Good morning. Good morning, everyone. Thank you for coming to our Analyst Day. Thank you to everybody on the webcast as well, listening to this live and on the replay. My name is Satya Kumar. I'm the head of investor relations for Lam Research. I'm just going to take a few minutes to give you a quick overview of the logistics for today. I'll start with the important stuff. There are two espresso machines in the back, so if you aren't sufficiently caffeinated, please feel free to grab a cup of coffee. We'll have three presenters for today. Martin Anstice, our Chief Executive Officer, Tim Archer, our President and Chief Operating Officer, and Doug Bettinger, our Chief Financial Officer. The prepared comments today will last about 90 minutes. We'll have a Q&A at the end for approximately 30 minutes.
I would ask that you please hold your questions until the Q&A. We don't plan to take any breaks, and we'll post the presentation after the end of the Q&A on our website. Our objective today is to talk about the long-term opportunities for Lam Research. In that vein, I'd appreciate if you keep your Q&A focused on the long-term objectives. There will be box lunches available for you at the end of the Q&A at 11:00 A.M. Let me alert you to some important information about our safe harbor. Today's presentation and discussion includes statements that are not historical in fact, or forward-looking statements that are covered by the Safe Harbor Act. Such statements are based on current expectations and subject to risks and uncertainties reflected in the risk factor disclosures in our SEC reports.
Please be aware that the actual results may differ from those forward-looking statements. The company undertakes no obligation to update them. Please see the accompanying slide in the presentation for additional information. In addition, today's presentation and discussion includes financial results presented on a non-GAAP financial basis. A detailed reconciliation between GAAP and non-GAAP results can be found in the appendix to today's accompanying presentation. I just also want to take a couple of minutes to talk about a few terms we'll use today, and want to make sure it's a level playing field, all of you know what we're talking about. It'll be very simple, I promise. One is you'll hear the term WFE. If you hear that means wafer fab equipment spending. That's a portion of our customers' capital spending that's spent on buying equipment from companies like Lam and our peers.
Another term you'll hear is SAM. That stands for served available market. That is a portion of the WFE market that we actively compete for. Another one that you'll hear is CSBG. You'll hear Tim talk about this. It's actually referring to our install-based business. The acronym stands for Customer Support Business Group. Just two technical terms for you as well. One you'll hear about, ALD, that's atomic layer deposition. I see nobody taking notes here, but there'll be a quiz in the end. The other one is ALE, that's atomic layer etching. That's all you need to know, okay? Let me introduce the speakers, but before that, I'll just go over the agenda. Martin Anstice will talk about value creation and capture. He'll cover topics on WFE growth and our vision and strategy. Tim Archer will talk about our execution plan for our performance.
He'll cover topics on our SAM and our differentiation plans, also talk about our install base. Doug Bettinger will then talk about our plan for financial performance and capital allocation. With that, let me invite Martin Anstice, Chief Executive Officer of Lam Research.
Good morning. I always hate it when I see a photograph like this because I look at it and go, damn, like I've aged in 5 years? We need to upgrade these photographs, I think, a little bit. Just to kind of maybe break the ice this morning, I thought I'd tell you a story about a conversation I had with my parents at the weekend. I've lived in the U.S. for 25 years, essentially, and I try once a week to call my mom and my dad or FaceTime with them, and that's what I did this last weekend. The customary conversation goes something like, "Hey, how are you doing? Is everything okay?" "What are you doing for the weekend?" Or, "What plans do you have for the week?" My mom indeed asked me that question.
She said, "What are you doing this week?" I said, "Well, I'm going to New York. I'm going to spend time with the investors and the analysts that cover our company." My mom said, "You still have to do that? That other chap doesn't do that now?" I thought, like, "What other chap? What are you talking about?" She said, "You know, the President." I thought, "Have I missed some kind of a presidential tweet or something?" I'm like, I don't know what she's talking about. My mom starts talking to my dad, eventually my dad says, "You know that guy, Tim." I'm like, "Oh, Tim. Tim Archer. Okay, now I got it." Indeed, Tim is the President of Lam Research, and I am the former President of Lam Research.
We live in a world of data, we live in a world where artificial intelligence is racing to deliver value. I'm sure eventually the transcript of meetings like this is going to record emotions and body language as much as it does words. Until we reach that point, I'm going to say something I just said again, which is, I am delighted to have an opportunity to spend time with the investors and analysts of our company, said Martin Anstice with a smile. Happiness and excitements. With all of that said, as Satya just indicated, the focus that I'd like to have this morning is a focus on demand. When it comes to investment levels in the short term or a conversation about sustainability, the most important conversation to have is the demand conversation.
The thesis that we're characterizing today, in many respects, is the belief that the future of Lam Research is inextricably linked to the world of a data economy. Our belief is that the future of a data economy and the role of silicon particularly is quite extraordinary, and Lam's place in that is a place of opportunity. The mission of Lam actually is pretty simple. Our mission is to create value for our customers, and we create value for our customers through the delivery of enabling technology, and then to capture our fair share of value in the ecosystem for the contributions that we make. We believe that the context of a data economy is quite fundamental. We believe that data is the fuel in the engine of the economy that we now live in.
Specifically, we believe that the role of the data center, the cloud, the enterprise, is now a fundamental accelerator of that data economy. In summary, we live in a world of opportunity. The messages that I'm going to choose to focus on in the next 30 minutes are really structured around this conversation around innovation-led demand. I'm talking about applications innovation as much as I'm talking about device innovation or even hardware innovation. The context of demand is everything about the context of silicon capacity supply. I'm going to talk about the relevance of silicon from our perspective. Everything we see today in the applications innovation space has its origin in the performance of silicon. Has its origin in the capabilities of the world to collect, to communicate, to store, and ultimately process data.
That foundation that silicon has delivered the world is everything about sustainability of innovation in applications and services and value in our future. The last part of my presentation today will introduce elements of the vision and the strategy of Lam Research that through the presentations that Tim will host and Doug will be characterized in more detail. Appropriate headlines for our industry and our company might be extraordinary and exciting opportunity. Might be demonstrated reality, yet the early stage of growth in a data economy. They might be a headline of a virtuous cycle of demand and value creation. They might be strengths, momentum, and sustainability. In our industry, we believe Lam has an unmatched position and an unmatched opportunity, and that's the primary basis of our conversation today. This is an extraordinary moment in the history of our industry.
One simple characterization of that actually is this slide. The reality of our world today is that several forces have come together at one time to create this platform for applications and service and value and business model innovation that is pervasive in the global economy. We live in a world of ubiquitous connectivity. There are more than five billion connected people in the world today with a portfolio of devices to connect them. In the last five years, the world was delivered an economic and performance platform called the cloud, and the cloud creates a massive scale for storage and computation. In many respects, the data center, the cloud, is analogous to the impact of the foundry business model when it was launched on the world 25 years ago.
25 years ago, when the foundry business model was created, it created an opportunity for an acceleration of innovation in a community of fabless companies that were no longer burdened with the responsibility to add manufacturing capacity for device production. The cloud has the very same opportunity, and it has a very same opportunity not just in the limited population of semiconductor companies, but it has that very same opportunity to be an accelerator of innovation in creating value for every single element of our economy. The compute world today is facilitated by our ability as a semiconductor ecosystem to pack tens of billions of transistors onto a single piece of silicon. To put tens of terabytes of storage capability onto a single solid-state drive capability.
Last but not least, in addition to connect and cloud and compute, our industry has a history and continues to deliver on the theme of productivity and cost. The reason that's so important in the scheme of our industry is in many respects, our customers and our customer's customers live in a world where there is inherent elasticity of demand. That's not true in all industries. In some industries, the word cost reduction is associated with reduced opportunity. That is not true in our business. The delivery of underlying technical competency and the delivery of economics have the potential to literally transform all aspects of the economies of our world. There are more sensors and more storage and more processing today than ever before.
There is more value capture opportunity from the totality of technologies that exist in the world today, in large parts, through the contributions of the semiconductor ecosystem and the contributions of Lam. There is always context for a conversation about technology and products and services. From my perspective, this is about as helpful a slide to understand context as any we could share. No doubt, there is tremendous uncertainty in the world that we live in, but there is certainty that the scale of the socioeconomic challenges are enormous. They will not get solved on their own, and they will not go away. There is certainty that every aspect of our lives can be disrupted and enabled through the delivery of technology to improve quality.
One really important piece of context here is that the fundamental scale of these problems, the scale of information data collection, analysis, synthesis, extends beyond what is possible without artificial intelligence, big data, and cognitive decision-making capability. The context is there is no known alternative to the solution sets of AI to the scale of problems that are summarized here. Whether our role in the delivery of solutions to these problems is narrow or wide really doesn't matter. If you look at transportation, the world is facing the challenge of global gridlock. We need smart infrastructure. We need autonomous vehicles in the mix of solutions. That will only be possible through continued innovation and continued investments in the world of semiconductors and silicon. The same is true in the area of food and water, specifically smart agriculture, which is necessary to feed the world of an increasing population density.
The world of healthcare, the opportunity to deliver personalized and predictive healthcare for an aging population of the world. The context for this conversation is compelling. The context, we believe, is really important. Perhaps only five years ago, when we were standing on this stage talking about the success of our customers, talking about the vision and the objectives of our customers, we might have been talking almost exclusively about the performance of a piece of silicon. Because in large part, our customers define success measured in terms of a faster-performing processor, a memory device with greater density, a component of silicon at lower cost, a component of silicon with more flexibility from a form factor or power management perspective. Today, our customers have completely redefined their purpose and their vision.
Today, our customers, in one way, shape, or form, have an aspiration to change the world, to make the world a better place, to contribute a set of value propositions to every element of our life. Certainly, our personal digital footprint is increasing and will continue to increase, and the numbers don't matter so much. Whether there is a $1 trillion growth opportunity or $2 trillion growth opportunity in the world of AI or cognitive learning really doesn't matter. The consensus here is that the context of the problems I just characterized and the relevance of silicon and the relevance of AI to that solution set creates an opportunity that is literally staggering in size, scale, and scope. Value creation opportunities are numerous.
The world of algorithmic investment trading, the world of image recognition that shows up in many elements of our life, the world of predictive maintenance on a whole range of devices from aircraft engines through to smart factories. The processing and personalization of healthcare information are all real illustrations today, and we are literally just scratching the surface of what is possible in the world of data. Data is a disruptor. That is a major theme in certainly my part of this morning's conversation. The growth of relevance of data is quite extraordinary. IDC reported that the digital universe, the universe of collected data, was five zettabytes of usable data generated in 2013. That's a mere 21 zeros, by the way. By 2025, they predict that we'll be living in a world of 160 zettabytes of usable data collected.
That's a 30 times increase In the amounts of data sensing, in the amount of data storage, in the amounts of data processing and value creation in the world. That trend, as I have spoken, is just beginning, and it is a trend that is a virtuous cycle. The more we collect, process, and harness value from, the more we want to collect and harness value from new data. The data economy is no longer just about collecting, storing, analyzing, and synthesizing. The data economy that we live in today is a world of intelligence. What that means is it's a world where the value proposition is improved insights and enhanced quality of action. For that reason, we believe that the fundamental future of Lam Research is inextricably linked to the rising importance of a data economy.
When the objective in a data economy is better insights or higher quality action, there's plenty of independent analysis that provides support to the story that we're sharing today. Intelligence, artificial if not human, increases with the quantity of data collected, no doubts. Secondly, the enablement of value creation is massively enhanced through advanced compute. That's what these two slides, these two graphs tell you today. Why is accuracy important? Well, the answer to that question is, it's all about the use case. Having an autonomous vehicle that is accurate 90% of the time is probably not going to meet the standards of safety that are relevant to our expectations as users and consumers. The world of investment trading with an accuracy of 50% might not meet our expectations for fidelity of investments. Perhaps for some that might be an improvement, I don't know, on investment performance.
Clearly, there is a very important connection here between the value proposition which requires action, and the value proposition requires insight with quantities of data and processing for the purpose of enhanced quality. Worthy of note, this world of data is no longer simply a world of structured information in simple databases. Now, the need is the need for real-time processing of unstructured data. That is a dramatically different technical challenge. That is a leading-edge technical challenge in memory and storage and in compute. AI, artificial intelligence, cognitive decision support, is hugely enabling of new services. It is hugely enabling in potential for leaps of productivity. It is literally a revolution. All sorts of devices in the world today are becoming sources of data and processes in their own right. A cell phone is no longer just a cell phone.
A car will soon no longer be just a car. The evidence of this transition is everywhere. There is literally an explosion of data. Not surprisingly, as stakeholders, we must together take custody, not just of the technology challenges, but of the ethical and social transitions in this revolution also. With that said, the opportunity we believe is limitless. When you look at the illustrations of value here in our work life and in our personal life, the growth curves that are relevant to the creation and acquisition and processing of data are literally three times, four times, five times for every single one of these applications. Some of the pieces of this system are already in place. The cloud's capability is in place. The cell phone as the primary vehicle of personalized connectivity is in place.
Many other elements of this are yet to come, hence my reference to this opportunity is only just beginning. This change is explosive, and the role of silicon is very fundamental. At a holistic level, when we answer your questions around opportunity for silicon and opportunity for our company in the context of our customers investing in capital, we typically speak to the industry holistically. I'm not going to do that today. I'm going to be very specific to some examples that I think tell the story of opportunity for our company. In case you're interested, more or less, we align to the consensus of the industry around bit growth and investments in WFE.
We don't show up here today with a different story to the story of our customers, nor do we show up with a different story to the story that many of you are communicating to the world in your role as analysts. The data center is quite fundamental, and it's fundamental in several ways. It's big. Wall Street Journal reported last year that Amazon, Alphabet, and Microsoft combined invest in the range of $30 billion in capital leases in a 12-month period associated with data center. It's big. More, it is the platform of innovation in memory storage and also compute. The irony here is our industry has gone through a consolidation of industry participants, but at the very same time, it's gone through a diversification of end-use demand.
On this stage five years ago, if we were having a conversation about the future of our company and the demand for capital equipment, we would have been simply talking about IC units, integrated circuit units. We might have been talking about PC units. We certainly would have been talking about cell phone units. That conversation today is literally irrelevant in trying to answer the question about the sustainability of investments in capital. The story today is not at all about units. The story today is about density and it's about content. The story today is about applications in innovation, which is driving the appetite for content. Some very simple illustrations here. A decade ago, PC DRAM was seven times bigger than the demand for DRAM from the server. Today, the server is twice as big as the PC in its demand for DRAM.
In the same timeframe, looking at NAND flash mobile, NAND flash was 10 times bigger than solid-state drive demands. Today, solid-state drives in enterprise exceed the demand for mobile NAND flash. In addition to a product diversification and an entirely new catalyst demand, the NAND flash world has transitioned from a consumer-oriented marketplace to an enterprise-oriented marketplace. That is a fundamentally different value proposition for our customers, and it's a fundamentally different world of economics in terms of ASPs and ultimately sustainable profitability. The second example I'd like to give you today is the autonomous vehicle. I use this example because there's a lot of press about autonomous vehicles in one way, shape, or form, and it's a great illustration of opportunity. First, let's answer the question, what is an autonomous vehicle?
At some level, it's a device that is capable of sensing an environment, of processing real-time to navigate without human intervention. It's also a mode of transportation. In the world of silicon, it's a device capable of sensing an environment and real-time processing for the creation of action and predictive action. It is also a complex leading-edge technology system, in large part based on silicon, that includes radar, lidar, GPS, odometry, vision, advanced controls, storage, and edge computing. The headlines here are actually pretty simple. Autonomous vehicle roadmap today has the potential to collect more data than the average server today has capacity to store. There are eight times as many automobiles manufactured every year as there are servers in the world. This is a high-impact opportunity for the world of silicon.
It is a high-impact opportunity for our customers in their pursuit of improving the quality of the world. I'm going to transition away from a demand focus into a conversation about supply, and maybe in some respects, I'm going to transition into a conversation about the sustainability of investments in our industry. On one metric at least, the equipment investment in leading-edge technology is increasing. The measure we've chosen to focus on here is the cost of investment in equipment, this reference to WFE, to achieve a single percentage bit supply growth. The headline is actually pretty simple. The headline is that in DRAM, compared to a 2012 baseline, the cost today per DRAM bit is three times greater than was true five years ago. In NAND flash, it's five times greater. The obvious question is, why is that okay? Why is that sustainable?
Why is it not a leading indicator on an impending adjustment in the investment levels or our customers in a negative way? The answer is very simple. It's all about value. The use case that I've just spent 20 minutes talking about is an entirely different value proposition for our customers and their customers than ever before in the history of our industry. The use case, the basis of the investments, the ROI, is entirely different today than ever before. We should not confuse, in our opinion, the dollars invested today and the value that can be derived from that investment with the investments and the value of yesteryear. To be clear, rising capital intensity can and does coexist with declining cost per bits in the installed base. True investment levels are higher, but again, the value proposition is increasing faster than the cost of the investment.
There is health in the broad ecosystem. The fundamental reality of device scaling, while more complex than ever today, we no longer live in the world where productivity and performance from the silicon roadmap came from traditional planar shrink. In many respects, that is the minority of technology inflections today in the industry. The world we live in today is a world of materials transitions and a world of new device structure. 3D architecture is illustrative. That is a world where the process steps of etch and deposition are central to the roadmap for silicon. Just to focus on the two examples with a little bit more specificity, our estimate for investments in DRAM in calendar 2018 is approximately $13 billion ± a billion. That number actually is almost the same number compared to 2007. In 2007, that investment would've gotten you 90% bit growth.
It gets you about 20% today. Hence my points around value and investments are an entirely different baseline today than they were a decade ago. What's really important is in the world of DRAM, in the world of the DRAM installed base, in large part from not just the additions of capacity, but the upgrades and conversions to part of the industry that's the most efficient of any segment of the industry, the cost per bit is now 20 times smaller than was true a decade ago. Our estimates in NAND flash, non-volatile memory, next generation non-volatile memory for investments in calendar 2018 is about $15 billion. Similarly, the bits out from that investment are less today than they were a decade ago.
Importantly, fundamentally, from the perspective of supply and demand balance and sustainability of investments, the cost per bit in the installed base is 50 times smaller today than was true 10 years ago. As I said a few months ago, the fundamental reality of device scaling, while more complex, is alive and well. In fact, in many respects, it's accelerating. The message here is a message of the prominence of silicon, we've chosen to represent that with a characterization of the proportion of global GDP that is represented by the revenues of the semiconductor ecosystem. Not surprisingly, when value is getting created in the 1980s and 1990s, that value creation was about the proliferation of personalized computing. There was a doubling of relevance of the semiconductor industry in the global economy. We've lived in a fairly stable environment, a plateau of relative contribution for some time.
The world of data, the world of big data, the world of analytics and AI and cognitive decision support is now reestablishing growth of relevance for the semiconductor industry, for the ecosystem of the semiconductor industry, and for the role of silicon. This is happening as a byproduct of value, there is a reason why the revenue forecast for so-called AI insights-driven public companies is currently around 30% per year. Our customers in this context have an ability to sustain higher prices without compromising the health of the broader ecosystem. It is true to say, in the last five years, that semiconductor revenues have increased by $100 billion. That's the scale of revenue expansion of the top five companies in the industry that we serve.
It's also true to say that the top seven cloud companies in the world have added over $200 billion in incremental revenues in the same period with a growth rate of 20% in that period, with an operating profit of 30%. This is a commentary on ecosystem balance. This is a commentary, we believe, on sustainability. At the end of the day, the most fundamental index that is relevant to predicting the sustainability of investment in silicon, aside from demand, are the economics of our customers. While it's valid to conclude that our customers have never invested more in silicon device manufacturing capacity in calendar 2017 or calendar 2018, what's also true is that our customers have never invested a smaller proportion of their profits in any period in the last 15 years. That healthy ecosystem, that healthy investment profile shows up in every segment of our industry.
It is actually possible, in my opinion, that we show up in five years' time, this slide is not here. There is no segmentation of semiconductor companies because the semiconductor community is trying to harness value from holistic systems integration. The world of embedded logic and memory is illustrative. We see stronger technology roadmaps today than we communicated with you 15 months ago in our last version of this meeting. Every element of this ecosystem has discipline. We continue to believe that our customers will be investing with discipline, that there will be a prominence of conversion, upgrades, and reuse in their strategies.
At the same time, we believe in this timeframe, 2017 through 2021, there will be an addition approximately of 1 million wafer starts per month of capacity in the semiconductor ecosystem to support the roadmaps of the data economy and the world of AI that I just communicated. One of the realities for Lam Research is we have today an engagement with this ecosystem that is so much stronger than at any point in our history. We just completed a calendar year where we reported record shipments for the company in the microprocessor logic space. The year before last, we reported double-digit market share growth in Foundry for the first time in the history of the company. Our engagement in every element of the semiconductor industry is higher today than at any point in our history.
In addition to that, the fundamental origin of strengths in the company, the fundamental strength from the platform of memory, is becoming more valuable. It is becoming more valuable in the world because of the roadmap of the data center, because of the roadmap of embedded memory and logic. The history and future of embedded memory and logic, and a world of memory-enabled neuromorphic computing are simple illustrations of why a strength in memory is an underlying asset. To make that point more clearly, perhaps, if we look at the last five years, the wafer fabrication equipment's CAGR in memory has been 24%. Over that same period, the growth of wafer fabrication equipment in logic foundry has been 2%. The role of memory, the role of the data center, the position of Lam Research, is increasingly more strategically relevant and important.
It is increasing in scale of opportunity, we believe. Now to transition into a few comments about Lam Research. As I mentioned at the outset, our purpose and our objective is actually quite simple. It's very hard to do, but the orientation we have is actually quite simple. We have a genuine commitment to enable and support the success of our customers. Without customer success, there is no opportunity for Lam or any other member of the ecosystem. Our focus is always on the most difficult challenges for our customers. We have an ambition not just to grow the size of our company, but we have an ambition to grow the strategic relevance of our company. All applications and all opportunities are not equal.
They are not equal in terms of their impact to the success of our customers, they are not equal in terms of opportunity for financial return for an equipment company. Our focus as a firmly established multi-product and services company, our focus as a results-on-wafer equipment company, our focus as a complex systems engineering integrator of hardware, software, process control, and materials is on solving the most complex problems of our customers. Creating for them value, capturing for us fair value, investing in our future. We have, whether we're credited for it or not, we have, as a management team, a responsibility to think long term and manage long term the sustainable, profitable growth of our company. In that context, one other element of our focus in Lam is on the optimization of capital allocation.
Every single decision we make, whether it's an organic investment, a non-organic investment, an investment in our own stock, is made with one simple test. It's the test of equity accretion over a long-term horizon. The context for solving the customer's problems should not be lost on us. It is the context of the so-called grand challenges. This is a slide you might have seen characterized by Lam before. A version of this slide shows up in every equipment company presentation of this nature. What's really important here is the challenges of atomic-scale precision and atomic-scale processing. The challenges of surface integrity, the inflections of high aspect ratio structures in 3D device architecture, the world of RC metal contamination killer defects. That world is placing a greater emphasis and a greater value proposition on the fundamental technologies of materials deposition, dielectrics and metals, and selective removal.
That is the space that Lam Research enjoys in the semiconductor ecosystem. Through the delivery of enabling technology, trusted productivity, and speed to solutions, we make possible in contributions with other companies the resolution, the overcoming of these challenges by our customers. That's what creates business success on the supply side of this conversation in the context of the demand side that I characterized before. Our focus is on investing in the quality of the team, building the most competent and capable etch and deposition company in the world. Our focus is on constantly broadening the product and services portfolio of the company. In the last five years, there has been a dramatic change around the breadth of the portfolio of Lam Research.
On the financial side of our definition of success, the more traditional definition of success is defined around having a product and services portfolio that competes for a greater proportion of customer spending every year, ultimately winning a greater proportion of our customer spending every year. Calendar 2013 happens to be the first full year after the integration of Lam and Novellus. You can see in that year, the shipments of Lam Research represented about 14% of overall wafer fabrication equipment spending.
Since that time, through the investments we've made broadening a product and services portfolio, since that time from the engagements that we have increased in memory to establish very strong leadership positions in critical enabling technology inflections, through that time, the position we've established in the world of microprocessor logic and foundries has enabled us to compete for and win, measured by the proportion of WFE that our shipments represent, now 21%. A 50% increase in a five-year period of the proportion of investment by our customers that we have competed for and successfully won. Needless to say, we're not resting on our laurels. We're investing in our future, and we expect to continue to increase the proportion of wafer fabrication equipment spending that we're competing for, and over time, winning.
In addition to that, our plan over a five-year period is to return at least 50% of our free cash flow to our shareholders through dividends and capital repurchase. Doug Bettinger will share more specificity around that a little later today. Through a comprehensive corporate strategy that we will share at a very superficial level today, as time only allows that, is based on a genuine commitment, as I've said, to the success of our customers. We have invested over multiple decades in building a company with an effective execution capability, with good strategy, with a culture that is truly enabling to create effective teamwork across organizational boundaries, to create value propositions that have their origin in the creation of an environment in our company where people achieve their full potential.
Unit process focus, the business unit focus, the focus on having a better deposition system or a better etch system in a certain application than any other company in the world, actually has been the primary axis of strategy that has gotten us here. That continues to be a very relevant part of our future. Architecting competitive differentiation on the axis of unit process, on the dimension of applications of wafer fabrication equipment processing in the fab. We are supplementing that strategy in our future with a focus on multi-business unit value proposition. Two simple examples.
With the breadth of the product portfolio that we have today, with the engagement that we have across the entire ecosystem, with the adjacency of etch and deposition process in the device manufacturing world, we have adjacent process co-optimization and even integrated product opportunities that are very significant growth trajectories in our company. Our vision isn't to deliver that with the insularity of our business. Our vision is to do that with the perspective of enhancing solution sets and creating optionality for our customers. We also have a tremendous opportunity to harness the value of our culture, our values, the strength of our demonstrated performance with ecosystem partners, with the academic community, with the world of consortia, with our customers, with our suppliers, sometimes even with peer companies, of harnessing enhanced market penetration, enhanced competitiveness and share through leveraging ecosystem value propositions.
We are invested more in ecosystem partnerships today than we were five years ago. We will be invested more five years from now in ecosystem partnerships than we are today. We also have an opportunity to harness a period of strength in our company, not just prospectively, but also in the future, to leverage the core competencies of Lam Research, the assets and the strengths of our company beyond the definition of today's products and today's marketplaces. As I said on a prior slide, the decision-making framework that we use to make the decisions on investments organically, non-organically in our core products and markets or adjacent markets, is defined around a focus on the dollars of EPS growth, on free cash flow generation, and on the equity accretion test over the long term that I summarized a few moments ago.
In summary, we believe in a very healthy industry with an ecosystem that has never had more balance than it does today. Never has more sustainability, in our opinion, than it does today. At Lam, we have a very disciplined focus on building competitive differentiation and contributing to the success of our customers. That is the essence of the vision and the strategy of the company. We have clearly achieved a record of outperformance in the last five years that is second to none. We have grown our company over a five or six-year period with a shipments CAGR of 24%. We just completed a calendar year where our shipments grew by 50% compared to the baseline of the industry of 30%. Outperformance is a reality in our history. We intend outperformance to be a reality in our future.
It is a privilege to lead a company, and that is even more so true with the special elements that I believe, that we believe exist at Lam Research. I'd like to take this opportunity to thank our customers, our global employees, and our partners around the world for making the history of Lam possible and making the future of Lam extraordinary. We are proud of our accomplishments so far, but we stand on this stage today inspired by an opportunity to show the world that our best is yet to come. We believe that we live in an entirely different global economy, with an entirely different set of value propositions associated with the role of data and the value creation opportunities that ultimately enhance the lives of every person on this planet. We believe in the engine of data and the prominence of silicon.
We live in a different industry today than ever before. We are a different company. With that, I transfer to Tim Archer, President and Chief Operating Officer of Lam Research. Thank you.
Thank you, Martin. Good morning. I am quite certain this is the first and maybe only time I'll be introduced to a presentation as that other chap. It doesn't come across nearly as well without Martin's British accent. In any case, you've heard from Martin about the data economy, and it's a compelling story for increased capital spending by our customers. I'd like to spend the next 30 minutes talking more about Lam Research and specifically about how the technology innovations that are required to enable this race to the data economy are creating tremendous opportunity for us.
I'm also going to talk about strategies that we're employing to differentiate our company, to differentiate our products, our services, and as Martin just mentioned, our collaboration model, not only with customers, but with also other partners in the semiconductor ecosystem as a means of ensuring that we win in these expanding markets. I'm also going to talk about actions we're taking to increase the stability of our business model by delivering new products and services to our rapidly growing installed base as a means of delivering significant value to our customers by helping them manage the performance of Lam equipment over the entire life cycle, and at the same time, deliver for Lam a growing and recurring stream of revenue. I'd first like to start with growth, and you just heard a pretty amazing story from Martin.
I'd like to go just a little bit further back in time and baseline you to Lam's track record of growth over the last 15 years. You'll see that in 2003, Lam competed for approximately 10% of the wafer fab equipment market, as Satya defined for you, WFE. In the ensuing 15 years, through the execution of strategies that have included acquisitions into the clean, the deposition, and process control spaces, as well as the development of a broad portfolio of products targeted at some of the most critical and fastest-growing technology inflections to have occurred over this period, Lam has been able to grow its SAM as a percent of WFE to 36% as we exited 2017. I think it's fair to say that no other company has been as successful in our industry growing their served market over the last 15 years as Lam.
I think our story is quite unique because we actually don't see this as the limit of our opportunity. There's at least one company in our industry today that competes for 60%-65% of WFE. Given the technology inflections that are required to enable all of the innovations that Martin talked about, as well as looking at the pipeline of new products that we have in our own development, we see a clear path to compete for at least 40% of WFE by 2021, with continued room for growth beyond as we identify the next wave of inflections. You might ask, what has to change in order to keep Lam on this track record of growth that we've built over the last 15 years? The good news is that from Lam's perspective, we really don't have to change much.
We're going to continue to identify and invest in technology and productivity innovation that helps our customers achieve timely and cost-effective device scaling. There is one thing that's changed and will continue to do so, and that is the criticality of the role that etch and deposition play in device scaling. In years past, maybe in the world that Martin talked about of 2D shrink, lithography was generally looked at the primary technology that determined the rate and pace of scaling. Today, etch and deposition play that same central role in determining node-to-node transitions. Perhaps even more so when you really look at the technology innovations that are required for advanced memory and logic scaling. I would like to dig just a little bit deeper and look at the specific competencies that a company needs to be successful considering this next wave of inflections currently underway.
Just looking at this list, there's maybe just one thing that you should capture, that is they're all etch and deposition-centric inflections. Hence, we see opportunity for Lam leadership broadly. Take vertical scaling, for example, especially for 3D NAND. We're now talking about film stacks 7 to 8 microns high. We're talking about aspect ratios for etch in the 70 to 80 to one regime. To succeed in these markets, you have to have fundamental core competencies. You have to have technology leadership and high aspect ratio etch. Lam does. You have to have fundamental core competency in designing high-productivity deposition tools for these thick stacks. Lam has that history.
For atomic scale processing, of course, for ALE and ALD, as you heard from Satya, it helps to start as a technology leader in etch and deposition as we evolve those processes to the point where we can control them at the atomic level. It's tremendous opportunity for us to focus on technology as a means of winning. Technology is not the only thing we focus on as a means of delivering value for customers. We're also innovating productivity for our installed base to ensure that our customers continue to receive more value as they use that equipment over its entire life. Rather than just hear from me, I wanted to take this opportunity to allow some of our other Lam leaders, both from business and technology, to tell you about our growth opportunities in their own words.
I'm going to play a short video. In this video, you'll hear from Sesha Varadarajan. He's the General Manager of our deposition business unit. You'll hear from Vahid Vahedi, the General Manager of our etch business unit, Pat Lord, the General Manager of our CSBG business unit, that's the installed base, and Rick Gottscho, our Chief Technology Officer. Let's play the video.
There's a worldwide driver in the end markets for increased data collection, storage, and manipulation, which in turn drives higher capital intensity in the semi industry. Embedded within this macro trend is a deposition SAM that's outgrowing the WFE by almost 2x from a historical basis. This outperformance is driven by increasing intensity of CVD ALD for enablement of vertical scaling and 3D NAND, increasing adoption of multi-patterning, and also new performance materials to enable device scaling.
Our etch products are becoming more strategic relevant in our customer wafer flows as we enable vertical and horizontal scaling. Vertical scaling is in production in 3D NAND at multiple customer sites. Vertical scaling also showing up in advanced logic as well as chip packaging. Patterning is really key for horizontal scaling in both DRAM and logic. With our advanced product offerings, such as Hydra patterning solutions, we're very well positioned as we enable horizontal scaling.
Over the last four years, we have seen our CSBG revenue cumulative annual growth rate grow 50% faster than that of the install base, demonstrating a steady improvement in the annuity we derive out of every Lam tool in the install base.
Our strategy for future SAM growth is to focus on solving our customers' toughest problems. As we work with our logic customers to transition them to the five-nanometer, we see an increase in etch intensity in critical transistor applications. We also see a strong demand for our advanced equipment process control that are designed to reduce variability as well as enhancing the yield.
Lam is also enabling SAM growth by a couple of different new strategies. One approach that we are taking is towards offering integrated removal deposition systems, which in turn allow queue time control for our customers. Another trend of note is around RC management in the future. As devices shrink, it stresses the natural limits of some of the materials we use to form wiring in our devices. This in turn drives the need for many novel new materials deposited through plating and ALD technologies to continue scaling the devices. These are core competencies of Lam and serve to increase our SAM. Another trend of note is the requirement to change most of the critical materials in a multi-patterning stack to improve its fidelity performance needed for the future device performance. This in turn requires new ALD and CVD materials to be deposited and delivered to our customers.
We will continue to grow opportunity in CSBG because we are focused on creating innovative new products and services to help improve the performance of the install base.
This is a really exciting time to be in the semiconductor industry. I don't recall any other time where every element of the technology, from the transistor to the package, is inflecting. What's particularly exciting about the future with all of these inflections going on is the overarching inflection associated with the convergence of logic and memory. The memory chip will actually be the logic computer of the future, and Lam's in a fantastic position with leadership in deposition, etch, and clean to exploit those new opportunities. Hopefully that was some helpful perspective for you from our other leaders. A lot of important information in there about different technologies, different products.
If there's one takeaway from what I've said this morning and what you just heard in that video, it should be that etch and deposition now play a very central role in device scaling, and the opportunity for us to innovate is tremendous. In fact, if you look at the technology innovations in our product portfolio and our customer engagement broadly, we believe Lam's opportunity to grow is unmatched in the industry. One reflection of that growth potential is shown in this set of charts. These charts show normalized SAM per wafer for Lam in a greenfield investment context. If you're looking for one headline from these charts, maybe I would offer that it is that Lam SAM grows across all device segments, across all technology nodes for at least the next four to five years shown in the time horizon of the charts.
If you wanted to dig just a bit deeper for another headline, it would be that Lam's SAM grows in all of these cases, even with the backdrop of EUV adoption, and our adoption assumptions are consistent with industry consensus. Lam SAM grows with the backdrop that it's in this timeframe that the 2D to 3D NAND conversion cycle also completes. You might ask if this data is kind of relevant to how the investment profile will look in the next four to five years. It's an important question. We believe it is. We see that, from our view, the majority of WFE over the next four to five years will be spent in the context of a greenfield investment.
Even if we're wrong a little bit on that point, we've said in the past that, for instance, in NAND, Lam's market opportunity actually is relatively agnostic to whether the investment is made as a conversion or a greenfield. Hopefully, this gives you some sense of the power of the technology inflections, the power of our product portfolio development, and some sense of the growth opportunity in front of Lam. I'd now like to shift gears just a little bit and talk about the foundation of how we differentiate in our markets to ensure that we win in these growing areas. I'd like to focus on four things: technology leadership, productivity, speed to solutions, and collaboration.
Since I said that these are the foundation of how we differentiate, I thought it might be helpful to start by talking about the culture and core values, the DNA of who we are as a company. Spent a lot of time this morning talking about our focus on technology innovation and new products, and that springs from a culture in Lam that rewards innovation and continuous improvement. I talked about how we've been successful at acquiring and integrating new companies into Lam to expand our product portfolio, and that comes from a culture that values mutual trust and respect, that fosters teamwork and collaboration, not only within Lam, but also, again, with customers in the ecosystem.
We have a culture that teaches every employee to think customer, company, individual, which I think differentiates us because it sets us all to one mind that our success is only achieved through the success of our customers. We have a culture that commits us to strategic, substantive, and focused investment in R&D to ensure that our product development pipeline always remains full. Now, to be very clear, it takes about two minutes to describe our culture. It's taken us decades to build it across 10,000-plus employees. What does it mean for our customers? Well, from a technology leadership perspective, it means that customers have come to trust that they will get from Lam a constant stream of new products that target their most difficult challenges.
Just take, for example, a couple of the tools on the list, the Flex high-aspect-ratio etch tool or the ALTUS atomic layer deposition tungsten tool. Both of these tools are technology leaders. They are the benchmark in their fields, and they're both enabling the 3D NAND vertical transition. For productivity, it means our customers can count on us to continue to drive to help them lower their costs and improve output from their fabs. For instance, the VECTOR Quad Station deposition tool that's used for the manufacturing of those thick stacks in 3D NAND. This tool sets the productivity benchmark for wafers out per square meter of fab footprint, which is becoming an increasingly important metric for our customers. From a speed to solutions perspective, it means they can count on us to continue to make investments like we did in the recent Coventor acquisition.
Our customers use Coventor software to simulate complex device structures to identify problems before they commit those designs to the costly and time-consuming manufacturing process. From a collaboration perspective, it means that customers and partners in the ecosystem can count on us to be the most engaging, the least threatening, most genuinely committed collaboration partner. Now I'd like to just give those same guys, Rick, Pat, Sesha, Vahid, a chance to tell you about how they see us differentiating and winning in our markets. Let's play this video.
Lam leads the industry in solving difficult metal-filled solutions, depositing challenging conformal films, and delivering innovative etch top layers. All three of these technologies are critical to enable continued device scaling of our customers.
Lam is a leader in critical etch applications. We focus on solving our customers' toughest problems. For example, we have 100% share in transistor hole etch for 3D NAND. Lam has also led the market with innovative products, such as the Hydra patterning system. We have 1,000 chambers in the install base with that feature on the tool. We're also introducing our Corvus for extreme etch yield enhancement to be going out to our customers this year.
When customers buy Lam equipment, they can rest assured that we will have not only roadmaps to ensure that their equipment is state-of-the-art and ready to manufacture at the leading edge, but also obsolescence and life extension solutions for our customers with older generation tools. We have developed advanced analytical tools to allow our customers to use big data generated in today's connected fabs to fine-tune their system performance. We are also increasing our data analytics solution for preventive maintenance and virtual metrology. Productivity means that when we ship a new system, it needs to be up as quickly as possible. When it's up in high volume production, it needs to be running at the highest availability, with the lowest possible number of defects, and at the highest possible throughput.
When it's time for maintenance and to replace parts, we need to have the right expertise and the right spares at the right time, at the correct location. We are also innovating by releasing new advanced productivity upgrades and solution that focus on improving our customers' tool fleet productivity.
Fueled by our Quad Station architecture, where essentially our engineers have put four processing stations within one envelope, Lam sets the benchmark for both raw output on the system as well as output measured as a metric of unit space that we use. When you combine this intrinsic performance with world-class service products, world-class tool matching products that we offer, this means we deliver by far the lowest cost per wafer for our customers for the most challenging applications.
Moore's Law in 3D NAND has gone from an 18-24 months cadence into 12-18 months cadence. It is actually accelerating. Given that we're the providers of critical solutions to our customers, focus on speed to solution is key in delivering those solutions to our customers at the right time. We recently acquired Coventor, a provider of virtual fabrication and software modeling. By integrating Coventor into our process development, we're trying to improve the time to solution to deliver the critical solutions to our customers faster.
Considering the criticality of etch dep co-optimization in almost all of the customers' device roadmaps, our extensive investment in internal optimization of etch and deposition results in faster solutions to our customers. This is not only a market enabler, but frequently also enables time to market improvement for our customers.
Collaboration with our customers and partners is key to our growth. Collaboration with our customers at the top R&D sites enables us to be the partner of choice for our customers for the most advanced applications. We are very uniquely positioned in the ecosystem with lithography and advanced process control and other suppliers to deliver the most advanced and most cost-effective patterning solution to our customers.
As we look ahead to this brave new world of new applications, atomic layer processing is a huge challenge. In addition, we see a huge challenge in maintaining the integrity of surfaces and interfaces, again, with atomic level precision. The last set of big challenges are all around affordability, and can we solve these problems in such a way that we have sustainable solutions that are benign to our planet's well-being? To attack these grand challenges, we are investing heavily in artificial intelligence, in machine learning, in big data mining, and additional sensors, essentially the Internet of Things. Relying on the new technologies that we help to enable, we are very optimistic about our ability to differentiate in overcoming the grand challenges that the industry faces.
Okay, that was the last video. You've now heard about technology and how it's driving growth in our markets. You've heard about differentiating to win. Maybe the headline of this slide gives away the next punchline, but what does it mean for our market momentum? We exited 2017 with our deposition share in the low 40s, our etch share in the mid-50s. Through execution of the strategies that we've just spent the morning talking about, we plan to add 4-8 points of market share to both of these businesses by 2021. I debated whether this should be my last slide. It's a powerful statement, both of the growth opportunities that exist in our markets, as well as our confidence of winning those opportunities. Remember, I said I wanted to talk about our installed base business as well.
Our installed base business is a business that doesn't get discussed a lot externally. We've said in the past that it represents about 25% of our overall business, but it also represents a tremendous opportunity for us to create value for our customers and capture value for Lam. We wanted to make sure we gave it ample time today. I promise I have a bigger headline than just market share as my last slide. The installed base business, it is a key element, as I said at the beginning, of our strategy to increase the stability of our business model. The reason for this is that the opportunities in our installed base are to first order decoupled from wafer fab equipment spending in any given year.
The reason for that is that the opportunity from the installed base stems from a set of equipment that we've been working to build now for decades. That installed base now numbers more than 50,000 process modules for Lam. You can imagine when your installed base grows to this size, the customer challenges vary. Many of them are unique. The Customer Support Business Group, CSBG, had to become very good at developing specific solutions to specific customer needs to deliver very specific customer value. For instance, for DRAM and NAND customers, the development of productivity upgrades for their installed base or the development of spare parts lifecycle management solutions, these types of activities help them meet their key challenge of continuously reducing running cost of the installed base. In the case of foundry and logic customers, Martin just talked about artificial intelligence and big data.
This is an area where Lam's actually leveraging artificial intelligence, big data, and advanced system analytics as a way of helping these customers meet their very specific need for reduced variability. We do this through improving chamber-to-chamber matching, system-to-system variability, using those new advanced tools. In the case of IoT, where there really are very many applications in this space, we have a Reliant business that helps deliver non-leading edge tools at a price point that helps these customers often meet the cost challenges of their markets. I thought it might be also helpful to explain the lifecycle of Lam equipment and the numerous touchpoints that we actually have to create value for customers and capture value for Lam. It starts with, of course, the shipment of a new system, after which we have the opportunity to sell spares, services.
A couple of years down the road, maybe three, four years down the road, customers typically look for us to provide upgrades to the tools, either to extend the technology to the next node or to improve the output or running cost. Another opportunity for us to create value for customers, capture value for us. Advanced services, I just mentioned the new analytic tools now engaging on the installed base. 10 to 15 years down the road, customers will often look to repurpose a tool from a leading-edge fab, either to another fab within their own network or to the external marketplace. This is another opportunity for Lam and our Reliant business to step in and offer refurbishment services, tool relocation services, or if desired, resale services. Once that tool is in its new home, the whole cycle starts over again with spares, services, upgrades, and advanced services.
Maybe there's two takeaways just from this lifecycle that I'd like to make sure you capture. One is that a tool that's shipped today, with our growth, we've been, as you know, shipping a lot of them, a tool that's shipped today will likely generate a predictable stream of revenue for Lam for decades. In fact, if we look, approximately 20% of the modules that are in that 50,000 module installed base have been out in the installed base generating revenue for Lam for more than 15 years. Second takeaway is that by innovating to create value for customers over the entire lifecycle of Lam equipment, we actually have the opportunity to capture, on average, more revenue over the lifecycle of that piece of equipment than we do at the time of initial tool sale. It's a strong business for us.
Maybe another way of looking at the quality of the business is in how it's performed relative to the underlying volume drivers. This chart shows from 2003 to 2017 that industry wafer starts increased by approximately 150%. We know Lam outperformed. Our installed base grew more than 300% over that same time period. Lam's installed base business grew even faster, increasing by more than 400%. If we look in the more recent period, and you heard Pat talk about this on the video, where we've been focused on developing additional products and services to expand our portfolio as a means of delivering more value to customers, we've grown the installed base business even faster than the installed base, 50% faster in the last four years. Hopefully that's a little extra information on the installed base that you'd find valuable.
I'd like to now sum up. You've heard three things today from me. You've heard how the technology growth drivers are creating new opportunities for Lam, expanding markets. You've heard how we're differentiating to win, and you've heard how we're trying to create a growing, recurring revenue stream by delivering value to customers through our installed base. What does it mean? From the technology growth drivers, we plan to add more than $1 billion of incremental revenue by 2021. From the differentiated solution strategies, we plan to add more than $1 billion of revenue by 2021. From the installed base business, we plan to add yet another $1 billion of revenue by 2021. On flat WFE, flat to 2018, we plan to deliver $15 billion of revenue in 2021.
With that, I'd like to turn it over to Doug Bettinger, who will give you more details on the financials.
Good morning, everyone. It's great to see a lot of familiar faces. It's also great to see some new faces. It's exciting to have the room full here. It feels like every Analyst Day or Investor Day, we get to this point and I look back on Martin's presentation, I think, man, he had some really cool graphics. Interesting stuff. I look at Tim's presentation. Tim had some really cool videos. Had atoms bouncing around on the video. Every year I tell myself, or I tell Satya, "We need to get some cool stuff in the CFO presentation," and it never happens. All I have in these materials, that I'm going to take you through, are some boring charts and boring numbers.
I'll be honest with you, to me, these boring charts and boring numbers tell an amazing story of an awesome performance of Lam Research and a great industry. All kidding aside, I really have two things I'm going to take you through, and I know you're here to hear about the updated financial model. Tim foreshadowed the top line, and I'm going to give you a little detail on the capital return program that Martin foreshadowed in his material. Before I do that, let me take you through just a little bit of the historic performance of Lam Research. 2017 was an amazing year for the industry. In 2017, WFE was up a little over 30%, went from $36 billion in spending to $47 billion. Memory led the way. It was a memory-driven growth. A great industry.
Lam was up nearly 50% in terms of shipments and revenue, roughly 50%. That's 1.5 times the growth of the industry. Amazing result when I look at it. Operating income, and this is my favorite part, we delivered great leverage to the bottom line. Operating income was up 1.8 times faster than the growth in revenue. We expanded operating income by approximately 6% to 29%. Earnings, we couldn't quite get to 100% growth. Pretty close, though. EPS went from, again, non-GAAP EPS, from $7.06 a share to almost $14, $13.70. Great performance. More importantly, we had amazing operating cash flows. Over $2 billion in cash generated, which is very relevant to the capital plan that I'm going to talk with you about in a minute. A great company in a great industry, in my humble opinion. This wasn't a unique year, though. This wasn't a one-off year.
This has been a great period for Lam. It's been a great period for our industry. Again, let's start with WFE. WFE in the last five years is up a double-digit CAGR, 11.1%. It's grown from 2012 in the high 20s of billions to the low 40s of billions , and Martin showed you that. This is an absolute affordable level of investment in the industry. A great industry, quite honestly. A great company and a great industry, Lam Research. Our CAGR over this timeframe, 24.7%, two times faster than the growth in WFE.
As Tim shared with you, it came from being in the right place at the right time, our addressable market got bigger, and we executed extraordinarily well and gained share as a result of that execution. If you look at it, we're the fastest growing semiconductor capital equipment company of our size in the industry. This has been awesome performance, in my humble opinion. Over this time period, our EPS CAGR, 45%. It's gone from a little above $2 in earnings to nearly $14 in EPS. Most importantly, we generated $6.7 billion in operating cash flow. That's super exciting when you think about what happened in December with tax reform, that cash is now more readily available domestically for us to do different things with it than we've been able to do before. I'll talk with you in a moment about that.
Let me take you through a little bit of some trends. When you're showing financial slides, you always want to show a graph that's up and to the right. These graphs are beautiful in terms of the up and to the right, again, in my humble opinion. Shipments over this timeframe have grown very nicely from 2014 to 2017, doubling, nearly doubling. Similar with revenue. As I was thinking about what story did I have to tell, when I was going to present this chart to you guys. I thought back to, this is about the time period that I've been with the company. I've been here a little bit longer, 5 years. I just came up on my 5-year anniversary. I went back to that earnings script from 5 years ago, and what did I have to say?
The observation I had was, in March of 2013, the shipments of our company, it was $896 million. It was a good performance. We were pretty pleased with that back then. Shortly after that, we crossed the billion-dollar mark, had a celebration for the employees of the company. We were very pleased with how we were progressing. Fast-forward to the guidance that we provided for the current March quarter, 5 years later, $3.175 billion. That's 3.5 times the growth of just 5 years ago. We're not a startup company. We're a mature company in a mature industry, growing at rates that are actually pretty impressive. Like I said, on the chart, you can see shipment and revenue double over the 2014 to 2017 timeframe.
We've made significant investments with our supply chain customers to scale the company up with our own manufacturing capability, our own field resources to make sure we've been able to support that. Over this timeframe, operating profits are up 2.9 times, and operating margin has expanded by 10 percentage points. Pretty impressive, again, in my humble opinion. Like I said, we delivered nearly $14 in EPS, which was up 200% since 2014. How did we do this? Just thinking through, how were we able to deliver this? Tim talked with you already about our SAM as a percentage of WFE has expanded. Our market share has expanded because our execution has been strong. We're not done delivering on that growth. We're going to continue, and I'll share that with you when we get to the financial model.
I guess not always do you want to see your charts up and to the right. If you're looking at expenses, down and to the left is a good place to be as well. Operating expenses, a % of revenue declined 800 basis points from 26% in 2014 to 18% in 2017. We delivered nice leverage to the bottom line. At the same time, we have significantly increased our investment in research and development. In fact, we've grown our share of R&D spent by the top 5 semiconductor capital equipment companies in the last 5 years. In 2017 alone, we spent over $1 billion in R&D. I think maybe what we're most proud about is we've grown R&D at a rate double the rate of SG&A at the company. It increases the likelihood of success that we're going to deliver future revenue growth.
It drives the execution of those market share gains that Tim talked about, four to eight points in both etch as well as deposition. It positions us well for that SAM expansion that you see on the left side of the page. You know, as a CFO, I don't think any CFO would be unhappy with that chart of return on invested capital. We're very good stewards of your capital. We think about returns when we make our investments. We think about the areas of the best return when we're spending money. An ROIC of 70% is six to seven times our cost of capital. That's value creation. We think about this when we place our R&D investments.
I thought I'd spend a little bit of time and talk about the business model, because I think this is important to understand, and oftentimes I think it's lost, and perhaps we don't always talk about this as much as we should. In 2014, roughly 70% of our costs were variable. In 2017, roughly 70% of our costs were variable. There is a high variability in this business model. We've maintained this flexibility, this variability, even with that doubling of shipments over this timeframe. It's important to understand. How does this happen? Materials costs are a big part of cost of goods sold. We have a highly variable labor force in the manufacturing environment, a lot of temporary workers that can flex up and down with volume. We use local service partners to do installation and warranty work in the field.
A decent amount of the compensation of the regular full-time employees at the company varies with the level of profits of our business. When you think about that, we've got a high variability. It's enabled us to scale up. I know a lot of you worry about, well, what if things turn the other way? We'll see that same level of variability. One of the things I thought I would just preempt your question, because we get it all the time, it drives me crazy, but I'll answer the question anyway, which is: what if business turns down? We thought about, what if WFE went back to $40 billion next year in 2019 or beyond?
As I think about this variability and think about the growth in SAM and share that Tim described, as well as the growth in the installed base, I think earnings, if that's the environment in 2019, would look very similar to what they looked like in 2017, plus or minus. I'll preempt your questions that are going to potentially come later, which is, earnings are pretty solid no matter what the spending level of the industry is. By the way, that's not our view. Our view is that WFE actually continues to be strong for many years because of what Martin described. I wanted to answer the question that we get all the time. I know you guys and gals are interested in the business model. I decided we're not going to do a new business model. I'm just going to give you 2017.
You guys can figure it out. No, we're not going to do that. I wanted to provide 2017 on the page, though, to give you an anchor. As we thought about this and had a debate, what WFE should we use? What are we going to do? I'm going to give you a 2021 model in a moment, and the assumption we have is flat WFE to 2018. It's just an assumption. If you truly believe in the data-driven economy that Martin Anstice describes, I personally believe WFE will be stronger than that. I didn't want to debate what is WFE in 2021. The assumption in the numbers I'm going to show you in a minute is, it's just flat to 2018. It's the $3 billion that Tim Archer described to you. What does the model look like? Here it is. Again, WFE flat to 2018.
Think revenue's in a range of $14.5 billion-$15.5 billion. 15 plus or minus 500, call it. You've already seen the leverage that we've delivered in the last several years. We can keep doing this. If you look at operating income, again, anchoring off 2017, we were nominally at 29%. We think we can keep going into the low thirties. 32%-33% is our view right now. If you do the math on all that, you get to an earnings per share that we think will be in the range of $23-$25. Importantly, the free cash flow generation in this scenario, we think, would be between 28% and 29%. It's a pretty CapEx-light business. 3%-4% of revenues go to capital expenditures. The free cash flow generation capability, we think, is in the high twenties.
I can see you all doing the math on that. If it really is $15 billion, that's around $4 billion in free cash flow. We generate a lot of free cash flow. Importantly, what are you going to do with that cash? Martin Anstice's already shared some of that with you. First, I thought it would be important to remind you of what we've done. Our priorities for cash are always going to be the profitable investment in the business. Always has been, always will be. When you're generating return on invested capital of 70%, absolutely the right place to put the dollars. When you look at this, the cumulative investment since 2012, I'm using 2012 in my slides because that's kind of the timeframe when we brought Lam Research and Novellus together, etch and deposition.
The cumulative investment in R&D over this timeframe has been $6 billion, about $6 billion. That's priority one. Even with those investments, though, we generated over this timeframe $5.6 billion in free cash flow. We're investing in R&D to deliver on that long-term vision of sustained outperformance, to take advantage of that growing addressable market, and to execute well to gain the shares that we intend to gain. I'll remind you, in 2014, we initiated the first dividend in the history of the company at $0.18 per share. We've grown that dividend fairly significantly. I think our first dividend raise was 67%, the second raise was 50%, last year it was about 10% or 11%. We've been paying attention to that dividend. We know you guys and gals want to see a growing dividend.
On top of that, we've supplemented the dividend with share repurchases, $4.4 billion over this timeframe, which has resulted in returning 90% of the free cash flow to shareholders in these two forms. Now, full disclosure, those of you that followed the companies for a while, you will remember that shortly after the Novellus acquisition, we did a big buyback that was kind of part of the deal. That was $1.6 billion. If you take that out, 90% would've been 60%. Still a meaningful amount of cash returned to shareholders. What are we going to do in the future? Martin's already foreshadowed this. We plan to return at least 50% of the free cash flow of the business over the next five years.
Strong capital return, highlighted by our confidence in sustainable market growth, it does align with how we think business performs around revenue growth, operating profit, and so forth. How are we going to return this? We're pleased to tell you today we're going to meaningfully increase the dividend with the next dividend. We plan to increase the quarterly dividend by 120% to $1.10 a share each quarter, with a bias to grow this over time. Now, while not directly tied to the installed base business that Tim shared, I can tell you that the sustainability of this dividend is comfortably supported by the profits and cash generated by the installed base. I almost think of the two going together. Very comfortable with the ability to sustain this dividend with the profits generated from the installed base business.
On top of that, we're pleased to announce an increase in our share repurchase program, an incremental $2 billion on top of the $2 billion that we announced in November, to $4 billion in total share repurchases since November of 2017. I continue to view buying our shares at these current prices as good value. If we deliver on that financial model that we just shared with you, I hope we're going to be pleased that we're buying stock back at these prices. I'm going to kind of summarize things with something kind of fun. A friend of mine used to do haikus at their Analyst Day, and he's now retired. I decided I'm going to do a haiku and take that as part of what we're going to do with Lam.
I asked him, I said, "Hey, what do you think about this?" He's like, "Yeah, go for it." My haiku, our haiku. Growth in WFE. SAM share and CSBG. Earnings growth you'll see. 17 syllables, five, seven, five. Let me just summarize. I think we just finished an extraordinary year. Really pleased with the performance of the company. We're continuing to invest, and I'm pleased with our capital allocation strategy. Before I go to Q&A, let me try to summarize what we tried to communicate today between Martin, Tim, and myself. We truly believe this is a great industry. We're managing a great company. I believe we've delivered great results in the past. We plan to continue to do that. The emerging data economy is enabling a very healthy industry. We've got new demand drivers that's delivering more sustainability.
We've gone from PCs to mobility, content growth, to new demand drivers in the cloud enabled by artificial intelligence. We've delivered great performance, in my opinion, between SAM growth and share from our execution, and we plan to continue that. I still believe, and my favorite part of the business is the installed base. It's an underappreciated growing annuity. It's a cash-generating machine. We believe business will continue to grow, and we will continue to be good stewards of your capital. With that, Martin, Tim, Satya, please join me on stage, and we'll be happy to take your questions.
Just before we take a question, I was kind of gratified that we went from three pitches to four, because I was the worst-looking progression on the three. In two and a half years, Satya now claims the worst progression with four pictures. Not only did he have so much more hair, it was black then.
It's a sign of wisdom, Martin.
See, this has every opportunity to get competitive.
Before we start the Q&A, let me remind you what we said at the beginning. We are here to talk to you about our long-term objectives, and I'd appreciate it if you keep the focus of your questions with the spirit of that objective. Roman in the back.
Thanks very much. Excellent presentation, and congratulations. I guess what stands out is the expectations you guys have set for share gains in both deposition and etch. I guess I just haven't seen a lot of examples of companies with call it 50% share that are talking about 4-8 points of share gains inside of 5 years. Could you just spend some time talking about what's exactly driving those gains, where you think those gains are going to come from? Then I had a follow-up.
Okay. Well, let me try to start, then I'm sure Martin will supplement as well with his views. We talk about 4-8 points of share gain. Clearly, it is in the context, even as you heard Doug say, it's about being at the right place at the right time. We have positioned ourselves and our products into some of the fastest-growing applications in the market, those are just starting to bear fruit. We talked about our atomic layer deposition tool, our atomic layer etch tool. These are opportunities that are just starting. They represent share gain chances for us. We continue to see vertical scaling increase in the number of etch and deposition layers. It's share gain opportunity for us. Share doesn't always mean going head-to-head and having to knock the other guy out.
Share means taking a larger share of the applications that continue to evolve and appear in the process flow. That's what we spend a lot of time doing, is trying to identify those next areas where we can be first mover into the space and gain share as those basically inflect and remove another application.
Yeah. The only thing I would add is, as we've said many times, we don't win everything. If we keep winning more than we don't, then the opportunity that's in front of us relative to continuing to build our engagement with the industry, with the breadth of a product and services portfolio, has every potential to deliver competing for the proportion of WFE that we've characterized and winning it. At the end of the day, it's about building trust. It's about building trust that we're genuinely invested in the success of our customers and having differentiated technology with productivity and time to market, to support the success of the customer. It's not easy, but it's possible.
Okay, thanks. Then follow-up for Doug. I just did some quick math on the target model.
Sure.
$23-$25 per share, according to my math, implies about 170 million-175 million share count. Is that?
I think it's a little bit less than that, Roman.
Can you just talk about the dynamics? It's not too far from where you guys are today.
No, it's down a bit. There's assumptions on other income and expenses in there, on tax rate in there, and there is a declining share count assumption in there that's based on that 50% capital return assumption, some of it going to dividends, some of it going to the buyback. That's how you get there, Roman. There is a declining share count baked in that.
You driving?
Okay. Pierre Ferragu?
Yeah.
Okay.
Pierre Ferragu, Bernstein. Martin, you mentioned showing the slides about NAND, DRAM, and logic, that this slide might not exist in a few years, and that the boundary between logic and memory is going away. I'd be curious to hear from you how you see that happening at your clients. Most of them tend to be very logic or very memory. Only one of them is really about equal balance between the two. How are they thinking about it? In what kind of discussion do you engage with them on that topic? And in particular, from where we stand today, how fast do you see that playing out, and what are the big milestones?
Yeah. I think the speed of the transition is all about the delivery of value. If there is enablements and/or economics that come from embedded memory solutions, systems integration of memory and logic, or at the applications level, this kind of vision for neuromorphic computing and the role of memory that Rick Gottscho, our CTO, describes in processing, it's all about value. We don't actually have to have the conversation with the customer. We do at some level, but what is really important for us to the customer is, what devices do they intend to make, what technology inflections are relevant to those devices, and what technology investments are relevant to us to be proactively engaged and supporting them in that roadmap. Whether it plays out or not in the way I described it, time will tell.
I happen to think it will, and I think the agenda of systems integration, which introduces value but often complexity, is opportunities for competitive differentiation and opportunities for differentiated technology. It's a roadmap dialogue more than a market transition dialogue at the customer interface.
CJ?
Thank you. C.J. Muse, Evercore. Wanted to, I guess, follow up on Roman's question. 30% increase in WFE is pretty damn impressive in terms of your target model. Curious, how much of that is incremental spend on new areas, versus pure share gains on existing areas? Would love to hear greater thoughts there. Then as a follow-up to Doug, can you comment, as part of the buyback, are you contemplating taking out one or both of those converts? And on the convert coming to you in May, do you have the option there to settle in cash? Thanks.
You want to deal with that piece first?
Maybe do that one first. C.J., the way those converts work is they'll settle when they mature. They net share settle, so shares get issued. It's a net share settlement. We've got one coming in May of this year, and then we've got that long dated 2041. One of the things, though, to understand is as those converts are so deeply in the money, we've begun to get, and this surprised me a little bit about a year ago, requests for redemption. It's not at our option, taking those out. Actually, the bondholders, and I can talk with you about why they're doing this, I think.
They're asking, "Please redeem me." I think the phenomenon there, the economics of it is the share price gets higher and higher, the incentive for them to continue to hold it because they're bond investors, they have to hedge the equity component. They actually don't like that, right? They're beginning to redeem. That's that piece.
I think actually the answer, C.J., to your question about where does the growth come from, actually was nicely presented by the slide that Tim characterized that show these kind of three individually $1 billion building blocks. Right? We were referencing what you know today as an annualized run rate of the company. If you look at the guidance for the March quarter of this year and just multiply it by four, which is not a forecast, but it's just plain simple kind of math. If you take that $12 billion reference, the $15 billion that we characterize comes from each of those $1 billion building blocks. There's a SAM expansion conversation, which is all about the positioning of differentiated products and technologies to support roadmap of customer. There's a market share conversation, and there's an install base conversation.
They're all key elements to the vision that we've characterized this morning.
Farhan?
A couple of questions. One, on the free cash flow, you committed 50% of the free cash flow back, returning over the next five years. What are you thinking about the cash on the balance sheet and the rest of the cash that you will generate? Are you thinking about some M&A, and is that in your plans also? Secondly, on the WFE, you laid out growth drivers, which basically had a lot more wafer starts added over the next four or five years than we have seen historically over the last four or five years. Opportunity on per wafer is also growing a lot more.
The question I have is shouldn't the WFE growth over the next four or five years, just because the wafer starts that are being added are much more, could that potentially be similar to what we have seen over the last four or five years or even higher?
I want you to answer that piece. I'll take the first piece. I want to be clear that we're capturing this message around capital allocation. We said we plan to return at least 50% of our free cash flow to shareholders. We didn't say 50%. We said at least 50%. What I said in my prepared comments is that the investments thought process for the company, whether it's an organic and inorganic investments, whether it's buying our own stock, it's a long-term accretion test, right? It's pretty straightforward. I think there's plenty of data now, unfortunately, painfully acquired, that big M&A in our sector is probably not going to happen. Our focus in terms of acquisitive non-organic opportunity is around disruptive technologies that's relevant to the outlook that we've presented. There are new challenges, and new technologies relevant to solving those challenges. We're interested there.
That is a perpetual focus in the company. If we see opportunities to harness the strengths of the company, whether they're technology strengths of the company, whether they're operational strengths of the company, whether they're channel strengths of the company or all of the above, in what we call adjacent markets, Adjacency is defined around what we are. Right? I think it's really important to recognize what we are. To say we're an equipment company dramatically misses the point. We're a results on the wafer company, Actually what is really important when you think about adjacency is recognizing that we're this incredibly well-qualified innovator of extremely complex systems engineering. We're a hardware, software, process control, and materials integrator. We do all of that in a nanoscale applications environment.
A nice example of adjacency, or at least, a foray into adjacency, was the joint development project we announced with NanoString last year in the space of molecular profiling. The technical attributes and the market attributes look very, very, very similar. How do we make those decisions? We make them in the way I just described in my presentation. Either those investments are more accretive in the long term or they're not. We're trying to architect with a vision for 10, 15, 20 years, sustainable, profitable growth and return for our shareholders. We feel not only do we have opportunity, we have responsibility to harness every strength and asset of our company.
To your question on WFE, Farhan, if you listen to what Martin said, that the vision for a data economy, the growing strategic relevance of semis, and the things that he talked about on the supply side, that's what translates to those opportunities for the more than a million wafer starts of capacity across DRAM, NAND, and logic foundry. Certainly would translate to not just a level of WFE that's consistent with today's levels, but growing over time.
Who are you picking?
Oh, Harlan. We'll get to you next time. Peter?
Great. Thank you. Great presentations. Martin, my first question is a supply and demand question for the memory space. If you look at your cost per bit and your bit supply growth for NAND, it seems to predict equilibrium at about a 50% growth rate on each, while DRAM, you suggested a delta, where 20% cost per bit versus the 30% bit growth. What do you think the implications for wafer fab equipment are in light of that? Two quick follow-ups. Tim, one question that would be helpful to think about is maybe if you sell $1 of equipment today, what is the either sales or free cash flow, however you think about it, that you would expect to get through the life of that product today? What was it historically, and where is it going?
My last question, Doug, you increased the buyback from two to four. Is there any timeframe on that that you're willing to share? Thank you.
Peter, you asked so many questions. Like, what the hell was your first question? I think I remember it. The investments in any part of the marketplace, the investments in memory and in DRAM and NAND that you just referenced, is all a byproduct of an ROI for the customer. Yes, it's fair to say that the DRAM market and the NAND flash market have actually quite different kind of roadmaps, and in terms of the series of steps that are in front of us in terms of technology transition and the economics of each step. Right? I think common knowledge, the DRAM segment of the industry is the most challenged in terms of economic scaling. The consequence of all of that ultimately shows up in this investment balance with value.
If there's substance to what I articulated around the world transitioning into value propositions that justify different levels of investment, the sustainability dialogue is going to play out. Maybe you don't believe the data kind of strategy dialogue, you don't believe the prominence of AI. If you don't, you shouldn't believe most of what we've talked about. If you do believe in the prominence of data and the role of AI and the prominence of silicon, you're going to endorse the fundamental market attributes and the role for Lam as a primary supplier of etch and deposition equipment. Second piece.
On this the installed base piece, I'll let Doug jump in if he wants to add something on free cash flow. In terms of the opportunity created, maybe just think back to the two slides I showed. Historically, I guess if we look at a historic perspective, today we are generating more revenue and from a piece of installed base equipment than we have historically. That was referenced by the acceleration of installed base revenue versus installed base growth in the last four years. That's been from a deliberate strategy on our part to engage with customers to find ways to create more value. I talked about a couple of them, the chamber matching, using some of the artificial intelligence and big data that we get off of our tools. A customer today, one thing that's changed is the fabs are getting very large.
Obviously, there are pressures to get those fabs up and achieve time to yield, time to money more quickly. One of the ways we can help is through the rapid startup and matching of large fleets of Lam tools. We get paid for that service. I think that we will continue to innovate, find new ways to help customers reduce their costs. I think going forward, you can expect it to increase. How much? Probably not by that 50% faster than the installed base. What we've said is we have a long-term objective to continuously innovate to grow the installed base revenue faster than installed base growth. That's probably the best way to think about it is it will be faster, but not quite to the same degree.
Over the next four or five years, that $1 billion of incremental revenue is probably the best way to think about impact to the company.
Free cash flow. This part of the business generates just awesome free cash flow. It doesn't require a whole lot in terms of investment. It doesn't require a whole lot in terms of R&D. I'm not going to quantify it for you, but it's a very nice generator free cash flow. Yeah, thanks for that question. I actually should have communicated the timeframe, 12 to 18 months for the buyback.
We're going to try and speed this up, so if you can kind of keep a hand up.
Thanks for hosting the event. Harlan Sur, JPMorgan. Question for Tim. As a part of the growth in multi-patterning, you highlighted the potential for advanced process control capabilities. Could you just help us understand what are some of those capabilities, and is this a future functionality that customers expect, or is this a feature and a functionality that customers will actually pay you better economics for?
Okay. I don't actually recall mentioning advanced process control specifically for multi-patterning. There was one thing that you might have heard in the video from Vahid about the importance of certain new capabilities, like our Hydra Advanced Patterning chuck. It is advanced process control that allows us to feed back the results and produce better CD control. It's integrated into our etch tools and has become a critical enabler for our customers to achieve better variability control through multi-patterning. That's probably the best example. We continue to look for opportunities to use data off of our systems to improve results. One trend I think you'll see across all equipment is increased sensoring of those pieces of equipment. Wherever we can get useful data off and use it, we are.
Just to continue that thought, Harlan. For those of you that are familiar with ASML and the pursuit of a so-called holistic litho kind of roadmap, this is exactly the same conversation. It is identical.
It exists with two objectives in mind. One of them is enhancing the competitiveness and the applicability of underlying process capability to the needs of the industry, and the second is to create new opportunities for new product and service revenues. A nice example for us is one of our acquisitions. We acquired a company in this space four years ago, and we have reported a tenfold increase in revenues in that four-year period. The word process control we articulated was fundamental to the vision of the company when we were standing on this stage talking about KLA-Tencor. The strategy's changed, the objectives have not.
Maybe just to add one more thing, because it's an important tie-in with the installed base. I mentioned the fact that after a few years, we generally get the opportunity to sell an upgrade to a customer. I think the Hydra Patterning chuck is one of those great examples, because as the customer evolves and now needs tighter CD control at the next node, that's the type of upgrade that can then be purchased by a customer, offered by Lam, and extends the life of the tool.
Atif.
Thank you. I have two questions. First, Tim, you showed a slide where your SAM at 7 nanometer logic and foundry is growing 1.5 times per wafer over 10 nanometer. Investors are just trying to reconcile what some of your customers, like TSMC, are talking about a flat CapEx, and within that, capital intensity is coming down, and obviously EUV opportunity is growing. Can you just reconcile for us how can we have a flat CapEx when your opportunity is growing? Then the second question is on the NAND market, and Martin, thank you for breaking out the NAND and DRAM WFE for this year. If my math is correct, NAND is down like 25% year-over-year in WFE. Can you just talk about where we are in terms of 2D to 3D install base upgrade cycle?
You can get that.
Both, Martin?
Well, you want to do both? Yeah.
No, no. Do you want me to do both or just the second part?
Second part. First part.
First part.
I'll do the second one.
I'm going to let him help with the first part as well. I think that relative to this comment about seven nanometer versus 10 nanometer, our SAM is clearly a statement of etch and dep intensity, more so than overall CapEx required for the node. I will let Satya, who's into the fine details of that modeling, speak. You can just think about it in terms of our look at the new materials, the new films. I didn't talk about everything that was on that technology inflection, but you saw things on there like transitions in the cobalt space for interconnect. These create new opportunities for Lam as we move from 10 to seven to five to three. That's just an example, patterning as well. Just think about it as a statement of dep and etch intensity in our view.
I don't know, Satya, if you
No, I don't have anything more specifically to add to that. Just think about the things we talked about in the video. There is a significant increase in that etch dep intensity as you go down from seven to five nanometer. We reflected some of that. Relative to your other question, NAND CapEx, I think your numbers are a little different than what we have. NAND CapEx, as we indicated in our last earnings, is up modestly this year. The data that we shared with you is consistent with that. Remember, we're talking WFE. I think we have Edmund.
Thanks. Thanks. Taking my questions. Two questions. First-
Mic's not working.
Hello?
Yeah, there you go.
Thanks. Taking my question. First question is, if I look at your growth, you say $3 billion, $1 billion on install base, and $1 billion on SAM, $1 billion on new product, right? If I think about that, the $2 billion more in product, $1 billion in install base, does that mean that your mix is shifting? Got a specific question for you, too. Does that affect your margin positive or negative? Is that a way you can think about that if I'm correct about the mix shifting? Then just quickly on the new product, how much of that is driven by just this atomic layer, like ALD/ALE versus just extension of existing product? Is there a way to kind of separate those two to kind of think about how much is at the atomic layer, the ALD/ALE versus extension?
There is a way to separate it for us, but not for you. Every significant aspect of the grand challenges I just described are relevant to the roadmap of the company and the opportunities for differentiation. We're never going to be able to accommodate the specificity of that disclosure, but you're right in isolating atomic-level process capability and control and repeatability as fundamental elements of differentiation for the company. On margin, it's materially irrelevant as a commentary on mix in the company. The profile of investment is quite different, but at a fundamental OpEx level, the businesses are similar.
In the model I just gave you, it's all in. We've comprehended everything we think is going to happen. Just go to the model and use that.
We'll take two more questions.
Yeah. In the back, Mehdi?
Yes. It's Mehdi Hosseini from SIG. Two follow-ups. It seems like the install base offers you tremendous opportunity. You talked about more than $1 billion of incremental revenue. What would it take for the company to break out the revenue mix by system and services so that we could better track that and better model that? You could also argue that that could help us to decouple companies' opportunities from WFE. Most of the questions are WFE related. If you enable us to better model this, you can essentially decouple that. Number two question has to do with the 4%-8% share gain. Everyone in the semi cap space talks about a share gain. Isn't that more a reflection of your customer mix, especially some of the customers that have announced new fab projects? Isn't that the primary driver behind the confidence?
It's not so much of beating out the competition, it's more a reflection of what your customers have already announced publicly, longer-term projects. Could that be the way to think about this 4%-8% share gain?
It could be, we'd argue that that misses the point. Clearly, customer mix for every participant in the marketplace is going to be sometimes a little favorable and sometimes a little unfavorable. That's part of the story. There are fundamental investments in differentiated products and services that are part of the performance we've reported and part of the performance we've just outlined prospectively. I think what's really important to recognize is that not all market share is equal. When people talk about market share and they don't talk about profitability, it begs the question what profitability is associated with the market share that you're reporting, which is why we're choosing to focus on competing for and winning a proportion of WFE, and we're choosing to focus in our disclosure on operating income %. We're not going to shy away from disclosure around market share as well.
We'll tell you, it tells only a small part of the story. The part of the story that's really important is the proportion of WFE that you are competing for and winning. It's the proportion of fundamental challenges for the customer and strategic relevance that you're addressing. We think some part of that is being in the right place at the right time, to your point. Etch and deposition today is central like it has never been to the technology roadmaps of the customer, but it still takes execution, and there's a reason why I have as much gray hair as I do. Who's next?
Oh, did you want to do it?
I think there was a question on CSBG as well.
Doug, do you want to?
CSBG segmentation.
Oh, maybe.
Actually, I'll answer that. We're not going to.
Yeah.
We're not.
Thank you. That was more direct.
We're not going to do that because, and there's a business reason. When I think about disclosure, I think about you, but I also think about running the company. Our job is to harness the totality of value from our company, and there is a razor and razor blade model, and to have disclosure would restrict the flexibility we have to do the right thing for our customers and the right thing for our shareholders. That's why we are what we are in terms of disclosure. You get to pick the last person, Satya.
There's no more questions.
Make it a good choice
it looks like. There's one in the back. Joe Moore.
We'll do two. We'll finish with Robert because he put his hand up three times.
I was wondering if you could make us feel better about the $13 billion DRAM WFE. Do you think there's a period of time where that means significantly higher supply in the future? I guess similarly, when I looked at your DRAM etch and dep intensity, it's actually going up less than in the other areas. Just how do we feel about that higher number of DRAM at WFE?
That's not a surprise, I hope. The kind of roadmap in DRAM has been talked about for some time. It's getting more challenging to deliver the economics. There are three to four known generations today in the roadmap that will survive the next probably six or seven years of technology migration. I get back to first principles again. Either there's value or there's not in capacity addition. Our customers will do the right thing. They don't spend money when they don't need to. When there's legitimate demand for chips out of DRAM fab, whether it's conversion or addition, our customers will find the optimal way to go do it. That's what they've done. I think that's what they continue to do.
Just want to add a quick thing. Make sure you think about not dollars of CapEx. That's the reason why we share the economics of the capacity additions. You have to think about the cost that it takes to add supply growth that's very different today than before.
Okay. Yeah. Huge responsibility, Robert. It's the last question. It needs to be a really good one. A positive one.
No problem. A question for Tim and a follow-up to some prior questions. First question for Tim is, back in the bad old days, the excuse for outsourcing was because of the cyclicality of the industry. If we're true believers now that the industry is more sustainable going forward, shouldn't we be insourcing because we're leaving some amount of gross margin on the table? That's the first question. The follow-up question is on the inorganic growth side. Clearly, there's been a lot of water under the bridge. We've had a change in administration. There's a $100 billion semiconductor deal out there. It would seem that while large deals may be off the table, there is an opportunity for roll-up of small and medium-sized companies, not just technologically inflection point companies.
While maybe not able to move the needle as much as large transaction, you could add up and certainly be additive to the company. I'm curious as to your view on that.
Yes, I'll deal with that first, and then Tim, you can close it. Certainly, I think posing the hypothesis that there are opportunities for some level of consolidation still in the equipment industry is valid. I'd put it in the possible camp, not necessarily the probable camp. I'd say that you should expect us to be part of the system evaluating opportunities just as much as anybody else. Again, I go back to what I said earlier. We have an orientation not just to the size of the company, but we have an orientation to the quality of the earnings in the company, which is about strategic relevance, which is why we lean into the more complex technical challenges, and it's why we message the importance of disruptive technology.
We're running a business, and we have as responsibility, as much as anybody else, to evaluate all of the opportunities to architect good deals with good value creation. I think we've demonstrated operational capabilities, not just in an organic context, but an acquisitive one as well that makes it at least not irrational. We'll see where the future takes us. Tim?
Okay. I'll try and answer this as positively as possible to end this, but it's a great question relative to stability of spending the growth in our business, outsourcing versus insourcing. We continuously have that debate. I think that what we have to keep in mind is that we are continuously looking at how outsourcing or insourcing affects our ability not only to respond in a down cycle, which is kind of traditionally how we've thought of using it, but I will tell you, through this recent period of tremendous growth, having access to qualified outsource partners has helped us meet a lot of the demand from customers on very short notice. I think that we want to maintain a very flexible manufacturing and supply chain organization. We are having that debate as to next four or five years, 10 years looks more stable.
Should we change the model from what it was 10, 15 years ago?
In closing, I'd like to end where I started, which is to say a big thank you. We really appreciate the fact that you take your time, you show up to engage with us, and in some respects, to guide us. We are here, frankly, for you, our owners and our investors, and our mission and vision is pretty simple, to deliver enabling solutions to our customers and capture a fair share of value ourselves. That is our guiding principle. That will always be true. Thank you for your time. We appreciate your kindness, and we have a number of meetings scheduled, probably with some of you in New York and in Boston in the next few days. We appreciate very much your time. Thank you.