Good morning. And on behalf of the Lam Research Management team, thank you for joining us on our 2020 Investor Day. Thank you to everyone on the webcast as well and for those of you who are watching it on the replay. My name is Tina Correa, and I'm the Head of Relations at Lam Research. I'm going to first start off by covering our Safe Harbor statement.
Today's presentation and discussion include forward looking statements that are subject to risks and uncertainties reflected in the risk factors disclosed in our SEC public filings. In addition, today's presentation and discussion include financial results presented on a non GAAP financial basis. Please refer to the statement on the screen for additional information. A detailed reconciliation between GAAP and non GAAP financial results can be found in the appendix to today's presentation. I'm going to take a few minutes to give a quick overview of today's agenda.
Starting off this morning will be Tim Archer, our President and CEO. We are also joined by 4 members of our executive team that will cover a few exciting developments about their respective business units. Pat Lord, who leads our customer support business group or CSBG, will cover the sustainable growth of our installed base business. Vahid Vahidi, Head of our Etch Product Group, will talk to you about how we're extending our Etch leadership with a groundbreaking new Etch platform. Sesha Ordarajan, Head of our Deposition Product Group, will give an overview of the differentiated solutions we have for atomic layer deposition or ALD.
Our Chief Technology Officer, Rick Goccio is going to present on a new technology we have developed to enhance EUV solutions. And finally, to cover our financial overview, we have our Chief Financial Officer, Doug Bettinger. The prepared comments will last approximately 2 hours, and we will have a Q and A after the presentations. Please hold your questions until we get to the Q and A. And one final logistics item, there will be lunches available for you at the end of the presentation.
I am now pleased to welcome to the stage our President and Chief Executive Officer of Lam Research, Tim Archer.
Great. Thank you, Tina, and welcome, everyone. I know there's a lot going on in the markets right now, so I do appreciate that you're joining us today to hear about our long term growth story and why we think it's an amazing time to be in semiconductors. I've been in this industry for over 30 years and I can truly say that I have never seen a more exciting period of innovation and opportunity. At Lam, we are driving the semiconductor breakthroughs that are at the very core of the digital transformation of our society.
And today, we will frame for you our strategy for outperformance in this era of unprecedented end market drivers. As you just heard from Tina, we have an outstanding slate of presenters for you today, who will give you very specific examples for why we believe that our product portfolio and our development pipeline have never been stronger. And why we are confident that there is no company better positioned for the coming inflections in semiconductor manufacturing than Lam Research. We're hosting this meeting at a very special time. You might have seen on the opening video, this is Lam's 40th anniversary.
And our success today is a culmination of investments we've made and experience we've gained over the entirety of those 4 decades. Of course, a lot has changed over all those years. Electronics have become integrated into nearly every aspect of our lives. Semiconductor devices have become dramatically more powerful and manufacturing processes ever more complex. As just one measure of how far we've come, memory bit density and transistor logic counts have increased by more than 5 orders of magnitude since Lam launched its first etch system into the market.
And today, we see this change only accelerating as new semiconductor applications are being created every day. And this embedding of electronics and semiconductors into everything we do has had a profound effect on the dynamics of our industry. In fact, I have never seen so many new end markets, device types and process nodes all growing at the same time. Yet despite this dramatic growth, we actually feel we're only at the very early stages of what is possible as both industries and consumers harness the power of semiconductors to generate, store and analyze data for social and economic benefit. Said a different way, data has become the fuel of the new economy.
Semiconductors have never been so vital to so many industries or the global economy as a whole. You can see in this chart how silicon wafer area processed has dramatically grown relative to real GDP over the last 30 years. And we believe that the long term fundamental dependence of the $85,000,000,000,000 global economy on semiconductors will continue to drive secular growth that transcends short term supply demand cycles. Now, I'm not saying that cycles are going to disappear. So if there is one takeaway you write down from this meeting, that's not it.
So please don't quote me on that. Cycles will still exist because as an industry, we will never get everything right all the time. But the important point is that strong secular growth in semiconductors compounded with the challenges of increasing complexity and device scaling challenges will drive sustainable growth in WFE across those cycles. If you look at the last 5 or 6 years, we have seen this powerful combination of forces come together to cause WFE to inflect significantly higher. From the chart, you can see it wasn't too long ago that we were talking about WFE spending levels in the mid-thirty billion dollars range.
However, today, as new end markets are being created at an ever faster pace and process complexity continues to grow, we have seen WFE rise to what we believe is a new higher base of spending. And in our view, a $60,000,000,000 WFE in the next few years is not unreasonable to assume. But what is most important for Lam Research is how we execute against this positive backdrop of rising WFE. Our guiding principles running the company are fairly simple. We put our customer success first.
We prioritize R and D spending to create best in class products and we build customer trust that spans the entire product lifecycle. While this strategy sounds pretty simple, our differentiation has been in our disciplined execution. If we look at the 6 calendar years ending in 2019, this is the 6 full calendar years since we brought Lam and Novelis together as one company. WFE has grown at a compound annual growth rate of approximately 8%. Over that same period, Lam's revenue has grown 2x at 16% per year.
From our perspective, etch and deposition have been the best markets in semiconductor capital equipment, And they will continue to be as three-dimensional architectures and new materials play an increasingly important role in every future device road map. And I believe that investors are beginning to recognize the long term value of Lam's leadership in these important etch and deposition markets as well as our track record of outperformance that I just showed. In Callard 2019, Lam was the 2nd best performing stock in the S and P 500. And over the last decade, our market capitalization has increased nearly 15x. So I want to thank you to thanks give thanks to everyone who supported us on this incredible journey.
And what we hope that you take away from the presentations you see this morning is that we believe this incredible journey continues for the industry and especially for Lam. The framework for the rest of this morning's talks will actually look quite familiar to those who saw my presentation at Investor Day in 2018. At that meeting, I laid out a strategy for outperformance that I believe has been both effective and resilient. It had 3 components: 1, focus on revenue capture from our expanding installed base to create a foundation for sustainable growth 2, gain market share by creating best in class products that deliver advanced technology ready for high volume manufacturing and 3, expand our market share by identifying and leveraging new technology inflections. One of the things that I really like about this strategy is that each of the elements is complementary to the others.
If we support our installed base well, our customers are happier and the likelihood that we'll gain market share increases. If we gain market share, our installed base grows faster and so does our opportunity to increase our annuity. And if we identify new markets and expand our SAM, we can gain share and increase our installed base at an accelerated rate. Because of the synergistic nature of each of these three elements, you will see in our talks this morning that we're actually committed to balanced success across all three elements. We're going to start by talking about the installed base.
I'm often asked what is something that not all investors understand about Lam's business. And I typically respond that many people still do not fully appreciate the power of the installed base business to drive cross cycle stability and growth. What is often missed is that our installed base grows every year. It may grow a bit slower or a bit faster depending on WFE in any particular year, but the installed base will grow every single year. You can see this on this chart, where in 2019, WFE fell by about 10%.
And yet in that year, we added approximately 5,000 more chambers to our installed base. That's 5,000 more chambers that will generate revenue for Lam by way of spares, services, upgrades and refurbishment for the next 20 to 30 years. And by investing to create innovative new products and services that help our customers extract more value from the assets they've already purchased from Lam, we end up with happier customers and we're able to extract more revenue from every tool in the installed base. And you can see that between 2013 2019, the average annual revenue per tool in our installed base has increased by approximately 50%. And as our customers benefit as well by way of reduced capital spending and lower cost from these investments.
We believe that as manufacturing complexity continues to increase, we have an even greater opportunity to grow our installed base revenue and deliver differentiated best in class support for our customers. And so to talk more about our installed base business, I'd like to invite up Pat Lord, our Senior Vice President and General Manager of our Customer Support Business Group. Pat, come on
up. Thanks, Deb. My name is Pat Lord. So good morning first. It is my pleasure to be here to tell you a little bit more about our customer support business group at Lam Research.
CSBG for short is really the business post equipment shipment and installation that focuses on driving our customers' install base of LAM tools to world class levels. It's really about delivering an experience to our customers that builds on trust. Customers it's a simple premise, customers that have a good experience with LEM tools are more likely to buy more LEM tools in the future. What we're showing you here is a measure of this customer trust. It is how our customers are ranking us using their own scorecards among all of their suppliers.
What you can see is that we systematically are being ranked pretty high and not only have we made progress over the years, we're able to sustain that level of performance with our customers, continuing to build trust. It is a rather sticky business if done right and it's only sustained by continuously and proactively delivering the maximum value out of the installed base to our customers. And it's not just about delivering it once. It's really about continuously doing this year after year and for decades, if you want to build that trust with our customers. So how do we do it?
We're not simply in the customer support business. We're really in the business of making our customers tools perform at a different level in yield, output, predictability and cost per wafer. We're in the operating solutions business across spares, services, upgrades and Reliant, our non leading edge equipment business. As Tim noted, some of the tools have been in our installed base for over 30 plus years And each is requiring annual spares and services, but also upgrades as well as refurbishment and repurposing towards new application over the entire life cycle of the equipment. And it's only by adding substantial value in these 4 business segments that we become more of a strategic partner, strengthen customer trust and then are able to share in the value delivered to our customers.
It's fundamentally monetizing that value and that customer trust. So as a result, CSBG is more of a business based on the annuity created by the utilization of the overall LAN install base and the value of the solutions we offer to our customers than on the cycles of wafer fab equipment investments in the industry. So let me talk about each of these business segments and I will start with spares. It's more spares is more than just having the right part in the right place at the right time. Yes, we do have a full portfolio of spare solution across our consumables, our non consumable parts, electrostatic chucks, including parts repair, refurbishment, re cleaning and recoating.
But where our spares offerings bring significant value to our customers is in 5 areas. It starts with reduced lead time. We offer a very specific contract with guaranteed availability and delivery timing to enable speed to our customers. It's about helping our customers manage their OpEx with predictable spending, comprehensive program based on RF, minute utilization of our tools or cost per wafer pass help our customers predictably work on their OpEx. The more they use, the more they pay, the less they use, the less they pay.
It's about giving our customers the ability to optimize the capital spending they've invested in Lam and in their fabs. We deliver spares with technology in the manufacturing that help extend lifetimes and thus extend the time between required maintenance. In terms, this dedicates more time towards production, increased output and therefore significant tangible productivity and cost of ownership value to our customers. It's then also about obsolescence engineering. Lam spends quite a bit of engineering effort ensuring that this equipment will continue to work and perform throughout the entire lifecycle of the equipment.
You can imagine over a 30 plus year lifetime for a piece of equipment, how important that can be for our customers to ensure that their productions are never interrupted no matter where they are in their product development over time. It's all about reduced risk. It's guaranteed quality and availability, especially when our customers are ramping and they don't want to worry about disruption in their supply chain. Overall, these five items are what our customers value most and it does help secure for Lam an annuity stream from each of our tools in the installed base. It's an annuity stream that keeps on increasing year on year, in part because the installed base is growing every year and we keep on adding tools to the installed base, but also because at the leading edge of memory, foundry and logic, the manufacturing challenges provide an opportunity to create spares offerings with embedded technology in them that extend lifetime as well as on wafer performance with defect quality and yield as a result.
So transitioning to service, our focus is obviously on addressing our customer needs. Not only do customers expect routine maintenance to be done correctly as well as expert troubleshooting talent from us, but they also need Lam's help with tool domain knowledge expertise to optimize productivity and yield. To address these challenges, Lam offers a suite of data and analytics, machine learning algorithms and artificial intelligence knowledge management solutions we refer to as equipment intelligence. It starts with our tool data optimization, the right data in the right place. It's all about big data and accessing properly curated information.
Then it's about knowledge management, be it installation, maintenance or troubleshooting, it's about aggregating the entire installed base knowledge and enabling real time point of use access of this information for our customers for our field service engineers. We do this through our mobility tablets and we call this the digital worker optimization. The next step is ensuring our customers are able to maximize the output from the fleet of LAMP tools that they have. It's fundamentally allowing them to achieve their entitlement in terms of production and we deliver this through simulation that automatically maximize the output of the tool depending on the scheduling and load that our customers are putting on our tools. Obviously, our tools mainly require routine maintenance.
It's all about doing it right the first time and doing it repeatedly, minimizing human variability, especially for all troubleshooting and service tasks. Our customers have many tools. They usually have fleets of tools. When you have 400 chambers or more, it becomes very difficult to manually manage these chambers. It's through the use of our data rich environment and AI enhanced multivariate analysis that it is then possible to match these fleets, make sure that they're performing optimally and then do predictive maintenance as necessary when you detect that the tool is not performing correctly.
Overall, this finally creates a predictable and stable environment that help us then leverage advanced technologies such as advanced process control or virtual metrology to optimize the process and ultimately on wafer results for the most challenging processing applications. Just as an illustration, a tool that does not qualify back to production after maintenance is extremely disruptive to fab output. It's about doing it right the first time. To this end and to ensure predictable maintenance, Lam has been implementing cobots that can be rolled into the fab, docked to our tools, working alongside our field service engineers to minimize human variability for repetitive and sensitive maintenance operations. What you're seeing in this video is 1 such cobot taking on the task of securing a flex tool chamber top plate assembly.
This task requires precise torquing of over 40 bolts in a very precise patterns and you need to do it 4 times for each bolt, 160 measurements and 160 actions that are very important not only to ensure that the chamber ultimately does not leak, but also to ensure that the impedance is correct for the chamber and therefore the on wafer results are what we want and what we expect and match to the rest of the feed. Cobots are ideally suited for these types of maintenance activities and they ensure repeatable, predictable quality maintenance on our tools and thus ultimately enhance fleet matching. What that means for our customers, it's really enabling optimal quality, maintenance cycle time and lower true cost for our customers. So I've talked about the impact of equipment intelligence for our customers, but what does it mean for Lam? We are leveraging the use of equipment intelligence to offer products and services that migrate us from standard services like engineer on-site labor contracts, that's your Maytag repairman waiting with a wrench for something to go wrong in one of the tools.
2, more comprehensive results based contracts, productivity solutions as well as predictive smart solutions. These equipment intelligence, productivity solutions and advanced service products truly help differentiate our service offerings and they also deliver higher value and higher return on investment to our customers. So you may ask why is this good for Lam? Well, in addition to delivering a service experience to our customers that's unparalleled and very differentiated, Our equipment intelligence based services help us share in the higher value delivered to our customer. They resulted in a very strong revenue growth trajectory with an improved profitability outcome for Lam.
There is indeed much more value in our domain knowledge based equipment intelligence service solutions and our customers are willing to pay for it. And it also builds on customer trust as our customers then tend to rely and engage on us more to solve their most challenging install base problems. It's an important part again of achieving that top supplier ranking. It's that value that you had in sorting your customers' most difficult problems. The installed base of LAM tools presents an opportunity for LAM to engineer and uniquely engineer upgrades our customers can purchase throughout the entire lifecycle of the equipment.
This comprehensive upgrade portfolio enables our customers to enhance the capability and productivity of their toolset, maximizing the value of their existing investments and enabling solutions to new wafer processing challenges. Upgrades come in 2 flavor, technology and productivity. Technology upgrades extend to a lifetime by upgrading existing equipment capability to meet process requirements for the next generation products. For example, upgrading your TiO FX to the next generation capable TiO GX. Productivity upgrades enhance system output by increasing uptime, throughput and or yield at much lower operational costs.
Again, to illustrate, we are showing you an animation of our Corvus upgrade. Originally developed as a productivity solution feature for our latest generation Keyyo tools, which have now demonstrated over a year running in production without requiring the chamber to be open for maintenance. Corvus can also now be retrofitted to the LAM tool install base as an upgrade. What does Carver's do? Carver's increases production output by robotic replacement of worn out consumables without the need to open the chamber, significantly reducing recovery time from days to minutes.
This is an activity that used to take multiple service engineers over 2 days to do. You would have to take the chamber down, you would have to open the chamber, manually remove the worn out ring, replace the ring, center the ring, close-up the chamber, making sure the chamber then requalifies into production. With Corvus, we're able to transform this operation from a 2 day operation, 48 hours, down to less than 30 minutes. As a result, Corvus delivers over 15% more tool productivity to our customers, more tool time dedicated to production. So this is just an example, but our ability to uniquely engineer these technology and productivity upgrades has allowed us to grow the subgrade business significantly faster than the installed base, actually 50% faster than the installed base.
And we anticipate to be able to continue to do so over the coming years. Last but not least, our Reliant business taps into the opportunity created by a large installed base of tools as well as the technology know how we have developed over the years through our R and D investment in both tools and process development. It's fundamentally leveraging some R and D. Part of the equipment lifecycle includes the repurposing of equipment for different applications. The R and D invested at the leading edge of logic, foundry and memory will find over time a role in the fabrication of chips, both specialized and technology advanced, but not requiring leading edge design roles and lithography.
These chips are part of the specialty technology market, driven by ubiquitous applications in wearables, smartphones, virtual reality, artificial intelligence, augmented reality, smart homes and automotive, especially with the connected car. Just as a reminder, for every smartphone main processor built at the leading edge of the foundry nodes, there are in order of 30 to 50 incremental chips from the specialty technology market that address specific applications, be it we'll start with power management with IGBT and BCD devices, Everything has a battery, whether you're charging it or discharging it, you want to make sure that it's managed correctly and done properly in an optimal way. Sensing, with MEMS and CMOS image sensors, obviously phones with multiple cameras, some with face recognition capability, sensors like gyroscope, accelerometers, fingerprint readers and microphone as well as now the evolution with communication RF MEMS with the adoption of 5 gs. Finally, a lot of these peripherals and accessories need to be controlled with specific dedicated chips and this is usually done with ASIC and MCU dedicated devices. So this market has grown actually significantly faster than the installed base.
I don't say it's grown significantly faster than the wafer fab equipment over the past several years. And we think it will continue to do so at a rate of 2x to 3x the overall wafer fab equipment industry over the next several years. And this really creates a strong growth opportunity for our Reliant business. So what does this all mean for Lam's CSBG business and for you, our investors? What we are showing you here is the normalized growth of our installed base as well as of our CSVG business as of 2019.
What you can clearly see is that by continuously and proactively building customer trust with differentiated install base experiences and through the added value that we're offering through our spares, services, upgrades and Reliant, I just described, we have been able to grow our CSBG business faster than the installed base over the past 6 years, and we anticipate to be able to do so moving forward. Actually, we think we can grow this business at least another 40% by 2023. So I hope I was able to provide you with a little bit more insight on our CSBG business here at Lam and I'll leave you with a steady growth business based on our install base and customer trust, less susceptible to wafer fab equipment investment cycles, growing faster than the install base because of the increasing annuity per tool we're able to engineer through our unique spares, services, upgrades and Reliant offerings. And hopefully with some confidence that we can grow this business over 40% by 2023. So with this, I thank you for your attention and we'll turn it back over to Tim.
Great. Thanks, Pat. The CSBG business that you just heard Pat describe, it's an outstanding part of our company. As you just heard, it establishes a strong foundation for stability and growth. It's a business that we expect to grow every year.
Very importantly, our customers benefit from the investments we make to improve the performance of our installed base. And just as important for all of us, it's a business we expect to grow revenue more than 40% by 2023. So it's again, it's a very important part of our outperformance strategy, and I hope what you heard today was helpful for your understanding. So now I'd like to move on to the second element of our strategy, gaining market share. And I know that in this area, investors are often skeptical about companies' plans to gain share.
It's tough to do. Every company says they're going to win until they don't. But I think here again, Lam has built a credible track record and we'd like to explain how we approach this challenge. First, we begin every year with our top priority being to defend every position that we have worked so hard to win. And you just heard from Pat about the tremendous learning that we gained from our installed base.
If we can take that learning and we can feed it back into our CEP programs and we feed it back to our development teams, we can accelerate the performance improvement of our installed base relative to our competition. So what is it that we need to do to be able to execute this strategy? 1st and foremost, we need to run a lot of high volume production wafers And that's exactly what we do. Take, for example, 3 d NAND. On this chart, if we total up the cumulative wafers run through the 3 most critical applications required to build a 3 d NAND device, LAM etch and deposition tools have run more than 90% of all the high volume production wafers.
This has helped us improve our tool performance much faster and it has created a very deep learning mode around our most important applications. But obviously, our objective every year is to do more than just defend our existing positions. And so we approach market share gains from 2 different angles. First, our customers generally view Lam as a go to supplier for their most critical applications. They look to us to solve their most difficult problems.
And the more we know about those problems, the more effective we can be. I just showed you how we have cumulatively run 26,000,000 more wafers than our competition through the 3 most critical applications required in a NAND device. As a result, we are the supplier first to see the problems on the horizon and this gives us a very important head start to design in new capabilities, to improve our performance and to position ourselves for further share gains. In the past year, you have seen this lead to innovations like our Corvus Edge Tuning solutions to improve yield and also our Vector DT to compensate for wafer bow and 3 d NAND. We think with greater manufacturing complexity coming, our opportunities to gain share in etch and deposition by leveraging this learning in critical applications is available to us.
But this increased complexity, it often leads to cost, and this creates another opportunity for us to gain share in the less critical applications. Throughout this year, I talked about our increasing focus in the semi critical space. When you look at the cost to build 128 layer 3 d NAND and beyond, it is becoming increasingly difficult for us to stay on this node to node cost reduction curve that has been the cornerstone of our industry's success. So while this is a challenge, it's actually also an opportunity for us for share gain. Disruptive productivity solutions are needed to help us stay on this curve.
They are needed to drive significant cost reductions in the less critical layers in the device. And so when we look at these opportunities, both in critical and semi critical, you actually see that we view our opportunity in both to be relatively balanced, an opportunity gained by continuing to solve our customers' most critical challenges and an opportunity gained by delivering disruptive productivity solutions. But to fully capture the opportunity in both segments, we actually have to be willing to disrupt our markets with great new products when the time is right. And we see every device roadmap across NAND, DRAM and foundry logic right now calling for greater etching capability with higher productivity. And so today, we stand here as the market leader in etch with the broadest and strongest product portfolio.
But what we're very excited to show you next is how we intend to widen that lead with the first completely new etch system from Lam in 20 years. So I'd like to invite up Vahid Vahidi, our Senior Vice President and General Manager of the Etch Product Group, to tell you all about this new tool. Fahid?
Thank you, Tim. Good morning, everyone. Very happy to be here and very excited to have the opportunity to introduce our new X system to you. But before I do that, I'd like to share with you a little bit of the Lam history and my own journey with Lam. I've been at Lam for over 20 years, and I still remember when we introduced our current edge system, the 2,300 to the market.
At the time that we introduced it, it had a very groundbreaking architecture and it allowed us a lot of flexibility. With that flexibility, we're able to deliver a lot of enabling solutions to our customers in all segments of the market. Our customers were happy, they rewarded us with more applications and we shared gain. Well, I'm here to tell you we're about to do it again. Our other ingredients of success in EDGE included partnering very closely with our customers in R and D and developing best in class products.
Partnering closer with our customers in R and D gives us a lot of visibility into the new technologies that are required as well as future inflections. And best in class products, it's all about delivering enabling solutions to the customers in time or ahead of when they actually need it, so that we can stay ahead of the curve. What you see on this slide are a couple of examples of these enabling solutions that we have delivered, such as hydro patterning system that enabled our customers to transition from 2 d from double patterning to quad patterning and high aspect ratio etch capabilities with KEO and Flex that enabled our customers to transition from planar to 3 d NAND. The name 3 d is used very much synonymously with 3 d NAND. But as Tim said, in reality, all segments of the market DRAM, logic and NAND are going to 3 d to take advantage of the vertical scaling.
This vertical scaling is a very strong driver of growth for us, and it requires a very precise high aspect ratio etch capability. As those of you in the audience know, and as you're familiar with the physics of etch technology, the higher the aspect ratio, the more difficult it is to deliver the precision and the capability for that high aspect ratio etch. To expand on what Tim said earlier, let's look at one of examples of these high aspect ratio etches, memory hole etch in 3 d NAND. For our customers that today manufacture more than 90 layers in production, And memory hole etch for that structure amounts to etching a hole to an aspect ratio of 50:one. That is 5 times higher than the tallest building in the world.
Now let's remember that these holes have a diameter of oneone thousandth of a human hair and we have to simultaneously etch more than a trillion holes on each and every wafer. And as you know, Lam is a very unique position and we're the only ones that do this etch in production. And we are running more than 1,400,000 wafers per month in high volume manufacturing for this particular application. So you can imagine the learning that we get from this etch in high volume manufacturing gives us a lot more knowledge about what's required to run this type of application in high volume manufacturing. I mentioned to you that we have very close collaborations with our customers as we develop the next technology nodes.
We mentioned about the etch challenges that are ever becoming more complex. The aspect ratios are getting taller, selectivity requirements are going up and the dimensions are shrinking faster and faster. Our customers are developing their products and they want to transition it from R and D into high volume manufacturing as fast as possible. The cadence in 3 d NAND is actually not slowing down, it's speeding up. And customers want to put their products into production as fast as possible, in order to recoup their R and D investment.
So they need a tool that has a lot of flexibility in R and D, but they can transition it to high volume manufacturing as fast as possible. And as the etch intensity increases, the fab space associated with these tools is becoming ever more challenging. And these fabs, as you can imagine, cost a lot of money to build, and the fab managers want to optimize and to get the maximum wafer out from these tools. So building on the successes of the past 2 decades, I'm proud to introduce to you our new edge system for the challenges of the next decades. Introducing Sensai.
With the legacy of Etch leadership, Lam Research has once again revolutionized the chip making process with Sensei, a groundbreaking plasma etching system. Sensei introduces redesigned process modules repeatability required to support advanced logic and memory device roadmaps well into the future. Sensei provides chipmakers with improved critical dimension uniformity and etch profile control for more complex 2 d and 3 d structures. These capabilities become increasingly important as dimensions shrink and aspect ratios increase, providing greater extendability to support future technology inflections. The platform and process modules have been dramatically transformed to enable the highest wafer output per area.
This space saving architecture creates more room for capacity expansion and node to node technology conversions, enabling chipmakers to increase output density by more than 50%. In addition to these transformative advancements, Sensei incorporates equipment intelligence, Lam's comprehensive smart fab solutions, making it the smartest semiconductor process tool available today. Sensei applies rich sensor data to continually monitor performance and subsystem health, while groundbreaking automation systems perform autonomous calibration and maintenance that reduce downtime and labor costs. Using advanced machine learning algorithms, SenseEye quickly adapts to changing conditions to minimize process variations and maximize wafer output. With Sensei, Lam Research has transformed plasma etching, providing chipmakers with the advanced functionality and extend ability required for the next 2 decades of innovation.
Lam Research, redefining the future of wafer processing equipment.
It was a very brief video, but I hope you actually got a glimpse of what this tool is designed to be able to do. We're very, very excited about the introduction of this tool and we're engaged with all of our top customers with this tool. It will break new grounds in both technology and capability in all segments of the market. The name Sensei itself comes from 2 concepts. Sense, for having the greatest sensing and monitoring capability on the tool and I, for intelligence in data manipulation and transforming data into information to give our customers the opportunity to get higher productivity out of their tools.
So as we redesigned this tool, we had the opportunity to redesign the chamber from ground up to deliver the best process performance so that we can enable our customers to continue the device scaling that they require. With powered by equipment intelligence, we're able to get the highest productivity out of this tool and have a system that is intelligent required for our customers in high volume manufacturing. As I mentioned to you, I've been at Lam for over 20 years and I can vouch you that everything that we have known and learned in the past 20 years have gone into the redesign of this chamber. We are delivering the largest process window available to us for this node and the future nodes. We have redefined and redeveloped the plasma uniformity control to deliver the best uniformity across the wafer, increase the ion energy to enable high aspect ratio etching, increasing and delivering faster etch rates and reducing our wafer cost to our customers.
And more importantly, we have redesigned this chamber with extendibility in mind, so that as we learn more in volume manufacturing, we can deliver the needed solutions quicker to our customers to enable the cadence that they have for introducing their next generation nodes to volume manufacturing. The tool is powered by equipment intelligence that Pat spoke about before. The self aware Sensei can actually manipulate the data and help our customers understand patterns as well as trends that they see in order to optimize the tool to get the maximum performance out of the tool. The self maintained, such as the ring exchange in production is something that's been incorporated into the system so that we can get the required improvements in productivity of greater than 15%. We have also delivered more and more self calibration of our sensors so that the tool can actually run and maintain itself in production and achieve a higher and higher uptime.
And with the algorithms that we have developed on the tool, it is very adaptive for our customers to improve the performance, not just from within a chamber, but chamber to chamber and tool to tool across the fleet in an entire fab. The footprint density optimization is something that we had to work very hard to achieve And it is something that while it makes this tool very, very attractive for all the critical applications, it gives us a great capability and our customers to be able to use it for semi critical applications as well, thereby increasing our opportunities to be able to pursue both critical and semi critical applications. So as you can see, it's a very groundbreaking tool and we're very, very excited about the introduction of this tool to our customers. And as I mentioned to you, we have engagements with all of our top customers for on Sensei this year early next year. And you don't have to take my word for it.
Here is a quote from one of our customers that is very excited about this tool, who come and go into the for their R and D development. And with that, I'd like to turn it back to Tim.
Great. Thank you, Vahid. I don't think we quite checked the thesaurus quite well enough to know how many words we could use other than exciting. So I will say the same thing again. It's a very exciting product launch for Lam Research.
And we're extremely pleased with the strong customer pull for this tool. Sensei, we believe is exactly the type of intelligent adaptive tool that will be needed to be able to deliver advanced technology at affordable cost, which as I showed you, we think is really the key to our ability to gain share with a single tool in both the critical and semi critical space and it really is a very important part of our share gain story in EDGE. I'd like to now move on to the 3rd element of our strategy, leveraging new technology inflections to grow our served market. I think most investors understand that Lam's business has benefited greatly from our leadership through the multi patterning and 3 d NAND inflections. Looking forward, we believe our strength in etching deposition puts us in a prime position to continue to play a key role in inflections that are coming for 3 d architectures for both devices and packaging, inflections for EUV patterning and inflections to new materials.
The importance of etch and deposition, both in the past as well as the future, can be seen in these charts, which shows that LAMSAM continues to grow at every technology node across every device segment. The spending mix in any given year, I'll point out, does affect our overall SAM for that year because still today, etch and deposition intensity in each of these segments is not quite the same. NAND has the highest etch depth intensity, foundry logic the lowest for Lam. But you will note that actually as a result of our efforts and the increasing complexity in foundry logic, you have seen Lam's SAM increasing quite significantly, especially for advanced nodes of foundrylogic. And some of that was reflected, I think, in what you saw as improved foundrylogic performance by our company in 2019.
We look to build on that momentum going forward. Also, from a forward looking perspective, we would expect that in the coming years, the memory foundry logic mix will also start to normalize, which provides additional benefit for our business going forward. But the important points on these charts, again, Lam, SAM continues to grow in totality. And when we add this up, our view is that our existing product portfolio is sufficient to power our SAM from approximately 36% of WFE today, given the current mix of memory and foundry logic spending to greater than 40% by 2023. But of course, I think you're starting to get the sense of this meeting.
We'd like to do more than that. And so our technology teams working closely with our customers have identified several multibillion dollar markets that we believe we have the opportunity to inflect over the coming decade. These are markets that are today served by conventional processes and equipment that we believe will ultimately face severe extendibility issues when faced with the device challenges of tomorrow, whether those challenges are step coverage, gap fill capability or fundamental material properties. And we estimate that if we are successful pulling these inflections into our served markets, it could expand our SAM by an additional 5 to 10 points of WFE over the coming decade. This is a tremendous opportunity for Lam to drive the next leg of growth in our served markets.
And so you can see on these charts 3 of the opportunities that we've identified. And for these three processes, we have created a suite of deposition and metal metal and deposition solutions that are based on our highly differentiated enhanced ALD concept. Our goal is to help our customers accelerate the conversion from these older conventional processes, many of which have been around for decades to a highly productive atomic layer deposition platform designed for extendibility. To tell you more about how we plan to capitalize on this opportunity, I'd like to invite up Sesha Bharadaran, our Senior Vice President and General Manager of our Deposition Product Group.
We are glad to be here talking to you about our opportunity to expand SAM through our differentiated ALD solution. So I've been with Lam similar to Wahid for over 20 years and been part of the team that actually worked internally to drive our company into the ALD space. Before I get into what we plan in the future, I thought we'll take a quick look at where we have come over the last several years. We first entered the ALD market around 2012 and over time, we have built a robust portfolio of applications and processes that address different segments of the market. We first actually did packaging.
Subsequently, we addressed the multi patterning space. And over the last few years, our focus has been more on the 3 d architecture enablement across both NAND and as well as logic as well. Along the way, we have pioneered really critical products like, for example, the word line gap fill. This is essentially the technology that's used to put down tungsten for 3 d NAND and every 3 d NAND wafer that runs in the world right now uses our replacement gate technology for built on our ALD. So, we did all of this on a pretty unique approach to how to do an ALD platform.
It was neither a single wafer nor was it a furnace. It was actually what we call as a quad station architecture, which has been a strong legacy for the last couple of decades in how we have approached productivity. We did the highest output understanding of processes to build those applications I talked to you about, it has let us grow our share in the market significantly. Over the last 5 years, our revenues in the ALD space have increased 5x. And if I look forward over the next 5 to 10 years, we have an opportunity to continue growing in this space.
It's going to require us to do some innovative things, like Tim mentioned, which is to look at where our customers have high value problem statements that are effectively coming to a standstill from a scaling perspective and then leveraging our technology, apply those problems and try and solve them. So in today's upcoming few minutes, I'm going to take you through 2 or 3 examples where we have done precisely that and also the consequence for us in terms of growth. Our first example actually we picked was from the DRAM space. This could also be logic, but in most devices, there are critical spacers that are deposited around the cell. And these spacers have many functions, but one of the things that they do is they control the device speed because they set the capacitance in the cell module.
So typically, it's done in a batch process. It tends to have a high dielectric constant. And the way the films are deposited thermally, it tends not to be stable through subsequent processing. So even if I deposit a HiK film of around 5, it may actually incrementally degrade further through the processing and leads to slow transistor speed. Worked well for the industry for a while, but clearly limiting its extendibility now.
What Lam has done is engineered a very novel atomic deposition source that lets us deposit low k materials that are very stable through its processing. So here, upfront, we are getting a k value, which is lower, so good for device speed. But through the subsequent edge steps, the K is retained and that has to do with how strong the material is, which again ties back to the source that I talked about. This results in higher transistor speed and also extendable for the future nodes. Once again, SAM growth through our innovation, but also enabling value for the customers through the higher transistor speed.
2nd example is around NAND, but again, this type of problem statement exists in multiple segments. So typically, dielectric gap fill, which is effectively getting some sort of material like silicon dioxide into a hole or a trench, is done by processes like spin on dielectrics or high pressure CBD. These have worked fine. They've been workhorses for the industry that has brought us through the last couple of decades, but they suffer from some fundamental limitations. One of them being that they tend to shrink upon exposure to incremental processing.
So graphically shown there, so after you heat it up, it shrinks. When the devices were small in the pre-3D NAND error, Volumetric shrinkage, few percent, no big deal. It didn't affect anything subsequently. But in 3 d NAND, as the device is getting really tall, the same percentage volumetric shrinkage in absolute volume terms turns out to be a pretty big number. It causes lots of problems.
For example, it could put so much stress in the device that could induce pattern collapse as well. So easy solution, let's go to atomic layer deposition, clearly deposits high purity materials, But it's not so simple. So what happens is in even though the cartoon on the left was really nice with straight walls, The practical nature of our industry is there are variations along the sidewall. So when you do a conformal process, you do solve the shrinkage problem, but you end up with a situation where you're so conformal, you may have a void, which is shown by the kind of the red slit in the future. So ALD itself represents only a partial solution.
So what we have done is engineer something we call as ICEFEL. Here, the goal is to keep the goodness of ALD in terms of material shrinkage and density, but also achieve complete gap fill. So this is how we achieve that. So we have engineered a source that instead of depositing, puts down geometrically selective, so only in some parts of the wafer that you want to control inhibiting agent. This effectively poisons the ALD reaction.
So in simple terms, where it's present, growth doesn't happen. By leveraging this and varying the depth of the inhibiting agent, we can actually achieve a pseudo bottom up fill, which in the animation, you can see the yellow plasma step that comes on, that's the inhibitor that we talked about that poisons and prevents growth from happening there. So this type of approach, we feel, has a wide range of applicability across multiple device segments to address the gap fill market. A third and the final example I wanted to illustrate is just conventional metal fill. I think people understand the way metal fill is done in our industry is some sort of PVD, CVD liner barrier followed by a gap fill process.
And as the device shrinks, the percentage of the volume taken up by these nonconformal PVD, CVD steps proportionally keeps going up. And what that does is it reduces the actual conductor area. So even though you shrink device to make it go faster, because the conductor area comes down, the resistivity goes up and you end up actually slowing it down. So it's clearly not extendable for the future. So by now you probably guess what I'm going to say given the team is about enhanced ALD.
So we have developed a conformal barrier and subsequent gap fill process that is exactly as it's shown in the cartoon, it maximizes the cross sectional area for the conductor. So effectively, you shrink the device, your resistivity still is what you would like it to be and it because it's conformal, it represents extendable solution for the future node. So I think I was hoping over the last three examples, you got a flavor for how Lam is approaching our opportunity to solve high value customer problems and create additional SAM for us. And the foundation is on the high productivity platform that I talked to you right at the beginning. New materials and new we can access up to 2 times the number of ALD applications we have access to today.
This will create additional SAM for us. And since in many cases, we are enabling the SAM creation, we are confident in our ability to participate at a disproportionate rate at the CREATED SAM, driving our shipments higher by 2.5x is kind of our model. So with that, I can turn it back to Tim to take us through the rest of the talk.
Great. Thank you, Sesha. So through those examples, I hope you got a sense from Sesha's presentation of just how versatile we think this new enhanced ALD solution set can be. And we'll look to help our customers accelerate some of these older technologies onto this very highly productive extendable platform. We also believe that as more markets continue to inflect into three-dimensional architectures, the critical importance of atomic layer deposition processes, both atomic layer processing both for deposition as well as for etch actually will be critically enabling technologies.
And in our companies between deposition and etch, we will invest in this technology. So to sum up our SAM expansion strategies, we really like being in etch and deposition because everything we've just shown you so far says these are the markets that are going to continue to grow faster than WFE. We have a great opportunity to accelerate the conversion of some of these older technologies to perhaps grow our SAM faster in the coming decade. And we are also continuing to invest and seek out new technology disruptions that allow us to enter into new markets. And on this third point, we have one more thing to tell you.
With our announcement last week of our collaboration with ASML and IMEC to develop dry resist technology for EUV, we're looking at the opportunity to pull the wet resist market for EUV into our SAM. We believe that this gives us an opportunity not only to deliver greater performance that's needed for the EUV module itself, but also to deliver a lower cost for our customers. Now if we step back, I want to make sure people understand, today we already do participate in a pretty meaningful way in the EUV patterning module. I often get asked this question, how does EUV affect Lam? In this module, you can see on this chart that we actually participate today through the deposition of high quality hard masks and also the performance of the pattern transfer and hard mask open etch steps.
But from this chart, you also see that there's an obvious opportunity for us to do more. We don't participate at all in the deposition or development of the photoresist, and that's what we're looking to change. With the introduction of the dry resist technology, we significantly increased our opportunity to compete for business in the EUV patterning module. And in terms of number of steps in which Lam equipment participates, it effectively doubles. And from a SAM expansion perspective, we really like this inflection because as the number of EUV layers increases into the future, so does our opportunity.
And so I already mentioned kind of the positive momentum we had in 2019 in the foundrylogic space and that our SAM is expanding in future advanced nodes, we see the development of this DRiResist technology as an opportunity to further accelerate SAM growth into the future. And so to tell you about this exciting new innovation, I'd like to invite up Rick Gaucho, our EVP and Chief Technology Officer for Lam Research.
Thank you, Jim.
Thank you, Tim. Good morning. Disrupting the industry through collaboration. Why do we want to disrupt the industry? From my perspective, it's fun.
And I'd much rather do the disrupting than be disrupted. And I think in the instance of EUV lithography, there's a great opportunity for disruption, as Tim just mentioned. But disruption comes with a lot of risks. There's an installed base, there's an embedded technology that's been out there for the last 5 decades in resist lithography, and that's wet processing. Wet resist, spinning it on wet development.
So customers are not just going to move blithely into a new space that they're not comfortable with. To mitigate those risks, it's important to collaborate. And so we've established, as Kim just mentioned, strategic partnerships with ASML and with IMEC. ASML, obviously, the lithography leader in the world. IMEC, the most successful collaborative enterprise in the semiconductor business where we can evaluate different approaches to integration schemes, device technology, and in this instance, lithography technology.
So EUV is in production. What now? Where are the challenges? Where are the opportunities? Why do we think that EUV lithography is the place where disruption from wet processing to dry processing can most effectively take place?
Well, as a reminder, in all lithography, there's a set of trade offs, often referred to as the triangle of death. Trading off sensitivity against line edge roughness, against resolution. Typically, if you want a more sensitive resist to get more throughput, it's going to come at the cost of lower resolution or more line edge roughness or both. And you can go around that triangle as much as you want and you're always going to get boxed in. This is one of the motivations, the driving forces behind wavelength change, and why we end up today at EUV.
And I can't help but point out that when I entered this business in 1980, I think the technology node was 1.25 micron, 1.5 micron and everybody was sweating about how we were going to get below 1 micron. EUV photons are at 13 nanometers well below the 1 micron limit. And of course, through etch and deposition technologies, lithography as well as computational lithography, lithography has long mastered the challenge of going below the wavelength. Nonetheless, for at any given wavelength, we have these set of trade offs. So what can you do about it?
If you're at ASML, you crank up the power of your source and just drive more photons to the wafer. So if I had to drive more photons to the wafer, I get around the sensitivity problem without compromising resolution or line with roughness, for example. And that's their roadmap and they're making great progress on it. But there's another approach that we can take and that is to look at the photosensitive material on which those highly energetic photons impinge. So by changing the material, making it more effective in terms of how it absorbs photons and creates a latent image and then how we develop that image affords another opportunity to attack the trade offs associated with the Triangle of Death.
And I learned a long time ago, early on in my career, I was sitting in a at a conference and there was a wonderful talk given by a guy at IBM who was pioneering the application of CVD in the semiconductor business. And he made a very simple point that stuck with me all these years. And that is the process is the material, or if you like, the material is the process. These things can't be separated. So in this case, the wet process drives certain material characteristics.
In particular, you need to that material has to adhere to the wafer. It needs to have the viscosity carefully controlled because this material comes in, it is spun off to the edge of the wafer in order to get a very uniform coating. Because it's wet, you need to control the viscosity, you need to control the adhesion to the wafer because you're spinning it. It's loaded with absorbers for the photons, but then those absorbers are chemically amplified by triggering a set of essentially a chain reaction to convert those photons to a latent image. By contrast, if we go dry, we don't need to worry about viscosity.
We don't really need to worry about deposition in the film in adhesion in
the film itself.
You have to
worry about adhesion, but you can tailor adhesion in the film itself. You have to worry about adhesion, but you can tailor the properties of the surface before you deposit your photosensitive material. Or you could tailor the properties of the material as you grow its thickness and you have tremendous control over the thickness of that film as opposed to if you're spinning something on. And the net result is you end up with a much purer material of essentially only absorbers. It's you don't have that other stuff mixed in.
So the dry resist technology affords us to create a much more effective material. It's also a greener solution, dry deposition of the resist and dry development of the latent image. When you spin on resist, it turns out more than 99% of the resist is just spun off the edge of the wafer and wasted. And this is expensive material. Those formulations are quite complex and it drives a high cost of ownership in the fab and you're using less than 1% on every wafer.
So that's a tremendous opportunity to save money and to save material reduced waste. Wet Development, the chemicals are less expensive, but they're comparably toxic, if not more toxic. So, waste disposal is a big deal. Again, the development is done in a very wasteful way. Most of the material was thrown away.
By comparison with dry deposition and dry development, I hate to say it, we still waste a lot of material, but it's 5 to 10x less wasteful than if you do a wet process. Energy reduction is also a benefit of the dry resist technology, because the film has a higher density of absorbers, it's inherently more sensitive. We're able to break some of those trade offs in the triangle of death without increasing the power of the light source quite so much. And we estimate about a 2x reduction per wafer pass in power. So we're enabling a greener solution for the fab of the future.
Now, the proof is shown here. Lithographers talk about a Z factor, which is a figure of merit that captures the variabilities and exposure dose and resolution in line with roughness. You don't really need to worry about the definition, but it's up there if you want to study it. And exposure latitude, which basically is a measure of how wide the window is, how much latitude I have when I'm printing these high resolution images. In a relatively short period of time, within the last 2 years, we've gone from essentially nothing to best in class exposure latitude and Z factor for EUV lithography.
So that's a big reason why we're excited about this and a big reason why all of our customers are excited about this and why we've been able to establish these strategic partnerships with ASML and IMAIC. Further proof is in the dry development area. One of the challenges in wet processing in general as aspect ratios have gotten taller and taller, And those aspect ratios have gotten taller and taller, higher and higher also in the lithography step, not just in deposition and etching. And so when you wet develop, you run the risk that the patterns can collapse because of surface tension. It's a well known problem.
You can see here that that's happening for these kinds of dimensions. But on the right, where we're showing dry development, the same pattern, no pattern collapse is observed whatsoever. That's a big reason why the exposure latitude is best in class for the dry resist technology system. So what's the disruption I'm talking about? It's shown here.
So the wafers will come into a Lam dry resist deposition system, the resist will be applied. They then can flow into the ASML scanner, come out of the scanner, they go into a dry development system also provided by Lam. We're doing this, as I mentioned, in collaboration, strategically with ASML and with IMEC as depicted here. And rather than me telling you more, I think it's very powerful if you hear directly from the CEOs of ASML and IMAIC and what they think about the new dry resist technology.
The industry has driven innovations in almost every part of society, and it will actually help our customers together with our customers and our peers to create some of the solutions that are needed to face and to find solutions for some of the biggest challenges that humanity faces. Semiconductor will go to the heart of those solutions.
Well, we've been working with Lam for many years, actually more than 25 years, and this partnership has grown. And now we're entering a very strategic partnership in the domain of patterning.
EUV is just at the beginning of its maturity curve. So it means we need a lot of innovations to actually bring it well into the next decade. And it also means that the challenges our customers have, we need to solve not only by ourselves, but through collaboration with partners in the industry. And this is where partners like Lam Research and IMAIC, they come into play. And one thing we've worked on is dry resist, which is a very exciting and new promising technology and innovation that would help the maturity of EUV in industrialization.
The benefit of dry deposit resist is that we will have more flexibility to optimize the composition of the resist. And in this way, we will have more knobs to turn on to
better integrate this resist into advanced device fabrication. Whether it's logic or DRAM, things are getting more complex. The flip side of complexity again is cost. So how do we reduce cost? This is why, for instance, the development in dry resist is very important because it will help the challenges on finer resolution, on more state of the art overlay, on productivity.
Lam brings the expertise over the position on the etching. And I'm very convinced that together with Lam and ESML, we will be able to optimize this process to get the best possible performance and to really introduce this process into large scale manufacturing as soon as possible.
I don't know if you know Luke, but Luke started his career as a lithography engineer. And he also spent a lot of time in etch. So he's super excited about this collaboration and the new technology. So as Tim mentioned, we are intent and we're in the process of disrupting a multi $1,000,000,000 market that's been mainstay of this industry for the last 5 decades and is clearly ripe for disruption to relax some of these constraints in the Triangle of Death. For Lam, we're projecting this means about $1,500,000,000 of new revenue coming from this new SAM over the next 5 years.
Besides our strategic collaboration with ASML and IMEC, we are engaged with our top four customers, all of whom are intent on ramping EUV technology and logic and memory. We've demonstrated you can see here, we're down we've demonstrated down to 18 nanometer pitch, which corresponds to the 3 nanometer, 2 nanometer node. You can see again the advantage of dry development over wet development. And we don't believe that's the limit. We have as Luke said in his testimonial, we have a lot of knobs to tune in terms of the material composition how we put it down, how thick we make it and how we tailor that film.
So we see extending EUV indefinitely through collaborative innovation. And last but not least, we're excited about enabling a greener fab for the future. Thank you.
Tim? All right. We're in the home stretch folks. Great to see so many familiar faces in the room. Thank you for joining us today.
You for joining us on the webcast for those of you that couldn't make it. I'm going to go through some of the numbers backward looking and then what a lot of you I think have come for, share with you the forward looking financial model. But I'm reflective right now. I feel like every time I come to an Investor Day, I get to this point in the presentation and I realized everybody had better looking content than I have. Great videos, Pat had robots in the fab, videos of our new tool.
Rick had wafers being bombarded by photons. All I've got is some serious looking numbers and charts. I actually need to have a word with my IR team, guys, next time, videos for the CFO. But seriously, while my numbers my charts might be less cool, I actually think the message I have for you is quite beautiful. It's all about financial outperformance at Lam through creating multipliers.
I'm going to talk about what I mean by multipliers as I go through my talk. But first, what have you heard from us so far? WFE. We think through cycles, WFE is a sustainable story, a growth cyclical some of you guys talk about. Pat shared with you some perspective on the growing annuity stream from our installed base.
I've been hearing for years. Tell me more about this business. Talk to me about this business. Hopefully, you feel like we were responsive to that. I'll share a little more as I go through my materials.
Lam for a while has been a story about growing SAM and growing share. Hopefully, you see a little bit about the strategic intent and how we're doing that. But when I think about this, again, I think about creating multipliers. What am I talking about? In the last 6 years, what has GDP grown?
2% to 3% something like that? Semis grew twice that, 5% to 6%. Semiconductors are a great industry. With this digitization of society, we all look at and talk about and Tim showed you some inflecting charts, we are enabling society in a more meaningful way than we have for quite a long time. Semi is our GDP plus industry.
Semi cap actually grows faster than semis at large. And actually that's largely because the subsegments of semis that are investing in equipment are growing faster than the rest of semis as well, foundry, memory. So Semicap is even better than the semiconductor industry. And Tim showed you, we have grown at twice the rate of WFE. That's because we layer these product multipliers on top.
We develop new products that grow salmon share faster than what's going on in the rest of the industry. I'll talk to you about how we do that from a financial perspective as I go through my slides. But when I think about Lam, I think about an excellent company and a great industry. So let me share with you I'm going to dive in for a couple of slides about how do we run this company and what do we do to try to enhance the value creation beyond just developing great products. I talked about what's on the left of this chart.
Product strategies to grow faster than GDP and to grow faster than the industry. On top of that, we layer on what I believe is one of the best operating models in the industry. This model provides leverage, sustainability, importantly, flexibility to weather the ups and downs of the industry and what it delivers for you, the investor, is resiliency. I'll try to convince you of that as I go through the rest of my talk. On top of that, on top of creating the operational multipliers, we add to that through what we do with the balance sheet, multipliers from how we deploy capital, how we drive earnings to maximize shareholder value.
I'm going to dive into this a little bit. But first, I thought I'd spend a couple of backward looking slides showing you what has this been able to do for you. So to kind of move through the ups and downs of what the last couple of years have done, what I'm showing you on this chart is the average of 2016 2017 compared to the average of 2018 2019. We had a little cycle in between. But if you go through this, WFE has been solid, right?
It grew at 17% over that 2 year comparison. What have we done? We've grown at 1.6 times the industry growth rate or 28%. That's all about multipliers, creating differentiation through product excellence. It's about pet growing the installed base faster.
It's about increasing monetization of every chamber year by year. That delivered a 28% growth for us. The way we run this company, and I'm going to talk to you about something that internally we call the land management system, we delivered leverage on top of the outgrowth in revenue. That's another multiplier you get from how we run the company. That was a 1.3x multiplier.
Operating income in this comparison grew 36%. And then finally, through a commitment to what I think is a fairly prudent capital allocation, we drove a 1.5 times growth in earnings per share or 54%. I could have layered on top of this how we've grown the dividend consistently over this time frame and framed it in terms of total shareholder return for you, but I think you guys know how that all works. So this has been phenomenal performance in my humble opinion. But importantly, what has it done over a longer time frame?
What this chart is showing you, the bar charts are our quarterly revenues. The line chart is our operating income percentage quarter by quarter. I'm purposely showing you this time frame because this is the period after which we brought Lam and Novelis together. So look at the chart. This is a story of revenue multipliers.
Nominally, we have tripled the size of the company in this time frame from growth in revenue. That's all about product differentiation, monetization of the installed base. But impressively, when I look at this is that continued improvement in profitability, the continued improvement in operating income. That's all about how we run this company. That's about operational excellence.
It's about the land management system. As I was reflecting on this a couple of days ago, looking at this, it's also a story of the management team that you've seen on stage today. At the end of the day, business is a team sport. It's about people working together for a common purpose. This is what this team has been able to deliver in terms of creating shareholder value through product differentiation and running the company well.
Look at the chart. We've taken a company that was mid to high teens operating income and transformed it or transitioned it into a company that's delivering mid to high 20s operating income. If you look at the chart carefully enough, if you look at that trough operating income in the last 3 years, it's actually a little bit better than the peak in the 4 years preceding it. So I feel really good about what we're doing with this company. We're running the company well.
So how do we do this? First, it's about differentiation. It's about focused R and D investments that drive that. I'm going to talk a little bit about that. It's about allocating capital to the right R and D programs.
The best use of our capital is deploying it in the products and services you heard about today. It's about developing a new tool like Sensei, first new etch platform from the leader in etch in 20 years. It's about creating new SAM with new highly productive enhanced ALD capabilities. It's about disrupting the industry through collaboration and partnership with industry fellow travelers and creating an enabling capability like Rick shared with you in Drive Resist. As the CFO of the company, in many ways, my favorite part of the company is the installed base business.
I love the annuity revenues that come from the installed base that Pat shared with you. It's about flexibility. We maintain flexibility in the cost structure because we need to. We've refined that over the years to be better and better at it. And at the end of the day, what does this do for us?
What does this do for you, the investors? It delivers resiliency. This is a high quality business model that is delivering high quality earnings. The real foundation of what this what makes this possible for us is our management system. It's been built on a foundation of core values that we've refined over the 40 year history of our company.
Let me talk with you a little bit about that. So first, key to what we do is delivering differentiation, delivering that multiplier to revenue growth. As I've said, this is all about a management team making tough decisions at times. First though, it starts with a proactive management of the entire P and L, which you see in the upper left hand portion of this chart. We're spending less as a percent of revenue on OpEx, but at the same time, which the chart in the lower left shows you, we're deploying more dollars to R and D.
We have grown R and D every single year as a company. We've done that by driving efficiency in other areas of the company, deliver the leverage, increase investment in R and D. If you look at that chart carefully in the lower left, we've moved spending in the past in 2013, R and D as a percent of our total spending was about 59%. Last year, it was 66%. That didn't come by accident.
It came by proactive conscious reconfiguration of how we do what we do, so that we can deliver those high value revenue outperformance metrics that I showed you and deliver higher return. That's what you heard from Rick, Bahid and Sesha, how we think about new product development, going after big markets. And the result actually that I'm extremely proud of is what you see on the right of this chart. Over the last 4 years, nominally, if you look close, we've delivered a return on invested capital of nearly 50%. That is 5 times our cost of capital.
That's what value creation is all about. We're purposeful and thoughtful about how we do what we do to deliver returns like this. So second, sustainability. What this chart is showing you is in green, the trend of our equipment revenue from 2,008 up until last year. What the blue line shows you is the revenue trend of the installed base of CSPG of Pat's business, consistent, growing pretty much every single year.
And you talked about how we're doing that, about our strategy to increase the monetization of every chamber in the field every year, growing from 1x to 1.5x and beyond. I've been saying for a while, I have a hard time envisioning this business not growing every single year. Now that doesn't mean you don't get some variability quarter to quarter, you do when utilizations move around. And what I've also said before, this business is nicely accretive to operating income. If I was investing in a business that looked like this does, it had a really high multiple.
And I know you guys have consistently been asking, guys and gals have been consistently asking, make this more visible. We've tried to today, but we're going to go one step further. Beginning in the March quarter, we will actually in the earnings release as well as in our SEC filings disclose quarterly revenue. So you can see how this business is performing as we go forward. So sustainability.
Flexibility is important in this business. Things move up, things move down. You have to be able to manage your way through those growth cycles and contractions when they occur. We have structured the company such that our cost structure is 2 thirds to 75 percent to 3 quarters variable with the level of revenue. That has been purposefully thought out.
This is how we go about making those products. But importantly, it's about how we have maintained this through the time period that I showed you since we brought the companies together. It's been about the same. Maybe it's gotten a little more variable actually. We don't have a lot of fixed costs in this business.
It's a CapEx light model. 3% to 4% of revenues go into CapEx. So this is about the flexibility and the variability from conscious choices we've made in how we do what we do. What do I mean by that? A couple of examples.
In manufacturing as an example, we choose the right partners, so that a lot of what we do actually is outsource manufacturing. We invest in those partners to make sure they're credible and capable and they can move along the journey with us. We partner very closely with suppliers. It's about how we staff the factory. We employ a fairly highly contingent workforce that can ebb and flow with business.
It enables us to ramp quick and pull back if we need to. That's been a purposeful focus on how we manufacture our equipment. We compartmentalize manufacturing in as small of components as possible so that we can train quickly, so that we can ramp up when we need to. We can bring in 3rd party labor, train them quickly and get them productive quickly. We do similar things in the regions where we're doing installation and warranty work.
We have 3rd party service providers that work with us to do some of the simpler installation work. It's been purposeful. It's been thought out. It's been refined over a 40 year history. A lot of companies say they do this.
We do it better than most. And a couple of metrics I put on the slide I think illustrate what you get for this. Relative stability in our gross margin, 3 times better than the peer group, consistently, predictability, stability. Inventory turns, best in industry inventory turns, so that that cash conversion cycle is quicker, so that we can generate that free cash flow that we need to deliver on that balance sheet multiplier. Sustainability, differentiation, flexibility, what does it get you?
It gets you resiliency in the business model. Illustrating this on the chart, I'm showing you two time frames on the left. The blue bar in 2012 WFE was down about 10%. In 2019, WFE was down about 10%. If you go back and look at our company in 2012, our earnings declined 3.7 times that decline in WFE.
Move forward to what happened last year. Yes, earnings declined, but a much lower rate 1.8 times. Resiliency. We've refined this model over time and we will continue to. Now you may remember in 2018, we were up here talking to you, I was getting a lot of questions about what if things turn down, how will the business perform?
And if you can transport yourself back in time to March of 2018, things just look great, right? It was up into the ride through the year until it wasn't. But we got all these questions. Okay, if it declines or if are flat, what is your business going to look like? And you may remember we said, I kind of think it will look like it did in 2017.
And I remember the night before Satya saying, we really need to think about this. We really need to talk about this. And again, I just saw things looking really solid in 2018. So we did. We penciled this out.
And I got comfortable saying, you know what it will look like in 2017. You know what, 2019 looked just like 2017. Our revenues were about $9,500,000,000 in both years. But actually, our earnings if you compare 2017 to 2019 grew by nearly 6%. A lot of that came from that balance sheet multiplier.
So that's the land management system in terms of the results it delivers differentiation, sustainability, flexibility, delivering resiliency. I talked about the balance sheet multipliers, what we do with capital allocation. The best return on the cash we generate at this company is to make the right investments in our products and services. That's always the priority. That's the highest return on anything we do with the cash.
That's why we delivered a return on invested capital 5 times our cost of capital. In the period 2013 to 2019, we've deployed nearly $7,000,000,000 in research and development. I told you it's a CapEx like business, it is, but it does need some capital. Those are the priorities for cash, always have been, always will be, invest in the right things to enable those revenue multipliers. But even with that, this is a very cash generative business.
And what have we done? You can see it. Significant buybacks. In 2014, we the first dividend in place in the history of the company and we meaningfully grown it every year. Over this time frame, if you go back and do the math, we've returned 92% of free cash flows.
And actually over the last 3 years, it's been 134%. Prudent use of the balance sheet delivering earnings multipliers. So now I think what many of you have come to hear us talk about is, okay, what is business going to look like in the future? First, our baseline, what did last year look like? WFE last year was in the $47,000,000,000 range.
I already said revenues were $9,500,000,000 We delivered $14,500,000 in earnings per share. Great cash generating year for us actually, 2nd best cash generation year in the history of the company. So pretty proud about that. So then maybe a little bit of a thought process on how we've put the model together this year. I went back and looked at every model we've put together, looked at what everybody else has done with their models and decided, actually what makes sense to me to us is we're going back to some scenarios.
For those of you that have been following the company, you remember in 2018, I tried to get out of the scenario business because every time I had previously given you a scenario, it was wrong. And so I was trying to be more intelligent about it. In hindsight, it wasn't the right approach in my view. So we're back to scenarios. So first, probably doesn't surprise you a $60,000,000,000 scenario.
What is the business going to look like in a $60,000,000,000 scenario in 2023, 20 24? I think we're $14,500,000,000 to $15,500,000,000 because of that growth in the installed base, the expansion of our SAM and the attainment of those market share goals. We'll deliver leverage through that land management system to 32% to 33% operating income and through prudent use of the balance sheet. We'll keep driving share count. And I think we can deliver non GAAP earnings per share north of $30 $30 to $32 is the range with free cash flow, 28% to 29%, which if you do that math quickly, that's $4,000,000,000 in free cash flow.
Pretty good result, I think. But again, I don't know. What I do know is by the time we get to 2023, 2024, WFE will be something different. And so I wanted to give you another scenario, so we can you can correlate, okay, if it's not that, you think it's going to be this. What if it's $70,000,000 Could I see $70,000,000 I could.
You just have to think that semiconductor revenues are $600,000,000,000 or so and $70,000,000,000 probably is about the right investment level. So at $70,000,000,000 dollars revenue is $16,500,000 $17,500,000 I think is where we end up. A little bit more leverage in the operating income line $33 to $34.35 plus in earnings. We're going to continue to refine that management system that I've been talking about. We're going to continue to drive improvements so that we can deliver incremental leverage.
Examples of things that if you follow us, you've heard us announce over the last 6 months is beginning of February, we announced a new factory in Malaysia, where we intend to grow into over time. Obviously, you know Malaysia is a lower labor cost region of the world than where we build our tools today. We'll grow into that, but it will enable us to do it efficiently. Late last year, we announced a new lab in Korea to be close to customers primarily, but it will also enable us to drive efficiencies in that R and D spending. We've been investing in India to drive efficiencies in the administrative side of what we do.
We're going to keep working on those things. We're going to keep working on the model to deliver the leverage we expect and I know you guys expect. So as we were putting the slides together, this is actually where I was going to finish. But then I realized, actually, I've got one more thing. And the one more thing is how to think about the balance sheet, how to think about that balance sheet multiplier.
I'm not here to announce a raise in the dividend. We just did that. But what I am here to communicate is we plan to deliver given the earnings growth, I think the model delivers a disciplined annual growth in the dividend. I've been getting a lot of feedback from you guys on the buy side. You want us to grow the dividend consistently every year, every 12 months.
That's our intention. We also intend to increase that plan. You may remember, in 2018, we said we plan to return 50% of free cash flow. As we stand here today, we now plan to return 75% to 100% of free cash flow. Now some of you are I think looking at the chart on the right saying, well, why aren't you doing more?
Why aren't you doing more than 100%? You don't need to be a math major to know you can't do that forever, right? You can only return what you generate. And we've kind of got the balance sheet where I think it should be. And so that's the approach.
That's the thought process around that balance sheet multiplier that you should be thinking about. It's how we think about it. And actually then I went back and I looked at that 2018 model. And those of you that were here remember how we closed things out. So actually it wasn't one more thing.
It's 2 more things. Our HIKU for the year. Revenue drivers, operating leverage, capital return. Lam multipliers create shareholder value and outperformance. This is a haiku multiplier.
Last time, we just did 1. We decided to do 2 this time. So anyway, thank you guys for your attention. Thank you for coming today. Thank you on the webcast for tuning in.
Let me invite Tim sorry, Tim, back up on stage to close things out.
Great. Thanks, Doug. Maybe we should have just finished with the haiku. But at the risk of sounding a bit repetitive, I wanted to make sure that we did capture a few key points today. First, hopefully, we made it very clear that we believe that secular demand for semiconductors continues and will drive sustainable growth in WFE across growth.
3, we believe in our own business and disrupting them with great products when we see the opportunity to widen our lead to the competition and gain market share. And finally, we're very excited about the opportunity and we believe that we can disrupt the equipment segments of others with new technology inflections as you saw from several of our presenters. But most importantly, since it is our 40th anniversary, I think maybe the best way to think about Lam Research is that we believe that we have 40 years of innovation and we're just getting started. So as Doug said, I want to thank everybody who chose to join us today either here in New York or on the webcast. And with that, I'd like to bring everybody back up now for Q and A.
Okay. So we've got about 30 minutes for Q and A. I'm going to kind of have Tina on the left of the room, Ram on the right. We'll go back and forth in the room, direct your question as you see fit. But before we get started, Tim, I thought maybe a couple of comments from you to kind of kick it off.
Yes. I think that before we get started with Q and A, I mean, obviously, we've kept the story very much about our long term growth objectives and programs. And so but I'm sure that top of mind for everybody is the coronavirus situation. A couple of comments I want to make. I mean, 1st and foremost, our top priority is the health and safety of our employees.
And that is something that we've been paying very, very close attention to. We have employees, obviously, deployed globally, and we put measures in place to safeguard their health and well-being. We are monitoring this situation 7 by 24, so you can assure that this management team plus the management team currently back in Fremont is all over this issue. It is something where, obviously, we've looked very hard at the business continuity plans for our supply chain. We have those.
Those are now active, and we are driving alternative options throughout the supply chain. It is something where in our earnings call in January, we tried to give you our best view of what we thought the virus impact would be. And at this point, we built that into our guidance. And in today's meeting, we want to keep this long term. We're not going to provide any further comments on our view about guidance for this quarter.
And so I think that the only thing I'd probably would finish then with saying is that, however, in our long term story and even in our thoughts about the year, there is nothing that I've heard from customers that changes our view for the long term story we just laid out or for the view that we gave you for the full year 2020. And so with that, we'll open it up for Q and A. Tina, Ram? Okay. Ram, why don't you start
on the right on the southern?
Yes, sure. Just one addendum, please keep to one question and one follow-up. We'll always circle back if we have time. So first question, CJ.
Thank you. Great presentation today. Really appreciate it. I guess first question on the installed base. The business was growing 20 plus percent CAGR over the last 7 years.
And obviously, your growth target of assessment? And is there upside to that number? And what would drive that? And then I have a quick follow-up.
So are you asking Pat if he'd like to increase the forecast he just gave you for the revenue in this business? I'll let Pat
go ahead
and I'd like to know what his internal forecast is.
Pat comment on that.
Our commitment is to continue to grow that business faster than the installed base. And that's what we're going to focus on. And the installed base continues to grow. It will continue to grow in the future. And you saw some of the accelerators that my colleagues presented.
At the same time, we're going to continue to engineer and offer new products and services that increase the annuity out of every single tool in the installed base. And that's going to be our focus. As we can see it today, as I said, it's at least 40%, and I'll leave you with that number.
Great. And then a question on the Resist side. I guess how disruptive is that to the traditional coder developer setup with each EUV tool? How do you actually fit in with the footprint? And is that an issue or not for currently constructed factories?
And then can you talk about where you are in terms of moving to high NA and perhaps the different challenges associated with that?
Yes. I think the jury is still out as to exactly how it will get implemented in the fab. We have still have a lot of learning to do, whether or not those tools are actually integrated with a scanner or they might be standalone units. So those are questions still to be determined. I think it's also important to notice that there is an installed base of spin coaters out there for EUV today, but the growth in EUV is still yet to come.
So there's a lot of opportunity to we don't necessarily have to worry about displacing what's already installed. The opportunity is what hasn't yet been installed. As far as high NA is concerned, we actually started this program with high NA in mind because there is no resist today that will be compatible with high NA. So you have this really super capable lithographic source, the scanner, but it can't print anything because the materials aren't there. I think we've already demonstrated that the dry resist technology will be compatible with high NA and is compatible going forward in that regard.
So what's happened is there's been so much excitement and customer pool from what we've done so far that while we originally thought the big opportunity would be at high NA, we now think there's a lot of opportunity before high NA.
Thanks for the question, C. J. Tina, you want to go to the other room?
How about Chris?
Thanks for taking my question. 2 of them. First, you spoke about the challenges in 3 d NAND going to higher layers and how it impacts that edge time and stress and warpage, etcetera. Do you think at some point litho scaling comes back to 3 d NAND? And if so, how does it impact depth and edge?
Rick, why don't you answer that one? It's a litho question.
That's sort of a litho, it's also edge and depth question. Yes, there's already evidence that horizontal scaling is playing a role in the future roadmaps for 3 d NAND as well as vertical scaling. So we see both of those things happening into the future. I don't know if Seshi Bahir NAND.
Seshi, maybe you comment on the some of the challenges and opportunities associated with vertical scaling of NAND as well?
Yes. Yes, I think as you grow taller, your question is whether at some point people stop growing taller and try to scale with little. We don't see that at least in the next 5, 7 years. However, customers will figure out some strategies working with us on how to get more 2 d density and that presents additional opportunities for Lam as well. So for example, there are some companies that talk about doing wafer scale bonding as a means to achieve more 2 d density.
And again, that's an area where we have some enabling technologies. Is
the Is the main comp JSR or whom are you like displacing today on the dry resist product?
I'm sorry, I didn't catch GSR?
Is it JSR? Or who are the incumbents there?
There are a variety of resist suppliers. The disruption, as we see it is as much in the equipment space as it is in the material space.
Thanks, Krish. Ram, you want to go to the other side?
John.
Yeah, guys. Thanks for the presentation. My first question is just around the target model and your expectations around WFE. I'd kind of be curious as to how you're thinking about the mix of memory versus logic foundry within that? And is the point around kind of the SAM expansion and the drive resist is that you're becoming less susceptible to that mix of memory, logic foundry?
And I'm kind of curious the $1,500,000,000 of accumulative revenue you see in 5 years in DRIV Resist. What's the linearity? How should we think about kind of the real kind of inflection point there?
John, you jammed 3 questions into 1. That was impressive. I'll take the target model and then hand it off a little bit. I went back and looked at the assumption we had when we put that 2018 model was a much heavier memory mix approaching 60% of WFE. As we look at what's going on in investments in foundry, seems stronger than I thought it would have been 2 years ago.
And so today, the model comprehends kind of mid-40s memory mix, something in that range, John. So that's part of what has changed. And as you know, we're really strong in memory. And so that's part of why the numbers pencil out the way they do.
So Doug, let me comment. I mean, I think that's a very fair statement. You asked the question, are we investing in some of the activities here? Will that sort of drive more balance perhaps between our exposure across different segments? Obviously, always in the back of our mind, we would like to continue to do better across all segments.
We absolutely love our exposure in memory. So I mean, let nobody make a mistake there. We think memory is one of is going to continue to grow. We talk about the digital transformation of society, the power of data. That all translates in a very big way into more memory in the future.
But at the same time, we see that added complexity and the difficulties of scaling to 5, 3, 2, whatever comes beyond that one, is going to open up a lot of opportunities for us. And I talked about atomic layer deposition and etch, you see the new EUV innovations. I think those are all great opportunities for us. And they'll probably be first used within scaling of foundry logic. But eventually, you're starting to see DRAM and other devices, especially as they inflect more 3 d.
They'll incorporate those same technologies. So rarely do we find ourselves investing in a product that only serves one of those three markets. But as you saw from that chart I showed, the opportunity for us to increase our SAM and foundry logic, we do think is accelerating as those devices continue to become more complex.
Okay. And Rick, the linearity of your $1,500,000,000 how do you think about that?
It's not exactly linear. But it will be driven 1st and foremost by logic and secondarily close second is DRAM, but they're both in play.
Back end weighted revenue aside, think about a chance? Yes. Yes. Tina? Harlan?
Great. Thanks for the team for putting on this event. First question, on your Sensai platform, you showed some pretty strong productivity metrics for leading edge, edge platforms, exactly what your customers are looking for as they drive their cost curves. When will we see customers put Sensa into their high volume manufacturing lines? And do you have a similar high productivity architecture in mind for your deposition platforms?
So maybe Bahid first.
Yes. As I mentioned to you, we have had a very strong pull from our customers for this product. The way you can think about it is that in memory, the cadence for development is about 14 to 18 months. So as we're getting engaged with our customers this year, it will go into their R and D for qualification for the next node. So we're expecting it to be having high volume manufacturing next year and the year after that.
Sachin?
On the deposition question, I think the sensor has 2 parts to it. 1 is the footprint density. So for that, as I showed you, we already have what we think is the highest density platform in the market with our quad layout. The second aspect of it is the Sensai and the big data capabilities that's embedded in that platform and that our company is proliferating those technologies from the Sensai program across all our platforms over the next few years.
Great. Ram, let's go back to the other side of the room here.
Does she in there?
Thanks for hosting the Analyst Day. I had a follow-up question for Rick on the dry resist solution. You talked about the benefits of your solution. Can you speak to some of the technical challenges that you may perhaps be facing in the near term? And how should we think about the potential insertion point of this technology?
You talked about logic being first and then DRAM being a close second. But in logic is it 7, is it 5, 3? And longer term could this be used for deep UV layers as well or is this primarily for
I think that was 3 questions.
These guys are good.
So some of the technical challenges, the biggest probably the biggest technical challenge is defectivity. You just can't afford to print defects. And we still have a lot of learning to go in that area. I'm not particularly concerned about fundamental limitations. It's just a question of getting the wafers through the tools.
We're a long way away from getting what was it, how many millions? 26,000,000. 26,000,000. Thank you. That 30,000,000 was the number in my mind, 26,000,000 more than the other guy.
So we're a long way from that level of learning. So that poses some risks. But we actually feel that one of the advantages we have is that the material is less prone to defectivity than the wet resist. So defectivity is one challenge that still lies mostly ahead. And the other one is this big question of, do we need to integrate with the scanner or not?
If we have to, that's a fairly large engineering undertaking. It's it doesn't, I think, require significant invention, but it is a lot of hard work that we would be doing in collaboration with ASML. Our customers prefer, I think, things not integrated. So that's actually an opportunity for us if we can pull that off, but that's a technical challenge lying ahead. As far as implementation is concerned, we think it could go in as early as a 3 nanometer node in logic.
I do think there's potential for back propagation into older nodes. And the technology, I think, is also amenable to longer wavelengths or pass nodes such as deep UV. We're not focused in that area today. We are focused completely on the EUV solution. But I think as we gain traction, there is that upside potential.
Did I get all your questions?
Yes. And if I could squeeze in a follow-up. Similar question on the ALD side. You talked about ALD disrupting CVD furnaces, some of the more conventional processes across, I think you mentioned DRAM, NAND and logic, which process nodes or which inflections should we be monitoring in terms of potential timing of ALD sort of taking over in a meaningful way? Thank you.
So I should maybe just start by commenting. I mean, a couple of those inflections are currently underway.
Yes. Yes. I think it's a range. It's not one particular answer. So that's why I'm struggling how to frame an answer.
But some of the things like, for example, the gap fill that I mentioned is something that's happening as early as second half of this year onwards. Some of the other things like metallization are in the years to come from. So it's a range of outcomes for your question.
Yes. I think what I'd add is, we don't underestimate the challenge. I talked a lot about one of the key points of stability of our business is the installed base and all of that learning has come. And so we don't underestimate the challenge of disrupting these markets. But the key point is when a technology hits an extendability wall, then all of that desire to sort of just extend what you have, it goes out the window.
And so I think that what will happen is as each of these more conventional technologies hits that wall, customers will start to pull those solutions in. As Sesha mentioned, a couple of those have already come in. The pull for dry resist is because some of those challenges are already being seen, especially as you start to look out on the future roadmap. And so it will take time. And what we tried to do today was give you a much longer and more forward looking view of how we see inflections rolling out over not just the 3 year period or 4 year period we usually use at Investor Day, but over a longer period of time.
And I think that you'll continue to see more of those extendibility walls get hit and more of those applications get pulled in through that time. So I think it'll just be something that we can continuously update you on as we continue to knock those opportunities off.
Okay. Great. Do you know this time?
Tim?
Tim, I'm expecting a great question because I know you traveled
all the way
to just ask the
Thank you. Okay, Doug, so at the last Analyst Day, you put up $14,500,000,000 to $15,500,000,000 was the 2021 model and that was
at like
$53,000,000,000 WFE. Now you're at the same you're at the same revenue at $60,000,000,000 WFE and you're 2 to 3 years out as well, so you should have more service revenue as well. Yes. So it embeds a much lower WFE share assumption. I think you were sort of implying 22.5% last time and this time you're like 17.5%.
So even if I account for the change in the mix between now memory being lower, it seems I'm a little surprised that your WFE share assumption is not higher given the narrative around SAM expansion and all that kind of stuff. So can you kind of speak to how you thought about that versus what the last model implied?
Yes, sure. I mean, when I looked at it, Tim, and I did go back and reconcile the whole thing, it wasn't 53%. At that time, in early 2018, I actually thought W3 was going to be higher than that. So that's when it turned out to be closer to 52, 53, but in early 2018 I thought it was going to be higher. So that's 1.
And then it's all memory mix, right? That is what it comes down to. It's like I said to I forget the question, I think John asked it. We were forecasting forward a 60% memory share of WFE and now it's more in the mid-40s. That's the biggest driver of why it's different.
And then maybe just one quick one for Rick. So, Rick, why now on the dry resist? Is it because I mean, the advantages were kind of always clear, but was it because you just didn't start working on it until you partnered with ASML? Why didn't someone start working on this technology earlier?
Actually, there were previous attempts over the last couple of decades that people have done various dry approaches to resist that never gained traction. It's not easy to disrupt an industry is one explanation. But what we saw, as I said earlier, was the high NA coming in where you have the optical capability of printing much finer resolution lines, but there was just no material solution anywhere in the offing. And even with today's EUV technology, the chemically amplified resist to me just aren't going to work. I mean, they're working today, but not at a cost structure or that's really going to be sustainable and certainly not going to allow you to scale.
So before this, while there might have been inherent advantages, and I think there are actually in dry resist, not the least of which is cost and the environmental advantages, you didn't have that compelling motivation where it just looked like there was no way to go forward other than something different.
Great. Thanks, Tim. Ram, other side of the room?
Yes. This I'll go to Joe here.
Yes. Thanks for taking the question. So you showed a chart that talking about your increased SAM for 3 d per layer count on 3 d NAND. I was just curious, can you talk about the growth in capital intensity there, just given that you'll be taking your assumption is kind of you'll be taking share there. I'm just curious because last week Toshiba discussed potential for some lithography shrink as enabling higher density per layer at a higher a lower layer count.
Just curious how do we think about that? You want to
try that?
Yes, sure. I mean, so you're referring to the chart that showed that our SAM increases at every node. So as layer count increases, etch and depth increases. But I think that there was another chart that I showed kind of coincident with this point of cost reduction node to node is extremely important. So I mean, make no doubt about it, we are partnered with our customers trying to find every way to help them reduce bit cost generation to generation.
And so in some cases, that will be through the introduction of brand new platforms that will help accelerate that very deep memory hole edge. Sometimes it's through the throughput action we may have on the Strata tool, which runs the majority of those stacked layers. And in some nodes, it doesn't quite work out so well. And so in fact, if you looked at that chart closely, you'll see that kind of layer to layer it actually varies a little bit depending on when some of those new productivity enhancements come in. But one thing we do know is whether it be 2 dimensional scaling, we talked a little bit about vertical scaling with some new materials, maybe even the chip stacking.
The SAM opportunity for us continues to grow. And it's just a matter of at what rate and what new opportunities are created for us to gain share through those inflections. And so that's just something we feel very confident with. And in many ways, we're very committed to making that industry successful through productivity enhancements.
Great. Tina, other side there? Maybe middle?
Thanks. Vivek Arya from Bank of America Securities. Thanks for the presentation today. I wanted to ask about the role of China in your growth prospects over the next few years. So first of all, starting from this year, I think you had mentioned a $2,000,000,000 to $3,000,000,000 incremental growth from China.
Any more color now that you are a few weeks after right your last presentation? I know you mentioned that you're not going to be talking about near term, but just is that $2,000,000,000 to $3,000,000,000 number for this year, is that still a solid number? And if you could give us some more color around NAND data versus foundry logic, how do you see that playing out in China?
Sure. Yeah. Maybe I'll take a shot and Doug can add. I think maybe the important comment I added at the end is we've heard nothing from our customers that change our view on the year. And so while obviously there is a lot of questions about will domestic China spending still be what we said just a month or 2 ago, I think that still remains our view.
China, if you look at it much longer term, obviously, the ambition is to build a strong domestic semiconductor industry, I think, will be unchanged by anything we're seeing here in the short term. And so our engagement with customers across NAND, DRAM and logic and foundry remains as it has been. Obviously, in the last few weeks, much more impacted in terms of our ability to engage. But I think in the long term, no real change. And as we said, I think, on our last discussion, we see the demand domestic China being quite broadly distributed across the different segments.
And so we feel like it's all segments, all opportunities that I showed you that we have there.
Got it. And for my follow-up, assuming there is some restriction on Huawei, I know you don't ship directly obviously to Huawei, it's indirectly. Is there a way to quantify how much indirect Huawei exposure that you have?
Well, I don't know how. I think Huawei not only consumes 4% or something of industry capacity. And if I think about it, if somehow they're not supplying it, somebody else is going to. It's going to take some time, maybe create a little dislocation, but smartphone demand is smartphone demand. Got it.
Thank you. Thanks.
Ram? I'll go to Sidney in the middle.
Thanks for doing the presentation. I want to ask about this opportunities for semi critical steps. I'm not sure if comparing the size of the bar charts the right way to think about how big that is. Can you help us understand what how big that market is? What kind of share do you have today?
And maybe give some Yes. So
I think that we
Yes. So I think that we actually probably gave you about as much information as we're willing to give you on those charts. So we're not going to tell you our exact share in each of the segments. But what we can say is we maybe in terms of just how we thought about the markets. You saw in the charts actually, we showed that our opportunity to gain share over the next several years is about equivalent in terms of market share overall impact to the company in critical and semi critical.
So that gives you some sense of how we view the opportunity. I think that in most cases, the semi critical applications, whether they be in etch or deposition, but those are steps where it's still required for you to deliver great uniformity and great defectivity and good uptime on the tools. But ultimately, in most of those spaces, there are a couple of suppliers who can do that in a way that doesn't allow you to differentiate so strongly in that area. And so in most of those cases, the decision gets made based on the cost of ownership of the process. So what you see with something like Sensai or you see something like with our tools that have very tight packing density and deposition, Customer benefits from the lower running cost per wafer.
And so we can expand our business in that area. But the real trick is you don't want to have and nor does the customer want a fab environment where you have to have all of these critical tools and all of these semi critical tools as well. And that's why Sensei is such a breakthrough for us is it brings together critical competitiveness and semi critical competitiveness on the same platform by incorporating all of these productivity enhancements and intelligent use of data all into one package. And that's why, as Sascha mentioned, we look to take that same type of tool intelligence and eventually port that over to deposition as well.
Okay, great. Tina, let's go to this side.
Robert Merer. Two questions. First one to Rick. As you pointed out, dry resist has been around for a while. There was some work at Berkeley a while back and Applied Materials has done some work.
They have a patent for deposited EUV photoresist that predates yours by a couple of years. What's the major differentiation between your product and other offerings that have been out there or other projects that have been out there by competitors that would prevent them from entering the market or give you a specific advantage? And my second question is to the other gentleman regarding Dep and Etch. We're moving to STACK CFET logic structure and the twelve eighty process that's going to be like 4 stack transistor cells and such. When the industry moved to a 3 d NAND architecture, you had an advantage over others because of the very high aspect ratio etch.
What advantages do you see as logic moves to 3 d?
So, I can't really comment on the problems that other people had in the past because I'm not privy to all the things that they encountered. But I do believe the material that we're using is differentiated from what was attempted in the past and it's tailored towards EUV versus longer wavelengths. The previous efforts were not as focused on EUV, I believe. And then the collaboration that we have with AS and L and IMEC, I think, is a differentiating advantage for us. We have the benefit of all their previous learning with whoever and their current learning.
So those are but the material itself is a different material from what's been used in the past.
So, Shaf, Mohit? So, 2 advantages we're seeking in advanced logic. 1, as Tim mentioned, with a combination of atomic layer etch and atomic layer deposition, we're seeking to deposit and pattern features with much higher fidelity and higher selectivity. The films are becoming very, very thin, so you have to remove and pattern some film without effectively touching the underlying film. So, ALE and ALD are 2 capabilities that in advanced logic that we're seeing being incorporated more and more.
2nd is, as the critical dimensions shrink, even if the film thickness doesn't grow, the aspect ratio still goes up. So the same aspect ratio challenges that people have in NAND for much, much higher aspect ratio also show up in logic and in DRAM as well. So all the technologies that we're developing for high aspect ratio also are beneficial in 3 d logic as well. Sesha?
No, I think that's a good answer.
Okay. We have time for one more question. Ram, you get to pick up. Greg, let's
here. Yeah. Thanks for taking the question. Doug, can I just ask a clarification? Sure.
With respect to the EUV $1,500,000,000 opportunity, how much of that is incorporated in the target model view that you have? And I would ask the same question about that 5% to 10% Sam Pullen opportunity that was identified. Is that also in the target model and to what extent?
In each case, I would tell you a little bit. Rick got the question earlier. Tell me about the linearity of the 1 $500,000,000 back end weighted about disrupting 5 to 10 points of other business. That is maybe even further out. So there's a little bit in the target model, but not a huge amount.
This stuff will manifest itself later on in a timescale. Okay.
Great. Well, I guess that wraps it up. So again, thank you very much for joining us today, both here in New York as well as on the webcast. We hope you found the presentations helpful.
Yes. Thank you for coming.