to the lowest end vehicles as well, a comprehensive product line for AV. Given that installed base of ADAS, we have crowdsourced dynamic real-time mapping.
We've collected 2.5 billion road miles, which are being navigated and updated in real-time, real-world conditions, another major competitive advantage. We've also driven the standards in areas like RSS, Responsibility-Sensitive Safety model that allows autonomous vehicles to enter society with confidence of improving over road conditions and road drivers of today. You know, a roadmap that's very compelling, purpose-built silicon with our EyeQ family, 176 TOPS, specifically optimized for autonomous vehicles and ADAS. Overall, an incredible business opportunity for us. The way we've architected this business, it's both consumer AV as well as the full robotaxi market. You can think of this as Tesla and Waymo, and our product line supports all of those use cases. Now, obviously, you've heard us describe our plans to IPO this business. That's well underway.
Because of that, I can't say a whole lot more about the financial characteristics of this business today. This is an exciting way for us to unlock shareholder value for the future. Now, IFS. In March, we said we're opening the doors of the Intel manufacturing network wide. We're gonna become a foundry for the industry. Since then, we've seen a strong pipeline of customers, all sizes across all sorts of market segments. You know, today, we're running test chips for customers, about 30 different test chips that we have underway for this year. You know, Intel 16 is well underway with customer interest. You know, we won the RAMP-C contract to build a U.S. government-trusted foundry service as well, which we're getting very positive responses from our customers in the government.
You know, the revenue today for this existing business and packaging Intel 16 just ramping. You know, we just announced our automotive foundry offerings as well. Stepping into the automotive sector. You know, we're excited about that today. The high-end car is 4% silicon. By 2030, 20% silicon, a 5x increase in that segment. We're launching our automotive foundry offerings. We also announced the IFS Accelerator, right? And we now have 16 partners, a billion-dollar partnership with our Intel Capital, and we expect steady growth as a result in these offerings. As we start introducing exciting capabilities like Intel 3 and Intel 18A, unquestioned leadership and excitement from leading-edge, high-volume capacity customers that we believe as we get to the end of this period will drive rapid acceleration in the revenue opportunities for the IFS business.
We expect that we'll have some of those first major foundational customer announcements later this year. We're doing this through the lens of smart capital. Dave, when he gets up in a little bit, will describe how we're sort of managing that we get efficient capital and ensuring the customers the capacity corridors that they need, and how we manage making sure that we do both of those, efficient use of capital while ensuring customers the capacity corridors that they need. You know, IFS has a range of compelling value propositions that process, you know, world-class packaging technologies. We've also created the broadest set of IP of any foundry in the industry with Arm, with RISC-V, and with x86. Of course, at scale manufacturing across geos as well. Tuesday was a pretty exciting day.
On Tuesday, we announced that we're acquiring Tower Semiconductor, and we're pleased to have Russell Ellwanger. Right. Russell, if you might stand up, the CEO of Tower Semiconductor with us today. You know, I'm so excited about this because we simply need more foundry DNA. We also need more foundry offerings on our menu. We're expanding the portfolio of services and capabilities. Tower Semiconductor has been the leader in, you know, what's called specialty foundry offerings for things like RF and, you know, displays and power devices. We've seen again, those are some of the most acute areas of shortage today in the industry. He will also help us in building, you know, a broader set of not only diverse offerings, but of extending our geographic reach with his footprints in, you know, U.S., Europe, and Japan, and Israel as well.
You know, a highly complementary transaction, EPS accretive, immediately helping and strengthening us in mobile, automotive, power, computing sectors, and accelerating our plan to be one of the largest foundries in the world this decade. This just makes us even more confident in our ambition. One of the things about, you know, and some of you have challenged me and said, you know, "So why are you getting into the foundry business? Help me understand this." Well, IDM makes IFS better. IFS makes IDM better. When you think about that, my IFS business gets all of those TD investments, those manufacturing networks, the IP blocks that we're creating for free. $Billions of R&D is being made immediately available to the Intel Foundry Services. Wow, that is leverage. IFS makes IDM better as well.
Our foundry drives us to create more robust PDKs, better EDA tool engagements, richer IP block availability from the industry, and the foundry customers are benchmarking us every single day. I ask my teams, "Are we being competitive?" You know, is Intel being competitive with the technology? Right. I ask my internal teams, but I also now ask our external foundry customers as well. We are benchmarking ourselves every day to guarantee we have the best manufacturing, the best cost, the best transistors, the best, you know, PDKs available in the industry every single day. IDM makes IFS better. IFS makes IDM better. It also gives us a powerful tool to better utilize our factory network as well.
You know, traditionally, we've ramped a factory, and we go through this extraordinary ramp it hard, get it yielding, get it all good, and just about the time the factories are running really good, we screw it up by rolling over that equipment to the next process node, right? Roll it over super fast. We create this natural inefficiency, right, as we move from node to node. You know, now with Foundry, we take those factories that we struggle to ramp, and then we continue running them much longer. You know, it's the gravy train of the rich cash flows of mature technologies over time. You know, I've called it the bug in the Intel business model, and IFS fixes that bug as we go to the future. More efficient use of capital and more leveraged use of depreciated assets than ever before.
A beautiful complement to the Intel business model of the future. To me, this is, you know, one of the exciting things. It isn't just that it makes it better technologically, it makes the Intel business model better, more resilient as well. Looking across our businesses, we see this robust path to double-digit growth. Our client business, steady, solid growth as we've done for six-plus years, larger market, and the products are getting better. DCAI, steadily improving our competitiveness, and the new roadmap, 2024 and beyond, gets really good. NEX, comprehensive portfolio of capabilities and a growing market uniquely positioned growing faster than that market. Our emerging businesses in graphics, accelerated computing, auto and mobility, and our foundry offerings, these are large, growing markets that we believe allow us to achieve, nay, overachieve. You know, I'm a meet, beat, raise kind of guy.
We've built resilience into these numbers. We don't need all of these to hit to be successful. That's the power of the portfolio, multiple engines of growth that we have. If you do the math of those six business units, it comfortably exceeds the 10%-12%. You know, we're confident that we have de-risked that portfolio that we have laid out today. You know, not all of the plans. You know, every strategy is perfect until you hit the field of combat. You know, you need some resilience and management of how it is. This is an unprecedented opportunity, and I feel confident in our plan to accelerate and deliver against that double-digit growth that we have described. We've laid out an investment strategy to get us there. You know, this is a little bit about how we're thinking about that.
You know, our IDM 1.0 model requires us to invest more in capital. We have underinvested, we have critical needs, and we said we need $25billion-$28 billion this year into that business model. Part of that is driven by what we'll just say catching up. We do not have enough capacity. We have backlog of demand. Build more capacity to satisfy the demand. Part of this is investing in our TD for the future as we've done with angstrom. Finally, that we need to create some shell space. We need some optionality out in the future. You know, in the smart CapEx, this shell space is long lead time at small capital investments. As you see, this is leverage between our IDM and our IFS business, substantial leve.age in these capital investments as well.
Our smart CapEx then says we're gonna use capital offsets, we're gonna use customer prepays, we're gonna use exciting new announcements like today we're announcing with Brookfield to create more capital flexibility for us in the future. We believe our base plan that we're laying out to you is very conservative in this regard, and minimum amounts of capital offsets are built in. We have an opportunity to get much, much more, and we're targeting much higher levels to further share the risk with our customers, with our geopolitical, partners, with our investment partners, giving us better margins as well as better free cash flow opportunities than what Dave is describing to you today. We do expect that this capital intensity peaks over the next couple of years and then drops off into that 25% of revenue in the subsequent years.
David Zinsner will tease this apart a lot more in our conversation today. Bringing it all back together, we've talked about the industry trends, we've talked about our beliefs that underpin our strategy, and we've given you the business case across these six business areas. We've embarked on an ambitious five-year plan that creates significant revenue growth and, we believe, shareholder value over time. Some of you have said, "How do I know that you're executing on this? Help, help build my confidence." You know, even our greatest skeptics are now saying it's the right strategy, but how am I gonna know you're executing? Well, throughout the day, we're giving you lots of proof points.
You know, these are the ones that we've laid out for our leadership products this year, tangible milestones that you're gonna be able to measure us against this year, and obviously products like Alder Lake, check, Intel seven, check, done, and a rich roadmap into the future. You know, we're building our core strengths, you know, expanding our capital network, you know, driving the acceleration of the build-out. We're ahead of schedule on our factory projects, you know, that we have. We're, you know, giving you proof points on Intel 4, Intel 3, Intel 18A, Intel 20A, and we're also re-architecting the company for growth. You know, creating new capabilities like Tower, IPO-ing Mobileye, creating, you know, foundry customer proof points as well into the future.
You know, we are just being innovative in every aspect of how we use capital, how we drive our products, how we build on our core strengths, and we're gonna give you unprecedented transparency. Starting in April, we'll give you the visibility into the six business units, and you'll see how all of them are doing and how we're executing against what we told you we would do. 2023, 2024, so much more. This business model, these investments really start to accelerate as we get into 2023 and 2024. Here we are on my one-year anniversary. I'm pleased with the progress, but we got a lot to do. We're indexed to high-growth markets, but we got a lot to do. We have deep, sustainable competitive advantages that we're building on.
We've laid out a robust strategy across our six business units, and we have extraordinary talent as part of this company. You know, we've given you creative ways of how we're going to unlock shareholder value. We believe this is creating a compelling picture for shareholders to look at us. My goal is that we double. I wanna double the earnings of this company and double the multiple of this company as you build confidence in what we're doing, a 4x increase in total shareholder value. The Intel turnaround train is leaving the station, and I hope you all get on board. It's an ambitious goal, but I am confident Intel's best days are in front of us. Thank you very much.
All right. Thank you, Pat. At this time, please join me in welcoming to the stage the Executive Vice President and General Manager of our Data Center and AI Group, Sandra Rivera.
Good afternoon. I'm Sandra Rivera, and I lead Intel's Data Center and AI organization. I know a number of you from the years when I was building and growing the network business. Networking today is a multi-billion dollar business for Intel and continuing to grow under Nick's strong leadership. Over the past eight months since I've taken this role, I've spent a lot of time with our customers, understanding their challenges, their growth plans, and their market opportunities. I've also spent time examining our own challenges so that we can be better partners to our customers. This afternoon, I'm excited to tell you about the ways that we're focusing our business to continue to grow the data center and AI business for Intel. To sum it up, my team and I are focused on three core pillars to accelerate our growth.
First, to maintain and grow our leadership position with hardware and software solutions that our customers need when they need them. Leadership starts with Xeon, and today, Xeon powers more than 85% of the world's data center infrastructure. Customers like Amazon, Microsoft, and Google continue to choose Intel Xeon for the differentiated performance and features that it offers to customers. We fortify our Intel Xeon franchise with a leadership portfolio of products that include FPGAs, AI accelerators, GPUs, as well as a rich software suite that can deliver to our customers a powerful platform for them to build on. Secondly, we will harness the ubiquity of Xeon and the x86 ecosystem to foster an open and vibrant environment for developers to innovate on.
The ecosystem of OEMs and ODMs, commercial software vendors, and platform solution providers built on and optimized for Xeon lower our customers' development costs, accelerate time to market, and improve overall total cost of operations. Our ecosystem and the decades of work by our partners to optimize software for Intel architecture is one of the most valued and differentiated capabilities we bring to customers. Third, we will leverage our unmatched strengths in AI and security to make pervasive across all of our products and all of the market segments. AI and security are two of the fastest-growing workloads, and we're going to infuse it into every aspect of technology. Our massive installed base with our built-in features of integrated AI and security put us in an ideal position to leverage these fast growth areas.
I'm encouraged about the future, and I'm confident in the assets we have to accelerate our growth. Now let me drill down into the data center and AI's silicon TAM opportunity. This is a fast-growing market over the next five-year horizons expected to grow in the mid-teens from approximately $30 billion in 2021 to over $65 billion by 2026. You heard Pat talk about being at the forefront of an era of sustained growth driven by the digitization of everything. We have three macro data center trends that add fuel and act as tailwinds to that growth. First is the explosion of data and the insatiable demand for compute in everything from client PCs to hyperscale data centers and everything in between. IDC projects that by 2025, the world will create and replicate over 180 zettabytes of data.
To put that into perspective, in 2021, the world created and replicated 80 zettabytes of data. You can see that we need to almost or more than double the compute capacity to be able to process all of that data. It's not just the amount of data that needs to be processed. It's also the different data types. From structured to unstructured, dense to sparse data, computer architectures need to evolve to keep pace with developer demands. Our customers are hungry for solutions that deliver higher performance in their real-world workloads. Developers want a consistent software stack that allows them to move across heterogeneous architectures that are required for specialized workloads like AI, graphics, and video. Last, the demand for AI and security is relentless. AI is required to intelligently and efficiently process all of those zettabytes of data.
Security is paramount to protecting the valuable business and personal assets, and particularly in the increasingly complex threat and regulatory environment we operate in. There's a huge market opportunity in front of us, and we're approaching that market from a position of strength and closely working with our customers to capture the opportunity in front of us. Capturing that opportunity requires product leadership, and that starts with Xeon Scalable processors. Last year, we introduced into the market Ice Lake, our third-generation Xeon Scalable processor, and it was met with high demand across all customers and market segments. In fact, in Q4 alone, we shipped more than 1 million Ice Lake CPUs, our fastest Xeon ramping to 1 million units ever. With Ice Lake, we deliver industry-leading capabilities with integrated AI, encryption, security, as well as networking features.
Ice Lake is available from all major OEMs, ODMs, and cloud service providers. In fact, AWS has introduced instances based on Ice Lake with their C6i and M6i instances that deliver more than 15% gen on gen performance improvement and twice the networking speed. Google's Compute Engine VMs also deliver greater than 30% gen on gen performance by incorporating the integrated acceleration and crypto features in Ice Lake. Next month, we will begin shipping initial SKUs of Sapphire Rapids to target customers. Sapphire raises the bar and sets a new standard in the industry for workload-optimized performance. For example, in the image recognition system market, we see customers will be able to experience a more than 6x increase in their ability to process images per second using the acceleration engines built into Sapphire Rapids.
Sapphire Rapids will also lead the industry in important memory and interconnect standards. For example, PCIe, DDR5, as well as the new high-speed cache coherent interconnect CXL, a standard that Intel led on in the industry. There's strong demand for Sapphire Rapids from every OEM to ODM, every cloud service provider, and every communications service provider are designing platforms on Sapphire Rapids. We've also secured high-profile wins with the Department of Energy, who are going to select and deploy Sapphire Rapids in their next-generation supercomputers, including the Aurora exascale system and the Crossroads supercomputer that will be available in the market in the coming years. Now I'd like to show you the type of AI performance that Sapphire Rapids can deliver as compared to a discrete GPU in a single-cell genomics use case, which is an important area for medical research.
Devon, why don't you show us what we've got here on the platform?
We're in the era of digital biology, where AI, HPC, and natural science are all coming together to kind of see end-to-end workflows, both on-prem and in the cloud. Our competition showed an analysis of 1.3 million cells on their RAPIDS platform, which is a library suite targeted at accelerating performance for data prep and machine learning on GPU platforms. On an A100 platform or an A100 GPU, this analysis took roughly 11 minutes, and this is presumably thirty times faster than on a CPU. What we did-
If you can wake up the system.
Yep, I think my mouse died. If someone backstage can click again.
This is what happens with live demo.
There we go.
There we go.
We ran the same simulation or analysis on software that's been optimized for CPU, and what we found is that on a single Ice Lake server processor, the same analysis completes in...
Nope.
This is the old one.
It got stuck.
Well, on CPU-optimized software for CPU on a single Ice Lake server processor, this completes in 489 seconds compared to the 686 that it takes on an A100, which is roughly 30% faster.
Yep
... on CPU. That performance only gets better on-
Sapphire
Sapphire Rapids. A Sapphire Rapids platform, a single CPU completes that analysis in 370 seconds.
Oh, here we go.
Yep, here we go. Here we'll see it. This will, on your right, go up to 489 seconds on Ice Lake. Or no, this is Sapphire Rapids again. Well, that makes it-
I think
That makes it roughly half the time of.
Yeah
an A100.
Yeah. I think the point is that we're delivering very performing AI workloads on a CPU that is very competitive with a discrete GPU. So.
Exactly.
Thank you for that, Devon. As Devon was trying to show, our integrated AI accelerators are able to deliver almost twice the end-to-end AI performance over a discrete GPU. We're looking forward to seeing all the scientific breakthroughs that Sapphire will be able to deliver with the integrated acceleration engines inside it. What we are offering to our customers is the ability to have AI in their Xeon that they already have, so that they don't need an additional expense and programming environment for a GPU. Now on to our roadmap. Today, I'm excited to announce that we are boldly expanding the Xeon roadmap. Our roadmap will put us in an even stronger leadership position by taking advantage of both our performance or P-cores as well as our efficient or E-cores.
Our roadmap is now better aligned to customers' needs. Our next generation product on the Rapids swim lane of our roadmap or our performance swim lane is going to be Emerald Rapids, which will be a socket-compatible CPU that will extend the benefits of Sapphire Rapids in higher performance, memory, security, and I/O capabilities. By being in the same platform, we're able to provide our customers an easy upgrade path. This next processor after Emerald Rapids to the market. Our efficient cores or E-cores are enhanced for the data center by delivering a higher performance efficiency throughput for low latency or high throughput applications. Xeon processors using our E-cores are built to support cloud-native applications. Sierra Forest will be our first E-core product or E-core CPU, which will be manufactured on the Intel 3 process.
It was designed with hyperscalers to ensure that we're further optimizing our TCO. Our next P-core product following Emerald Rapids will be Granite Rapids. Granite Rapids will also be manufactured on the Intel 3 process, taking advantage of the higher performing denser library. Granite Rapids and Sierra Forest will both be delivered in our next generation high-performing platform which share the same base architecture, making it easy for customers to have portability between the two platforms for improved ROI and making it easier for us to deploy with lower validation costs and accelerate both ours and our customers' time to market. With our new roadmap, we will extend our leadership in the data center with the fastest growing workloads, which include microservices, edge computing, networking, and storage. Now let me talk about our ecosystem.
Intel has the world's largest data center ecosystem, and our ecosystem helps customers improve their TCO, accelerate their time to market, and lower their development costs. There are more than 700 commercial software products developing and optimizing for Intel architecture, including Microsoft, Oracle, Red Hat, SAP, and VMware, and hundreds of open source communities that are writing code first and best on Intel. In fact, there are over 100 billion lines of code that have been optimized for Xeon and Intel architecture. There are over 100 million Xeons deployed in the network, in the infrastructure, today. This spans from on-prem servers running IT services to networking equipment managing internet traffic to cloud services running data analytics. Today, there are more than 58,000 cloud instances based on Xeon in the world.
The ubiquity of Intel architecture puts our differentiated features at the fingertips of almost every developer in the world, and we will continue to invest in and collaborate with our partners to maximize productivity and deliver better value with Intel architecture. Now let's talk about solutions. The solutions we build with our partners are a key differentiator for Intel. Intel-branded solutions leverage our broad portfolio of an expansive ecosystem and our in-house technical expertise to give customers a differentiated and competitive advantage. There are more than 500 industry partners designing, building, and delivering solutions to better serve customers with over 700 workload-optimized solutions in the market today, spanning everything from edge to cloud.
IDC recently conducted a study of customers using Intel-based optimized solutions, and among the findings were that the customers were experiencing a three-year ROI of more than 465%, and were experiencing application performance improvement of nearly 64% on average, and a faster time to deployment of almost 60% on average. Now turning to workloads. AI is the fastest-growing workload, and our AI strategy follows the same approach we've taken with other large technology transitions in the market. That is, we lower the barriers to entry, we increase the market participation, and we accelerate the rate of innovation and time to deployment. We have built-in AI into our platforms so that we can make it more available to customers.
We have a software stack that we upstream into the most popular industry libraries as well as the most popular frameworks, including TensorFlow and PyTorch. Let me walk you through the AI data flow and show you how the majority of AI data flow pipeline runs on Intel today. The first part of the AI data flow is the data prep stage, and this is where data is loaded, cleaned, and prepared before it's moved to the model training phase. This stage is particularly important, especially when data is unstructured. For example, when building a facial recognition model, you don't want the system to be focused on non-facial elements of the image. The next phase is the model training phase, and this is a mix of classical machine learning as well as small to medium complexity deep learning models.
Most of these models also run on Xeon CPUs. Deep learning today runs mostly on GPUs, and these would be the high-complexity models because deep learning training or high-complexity deep learning training models is a massively parallel workload, and therefore, they typically run on GPUs. But the most popular GPU runs in a closed environment. Our approach is different. Our approach is an open software approach, giving customers choice and performance in running their deep learning training models on a Xeon CPU. Lastly, we look at the inference and deployment phase, which is the largest growth market in the AI data flow. Today, over 70% of AI inference runs on Xeon.
Xeon is a rapidly growing market because it's really being fueled by the build-out of edge computing, a place where Intel has a strong leadership position with Xeon. As you can see, the majority of the AI data flow runs on Xeon, and we forecast that Intel's AI logic silicon TAM will be over $40 billion by 2026, with AI inference growing about 25% CAGR during that same time. We will capture this market through the AI, an AI software stack based on open standards and a broad portfolio of products, including CPUs, GPUs, accelerators, all with optimized software. Our approach is a winning formula to propel the next era of AI innovations. Now on to security, another high-growth workload. With Xeon, we protect data at rest, in flight, and in use.
Earlier today, Rick Echevarria talked about how we are accelerating security across all segments with unique hardware features that we are building into our products. Security rooted in hardware provides the best protection for tenant data. Our Ice Lake with integrated crypto accelerators increase performance for data at rest by 3.9x over previous generation, and that can protect exabytes of stored data. Tencent secured data in flight by over 5x in their networking platforms by using the integrated crypto acceleration features in Ice Lake. For data in use, we fortify Xeon with an SGX feature, which is a secure enclave that we create in the processor that isolates tenant data from other data, making sure that we're protecting their data in a multi-tenant environment.
SGX has fueled a new capability and service that's being offered by the world's largest cloud providers called confidential computing. We see Microsoft Azure, IBM Cloud, Alibaba Cloud, Tencent, Baidu, and OVHcloud all using SGX to deliver confidential computing to their customers. Our security leadership is based on integrating these features into our Xeon platform and also offering it ubiquitously over that Xeon portfolio with software ecosystem from our broad customers and partners. It's an exciting time to be at Intel, and I'm thrilled to be leading the data center and AI organization. I'm encouraged by the progress we've made over these past eight months with our unflinching focus on execution excellence and roadmap leadership.
We will be accelerating our growth from the mid- to high-single digits in revenue over the near term, and we expect to be growing revenue in the mid-teens in the longer term. Our path to growth is anchored on three pillars. First is to leverage our extensive leadership portfolio and deploy products at massive scale. By executing on our new roadmap, we will lead both in per core performance and also high throughput density performance. We will grow in our traditional businesses and accelerate entry into new ones. We can support customers at a scale that is unmatched in the industry. In December alone, we shipped more CPU processors to our customers than any single competitor shipped in all of 2021.
We were also able to meet key customer demands for hundreds of thousands more processors well outside of normal lead times by leveraging our IDM advantage. Second, we will expand our strength in hardware, software, and solutions and unleash the power of the ecosystem and spur innovation. Our extensive work with software leaders and open source communities allows our customers to tap into the differentiated features we build into our platforms and deliver greater business value. Third, we will make AI and security pervasive across our products and all of our segments. Our multi-generational investment and ubiquity of Xeon ensure that AI and security will run on Xeon, and we will continue to lead in the industry's fastest growing workloads. There is a massive opportunity in front of us as the world's data centers evolve to keep pace with changing business models and use cases.
We have an extraordinary set of hardware and software assets, technology, process technology, packaging leadership, people, and capacity to capitalize on this opportunity. We invite you, I invite you to join us on our growth journey. Thank you.
All right. Thank you, Sandra. At this point, please join me in welcoming to the stage Executive Vice President and General Manager of our Client Computing Group, Michelle Johnston.
Thanks, Tony. I'm excited to be here today, and I wanted to take a quick moment to introduce myself. I've had the privilege and the honor of working and driving the CCG strategy with our customers over the last seven years, and that has allowed me to develop incredible, transparent customer and partner relationships really built on mutual trust and feedback. Now I have this unique opportunity to bring that feedback and that strong connection back to the business unit to really further influence and accelerate the pace of innovation in CCG. As a hybrid worker myself who's always on the road, I'm intimately aware of the essential role the PC plays in all of our lives. In talking with our customers, we steadfastly believe that the PC will continue to rise in importance.
This has and will continue to drive an increased TAM as we see the number of minutes spent on the PC and the number of PCs per household continue to increase. This fuels an insatiable appetite for new and advanced PC experiences as we interact more and more with our PCs every day. People are no longer just using their PCs for work. It's a central hub in our lives, a portal for work, life, and play, making our vision to deliver purposeful of CCG, Jim Johnson. Welcome, Jim.
All right, let's dive into our client business. We're gonna talk about our view of the market and importantly, the decisions we're making to pursue that market, both in roadmap and platform, but also in ecosystem partnership and investments to deliver on experiences so we can fulfill our assignment to deliver market and financial growth to Intel. We're coming on six years of consecutive growth, and in these six years there's been many, many market dynamics. We've disinvested two businesses not meeting our expectation, and one of our customers left to their own SoC. We still grew, and we will grow. How do we grow? On the back of solid execution. In three years, three new core platforms on an annual cadence. Within those platforms, we deliver other product lines, Wi-Fi, Bluetooth.
Not only do they deliver the experience, they're delivering $1.5 billion of revenue at CPU-like margins. Then we bring these together with our partners and deliver platforms and experience. Our view of the market is, of course, it's more robust, it's much larger, and it will continue to grow modestly. That's important because it is driving our investment strategy, both in capacity and R&D. We see the growth sustaining itself because we're now growing off a larger installed base. That one's pretty straightforward. If you look at the increased penetration with acts like the U.S. Emergency Connectivity Fund, we have encouraged funding to a one-to-one PC-to-student ratio in the U.S. Those actions make a difference. They're making a difference also in other parts of the world where we're still at a 10-to-one PC-to-student ratio.
For those of us in our homes, that work in our homes, we still measure the PCs per home, when we believe it's gonna be a PC per person in the home. Let's hit pause and hear from Panos Panay, the Chief Product Officer of Microsoft.
Thank you so much for having me today. I am excited to be part of the investor conference this year. Intel has been the critical part of the recent success of Windows 11. Together, we built this product for the hybrid world we're in from our homes over these screens and on Teams. It's quite amazing. It's probably an understatement to say the last two years have had a profound effect on how each of us lives our lives. A huge part of that effect, of that change, is the role the PC plays across our work, our school, and home. Our lives have been impacted in such a more meaningful way. We've seen behavior shift from the way we communicate, to the way we shop, to the way we create, to the way we watch. Those changes, they're durable.
I mean, they are here to stay. These are a new set of behaviors that are driving a new era of PC use. These behaviors are born out of necessity, but they've evolved into habits. Usage on Windows as a whole has skyrocketed, with people spending more minutes per day on Windows than ever before. We're seeing people move from one family PC in their home to a PC for every person in their home. In fact, 26% of new PCs this year were either first-time buyers, which is just awesome to see, or an additional device. We've entered the next era of the PC right now, and we believe we will see continued growth post the pandemic.
That reliance on the PC for every part of our lives has inspired many people to upgrade to a more modern device and helps them take advantage of the new features and, most importantly, security that comes with Windows 11. Our partnership with Intel on Windows 11, specifically Alder Lake and Intel Bridge Technology, bringing Android apps to Windows customers, has played a massive role in the momentum we're seeing. As a team, we feel an immense gratitude and pride working with you, Intel, to deliver a product that has increasingly become a part of each of our daily lives. Now, as we look to the future, we'll continue to invest with Intel and all our other partners to deliver Windows experiences that enrich and inspire each of our lives. Thank you so much for having me today.
Even pre-COVID, we turned to our PC to contribute, to develop something important to us, and it's as varied as the people in this room or on the net. If it can be as simple as doing your kid's soccer roster and their schedule. It could be pursuing your PhD. It could be that great project at work that you need to get that proposal buy-off. But something's emerged in the last two years that's really important. We use it to connect. Personally, I prefer taking calls, notice I don't say phone calls on my laptop. I can see the face of who I'm talking to or the faces of who I talk to. We can share content. It's actually my phone has become my companion device, and it sits on my desk next to the PC while I'm there.
I'll come back to that a little later. Our imperatives, our strategy is, first and foremost, leadership products that our customers can build their business on. Secondly, investing in and partnering in a reinvigorated open ecosystem. Third, as we now are more remote, our content is getting delivered via the cloud. I get asked at this point quite a bit, what about the metaverse or the immersive web? That is just another example of insatiable compute as we do immersion or virtualization. Those use cases will be developed on a rich client. They'll be consumed in rich clients. We have some of this activity already starting with partners, and I'll share that in a second. Before I go into the products themselves, let me talk about the foundation of our roadmap. Two unique foundations.
First of all, breakthrough leading architectures. The hybrid architecture brought in efficient cores and performance cores to workloads. I'll show you an example. We then get to disaggregate that architecture into tiles, and what that allows us to do is each tile can be architected and designed for the needs of specific segments. We're doing new techniques for even more ultra-low power performance on the next step of the roadmap. If you come down on the page, how you construct it and how you build it is really important. Using unique packaging technology, we can take the hybrid architecture that launched on Alder Lake and will be refreshed with Raptor Lake and now disaggregate the tiles, and now the tile can also match up with the unique process node capabilities, performance leading nodes, I/Os and things like that, stable high volume nodes that you can depend on.
With this, we ask the question: how's it working? Alder Lake, known in the market now as 12th Gen Intel Core, we delivered 60 processors over four months, had 500 designs ready to roll, and that product is ramping incredibly fast. Secondly, when you deliver performance like this, not only do the end users value it, they pay both our customers and us for it. We've been seeing tech press and developers since we started launching the product, and it's just great to see the external feedback validating the architecture and the execution. Let's hit pause for the second time and get a demonstration. Hey, Chuck.
How's it going? The new H series processors have been delivering incredible real world gaming experiences, and we're not talking just one or two games here. Here you can see we're playing the popular game Hitman 3 against the best comp has to available in the market today. When we test this in our lab, we're showing 49% higher frame rates versus comp. We have Riftbreaker showing 27% higher FPS. We have Mount & Blade II: Bannerlord at 23%, and Total War: Three Kingdoms at 47%. As you can see, across the board, we beat all these games versus comp that are in the market today. Now, 12th generation has been a huge performance winner, but we wanna continue that momentum with our next generation product, Raptor Lake, right?
It's been performing so well today. We wanna give you a sneak peek.
The first publicly, by the way.
Yep. Very first one. Now, you can see I have the Raptor Lake system here. It's a desktop system. This system features eight performance cores and 16 efficient cores for a total of 32 threads. You can see that here on performance monitor right here. That is all 32 threads. I've loaded this system up with many of the applications content creators use today. One of those applications is Blender. Let's give those cores something to work on. I'm gonna go in here. I'm gonna use Blender to-
Notice all the boxes are white, showing.
Yep
No demand on the cores.
Yep. We're right now, they're pretty much idle. Let's kick this off. This is a very CPU intensive task. You'll see all 32 threads, they're gonna ramp up to 100%. You can see the
Which is reflected in the blue. That shows the utilization of the CPU.
Yep. They're all ramping up to 100%, and that's fantastic. That means we're gonna finish this task faster than ever before. What if you wanna work on something else at the same time? Well, that's where the magic of Intel Thread Director comes in. Let me show you. I'm gonna take this rendering, and I'm gonna just simply push it to the background. Now, you'll notice the bottom 16 threads, those are the 16 efficient cores, they're gonna stay maxed at 100%. They're working on that render in the background.
All you did to move the workload there is hit minimize.
I hit minimize.
Okay.
One simple click. Now, you'll notice those top 16 threads, those performance cores, they're ramping down to zero. They are now available for me to focus on whatever I wanna focus on, and they're there to deliver the performance I need. We're gonna go into After Effects, also very CPU intensive, and I've got a little video clip here, so let's go into that. Now, you see I've got this clip, and it's the clip I want, but there's a yellow car in the background. That car has to go. Let's minimize this. We're gonna go into After Effects. We're gonna do a content aware fill. Now, I've identified that yellow car. It's gonna go through every frame, see where the car is, determine what the background would be if the car wasn't there, and then get rid of it, essentially erasing it from the video.
That's very CPU intensive. If you look on those top 16 threads, those performance cores, I still have a lot of headroom left over. I could be going Word, Excel, PowerPoint, reading my email, whatever I need to do, but the bottom 16 cores continue to focus on that rendering in the background. Now, you can see it look. I think we're almost done here, so let's go into here. Yep, the yellow car's completely gone. Again, that's the beauty of our brand new hybrid architecture and Thread Director. We take your tasks, we divide them to the right core so you can do whatever you're working on faster and more efficient than ever before.
Thanks, Chuck.
Back to you, Jim.
Thank you. Architecture matters. As you can see, Raptor Lake is the next generation of hybrid architecture. It's gonna deliver up to double digit performance increases, more cores and more threads, enhanced overclocking features in our tools for the geeks that wanna crank it up all the way. We have an AI M.2 module. These are the modules that slip into the motherboard, like, say, Wi-Fi modules do, so people can start programming on AI, and I'll tell you the benefit in a minute. But it's important because this AI capability gets integrated into the package on the next generation. Packaging matters. If I simplified all the products of Alder Lake and Raptor Lake, there's like a package for the performance, high performance, gaming, creator segments, and there's a package for the more mobile oriented segments.
With Alder Lake, because of the supply constraints in the industry, we've second sourced most of the components with our customers on Alder Lake, and if they choose, they can drop in the performance capable products or the mobile sockets right into their designs, same board layout, and importantly, the same bill of material. There's pragmatism in action, and as you saw, it's powered on, it's working well, and it'll be shipping second half of this year. We take that architecture and we disaggregate it, and as I mentioned, we break it into tiles or design it into tiles. The designers wouldn't like me saying it that way. This allows us to do hybrid, but we can adjust how much hybrid are in E-cores and P-cores based on more fine granularity to the segments. We can do lower power capable workloads. We have a next-generation graphics engine.
We call it the tGPU, the tile GPU. We can bring a discrete tile from a discrete roadmap into the package, and we're also integrating AI so we can do tasks that are done on the CPU at about a tenth of the power. Now before I transition to the ecosystem, let's hit pause a third time and hear from John Solomon of Google.
Hi, everyone. This is John Solomon, and I'm the vice president and general manager of the ChromeOS team at Google. I'm excited to be here today, virtually at least, to talk about the strong partnership between Intel and ChromeOS. You know, actually this partnership goes back to the very beginning of the development of Chromebooks here at Google, where Intel was a key partner from the start, and we've really done a lot together. Intel's invested in this ChromeOS ecosystem, both with Google but equally importantly with the OEM ecosystem to help them ship devices, whether it's the most accessible, affordable, sturdy, reliable products for students or highly performant, latest technology devices to run enterprise operations. Intel has consistently invested in the ChromeOS ecosystem, as I said, both with us and with manufacturers.
We're very grateful for that support. It's been great getting us to this point, but we're even more excited about where this partnership can go going forward, and that's what I wanna talk a little bit about today. We're all leaning in to create an elevated experience and a higher performance experience with all of the security and simplicity benefits of ChromeOS. Within Google and ChromeOS, that an open ecosystem is really important, and we share this with Intel. This is really this shared value of openness in the ecosystem is something that really helps power the partnership. I would like to say thank you very much to Intel.
It's been really from an industry perspective, pretty unique partnership, and we're looking forward to continue to build ChromeOS into really the best possible computing experience in a modern cloud-centric world. Thank you.
It's one thing to deliver a product, but if we wanna deliver a leadership experience that meets our needs, we need to invest with our ecosystem, and this ecosystem is the only way to deliver the number of experiences and the quality of the experiences we want to deliver that no single company can. It starts with the OS. You don't get a hybrid architecture working without our Thread Director and their OS, all of these OSes, moving work to where they can be best performed. That's point number one. Secondly, more than just hit minimize, if you work closely with the top software developers in creator, in gaming, they can actually do much, much more in their product and architecture, which is how you get the type of performance that Chuck showed you on the slide.
Hardware side, independent hardware vendors, and these pieces of hardware, like the displays we use, the cameras we look at each other through, the sensors that anticipate our needs, they're hardware and software, so they develop to a rigorous platform specification we provide with our customers. They develop their products, and they come in every platform to one of our three open labs, China, Taiwan, or the U.S., and optimize their solution for Intel architecture. Then all this comes together with our key OEMs listed on the chart here and more, where we co-architect and develop the device and the experience. How does this show up in the market? The Evo platform is specifically specified and designed for mobile performance, whether you're on the go or you stay at home. Intelligent collaboration, video conferencing.
I don't know if any of you have a dog that barks during your conference call or have kids that or babies that cry. That AI acceleration we put into the chip can do that at a tenth of the power of general purpose computing. Presence detection is the use of sensors anticipating and reacting to our needs. One simple example, when you get up and leave, sees you present, it looks for only your face and logs you on immediately. Engineered for Evo is something we started with Thunderbolt accessories and docking solutions with our OEMs to make sure every generation they seamlessly paired with no problems. At CES, we announced taking on the next challenge of Bluetooth accessories seamlessly connecting to the PC.
The last one I'll talk about here is, remember I'm at my desk with my companion device, the phone sitting next to it? We are going to enable multiple devices, say watches, phones, tablets, with whatever OS they are running, come onto our client with whatever OS we're running, so you can take those calls, texts, and move your photos without ever leaving your machine as a first step, and they'll be shipping second half of this year in selected Evo designs. Just as importantly as our built-for-business platform, vPro has been a platform we've been co-designing with large enterprise and OEMs and their software applications for 15 years, initially for down-the-wire management, and now importantly, security. In fact, we have a security engine in our silicon that can detect 85 instances of ransomware today.
We don't only sit inside our walls, let alone our firewalls anymore. We now are working with partners in the cloud to deliver this capability right to your desktop or right to your notebook in your home. We announced a partnership with VMware, where we're co-selling these features as part of their Workspace ONE solution today, and they work both in the enterprise through the edge out to the client. We just signed the second phase of that agreement. We're gonna co-architect and co-design features for true chip to cloud security. Secure client, cloud delivered. Not another pause. Let me do a handoff to Michelle to summarize where we're at.
Thanks, Jim. I hope you're all feeling as inspired and confident as I am about CCG's future. With a growing market and leadership products, I am confident that we will continue to win back market segment share, and I will be driving the team to our seventh year of consecutive revenue growth. To quickly summarize why we believe this is all possible, we expect the PC to remain an essential tool people turn to to focus, create, and connect in ways that matter most to them. This will drive a sustained TAM growth. To embrace this growth opportunity, we will continue to deliver an annual cadence of products and platforms that not only deliver new and emerging experiences but take advantage of opportunities like navigating the metaverse.
We will also continue to build unrivaled partnerships within the ecosystem, many like you saw today with Microsoft and Google, that leverage the best of the cloud and the client to deliver the choice and freedom that people want. There has never been a more exciting time in the PC industry, and I am thrilled to lead Intel's client business moving forward. On behalf of Jim and I, thank you. With that, we'll bring Tony back up. Thanks.
All right. Thank you, MJ. Thank you, JJ. At this point in our program, we have a short break. For those of you in the room, there's refreshments out in the hall. For those of you joining us online, we'll be back in 15 minutes. Thank you.
All right, welcome back. If I could, have you take your seats, we're getting into the home stretch of the program. The last of our presenters for today is our new CFO, David Zinsner.
Please be seated.
I needed a little bit more time. There we go. Back, Dave. Back up, Dave. All right. I don't know when the live stream cut back in. Something had to happen today. We made it through most of the program. Again, we'll have one more presentation, and then we'll move into Q&A. I believe almost everybody in the room probably already knows him. I'll announce him anyway. Our new CFO, David Zinsner.
I was so excited for The Tonight Show entrance, and they blew it on me. All right. Well, thank you everyone for joining, and you're now in the home stretch here with my presentation. I'm between you and drinks and what have you, so I will be relatively quick. Just so you know, for those of you that don't know me, which probably most of you do here and on the webcast, I've been in the semiconductor industry 25 years. 16 of those years have been the CFO of a semiconductor company. Every one of those companies, there was some form of transformation necessary. It's kind of the thing I love to do and I get attracted to.
I think that Intel's transformation story, which we're really on, embarking on today, is the most exciting one that I've ever had a chance to be a part of. I'm just thrilled to be the CFO and just drive the 2.0, IDM 2.0 growth story to fruition. I look forward to working with the team as we take this journey to make Intel the premier company that it needs to be. Okay. As far as goals, and my presentation is somewhat aligned to these goals, these are the four goals for me. I've spent about a month here so far, just deep diving and getting an understanding of the financials and the strategic roadmap and the product roadmap and how the process technology evolution will form. I've got a...
Actually a fair amount of confidence in how we will drive the financial part of this story, and I'm gonna go through that in, like, a click down from probably everything you've heard so far today. Number 1 goal for me is to help augment the revenue growth story, but to do that in a way that we drive profitability. That's kind of the thing that really one of my mantras everywhere I go. I believe at this point we are structured for success in terms of this growth. I'm gonna click through some of the things that we'll do to drive that growth. 2 is obviously gross margin expansion. I know that everyone looks at the margins in these low fifties.
We have a clear path to get them back up into the mid- to high 50s% in terms of gross margins, and I'll show you that aspect of things. Third is the biggest thing for me is free cash flow. We have a very good story about how we're gonna return to significant cash flow for this business. Part of that is smart capital, which I'll go into details on. Part of that is just good financial discipline. Ultimately, you know, the goal is to drive strong shareholder return for the investors. Okay. Pat showed you this chart or a similar chart, which is the breakdown of every one of our businesses. Every one of these markets that these businesses are going after are large markets.
Every one of them is growing, some more significantly than others, but they're all growth stories in and of themselves. We have strong franchises in many of these markets. We will have strong franchises in other of these markets. You can count on me, we are going to make sure that we report to you in a transparent fashion every one of these businesses, and I get the opportunity, because they're broken out the way they are, to drive accountability in these businesses, make sure that they're driving the right level of profitability and the right level of return on invested capital. We're gonna instill financial discipline through this company, and this will be the beginning of it, the segmentation that we have that you have here. Let me walk you through a little bit on the revenue side. Okay, start with the markets.
I said that the markets were large and growing, and you probably, if you had been to any one of the sessions in the morning or you listened to the last two sessions this afternoon, they kind of walked you through their markets. Let me kind of aggregate it for you. These markets on a combined basis will grow in the double digits, we expect, and will be north of $450 billion by 2026. Huge market opportunity for Intel. I'll go kinda clockwise around and just touch on them a little bit. Graphics and accelerated compute, you probably heard from Raja if you attended the meetings this morning. That's a market that's growing in the mid-teens, $100 billion TAM by 2026.
The foundry business that Randhir augmented by the acquisition of Tower, that's gonna be a market that gets to $140 billion growing in the high single digits% by 2026. As Pat talked about, we can't talk a ton about the ADAS and AV market because of the impending Mobileye IPO. Rest assured, I think all of you know that follow the industry, this is an industry that's growing fast from a semiconductor perspective. We would expect a high growth rate. Coming down to network and edge, driven by 5G infrastructure buildout, driven by the digitization on the edge, we see that market growing in the low double digits% to $75 billion. Data center is, I think everybody knows, large market growing in the mid-teens%.
We expect that to get to about $65 billion through cloud buildouts, you know, AI workloads and so forth, to $65 billion by 2026. Lastly, the compute business, obviously already a very large business, our, you know, one of our major markets. We see that growing in the low to mid single digits as a percent, getting to $90 billion by 2026. If you look at the emerging markets, you know, those will be the markets that drive a lot of the TAM growth. Even the traditional markets we see growing to $240 billion by 2026. Robust growth.
I'd just tell you know, I think in some cases when I've talked to investors more recently about our assumptions, there was some concern that we were hanging our hat on the client business growing at, you know, high single digits or low double-digit growth rate as a market. We're not. We're expecting. We have very conservative assumptions around the client business at low to mid-single digits. Now, clicking into our business, Pat went through a lot of the details around the growth rates. I'll just kinda reiterate a little bit where we stand. We're given that we're in a investment phase and we're driving our product portfolio to leadership in areas where that still need to get there, we're building out our emerging businesses that still have to grow.
We see the growth rate being in kind of the low single digits for the first couple of years. Then we see that accelerating in 2023 and 2024 to mid to high single digits. Then ultimately, as Pat talked about, we feel confident that we can get to a growth rate of 10%-12% in 2025 and 2026. Now, from my perspective, in addition to just making sure that we deliver on that growth rate, there are things I'm gonna be focused on for me and for the finance department. I just put a few examples. There's obviously more, but let's just take data center. I wanna make sure that we make very focused investments that the guardrails around this business are very tight.
They're focused on the areas that are gonna be important to drive the growth, to drive the success in data center. In the client business, it's about making sure we exploit the great franchise we have in that business, make sure that we're getting the value from customers for the leadership position we have. Pat talked about Mobileye, but you know, Mobileye is an opportunity to unlock shareholder value through the IPO process, so we'll make sure we execute on that. In the foundry business, it's making sure that we expand, but expand in a way that is thoughtful and make sure that we maintain good profitability, good discipline around our CapEx intensity and good discipline around our free cash flow aspirations. On to gross margins.
We talked about, you know, we are gonna be in an era for the next few years where we're investing. Investing to accelerate our process technology, investing to make sure that we get our fabs and our shells in a place where we can expand at the rate of demand. That takes some investment, obviously, and that will impact gross margin. Gross margins we see as being in this 51%-53% range for the next few years, 2022, 2023, and 2024. I have high confidence that we can get the gross margins into the 54%-58%. In fact, I think that the product gross margins will run higher than this, and the only thing that keeps it in this 54%-58% is the foundry margins.
Because the foundry margins, obviously, even at the leading edge, are in the low 50s%. That will be in the low 50s%. The product margins should be above this. We see that the blend would be somewhere in the 54%-58%. I'll walk you through some of the things beyond just what you might think of traditionally as areas that we'll focus on. First, we're going through an accelerated pace of process technologies. We're trying to do five process technology in some four years.
When you do that does put pressure on the gross margins, but as we move the cadence back to a normalized cadence, as we get the benefits of EUV, as we get the benefits of being at the most advanced process node, that will drive an improvement in gross margins, and we expect that to be somewhere in the 100-200 basis points. Second is just pricing. You know, we obviously, in some areas, we do have leadership product, in some other areas we don't. But as our product portfolio improves, not only through this year, but through the next few years, we see a great opportunity to improve our pricing for the leadership products that we're producing for our customers. That's probably worth 100-200 basis points.
On scale, you know, given that we have high aspirations for growing our business, and as it accelerates into the 10%-12%, we'll get a benefit from that scale, but we'll also see improvement in mix. Some of the businesses that are emerging, that we expect to see high growth from actually carry gross margins that are higher than the corporate average. That will help lift the gross margins as well, and we see that somewhere in the 100 basis points. Lastly is the day one, exclusively, which is the fiscal discipline. It's me and my finance team really driving the operations, driving the businesses to execute to a world-class measure of cost. I, you know, I've only been here a month, but I've already seen a ton of different opportunities to do that.
I think I'm probably being conservative by guessing that it's 100 to 200 basis points. It's at least 100 to 200 basis points. If you add up all the high ends of that and, you know, roughly put us in the middle of the gross margin range, you'll actually get to a number that's higher than that. We're gonna shoot higher, so that at a minimum, we're well within this 54%-58% range. The one kinda headwind we'll have is inflation, which I'm sure everybody is conscious of. If they go to the grocery store, they're conscious of inflation. We're not immune to it.
I'd say in the semiconductor space where everyone is experiencing inflation, we're actually better positioned because we're vertically integrated, because more of our manufacturing is internal, because we have better control over the costs in that area. That gives us a competitive advantage, a better cost advantage relative to our peers. We see, actually, ironically, inflation as being somewhat of an advantage for us as we progress. Okay, now CapEx intensity. Pat, I mentioned in his remarks that we're shooting for 25% net capital intensity to be around 25%. You know, obviously we'll be through an investment phase in the first couple of years, which will push the CapEx intensity up to about mid-30s% of revenue. I feel very confident that we can get this back to a net 25% in the 2025, 2026 timeframe.
Part of that is just we're gonna manage that. That is the guardrail by which we will operate this business, is 25% CapEx intensity. I'm committed to make sure that we maintain the balance sheet and the P&L in a manner that we would expect. I think there's a lot of things that will help us there. You know, as we move back to a normalized cadence, that will help alleviate some of the pressure on CapEx intensity. As we get caught up, you know, one of the reasons why we have this higher CapEx intensity is we have to invest in shell space that we have been, quite honestly, under-investing in the last few years.
As we are at a point where shell space is positioned at the right time when we need it to manage our demand, we'll be in a better place to reduce our CapEx intensity. The most important thing I think for us that's in control, that we're in control of is smart capital. Pat mentioned smart capital. I'm not sure everybody knows what smart capital is, so I'm gonna try to define it a bit for you. In at least my view, I see it as kind of five things. Number one is this shell first strategy. It's building shells in advance of necessarily when you need them, so they're ready as demand comes that you can expand your capacity. It allows you to kind of modularly add capacity to the infrastructure.
The advantage is obviously the shell is you know, not all of the cleanroom's expense. In fact, it's a moderate amount of the expense, and so that helps moderate our CapEx. The other benefit to that investment strategy is when we're investing in shell space, we depreciate that over longer lives. We depreciate that over 20 years. So the gross margin headwind from adding shell space is actually relatively minimal versus what you would see when you invest in equipment. The second Pat mentioned but is government incentives. We have the CHIPS Act in the U.S. We have an equivalent one in Europe. We have states that are offering incentives. We have countries within Europe that are offering incentives. There are even incentives in parts of Asia.
We will take advantage of those incentives, as an offset, to some of our capital investment. It's good for us, obviously, but it's also good for national security reasons, for countries around the world. It's also a great economic engine for countries that invest in semiconductors. It drives a lot of job growth. Third is just customers. Randhir likely mentioned this. I think Keyvan mentioned this also. Pat mentioned this. You know, we will seek customer commitments, as we look to expand our fab footprint, in the form of prepayments. We've been talking to customers, I'm sure Randhir mentioned this, and I think we have good line of sight, to see prepays as we start to build out, in particular the foundry business.
Fourth is financial partners. We just signed or agreed on a memorandum of understanding, an MOU, with Brookfield. For those of you that don't know Brookfield, they're the largest investor in infrastructure projects in the country. They largely invest in energy projects and so forth, land development projects. We are developing a model, kind of a first-of-a-kind model for project financing with Brookfield that will help drive more efficiency in our capital investments. We'll have more to talk about. It's, you know, kind of a framework at this point, but I think can really augment some of these other capital offsets to help drive more capital efficiency in our investments.
I'll tell you, when we look at two, three, and four as those are the ones I would consider as the capital offsets, we think that we've modeled at least that will be about 10% of our gross capital investments to drive it to a net capital number. In reality, we are chasing 20%-30%. I think if it isn't 30%, I'd be surprised. There's literally $hundreds of billions available for us in these offsets. I'm really optimistic that we can actually do better in this regard. Lastly is just, you know, utilizing third-party foundry. Obviously, you know, those wafers come without the CapEx investment. It also allows us to be more flexible.
You probably heard Raja in his presentation talk about how he was gonna mix the internal and external foundry capacity. This is gonna be something that we also leverage. Also, I think Keyvan mentioned it as well as a part of his strategy. Okay, and now the one that really is near and dear to my heart is free cash flow. Our goal is to get free cash flow to 20% of revenue. Obviously right now we're in an investment phase, so it's not gonna be there. We see a clear path to getting to 20% revenue and free cash flow. Some of it is just what I talked about. It's revenue growth. It's gross margin expansion. It's leveraging the smart capital to offset some of our capital expenditures. Also, I think we have good leverage on the OpEx side.
Obviously we're at an elevated level right now, but as revenue grows at a more accelerated rate, we think we can grow expenses at a lower rate, and that should bring CapEx as a percent of revenue down, which adds to our free cash flow. Then lastly is working capital. You know, again, I'm only been here a month, but I see a lot of opportunity in working capital. It just hasn't been something that's been a big focus, and there's a lot of dollars locked up in our working capital. I'll give you one example. We have a very kind of a lack of linearity in our shipments. It ends up driving less of our cash getting collected in the same quarter in which we ship.
I think we can work on things like that that will ultimately be pretty beneficial to our free cash flow. From a balance sheet perspective, you know, I think it stays pretty much as it has been. We intend to keep our leverage at a modest level. We wanna maintain healthy liquidity. That's a strong cash balance augmented by our $5 billion revolver. Then lastly, just, you know, of course, both of those will ultimately yield a very strong investment grade rating. We want that flexibility with a strong investment grade rating to be able to invest and do the things that we need to do at the time we have to do them, regardless of what the economic situations are.
This will give us maximum flexibility to make those investments in a timely fashion. On capital allocation, I broke it into three areas. Number one, I think you probably hear this a lot, but it's invest in the business. Invest in the talent, which is the number one thing we've gotta do. Invest in technology, invest in infrastructure. You know, the goal is to maximize growth at the end of the day, and this is critical for us to be successful. In addition to the organic, we will augment it with M&A, inorganic opportunities. I would just tell you that, Pat and I have a very high bar as it comes to M&A. It's gotta have massive strategic sense, and it's gotta make great financial sense, and ultimately, it's gotta deliver very good return for shareholders.
We'll maintain that bar in everything we do. On the flip side, I'd also tell you that M&A is not just about acquisitions, it's also about monetizing assets. We just sold the NAND business just at the beginning of January. We got $7 billion. That gives us flexibility to invest in the business. We're going to take Mobileye, we're gonna monetize that in the market. That unlocks value for us now, gives us cash now, but also allows us to ride the future valuation expansion of the Mobileye business, which is, you know, kind of a win-win for us. In terms of M&A, I just did wanna mention about Tower. It's been, I think, mentioned a little bit over the course of the day.
Maybe I'll just reiterate from a financial perspective, this allows us to target this $140 billion market by 2026. They are one of the leaders in the specialty foundry market. That's a $15 billion market. But more importantly, they give us all kinds of know-how in terms of how to go to customers, how to be a foundry, how to deliver what customers are expecting. We see significant synergies. Synergies in terms of helping Intel get better, and synergies in helping Tower get better in terms of their cost structure. I think there's gonna be massive opportunity. The ROI on this thing is just going to be spectacular.
I should say that on day one, it's accretive immediately, obviously, but that pales in comparison to the accretion we'll get as we progress once it closes. Come back to the capital allocation. Invest in the business, utilize M&A for both monetizing assets and acquiring assets. Lastly, return of cash. We are absolutely committed to returning cash to shareholders. We intend to maintain than the earnings growth. Okay. That's kind of the prelude to the model. I'll start with the 2022 model, give you a sense for that, and then we'll go into the out years.
Okay, for 2022, Pat already mentioned that, you know, while the market's still very, very healthy, you know, the supply chain challenges that everyone's dealing with is a bit of a challenge. That is affecting the demand a bit. We're still gonna grow. We think we'll grow to $76 billion in 2022 with 52% gross margins right down the fairway from what the range we gave investors in the third quarter and the fourth quarter. We said 51%-53%. It will be right down the middle at 52%. We are making investments on the OpEx side to drive product growth. That is going to elevate our OpEx both in terms of dollars, but also in terms of percent of revenue.
That translates to a $3.50 EPS. Now, free cash flow, we expect to be slightly negative $1 billion-$2 billion. Obviously, we'll be managing that tightly. We talked about a CapEx range of somewhere in the $25 billion-$28 billion in the third and fourth quarter. We've kind of locked in at $27 billion on a net basis. You know, we have a relatively modest offset assumed in this number, and we'll obviously be looking for more, so there's a potential we can do better than this. But for now, that's the modeling assumption. Let me talk about what I would call the investment phase, which is 2022 is an investment phase, but also 2023 and 2024.
We start to see the acceleration on revenue that I talked about, mid- to high-single-digit growth. Gross margins remain within this 51%-53% as we still are in this accelerated roadmap of process technology and capacity expansion. I do think we can start to see the OpEx start to show some leverage, so that comes down into this 28%-30% that I have here. Net CapEx intensity that I mentioned before will be up in this mid-30s% as a percent of revenue, and I would expect free cash flow to be roughly neutral. Then we transition to the long-term model. Now growth is really accelerating. It's 10%-12%. We'll get to the margins that I wanna get to, which is in the 54%-58%.
OpEx continues to show leverage down to 25%-27%. CapEx intensity comes back to a more normalized range of 25%, and that should drive free cash flow somewhere in the 20% of revenue, which, you know, by industry standards would be a spectacular number. Really drives really great ROIC. I think if we deliver on this, which we feel confident, I deliver on the first part of what Pat's double-double was. He wanted double the earnings, double the multiple. I deliver on this double the earnings for sure with this model, so that will be our goal to do that. Okay. Back to the end of Pat's. Pat, it was ironic. Pat said it was his 1-year anniversary. It's actually my 1-month anniversary today, too, so that was somewhat ironic.
You know, I think Pat delivered these messages very well. You know, we're in very high growth markets, as I've shown. We have competitive advantages with this IDM 2.0 strategy that no one else in the industry has. This is absolutely the strategy that will make us successful. There is an investment period to make it successful, but the outcome is significant, both in terms of P&L and in terms of shareholder value creation. Pat touched on this leadership and culture, but just from an outside person coming in, it's real.
I mean, he is driving real accountability across the organization, and we have real kind of processes for that if you get a chance to talk to Christy, our head of HR, she can go into it, but it's different than what Intel has been managed to in the past. I just think the leadership team is an amazing leadership team, and we're coming up with ways to unlock shareholder value for everyone. But more importantly, I think for me, it's about making sure we do all these things and are always conscious of the P&L, always conscious of the cash flow, making sure that we deliver on the shareholder return aspirations that we all have collectively, both internally and externally.
I'll just tell you one other thing. You know, I went through these models again, having only been here a month. It's maybe easier for me to say this, but those are the models we're absolutely committed to, but I'm shooting for higher than this. You know, as Pat I think said, "Get on the bus" or "Get on the train" or whatever he said. You know, get on and come aboard 'cause this is gonna be an amazing story. With that, I will turn it back over to Tony.
All right. Thanks. All right, we're gonna move into the Q&A part of the agenda at this point. I'm gonna give the stagehands a little bit of opportunity to move some chairs out. So let me explain a bit how this will work. We're gonna have some runners with mics sitting in each one of the alleys, so just go ahead and raise your hand and they'll run a mic over to you. They will hand you the mic, and then, when you're done, you can just hand it back. We're gonna wipe down all the mics to make sure it's COVID safe. Because we're trying to get through a number of questions, I would ask that you ask only one question and then you pass the mic back and we'll try to get through as many as we can.
We'll go through this about 25 minutes or so, at which point we'll bring out the rest of the executive leadership team, and then we'll go through sort of the end of the hour. With that, I'll invite Pat to come back on stage.
It's good.
All right. Let's go. Why don't we just start down here in front? Brooke?
Thank you. Vivek Arya from Bank of America. Thank you for the presentations.
Just wanted to clarify if the roadmap for Granite Rapids is the same as what you had committed to last year. But my question, Pat, is that in the PC market, how risky is it to assume that PCs, which are kind of at a cyclical, you know, kind of back to their prior peaks, can stay at this 350 million and maybe even grow from here? Then within that, there is also emerging competition from Arm. I say that because PCs are such a big part, right, of not just your sales, but your fab utilization. So how much have you kind of tested the model to what happens if the PC market goes from 350 to 325 or 300 million units, or if Arm, you know, takes a bigger share of that? Thank you.
Yeah. On the first one, we did change the roadmap, as we said, on the server roadmap, right, with the Sapphire Rapids and Emerald Rapids into that platform. Granite Rapids is now a product that's on Intel 3, a new micro-architectural core going into it, so it's a higher performant product in 2024. We've changed the roadmap. We've made it stronger, hit this cadence as we described, and we added Sierra Forest, right? You know, a product that's in that same platform, but with the efficient core. It's a different roadmap than we were before, and we've, you know, literally in preparation for today, just taken all of our major customers through it, and they're very happy about the changes that we've made. Good response to the customers. On the PC market question, it's one that we get asked a lot, we probe a lot.
You know, we're, you know, I'll say the amount of time that we've spent on that exact question, right, in our market surveys, working with IDC, Canalys, et cetera, working with our ecosystem partners as well. You know, we're confident. First, we're confident in those market numbers, and there's nothing amazing here.
Right? You know, you go look at refresh rates, you go look at penetration rates, you go consider, you know, the continuing expansion in various, some markets going forward. There is nothing heroic in those numbers, and that's why we use the language a structurally larger market. You know, just one example is Windows 11 is gonna drive a major corporate refresh cycle. You know, we're overdue on a corporate refresh cycle, and new security capabilities are required for Windows 11. You know, we just keep testing that, and that's part of the reason we have Panos talk, as well, you know, 'cause they're seeing the same thing, and we've done a lot of validation of that with customers. Now, that said, you know, we do have other risks here that we're looking at, like the Arm market, and that's been super low end, right?
You know, largely the only reason that we've seen that emerge, 'cause the compatible PC marketplace, you know, we just don't go down that low. Supply constraints have allowed that low market to sort of, you know, exist and bubble around. We have to deliver better products, right? Because if the product is better, people don't wanna go through the painful work of creating new software releases and, you know, all of the other challenges of the issues that a long-tenured software ecosystem brings to the PC space overall. We have to be competitive. We're well on track to do that, and the roadmap that we've laid out, the strength that we've seen in Alder Lake, if anything, we're more confident today than we were having this discussion 30 days ago. We're just gaining momentum.
The IDC results showed us gaining share in Q4, and the market robustness has been constrained by supply chain issues, right? It was not demand. It's been supply chain issues. Otherwise, this number would be more like 370 or 380 as opposed to the 350 that we showed.
Thank you.
All right, everybody. That was like a question and a half, so I'm gonna ask everybody to kind of round down a little bit. Tim?
Hey, guys. It's Matt Ramsay from Cowen. Thank you, Pat, for making your team available for so much information today. I wanted to touch on a couple points on gross margin, and Dave, you can definitely chime in on this as well. We're starting at 52% in March, and that's kind of the near-term range. I guess there's some concern that as you roll out these nodes more quickly, that there'll be some pressures that'll take margins below that level, and you guys have expressed some confidence that you'll stay above 50. Dave, maybe within your month at the company, you could give your perspective on 50 really being the floor and what upward pressures would counteract some of those downward ones.
Yeah.
The second little point is I saw in the long-term expansion of margins, you had talked about ASP expansion. There's been a concern, I guess, that Intel might protect share at an absolute on pricing, and it sounds like you guys are talking about ASPs expanding. If you could talk about that'd be great. Thank you.
Okay. Before I let Dave comment, part of it is, and I just want to emphasize, you know, what he said at the end of this presentation. This is how we're gonna manage the business, right? You know, we have a lot of levers in how we manage the business, you know, how we go drive our product teams, how we go create cost structures, you know, how we go drive yield. There's a lot of levers inside of it, and I'll just say, this is how we're gonna manage the business, right? You say, "Hey, there's gonna be other headwinds." There's always freaking headwinds, right? You know, if you don't want headwinds, don't get out of bed tomorrow morning, right? That's our job, to manage the business to what we think is a conservative set of financial parameters and guardrails that we have laid out.
Everything that we've done over the last year is we're a meet, beat, race kind of culture that we are building. We're executing on the environment that we have laid out. Now I'll let Dave give details, but
I think he was telling me what I need to do. Yeah, I was gonna say the same thing. I mean, part of it is we're just gonna manage to that level. The one other thing with this year is we are going to see product launches as we progress through the year. We're pretty excited about those product launches and what those product launches will deliver in terms of updrafts to the gross margins. That will also help. I'd also tell you that we, you know, the biggest cost factor, well, two, one of the biggest cost factors that we have in terms of margin pressure when you're, you know, accelerating your process technology is this, like, you know, new startup cost that we have. We had a step-up in 2021. We're having a step-up in 2022.
2022 step-up is largely kind of flattish, and in fact, it's flattish for the next several years. Part of this is the tick-tock approach that we're doing, that it allows us to kind of more evenly spread that startup cost, which will also help. You know, a lot of it is, as Pat said, it's just us, you know, rolling up our sleeves and managing the business to that level of margins.
All right. Thank you.
Oh, the pricing thing, which I can answer for even though he
Oh, yeah. Go ahead.
Even though he violated the principle too. Yeah. On the pricing side, look, it's really not about taking a price and going, "We're gonna jump that up." It's really coming out with world-class products, leadership products that deliver something really important to customers, and when you do that, customers go, "I value that. I pay for that." That's the assumption that we're making and the confidence we have is that we will deliver on this product roadmap, and that begets, you know, better pricing over time.
All right. Great. Thank you. We've had a question and a half and a concatenated question. We'll keep trying. I'll go back to Ben here.
Hi. It's Ross Seymore from Deutsche Bank. I'll just make it a eight-part question for you. I wanted to ask about the competitive landscape. What you've launched in Alder Lake has done a great job of reversing market share loss, going to market share gains. The server side has been a little bit of a different story, especially on the cloud side. Results from some of your competitors recently were much stronger than you guys in that cloud market. Given how you've reworked the roadmap today, can you talk about any changes in when you would hit competitive parity and take a leadership position in that given Emerald and Granite push out? And why are you skipping the Intel 4 node completely?
You know, I don't know if that's a five-part question, but right. Sapphire is a leadership product. Okay? I'll say it again, and again, and again, right? It is the best product in the market that we're ramping this year. You know, it has the best power, performance, right, key features, and security, and AI, right? It is not as good of a part as we would want. I will certainly grant that, but it's the best product in the marketplace, period. We do expect to see AMD respond, right, with their part. The next couple of years are gonna be a bit to and fro, and that's what I've said before. It's not like we have the unquestioned leadership as we have in the client, where I've said AMD's in the rear view mirror, right?
We never see ourself falling behind again. Here, it's gonna be a much more competitive market for the next couple of years. We're, you know, executing against our roadmap gives us better competitive tools. The gap in, you know, in these competitive this is substantially reduced over what it has been the last couple of years, but we don't have that defined, unquestioned leadership until the 2024 products come, right? With that, we chose to accelerate our embrace of Intel 3 because it produces a better product. You know, customers in the server space don't want the same cadence that you can tolerate in the client space. They want this more predictable cadence, right, of platforms and products. It's a much longer adoption qualification cycle associated with it. We align the Granite timing, take advantage of Intel 3, create a better product.
You know, we're moving from Intel4 to Intel 3 gives us about 18% in process technology. We're also getting a new E-P-core, a new performant core inside of that. That's another 10%-12% on top of it. We've created a little bit later, much better product that aligns with our customer cadence. That's exactly why we've done it, and we've also created the split roadmap that gives us the efficient cores with Sierra Forest, and that really hits that cloud market. Since we have both P-cores and E-cores in our portfolio, our competitors only have one, we get both, we're now optimizing for the two segments much more effectively.
As we would say, "Hey, it's gonna be challenging through 2022, 2023, but when we get to the 2024 products, unquestioned leadership," and that's what Sandra has said, that we'll be accelerating our growth rate over that period of time. You know, as you go think about that, this is still a nicely growing business over the next couple of years. As we look at what we're doing with AXG and NEX respectively in those market segments, overall, you know, we're bringing a lot of capabilities to a marketplace that's in these other areas, growing very nicely, in this area, getting more competitive, and that's before we talk about security, software, many of the other things that we talked about.
All right. Kenji?
Oh, hey, sorry. Hey, Dave.
Hi.
Hey, Pat. Thanks for hosting this. Very, very informative. This is for you, Dave, right in your wheelhouse. Welcome back to the stage. The question is really on return on invested capital, and not capital allocation to shareholders, capital allocation internally.
Yeah
... when you're looking at investing in IFS versus IDM.
Mm-hmm.
I didn't get a good sense, and maybe I should have probed better in the sessions, but kind of what's the decision-making, what is it dependent on? What are the milestones that you are looking for, you and Pat are looking for when you're saying, "All right, this is the investment going into IFS. This is going into IDM"? And then how much of the capacity is fungible? That's kind of those were my, I don't know, 1.25 questions.
Yeah, no, that wasn't 1.25. That was at least four. Yeah. Okay, so maybe I'll start. I'm sure Pat will want to come over the top on this one. We do have ROIC measures, you know, obviously. I didn't mention it in here, but, you know, if we get to the higher end of our range here, the ROIC that that delivers is, like, north of 15%, probably likely 20% ROIC. I would tell you that internally, that's roughly in the range of what we're gonna wanna see, and we're probably gonna wanna be shooting higher in certain cases to make sure that, you know, ultimately we get to those levels of ROIC.
We wanna deliver economic value relative to our weighted average cost of capital, which is more in the kind of high single digits% of revenue. We're making those decisions all the time. I would tell you as it relates to the foundry business, for example, we just have a belief more than anything that that is a business that we should be in, that we, as Pat mentioned, IFS benefits from IDM benefits from IFS. We have a belief in that that almost transcends ROIC. It's a meta-level strategic decision that we're gonna make.
I would say more than anything, as we're looking at capital allocations in that regard, it's about demand, and it's, you know, methodically building out capacity as we see demand, whether it's in the product side or in the foundry side, what is it that we're seeing? Some of that, you know, kind of feed of information is gonna come from customers coming to us with prepayments, you know, wanting to secure significant capacity. That's a trigger for us to know that that's a good investment. I think if you can at any time deliver margins that are north of 50%, chances are the ROIC is not gonna be a big issue for you in terms of delivering to not only the company's ROIC metrics, but also to shareholder return ultimately.
Remember, there's three levers that we talked about inside of our IFS investment.
Right. One is leveraging depreciated assets, right? You know, we're, you know, extending life of fully depreciated fabs and capabilities. That's always been the sweet spot, you know, for foundry margins. We are gonna have a lot of fab capacity that we're gonna be able to leverage, extend life of equipment. You know, second was the unique capital offsets that we're gonna expect to see that go into that as well, right? We're saying, you know, that we're going to be in a position whether those are customer prepays, you know, other, government offsets, right? We see a lot of opportunity to leverage what they've talked about in terms of smart, capital there. You know, finally, as we go, look at that business as well, you know, it is highly leveraged.
You know, when we run 18A, you know, it may be a slightly different recipe for some foundry customers, but generally, it's gonna be 18A. You know, what's next? So on. You go, "So this is gonna leverage the R&D investments in a very powerful way," and that's what I was describing in my comments. You know, the IP libraries that we create for our internal designs get better monetized. You know, the R&D efforts, TD efforts, packaging efforts. You know, this is a powerful leverage to further augment, you know, the unique assets that Intel brings to the table here. That's why we've said IFS makes IDM better, IDM makes IFS better. You know, if you look at the TAM charts that David showed, that was the biggest number on the page. This is a great opportunity for us.
Finally, right, and I'll make the point again as I did in my comments, you know, a major investment in Intel today is a hedge against the massive geopolitical environment of the world. The world wants more distributed, more resilient supply chains for the long term, and we're definitely gonna lean into that very powerful tailwind.
Now, if you're asking about the milestones, by the way, you know, of course, in everything we do, we're gonna be monitoring how we're doing and, you know, obviously, you know, as we get further along and have better information, we'll make course corrections and modifications along the way. That clearly will be part of our model, no doubt.
Just remember as you get your question, to please say your name and what firm you're from as well. Brooke?
Yeah. Thanks. John Pitzer with Credit Suisse. Pat, thanks for all the information today you and your team put together. You know, I think many of us out here have an easier time underwriting the TAMs you put out there than your ability to actually go in and exploit kind of the growth opportunities. The one dynamic I would like you to talk about, and we've talked about in the past, is this idea within the data center and accelerated compute, general purpose compute versus sort of semi-custom and ASICs. What role do you see semi-custom and ASICs playing? How big a part of the market do you think that will become? Can you now participate in that market with IFS in a way perhaps that you didn't? I'd be curious as you answer the question, 'cause everyone has a six-part question.
It's interesting that you guys are now willing to license out x86.
Yeah
... and we've never seen, you know, an Arm x86 benchmark except with either Intel, AMD, and a third party. What's been the reception to sort of the x86 potential licensing?
Let me start at the back of that question, right? 'Cause I think it will help to inform the whole question. You know, as we are engaging, particularly with our largest customers, our cloud customers, I'd say probably a full third of our foundry pipeline is interested in x86. So we're seeing a very robust interest in what that can mean as we start to bring together our IP with their IP, particularly the big cloud customers, but much, much broader than that as well. You know, as we think about that then, you know, as we sort of reflect back, this just reinforces the strategic shift that we're now saying we are much less religious about which architecture is appropriate for which portion of the marketplace.
That said, there is no confusion that an x86 socket is by far the best margin thing, the longest tail, trillion lines of code run on this thing. You know, that is the best thing for everybody, us and the customers, 'cause reoptimizing software stack is a terribly hard, long, painful process. The other thing, and if I point out just one thing, John, from our discussion today, you know, we described Falcon Shores in Raja's presentation, and it's steps like that that make this even more pointed that we're now saying, you know, in a memory-centric architecture, we're gonna open up a coherent view for the software programmer of GPUs and CPUs in an almost seamless fashion, right, across that. We're gonna have optimized accelerators, other ASICs.
You know, if you look inside of Sapphire Rapids, we already have a number of special purpose accelerators as part of it, as well. Again, we're not confused. An x86 line of code is the best thing that we can possibly do, but we're also gonna open our canvas of innovation with our customers to take on more of these other workloads because in the, you know, in this insatiable demand for computing, you know, we've seen systolic arrays are very important for AI functions. Clearly in the networking space, specialized security, you know, other I/O hardware is more capable. We are gonna be embracing those more effectively, and the combination of IFS opening the doors, our IP and customer's IP gives us more flexibility.
The embodiment of x86 plus GPU plus Arm plus RISC-V plus our broader set of portfolios, and we're tying all of that together with one software architecture with one API. That combination of things, I think, is a very powerful one that just positions us for strategic capacity as Intel never had before and differentiates us from anybody else in the industry by a wide margin.
Okay, let's go back to this side of the room. Tim?
No. Thank you, Pat. Chris Caso from Raymond James. The question is about the degree of flexibility that you have with the spending, in the event the market conditions should change, you know, better or worse, as you go forward. You know, given that the spending that you're doing is really based on roadmap and getting the technology as opposed to just simply building out capacity, does that limit the amount of flexibility that you have such that if, you know, market conditions could soften 2023, 2024, that you'd run a risk of, you know, margin erosion or, more importantly, free cash flow burn?
Yeah, we're gonna manage it. We're definitely gonna manage it. Those are, you know, kind of guardrails by which we'll manage the business. But we have other levers. You know, it doesn't have to be, you know, turn it all off. It can be, you know, more aggressively in the smart capital arena. It can be, find partners. It can be, you know, manage our third-party foundry relationships a little bit more. We have a lot of levers actually to manage this beyond just, "Hey, let's just turn off the spigot." But we will manage to this, you know, mid-30s% of revenue in the early years, down to the 25% as we progress into the 2025 and 2026 period.
The other thing I'd add on top of it, I'd just say, boy, you know, we have so many people approaching us about building more capacity.
Mm.
Right? I mean, it is everywhere, right? That's why I say, boy, you know, $500 billion-$1 trillion, hey, are there gonna be some bumps along there? There always are in the semiconductor industry at that point. But there is so much demand across, you know, whether it's cloud customers, whether it's at edge and 5G customers, whether it's at new consumer experiences. You know, we're discussing with one of the cloud customers today, you know, that we're not, like, close to their demand. We're half of what they want to do by their 2024 build-out, right? And we're just saying, "Boy, how do we go catch up with this?" That's before we started to talk about the Tower opportunities that we're just adding to the portfolio.
I realize the validity of the question, but every comment that we get from our customers and the engagement with the market just says we are so far away from that. You know, maybe that's a question in 2027 or 2028, but I just don't see it in the next several years.
Ben?
Hi, it's Tim Arcuri at UBS. Thanks. I had a question on IFS, and that's obviously one of the key levers to the growth out in 2025, 2026. On a separate slide, you showed that you're not gonna start taking tools for IFS until 2025. That seems a little late to then have IFS become a big revenue driver out in 2025, 2026. Does that imply that you're gonna co-mingle some of the foundry business inside your own fabs? And if so, then if I'm a customer, I'm gonna think maybe skeptically about that because I think you wanna separate those two businesses. Thank you.
Yeah. You know, separating the two businesses only makes sense when you get to significant scale. What customers and, you know, at some point they're gonna say which fab and those kind of things, but the key question is, are we gonna commit capacity corridors, right? Just saying, boy, what's the wafer commitments that we have at different points of time. In the early days here, it's not like we're setting equipment in this side of the fab to run or setting up fab modules. It's highly commingled in the early days. By the way, that's the way customers will want it, right, at that point, 'cause they're making early decisions. They don't wanna have to, you know, necessarily make long-term commitments before we started to build some test chips, some momentum in the market with them, as well.
You know, as you think about that over time, obviously, we'll accelerate our plans if we get bigger customer, you know, prepays, bigger customer commitments, you know, higher volume customers come into it. We've committed a small amount of dedicated capacity and a small amount of smart CapEx for customer offsets for that business. We would certainly hope that we're taking both of those up over that horizon, and we're well above in terms of the revenue forecast that we've built in that are still modest for IFS, the capital commitments that we've built into this plan, which are still modest, but also the customer prepays and other capital offsets, which are still modest.
Hopefully, all three of those rise well above the plan, but we accomplished the 10-12 with a very modest set of assumptions around the growth of IFS over this period of time.
Okay, we're gonna transition a little bit here. If I could ask Pat and Dave to stand up, I'll have them move the rest of the chairs out. We'll take one more question here while standing. Kenji?
Gus Richard, Northland Capital Markets. Simple question. Pat, you've hired some great people, and they don't look money motivated. What's the sales pitch? How are you getting these people in the door? Is it just the mission impossible? You know, any color there would be really helpful.
Yeah. You know, people, as I commented, you know, we are on a mission. You know, there's almost an irrational commitment to this commitment of Intel as this foundational technology company for Silicon Valley, for the industry, for the world. You know, this team that we're assembling, I like to call it my five-year team, you know, right? Where we are kumbaya coming together because we are going on mission together to restore the most iconic company in the industry to deliver against, you know, just an extraordinary set of opportunities in front of us, to launch new disruptive businesses for the future. That's a journey that is unleashing incredible energy of the 120,000 people we have inside of the company, but this leadership team, as well as seeing that opportunity and that passion.
Now, you know, when you translate that, hey, we gotta take care of our people, right? We gotta compensate them well. Great talent shows up because they believe in the mission.
They say, "I wanna work on the hardest problems, and then I wanna work with the best people, and I wanna have a culture and an environment that I wanna be part of." "Oh, yeah, I wanna be compensated as well." It's in that order. What we are assembling is a team that believes in this mission, that Intel is the company that can deliver technologies that improve the lives of every single person on the planet. That's the mission that we're on, and we have a unique set of assets to accomplish it.
Great. Pat and Dave, you can take a seat. I'll now invite the rest of the management team to join us on stage.
Pat and I forget where we're supposed to sit.
We've made it easy for you.
Oh, wait. I think maybe...
We failed. Okay. Great. We'll run this the same way. I'll kind of go back and forth across the audience. I'll try to also moderate and send the question to the right person. I think, Brooke, do you have one on this side?
Hey, thank you. Loomis Sosa, Taiwan Capital Markets. Pat, touching on something you just said a second ago, 2025 is still sort of out there. I guess, originally, maybe I thought that revenue for IFS would start to come in sooner. Realize that it takes a while to build plans and it takes a while to do semiconductor things just like software. You know, do you want to accelerate IFS? And if so, what are the limiting factors?
Yeah. I'll start, but we should ask Randhir on that one as well. You know, I'll just say, "Hey, we're building a nice portfolio.
What's that?
I guess if you could answer the point. You didn't really answer it.
Okay. We're building a nice portfolio of business on Intel 16, on the packaging, on Intel 3. It's just nicely building over time. As we get out to Intel 3 and then Intel 18 in particular, these become whale customers, right? That's where there's a lot of variability in how rapidly that revenue ramps up in the 2025, 2026 time frame. You know, we haven't landed those customers at the scale that we see as possible, so we haven't built that into the plan or we'd be way over these numbers, as well. As we said, we expect some of those decisions to be made this year, right? Because if you look at the timeline for those big, you know, 18A, for major customers coming in a big way, they have to make their design commitments late this year.
We feel everything is sort of coherent in what we've put together. You know, Randhir, maybe you can comment a little bit more on some of the biggest customer activities here.
I think we have five customer wins. They range from high performance compute to mobile and to automotive. They range from the Intel 16 node all the way to Intel 18A. The timing for this is early, but as I said in my presentations, this year we have 30 test chips that we are driving for our customers. We see and 2022 will be a defining year where we'll start planning for the 2024, 2025 horizon in terms of our capacity ramps.
Okay. Tim, you have ones from this side of the room?
Hi. Jerome Ramel, BNP Paribas. Maybe a question for Sandra. Looking at Habana, do you see traction beyond AWS for Gaudi?
Yes, indeed. In fact, let me talk first about how well it's going with AWS. As I mentioned, they launched last year the EC2 DL1 instance based on Habana, and we're delivering for customers a better time to market in terms of training their models and also an almost 40%, or in some cases, up to 40% better price performance. We are going to be introducing to the market a little later this year, I won't say when yet, the next generation of Habana, and we expect that to also be deployed on other hyperscale cloud customers. In terms of the additional wins that we have, we actually have a supercomputer win with San Diego Supercomputer Center that has been deployed, and we have a number of enterprise customers that are also adopting Habana.
They've not been public about it, so we haven't talked about it, but there has been a lot of interest with this idea that we could have an open software platform for Intel-based products across the whole spectrum of both Xeon with additional AI accelerators built in with our GPU, Ponte Vecchio, that Raja is delivering to the market, and then with the Habana Gaudi portfolio.
Ben, how about from this side?
Thank you. Srini Pajjuri from SMBC Nikko Securities. I have more of a short-term question on the data center business, Pat. In your guidance for $76 billion, what sort of growth are you assuming for this year? And also, if you can walk us through what you expect the market share dynamics are gonna be. Obviously, you have Sapphire shipping. And when do you think Sapphire will have an impact, potentially kind of slowing down or whatever your expectation is, when do you think that'll come into play in terms of your market share dynamics?
Yeah. Sandra, you wanna-
Yeah. As I discussed, we are beginning to ship Sapphire Rapids at the end of this quarter, next month, and we expect Sapphire Rapids to continue to ramp throughout the rest of this year and then, of course, well into next year. I think that sometimes the market doesn't fully comprehend the fact that customers ship products for a long time. They have the products that we've deployed, as I mentioned, Ice Lake has ramped to 1 million units in Q4. Lots and lots of demand for Ice Lake. Sapphire Rapids comes in, we expect the ramp to continue throughout the year, and then that will continue throughout next year, as we then introduce Emerald Rapids as that SoC compatible refresh of Sapphire Rapids for customers, and an easy upgrade in the same platform for them.
We're gonna continue to see Sapphire, particularly growing in those highly differentiated workloads, AI and security. Sapphire will provide that broad portfolio of capabilities in terms of enterprise networking, data center, and edge use cases. We continue with Ice Lake, Sapphire, Emerald, and you'll see that multi-generational ramp across our products in the next couple of years as we get to 2024 with both Sierra Forest and Granite Rapids.
Yeah. Just to tie it together, you know, the 76 does not assume heroic assumptions on the part of any of the six businesses that you saw. All right, as I say, we're comfortable in that. You know, we do believe we're growing a little bit slower than the market in the data center for this year. Obviously, the roadmap that Sandra has laid out, we believe that we're, you know, doing better as the roadmap gets stronger over that horizon. It is a growth business that's adding into the 76 over this year, but in a modest way.
All right. Kenji?
Hi, it's Joe Moore from Morgan Stanley. I've been asking you variants of this all morning in the sessions, but where is Intel today from the capabilities of running a foundry business? You've talked about the process being in very good shape, but in terms of the broader organizational structure that you need to support foundry customers and the pricing and the manufacturing flexibility, you know, where are you with that, and how can Tower help you with that?
Randhir, do you wanna start on that one?
Yeah.
Yeah, I can. Thank you. We have, in the last 11 months, built a very strong organization. We have more than 60 experts that came with foundry experience. They are part of my management team, making decisions and interacting with the customers. Our customers, who we engage with, made the comments that, you know, this is the best foundry engagement experience they're having. So, we were very, very aware of this, and one of the first thing Pat said is, "Let's make sure that we build our core capability, focus on getting the organization set up." Secondly, as we close on the Tower acquisition, we will have about 6,000 experts who will be joining our foundry team.
I think I look at the time and where we are in terms of growing our capability, we feel pretty strongly about it.
Yeah. If I just add a little bit, Joe, to it. You know, while you know, sometimes you describe M&A as tech and talent acquisitions, this is a tech and talent acquisition in some ways because we're adding a bunch of technology to our portfolio and a whole lot of talent to answer that exact question. Because we do believe we need more foundry DNA. That's an important part of what led us to this conclusion, but we also bought a great business. You know, $1.5 billion business that we think we can grow rapidly from it. We believe we can expand the margins of it. Our supply chains are gonna benefit this. Our manufacturing capacity is gonna benefit this. Tower Semiconductor not only helps grow our foundry business, but we're already.
I'll say, you know, from the customer perspective, in the 48 hours since we have announced this, I'll say we've been somewhat overwhelmed by the amount of customer inbounds that we've already gotten, about adding this to the Intel portfolio. We see it as tech, we see it as talent, as a great business, and one that we can grow in a very, meaningful way over the horizon. We're quite excited about it.
Brooke?
Thanks, gang. Chris Danely from Citigroup. Pat, thanks for rolling everybody out here. It's like the Intel Brady Bunch up on stage.
Woo-hoo. My team.
Yeah. A philosophical question. You came out last year, you know, right out of the chute, it was, "We're gonna manufacture here. We're not going to TSMC. We're gonna get back to basics. We're going to fix manufacturing." It was like this back to basics Pat. Then in October, you announced, "Hey, we're gonna be a growth company." It was like back to basics Pat went to make it rain Pat. I guess, you know, you guys have this pretty big manufacturing crisis. Why not fix the manufacturing crisis first or at least get to parity before going after all these growth opportunities? I'd just appreciate your sort of thought process there.
Yeah. You know, I do think there's an aspect to that of walking and chewing gum at the same time, where, you know, we have to fix the foundations, right? We have to get the execution. We have to get the manufacturing. We're on path on the process technology. But even if we just like build on the last question, how long will it take a foundry customer to come into our factories? That's a three-year process if it's on a leading edge node. You have to start those engagements somewhat presumptively, right? Way ahead of time to start building out to some of those revenue streams if they're gonna materialize in 2025 and 2026. This is a long cycle. You make these bets long. The conviction that you bring to those decisions because they don't move around that fast.
This is long development cycles, long product cycles, as well. You know, we also fundamentally believe that Intel is not just a manufacturing company. We deeply believe in the technology and the manufacturing, but we're also a world-class product company as well. With that, you know, areas like Raja's business as we've talked about, you know, extraordinary opportunities that we've been working on for several years. You know, this is Raja unleashed into a market that is begging us to come in. You know, it's products that we've been working on for a while and we're now unleashing those capabilities. You know, the Mobileye business, you know, we're unleashing the Mobileye business. You know, unquestionably the AV transition is underway.
You know, this would be foolish of us from a shareholder perspective to not do what we've laid out in that area of the business. You know, we've, we're working in the area of networking for quite a while, but we've never assembled those assets in a way and now we have a world-class leader in Nick to go drive a fundamental network transition. You know, think where we were three years ago on networking, right? We're worried about Huawei and what's gonna happen in the world. That's gone. It is an open FlexRAN-based architecture that Intel created. Of course, this is something we must do in this period of time. As we look at all of those, you sort of say, "Wow, are we doing too much?" Well, I'd say in many of these cases, these build on each other, and that's the point of these business units.
They build on each other, they reinforce each other, they benefit each other, and we believe that we're trying to carefully manage that right balance of the right strategy with the right execution disciplines, with the right organizational structure. We believe we're well on the way to do exactly that, fix the foundations, launch growth businesses, and we're gonna show you the proof points this year that that is happening at scale in Intel and for the industry.
All right, let's go back over this side of the room. Tim?
Yeah, Pierre Ferragu, New Street Research. I think someone at Intel, I'm not sure, I didn't check like the primary source, but one day someone at Intel said something along the lines of strategies, capacity. The way I've always understood that is very simple, almost brutal, which is that if you have your capacity, and if you have more capacity than your competitors, excess capacity, you have access to very, very small, very, very low marginal cost, and that gives you a lot of competitive power to push back on your competitors. My question is that the way you understand it? Do you intend to play on that?
Within your guidance and the guardrails you have in your economic model, do you have room for that?
Keyvan or me?
Well, Keyvan, why don't you-
Yeah, no.
I think Keyvan should start and then David can come up.
Well, we live by that comment of capacity is strategy because when Pat says we're going to Groveian culture, that's the first lesson Andy Grove taught us, which was, look, one thing is the product leadership, the other one is you get your technology leadership, you have the product leadership, but you don't have capacity, and that's exactly what we are doing. We're building the foundation, which is essentially the smart capital that both Pat and Dave laid out, and that starts with having the shell space, which is a smaller portion of the total investment. The lion's share of that capital investment is really for the equipment, and that comes a little bit later, right? That's when we have all the capacity triggers.
What our strategy is precisely, let's make sure that as we are marching forward with this leadership roadmap, that we have the capacity to serve all of our customers all around the world, both the IDM as well as the IFS. We are not gonna be limited by capacity, and that's my job to make sure my friends here are not gonna come and say, "Keyvan, we have the customer, we have the leadership. Where is the capacity?" That better not happen.
I'd just add, you know, Keyvan and I are, first of all, locked arms on what we're going to do.
Yeah
We are confident in our model that this is the right model to deliver the right level of capacity for what we're gonna see in terms of demand, and we're gonna leverage the heck out of smart—this smart capital strategy, quite honestly. I'll go back to my original comment, which is that, you know, the middle bar of those which were capital offsets, we've only planned 10% in the model. We're shooting for 20%-30%. I have high confidence we're gonna do much better than 10 for sure. That also gives me, it gives me and it gives Keyvan some room to work in to make sure that we're not caught flat-footed in terms of capacity.
Yeah.
Ben, did I see one on this side?
Yeah, this, Pat, earlier you were talking about N3, N4 cadence, and to paraphrase Henry Ford, in the good old days you could have any cadence you want as long as it was Intel's cadence. It sounds like, did I hear you correctly that it sounded like you're gonna begin to modify cadence based upon different customer needs and different segments? If so, does that basically mean that IDM 2.0 is fundamentally more customer-centric than IDM 1.0?
You know, I'll ask Ann to jump on this as well. I mean, obviously in what Ann described, it's a more modular view of how we do process technologies. Clearly, Foundry has created, I'll say, a broader portfolio that we're trying to operate against, with this as well. You know, we've seen some of these capabilities actually in the product world in what we're describing as well. As we open up 3D packaging, we have different tiles on different process technologies to create more optimized product costs and feature capabilities. You know, it is in a hybrid disaggregated world. We have more flexibility through packaging, and we're also creating, as Ann talked about, the modular PDKs. We have more flexibility in the process technology recipes as well.
All of these taken together, we're gonna be far more customer-oriented in our technology development, both internally and for our foundry customers as well. Anne, if you'd-
Thanks, Pat. Dan, one of the things I'll add on, I spoke earlier this morning around our tick-tock-like model in terms of development. Our single major change will be in the Tick, and the next step will be in the Tock. The Ticks are every two years, and the Tocks are every two years. We're maintaining a key cadence in that while we're delivering the flexibility to enable both our foundry customers as well as internal customers.
Kenji? One from that side. If you don't have any, we can move back to the other two, or.
Thanks. It's Matt Ramsay from Cowen. Thanks for letting me jump back in. This question's for Pat and for Greg. You talked a little bit today and in our sessions this morning about more directly monetizing software. Maybe you could talk about how that fits in the five-year model that Dave presented, and if you have any targets there, that would be helpful. Thanks.
Yeah. Greg, you wanna start, and I'll pile in on that?
Yeah. First off, thanks for asking a software question and making me feel validated. I talked in my sessions that right now we sort of book a little over north of $100 million in software revenue, and that's without us really trying to sell it.
I think we have a great opportunity to actually just, as I said, go to the top of the stack and pull through the value that Sandra talked about and Pat talked about from the hardware capabilities we provide, particularly in the security and AI area, right, and deliver more of a trusted computing platform and the key-to-SaaS services that we need to achieve that, leveraging, you know, literally, I mean, tens of thousands of engineers worth and decades worth of investments in the open source ecosystem that the cloud vendors are currently monetizing. We can do the same thing. I think, you know, Pat's given me some guidance on sort of like walk, crawl, run. I'm sorry, crawl, walk, run before we get to the full path.
This year I'm gonna increase that to about $150 million, but we think we can do more over time.
Yeah. We also see it, Matt, as this critical differentiator for essentially every business that we're in.
You know, if you look at Mobileye, you know, it's a set of TOPS with a highly optimized software stack on top of it, and that's what's creating that very differentiated platform experience. We just acquired Screenovate into our client business to start driving the ecosystem experience for the PC, and we're gonna monetize that software, right? You know, Nick has you know, a very large set of his engineering is software associated with the stack that we have, you know, for our FlexRAN offerings as well, and those are gonna be licensed software capabilities. You know, Sandra has started to create a SaaS model for some of her features that are being pulled through in the cloud area. You know, Granulate, we just acquired to start having SaaS services associated with cloud gaming environments that we're hosting in a cloud environment.
You get these examples across every business that we're gonna begin layering in differentiated software that pulls through our platform value and creates SaaS service revenue on top of it. I'm gonna be holding us, Greg accountable to drive an overall SaaS and software strategy for the company. Again, on the $150 million bucks, you know, it's not much, right, at that level for a company our size. We're gonna grow it rapidly, but we're gonna pull through platform differentiation, right, which is gonna very disproportional business benefit even though the near term revenues are rather small.
Brooke, did I see one on this side?
Hi, Bob O'Donnell with TECHnalysis Research. Another software related question. Greg, you can feel better. Also for Raja and Pat. You know, in the past, Intel's talked a lot recently, in fact, particularly around XPU. Of course, oneAPI is strategically essential to do multiple XPU types of things.
You know, you didn't talk a lot about that today, and I'm kinda curious if that represents some sort of shift, because theoretically, oneAPI drives the ability to have multiple chips in each of these different devices. Curious to hear a little bit more about how you're each thinking about that element.
Greg, do you wanna start?
Yeah, I'll start there 'cause I did cover that in my session, in my two sessions. We continue to invest in the oneAPI system ecosystem. It's a collection of technologies, you know, oneDNN for deep neural networks, oneMKL for our accelerated math kernel libraries, you know, primarily for Xeon. But also working with Ponte Vecchio and our whole graphics portfolio 'cause the value that we've done is that all that technology, and we pushed out oneAPI 2022 in December with over 900 new features, over 42 different technologies. We've seeded all of that within the, for example, just the AI ML ecosystem, PyTorch, TensorFlow, ONNX from Microsoft. I mean, all the major PaddlePaddle, all the major AI systems, we've already plugged in all the oneAPI accelerator technology.
In Q3 alone, we had 8 million downloads of TensorFlow with the Intel accelerated libraries from the oneAPI ecosystem. Our strategy there, besides, you know, having a compiler that you can get and use, in fact, it's now our compiler from oneAPI is now available in the Microsoft Visual Studio marketplace. You can do a one-click download and install and replace the Microsoft compiler in that ecosystem and get better performance, for example, on Alder Lake if you use it to compile your game software. We've made all these investments and we've seeded the ecosystem. Now the value of that is, as we deliver new hardware, those systems are already primed, ready, and optimized for acceleration. We may have to do some additional tweaks to accommodate new accelerators or new hardware features to get peak performance.
By having that oneAPI ecosystem broadly deployed and embedded in every critical workload, not just AI ML that anybody wants to run, including OpenJDK for Java within the Python ecosystem, we've just got ubiquitous acceleration across all of our platforms, and that's really the value of that technology. A lot of work's been done over the last two years to seed it into the ecosystems that everybody else is using. I read one report, there's 24-25 million developers in the world.
IDC reported that 20 million of those developers are yet leveraging and utilizing Intel-accelerated technology, and many of them don't even know it.
Maybe, you know, since our networking guy might feel otherwise not loved as well. The XPU question, you know, maybe just talk a little bit about.
Sure
IPU as well.
Yeah. The IPU, our infrastructure processing unit, that we introduced last year, is really used for those who are building large data centers in order for them to run the infrastructure code that runs the data center inside that IPU on a sort of a separate, isolated set of cores so that the tenant workloads don't bring down the infrastructure. Now, this is a really critical piece for how data centers are gonna be built in the future, and we really wanna be able to open up that ecosystem, right? In order to be able to do that with the first introduced cloud scale commercial offering of a component like this that's really gonna transform the way that they're built, we wanted to make sure that that was open.
In order to do that, we're opening up software and getting behind an open source ecosystem called IPDK, the Infrastructure Programming Developer Kit. By getting behind that, we're gonna help drive an open ecosystem so that our customers don't get locked into a single solution. We just think that serves the industry better in the long run and sort of plays into our strength at Intel.
All right. I think we have time for one more question. Vivek, one question. Vivek.
All right. Actually, I had a somewhat just a financial question. I think, David, you gave us net CapEx and gross margins based on some net depreciation based on some unknown level of incentives from the government.
What are you saying?
I was hoping, right, if you could give us, you know, what is the gross CapEx? What is the depreciation? And what level of government incentives are you assuming? And will there be any restrictions, if you take those government incentives?
Yeah. I did give you. Yes, that is a net CapEx number. The gross I gave you also, though, that implied in our model was a 10% offset. You know? That was just, that was more than just government. That was government partners, and financial partners. Those three combined were 10%. That's what we assume. You can kinda back into the gross number that, you know, you'd get in every quarter or in every year. You know, I mean, yet to be determined exactly what gets written into the rules around restrictions. Early to say. You know, typically, there comes with some form of restrictions, but it's things that we can generally operate within the guides of.
Yeah. I'll just add, you know, I mean, we're very closely involved in U.S.-EU CHIPS Act, if we pick. You know, if we look at those in particular, we're very comfortable with provisions that are being associated with those funding offsets. We don't think it comes with any concerns in our ability to go execute our business. The reason we focused on net CapEx is because that's what's gonna flow through to depreciation, right? In that sense, we think we're giving you the right number, the pure number, you know, so that you can look at the business of what the real impact on our depreciated assets will be and how it will affect margins over time. That's the right number to look at.
We'll of course be updating as we hope to well achieve, and as we said, we've committed about 10% in the business, a very low number compared to what we see out there. We're hoping to achieve 20%-30%, which is gonna drive a much, you know, a lower CapEx implication, much lower depreciation over time. All of those things will flow through in very positive ways through the entirety of the business case that Dave laid out.
The other thing, Vivek, when you're modeling depreciation, just keep in mind, when we're doing that CapEx, there is an amount of that CapEx, a percentage of that CapEx, that is depreciated over a fairly long life.
Right.
It's, like, 20 years. It has less of an impact in any one year in terms of the depreciation.
Thank you.
Yeah. Thanks.
All right. Pat, do you wanna close us out?
I do. Thank you. First thing, thank you for joining us. You know, we appreciate that you spend the time to learn more about our business. You know, we have started this business, and as I described, right, the turnaround train is leaving the station. You know, we are well underway. At the one-year mark, we have accomplished a tremendous amount, but we also, in humility, realize we have to rebuild your confidence. I hope in the course of today that you have seen the strategy is taking shape, the execution is building against it, our commitment to leadership, but also transparency, so that you get to see the business that we're building and understand the business units, the implications. We've given you guardrails we're gonna manage the business against.
I have the honor of being the CEO of the world's greatest technology company and being the leader of the world's greatest technology leadership team, who gets to participate with 121,000 people that are on mission for this company, and we just hope that you see the opportunity to join the turnaround train of Intel, 'cause we're going incredible places together. Thank you for joining us today.