Good afternoon, welcome to Entegris' 2020 Virtual Investor and Analyst Day. I hope everyone is staying healthy and safe. Just a couple of housekeeping items for the webcast. For the best performance, as I explained in my email, if it's possible, if you could disable VPN and any other programs as much as you can. I do have a phone number backup if something happens to your connection and need to access by phone, just send me an email. There's a PDF slide handout of all the slides on the IR site and also on this webcast tool as well. You see three boxes in front of you. All three of those you can adjust the size. Then if you need to reset, you can hit that bottom circle on the bottom of your screen, bottom left, it's an arrow.
When you hit that, it will reset to the default. We will have a Q&A session at the end of the presentation. Please enter any questions you have. You can do it throughout all the presentations or during the Q&A. You can enter them into the box on the left, bottom left. Make sure to put your name and company name in there just in case. You can also send me any questions you want over email as well. We expect the program today to be approximately two hours. That would include Q&A. To be clear, today we'll be discussing strategy and long-term growth potential of the company. We'll not be addressing the current quarter. The speakers that are with me today, Bertrand Loy, President and CEO, Todd Edlund, our COO, Jim O'Neill, our CTO, and Greg Graves, our CFO.
Before we begin, I would like to remind listeners that our comments today will include some forward-looking statements. These statements involve a number of risks and uncertainties, actual results may differ materially from those projected in the forward-looking statements. Additional information regarding these risks and uncertainties are contained in our most recent annual report and subsequent quarterly reports that we have filed with the SEC. Please refer to the information on the disclaimer slide in the presentation. On this call, we will also refer to non-GAAP financial measures as defined by the SEC and Regulation G. You can find reconciliation tables in the presentation, which is posted on the IR site. With that, I'll hand it over to Bertrand. Bertrand?
Thank you, Bill. Good morning, everyone. It's been a while since our last Analyst Day, and since then, the company has continued to perform very well, meeting and in many cases exceeding the various objectives we had set out for ourselves nearly three years ago. Clearly, a comprehensive update was in order, and this is exactly what we want to accomplish with you today. I would start with a few high-level slides to orient you, number one, and then also to highlight some of the attributes making Entegris such a unique investment option. Next slide, please. Since its creation nearly 50 years ago, Entegris has changed a lot. With our revenue expected to reach $1.8 billion this year, we are the largest independently traded specialty chemical company serving the electronics industry.
Yet something that has remained constant for all these years is our mission statement. As you can see, at Entegris, we exist to help our customers improve their productivity and the performance and technology. We do this by developing new materials and very unique process solutions. What I want to say here is that the implications of such a mission statement are profound, and they tell you a lot about not only our business model, but also about our mindset as an organization. At Entegris, we want to be viewed as an extension of our customers' engineering teams. Our innovation is the result of very close partnerships with our customers. As we collaborate with them, we gain precious insight into their technology roadmap, and we can then identify areas where we can uniquely add value.
This insight informs the management decisions and guides both our organic development efforts as well as our acquisition strategy. The result of all of this is the platform we have created with very unique capabilities and expertise in a few key areas, specifically in materials, microcontamination control, and materials handling. Todd Edlund, our Chief Operating Officer, will describe more fully the various technology platforms comprising our three divisions, and he will discuss our expectations for these three businesses. The last point I want to make on this slide is that by the very nature of our solution set, we sell across the entire electronics ecosystem. 90% of our revenue is tied to the semiconductor industry, and as you can see, 50% of our products are used by fab end users, while the rest is sold to equipment makers and chemical suppliers. Next slide, please.
Another unique attribute of Entegris is that we have a very resilient platform. There are a few different ways you can think about this, as you can see on this slide. First, most of what we do are consumable products, and they can be materials, chemistries, filters that are used in our customers' daily manufacturing processes. Our solutions are sticky, and the reasons for that is our solutions are increasingly tailored to unique customer process conditions, and therefore, the switching costs are high. That means that once our solutions are designed into the process recipe of our customers, we can enjoy a steady stream of recurring revenue for as long as these fabs remain in operations. I will provide a little bit more context about the useful life of semiconductor fabs later in my presentation.
Another way to think about resilience and stability is that our offering is very broad. We sell approximately 15,000 products, and we only have a few platforms reaching $50 million or more in annual revenue. This means that we are not overly dependent on the success of one or two products. In fact, our opportunity pipeline is very broad and very well-balanced, as you will see later today. Finally, as you remember from my previous slide, we have a very diverse customer base, more so than any traditional supplier to the semiconductor industry. While we have very strong relations with every industry leaders globally, we only have one customer representing more than 10% of our revenue.
It is often the case that the resilience of a business platform comes at the expense of top-line growth, and I would tell you that it's really not the case with Entegris. We've been able to deliver on both fronts. Let me turn to the next slide to review the recent performance of our top line. As you can see here clearly, our top-line growth has been very strong in the past five years. We experienced a renaissance in our core semiconductor market, which grew well in excess of GDP. You can see that in the dark and light gray bars, respectively. More importantly, we were able to outpace the industry organically, very meaningfully, and that's in the light blue bar on the graph. We added another couple of points of growth from acquisitions we made along the way.
All this translating into an average top-line growth of about 11% and superior shareholder return over the period. Now, of course, none of this matters. It's all in the rear view mirror, and the question in everybody's mind on the call today is, can we do it again? The short answer to that question is yes. Over the next few presentations, we will detail for you how we intend to do this all over again. In summary, if you go to the next page, please, the conviction you will hear from our management team is based on the following factors. The semiconductor industry is poised for strong secular growth. The industry technology roadmap is evolving very favorably for Entegris. You will learn today that our value proposition centered around materials and materials purity is increasingly important to many new technology applications.
Our growth aspirations are ambitious, they are credible. They are backed by very concrete plans, a very capable team, an increased and intense focus on the customer, and substantial investments to compete effectively in the marketplace. Finally, we expect to generate attractive cash flow levels, which will give us additional optionality to create incremental shareholder value. I will cover all of these elements in more detail, but obviously all of this starts with growing the top line. The next slide illustrates how we are thinking about this. Let me start with the headline. Over the next three years, as you can see, we expect to grow organically 9%-10%. To put this in perspective, this is a growth rate very similar to what we have accomplished and delivered in the previous five years. There are two high-level ways to unpack this growth rate.
The first, and probably the simplest way to think about it, is the industry growing at twice the rate of GDP, and Entegris outpacing the industry by 300-400 basis points. The alternative way to think about it is as a combination of SAM expansion and share gain. I am sure that some of you are already trying to calculate our SAM growth, I will make it easy for you, and I will give it away. You can see that over the planning period, we expect our SAM will grow at an average of 8% per year. A little bit of share gain over the planning period, but mostly it is about SAM expansion, driven by three main forces. First one is an acceleration of chip demand. The second one is the growing Entegris content per wafer.
The last one is more wafers being produced at the leading edge. Let me dive a little bit more into each of these three components. If we go to the next section, we'll start with the industry lift. I don't think I need to spend too much time on this. If we go on the next slide, I think you've seen many versions of similar slides. You remember the various phases of expansion and stagnation of the semiconductor industry and how wafer start growth has been comparing against GDP growth over the past decades. What is exciting today is that the industry is entering a new phase of hyper-growth. The world is increasingly interconnected. Massive amounts of data are being generated at an exponential rate every year. Clearly, the digitalization of our lives is happening, actually very rapidly, as a matter of fact.
We are witnessing this first acceleration this year out of necessity during the current pandemic, with the emergence of new norms around work and learn from home. This is just the beginning, and things will accelerate even further once the 5G infrastructure is in place and once new applications take full advantage of the new network speed to transform our cities, our healthcare and transportation systems, and so much more. All of these upcoming trends will drive demand for semiconductors to new highs. In that context, we believe that it is reasonable to expect that the semiconductor industry could outpace GDP by a factor of two for the years to come. Let's go to the next section, please. As I said, the primary industry we serve, the semiconductor industry, will enjoy strong secular growth, and this is great.
This is a very solid foundation upon which we can build. Of course, you want, and we want to continue to grow faster than the industry. We will do this by identifying areas in our served markets that are expected to grow faster than the rest of the industry, and by gaining exposure to those emerging growth areas. In my next three slides, I want to call out a few of these exciting growth areas for us. I will try to quantify how our served markets per wafer is expected to evolve. Remember that the numbers you will see are directional and that our customers are constantly looking at ways to improve their yields and consumption rates. With that disclaimer, let's go to the slide that you already broadcasted, Anna.
The first slide illustrates why the next years will be rich in opportunities for our SCEM division as a result of the compounding challenges of miniaturization, high aspect ratio device architectures, and the migration to nanowires. All of this combined with a big push for better interconnect materials. Let me tell you, I have been in this industry for 20 years, and to be honest, I do not recall a period of such activity, such intensity in new materials development in a very long time, in fact, if ever. That observation is really what leads me to claim that material science will be increasingly driving the industry roadmap. As such, the opportunity, our SAM per wafer could be very significant across the various industry segments, and in particular in 3D NAND, as you can see here on these charts.
Jim O'Neill, our Chief Technology Officer, will discuss in detail how our solution set enables these tall structures in so many ways. At a high level, let me just say that there are two major areas of opportunity for us. The first one is in deposition materials, and it will be an opportunity for us to leverage the capabilities that we acquired with ATMI in 2014, capabilities that we strengthened with the acquisition of DSC and MPD last year. The second area of opportunity for us will be in selective etch. As an aside, you should know that the critical additives in these etching chemistries are developed and manufactured at DSC, one of the two companies we acquired last year. Clearly, all of these acquisitions have been very relevant, but also very timely for us.
The takeaway of this slide is that our SCEM division is uniquely enabling higher bit density, faster compute, and better energy efficiency. Next slide, please. Another very important theme behind our SAM expansion is the need for greater purity. Our customers are aiming at introducing smaller feature sizes and taller architectures, which is a very daunting combination. To accomplish this, they need greater retention across a wider array of atomic-level contaminants. Tight contamination control is essential for our customers for at least two reasons. The first one is to reach acceptable yields by removing killer defects, and the second reason is really to improve the long-term reliability of the chips by removing latent defect. Jim will talk a little bit more about that.
As a result, the opportunity per wafer for our liquid filters in the fab environment continues to grow nicely, both in logic and in advanced memory. This is a tailwind that you're familiar with, a tailwind that we've been enjoying for some time now. Based on our active collaboration with the industry leaders, we expect similar opportunities to develop for the years to come as purity requirements continue to become more stringent, more difficult to achieve, and as we continue to advance our technology. On this next slide, this is really, in fact, the flip side of the same coin. This is the second dimension to our purity theme, and the opportunities that it creates for us.
Essentially, what I'm trying to say on this slide is that if our semiconductor fab customers want to achieve ever greater levels of purity without compromising their throughput, they need to subject their chemical suppliers to much more stringent purity requirements for incoming chemicals delivered to their fabs. The reason this opportunity upstream in the ecosystem is accelerating for us is that a growing number of chemistries are required to achieve greater and greater purity levels. This need for cleaner and more exacting standards up and down the supply lines are opening many new opportunities, not just for our liquid filters, but also for our high purity packaging solutions.
Increasingly, I want you to think about Entegris as playing an advisory role with the chemical suppliers on how to best enable and control the purity, the stability, and the safety of electronics materials across the ecosystem, from bulk manufacturing all the way to the point of delivery to the fab. With volumes of chemicals used on the rise, as you can see on the upper left of this slide, this will be another very powerful tailwind for both our Microcontamination and AMH divisions. Moving to the next slide. This is really the final driver that I want to call out, in terms of SAM expansion. Let me be clear, it's a driver entirely outside of our control, yet it's very real and it's very powerful.
What you can see on the next slide, the two graphs, showing really the steady push forward, the steady transition of more and more wafer production to the leading edge. You can also think about these two graphs, by the way, as an illustration of the rate of obsolescence of process technology in both logic and memory. As you know, and as you can see here, clearly, we have longer life cycles for logic fabs, and we have shorter life cycles for memory technology nodes. A couple of takeaways for you here. As you can see, by 2022, approximately 25% of the wafer produced in logic fabs will be at 20 nm technology or below. In that same time frame, we expect almost all, 90% of the 3D NAND chips to have 96 layers or more.
This is, of course, very positive for us since, as you recall from my previous slides, we expect greater Entegris content per wafer at these advanced nodes. Now it's actually probably a good slide and a good time for me to make a few additional disclaimers. As I said earlier, number one, we do not control the timing or the magnitude of these technology node transitions, and yet they have significant impact on our performance. Point in case, in year 2020, when many customers, both in memory and logic, transitioned successfully and therefore rapidly to new nodes, you can expect us to handsomely beat the industry growth. Conversely, in years with fewer node transitions or more challenging node transitions, you should expect our growth rate to be closer to the industry growth rate. Let's go to the next page, please.
If I wanted to summarize my previous four slides, I would say that the semiconductor industry roadmap is becoming increasingly challenging and that these challenges translate into many new, exciting opportunities for Entegris. At the same time, our customers have higher expectations of their suppliers in general, and of Entegris in particular. You should know that as a management team, we recognize this, and we are very focused on strengthening our differentiation and owning the various facets of our value proposition. Let me touch a little bit more on that on the next slide. You can recognize on the left the flywheel, which is really spelling out how we think about customer centricity at Entegris. If you have been following Entegris for long enough, you have seen different versions of that flywheel. You may even be tired of it.
Well, let me just say that expect me to continue to use it because I genuinely believe that our constant and disciplined focus on this flywheel has been fundamental to our success. It highlights the critical steps we want our customers to experience when they engage with us. As a team, as a management team, we are constantly focused and constantly thinking about what we need to do to improve the customer experience at every crank of the flywheel. We are constantly reflecting on engagements we just completed and projecting forward on the new opportunities we are chasing with the objective to improve at every crank. As I like to say, it's a lot more than a model. It's really a mindset at Entegris, which ultimately leads to a very strong organizational alignment centered on a shared purpose.
On the right side of the page, you can see the three pillars of our value proposition: technology, operational excellence, and the required infrastructure to effectively support our customer-driven innovation model globally. As I said earlier, we are constantly assessing with the input, not only from our global teams, but also our customers, the capabilities that we need to add or strengthen in order to further differentiate from the competition. Todd Edlund, our Chief Operating Officer, will discuss these three pillars in more detail in a moment. Let's go to the next section. Greg Graves, our Chief Financial Officer, will review our capital allocation framework more comprehensively, but I wanted to touch briefly on our M&A strategy, which will remain an important focus for us in the years to come.
On the next slide, when it comes to acquisitions, our beliefs, areas of focus, and guiding principles will essentially remain the same. First, given the nature of our business model, we believe there is a strong correlation between creating differentiated customer value and creating long-term shareholder value. In other words, expect us to focus on M&A activities, and on platforms that can enhance our capabilities, enhance our value proposition, and ultimately enhance our brand. The second point I want to make on this slide is that we will focus on technologies and applications and markets that we believe can help sustain our top-line growth, out performance. Expect us to remain very focused on advanced materials and separation technologies primarily, and to a lesser extent, on materials handling technologies.
Finally, expect us to remain disciplined in our execution from the moment we screen, assess, and value potential targets to the care and speed that we put into the integration work, in order to unlock their full potential. On the next slide, we can probably skip ahead two slides, just introducing our financial objectives. By 2023, we expect our top line will reach approximately $2.4 billion organically. We expect to increase our ROIC to 20% and to deliver a non-GAAP EPS between $3.75 and $4. I want you to note that this EPS range includes both the organic EPS as well as the incremental earnings we expect to generate from our capital allocation decisions. Greg will review all of these numbers and the underlying assumptions in a lot of detail. I probably will not go a lot beyond that in my section.
Let's move to the next slide. In the next section, we wanted really to use our time together today to formally introduce a very exciting initiative, which is really our Corporate Social Responsibility framework. On the next slide, you can see that it's a pretty comprehensive program, and it's really a statement that both the management team and the Board of Entegris wanted to make. We recognize the impact and the responsibility we have to our people, to the environment, and to the communities where we operate. As an organization, we want that impact to be as positive as possible. The framework we are presenting today is a more formal commitment to many initiatives that historically have had a less structured focus within the company. We have chosen to build our framework around four pillars: innovation, safety, personal development and inclusion, and sustainability.
We chose these four pillars very carefully. We wanted them to closely connect not only to our value system, but also to the value proposition of Entegris and our business strategy. Safety is a good example of this intent. Safety is obviously a key focus in our workplace and on the Entegris shop floor, but the safety of our customers is also top of mind in our innovation process when we develop new product platforms for semiconductor or even for adjacent markets. A good illustration of this mindset is the modest, yet critical role Entegris is playing in the fight against COVID-19. A few years ago, we launched our Aramus high-purity bags for biologics. The properties of the Aramus bags are unique in the marketplace. Our bags are gamma sterilizable. They can also withstand cryogenic temperatures, which means they will not break in cold temperatures.
This means that our bags could play a very critical role in the effective distribution of the vaccines. By reducing bag rupture during transportation, we would allow more people to have access to a vaccine faster and more efficiently than using more traditional technologies. I do not have the time to cover each pillar in great detail, unfortunately. Please visit our website, learn more about our CSR program, get acquainted with our 2030 goals. For our goals, I just want to mention that we chose a 10-year horizon simply because we wanted our goals to be bold, and we needed to give our teams enough runway to break the mold and reinvent ourselves. I just want to say that in many cases, we expect we will be able to reach these milestones earlier than 2030.
For instance, the goal that our Board of Directors selected for themselves is a good reflection of that. We expect to achieve 50% diversity in the boardroom well before 2030, obviously. This will be a very exciting journey for all of us, a journey during which we will demonstrate that it is absolutely possible to do the right thing and to create lasting economic value. Stay tuned as we will, of course, update you along the way. In closing, on the next slide, I would like to summarize what I hope will be your takeaways from our 2020 Analyst Day. First, we participate in an exciting industry that will enjoy very strong secular growth. We will not spend a lot of time on this topic because our assumption is that you probably would not be on this call if you did not share that view.
The second big takeaway is that our exposure to emerging technology inflections is accelerating, which in turn translates into a very rich and diverse opportunity pipeline. Jim O'Neill will shed a lot more light on this. The third point is that we compete very effectively. As a management team, we are laser-focused on raising the bar constantly. Our quest for excellence is a journey, not a destination. Todd will discuss some of our areas of focus very shortly. Fourth, our business model is more resilient than most industry participants. It is, as I mentioned earlier, a function of the inherent stickiness of our consumable business model, as well as the quality of our execution. Point number five, our recurring revenue model is also the reason for our strong cash flow performance.
Lastly, I would want to say that we have earned the reputation of being a smart allocator of capital and a very effective acquirer. So in other words, Entegris is a value compounder, and this is something that Greg Graves, our CFO, will unpack for you at the end of our presentations today. So this concludes my section, and I will now turn the virtual stage to Todd Edlund, our Chief Operating Officer. Todd?
Great. Thank you, Bertrand. Happy to be here today to talk about this company that we've built over the last 50+ years and the great opportunities that it has not only demonstrated we can deliver, but we expect to be able to deliver for years to come. Over these decades, we've really built Entegris to be a very effective, trusted partner to the semiconductor industry, one of the most demanding industries there is. We've learned a lot about how to do that, and as Bertrand mentioned, we've already started to look for ways to do that in other places. Semi remains our focus. What we've really done is focus on semi and how we can evolve to serve the most important needs of the industry, as Bertrand described, and Jim O'Neill will also describe.
Really, we look for how can we find those places where we can bring a lot of value that will outgrow the general pace of the industry. It's materials. Materials performance is critical to the advancement of the industry. It's the safe and effective delivery of those materials, which is a key focus area of Entegris. How do you get that material safely to the fab? How do they use it safely in the fab, and how do they get the most efficiency out of the use of that material? Then very uniquely, Entegris also brings the ability to work on defect reduction through contamination control, contamination prevention, damage prevention, and deliver yield. Ultimately, that's what our customers are after, is yield. Move to the next slide. Bertrand showed you this. This is how we live.
It's really about these three pillars we've learned, and continue to believe, give us competitive advantage and are the focus of our investment to be that trusted partner to the semiconductor industry. I'll describe each of these in the slides coming up. We can go to the next slide. Starting with Technology. This is the number one important thing that we have to do to win, have the best solution and have it first. That's what our key focus is on, especially for the leading edge. That starts with making sure we have the applications expertise to be in the room when customers are describing their needs, when they're developing solutions. They can see us as a helpful partner to have there to actually help them get to solutions faster. Applications knowledge is one of the most important things that Entegris possesses.
You have to back it up with significant R&D investment. We typically spend between 7% and 8% of our revenue on R&D. We expect that to grow, as we'll describe later, during the plan period because it is the most important thing we have to continue this outperformance that we've been enjoying as these new nodes are launched. We have to also do that close to our customers. We have tech centers, technology centers in our most important markets around the world, that can provide places where our customers actually come and work with us to develop solutions to their needs. That's one of the most valuable things we have. The right people, the right capabilities close to the customers.
At the end of the day, they have to be able to count on us to take that new product and ramp it very quickly as they roll out a node. There's two examples shown here of that. On the left, Torrento II, a 2 nm filter, is one of our most advanced and more recent wet etch and clean filters. This shows you the steepness of the ramp we've had to go through in a one-year period to get from sampling all the way to high-volume manufacturing. On the right, also for metal precursors, one of our fastest growing businesses in the company. This shows you a very similar challenge that we also face with that business. Get it right, get the data on it, and be able to ramp it very quickly to match our customers' needs to quickly get to yield.
The technology portfolio is not just R&D, but it's how you turn it into a high-volume manufactured product. Go to the next slide. Global infrastructure. This is, again, as I mentioned, being close to the customers is key. 2020 put a stamp on this. It was the year where everybody in the industry was highly focused on supply lines, making sure that we could be responsive, not only to the restrictions on travel and movement of goods that COVID caused, but also on geopolitical uncertainties and concerns about trade. We were able to demonstrate the unique ability of Entegris with its global footprint and multiple manufacturing sites across Asia and across the United States to respond and adjust to our customers' needs.
I'm very pleased to say that when the industry turned out to be stronger than we feared it might be at the beginning of the COVID crisis, we were able to respond and keep our customers running. That was, as you saw from our growth this year, it was a significant challenge. We were prepared for it because of the way we built the company. We built our Asia plants, and you can see the growth and the revenue that comes from those Asia plants for our global needs, as well as our Asia head count, where 50% or more today of our customer facing resources, including technical people that interface with the customers, are near our customers.
That's all great, but the other thing that's really exciting about this, and perhaps most exciting to me, is that the growth of the company has led us to have a need for more manufacturing space. We have significant footprints in the U.S., in Korea, in Japan, in Southeast Asia, and we saw the need for a new facility with more footprint, especially for the strong growth of the MC division, the SCEM divisions, and so we are planning right now, and we made a commitment to a new Taiwan manufacturing center. This will be our largest manufacturing campus, and it'll be beginning in 2021. We'll be building it out over the next few coming years because we need the volume. That's the main thing we're after there. Of course, it's near our largest customer, which is very convenient.
Again, shows our focus on having tight supply lines close to our customers, reducing the dependence on logistics, and being able to serve our customers in Asia well. We're very excited about that, and there'll be more information coming about that plant in a few weeks to come. We'll go to the next slide. Then operational excellence. In this industry that we're in, customers are more focused than ever on the effectiveness of their supply chain and the dependability of their supply chains. Our customers' most painful excursions have generally come from excursions not safety that we actually can demonstrate we can have that down to our supply chains as well, so they can be confident we'll be able to continue to supply consistently over time. That's their main focus is consistency of product, consistency of quality, and the data to prove it.
A couple of the measures that Bertrand talked about is our focus on quality. You can see our sigma level has grown past 5 sigma. I'm very proud of that. It's grown a lot over many years' efforts to constantly push our quality reputation higher and higher and our performance to go with that. Our injury rates, our safety performance, you can see, has improved markedly. I would say in the last 18 months, we've become especially focused on safety and more and more proactive programs to build a culture of safety at Entegris. The other good thing about Entegris, because of the size of what we are, the scale we are, and the things that we do, we can actually influence materials development.
Our suppliers care about what we need to have, and they're actually willing to work with us to deliver the kinds of quality, purity, and performance that we're looking for. The next step of that is really continuing automation in our factories for quality and manufacturing, but also for applying our data. We have a lot of data that our customers expect us to gather, and we're putting that to use to improve our productivity. That all sounds good, but the most important evidence that this works is on the operational excellence. If you look at the awards at the bottom there, this is the thing where I think I'm most proud of, is we've, in the last 12 months, received awards from some of our most critical customers, both equipment manufacturers and two large device manufacturers that influence these awards.
Recognizing the value that Entegris is bringing, recognizing our ability to ramp production, and our ability to perform consistently. That's a very important endorsement from our customers that obviously has also manifested itself in our growth. The next slide. Now each time we do an Analyst Day, I like to give you a quick refresher on the divisions. I won't spend too much time here because you've seen a lot of this before, and Bertrand described them a bit in his remarks, too. The Specialty Chemicals and Engineered Materials division, this is the division that's largely unit driven, that's putting things on the wafer and touching the wafer with almost everything they do to help these processes and these steps in these processes yield and produce the performance and the output that they're looking for.
These are the deposition materials, specialty chemicals for cleaning, special etch chemistries for what Jim will describe are some of the significant challenges of those vertical architectures that Bertrand mentioned, specialty gases for implants, and then CMP chemistries for some hard substrates as well as some other components and used in semi CMP. You'll see the growth drivers are going to be familiar from when Bertrand talked about materials intensity, the need for novel materials. These are all the things that the SCEM division is responsible for, responsive to. The SAM that you see, I'll just note, each of the SAM numbers you're going to see are only semiconductor. The SCEM division in particular has many applications in other markets that have demanding materials needs. The SAM you'll see today is focused on semi, which is still our number one focus as a company.
The next slide. In the middle, you'll see the key growth opportunities, and those match up to the drivers that we've been talking about, that I talked about on the previous slide. Advanced deposition materials, our fastest growing business inside of SCEM. We've continued to win in both the memory and logic space and are enjoying ramps this year. You can see the big growth in revenue this year for SCEM. Ramps in those new nodes are especially helpful to this division because of our share position in these new nodes. selective etch chemistries goes with that. Advanced coatings, rolling out to help improve cost of ownership and reduce downtime of equipment such as etch chambers. We actually use it in some of our products, too, to prevent corrosion in how we deliver materials.
We win because we have the ability to not only get these gases pure, get these solids pure, but deliver them in a pure way, in an effective and efficient way. That's one of the most important competencies of Entegris, is the ability to deliver these materials. I'll mention lastly their chemical synthesis. Bertrand mentioned some of our acquisitions in the chemical space, such as DSC and MPD. These give us the opportunity to vertically integrate the development and synthesis of our chemistries from the molecular level. We're expecting this division to continue to grow. It's been growing about 7% a year. We expect it to continue to grow about 300- 500 basis points above the market rate. Next, please. The Microcontamination Control or MC division, this is our largest and our fastest growing division.
This is what really makes Entegris unique, that we can bring not only these materials, but we can provide the means to make them and keep them pure. It's largely unit driven, liquid filters and purifiers that are used in bulk manufacturing of chemicals as well as quite extensively through the fab. Environmental filters, I'll show you a little bit more about that later. Some of them are capital-driven, are gas filters that go into a lot of gas equipment and manifolds, and then larger purifier systems that the gas companies use to make sure they deliver the appropriate level of purity to the fab, usually at the fab. The growth drivers here is purity. High purity needs that the semiconductor industry has, and then the yield challenges that contamination causes.
We can actually make a big impact here, where a lot of other chemical companies aren't able to do that on their own. Next slide. The growth opportunities, everybody suspects leading edge filtration, and that's not surprising. That's an important part of our focus forward in development. As interestingly now, part of what's driving this growth is the flowback of those requirements from the device manufacturer through their chemical manufacturers. We work very closely across the chemical manufacturer ecosystem to bring them the latest capabilities we have in filtration and purification, and enable them to provide the fabs with the purity levels that they're seeking. That's been a big part of the growth, as those requirements that I mentioned earlier are being flowed down to sub-suppliers. We've been helping them to be able to achieve that.
Obviously, the semi market is our major focus, but we also see opportunities in non-semi markets. In some of our acquisitions, such as Anow, are focused more on some other markets besides semi and give us a chance to learn more about how we might grow in those as well. Separation science is key here, but the number one thing that allows us to apply that is our applications expertise. Our ability to work with the end users and the OEMs and the chemical companies to help them solve these problems, and have the knowledge to do that is extremely important. We expect this business, it's growing, as you can see, very quickly. This is as reported sales over these years. To continue to grow 300-500 basis points above market. A very exciting business for us at Entegris. Next slide.
In the materials handling, as Bertrand mentioned, this is part of the heritage of Entegris. Founded 54 years ago to help with wafer handling for a niche in semiconductor industry. We still are the leader in that today. This business has got a mix of unit-driven and capital-driven products. Even the capital-driven products, they're typically refurbished or reused after some period of time to maintain the performance that they need to have in these very advanced fabs. This is where you're going to see unit-driven products such as drums and containers for liquid packaging. As Bertrand mentioned, we've taken what we've learned in packaging things like photoresist and our materials knowledge. We started to apply that to some life sciences applications. You can see the biologics bag there, representative of what he was mentioning relative to our work around COVID.
Shipping of wafers, shipping of liquids, protecting them in that time on their way to the fab in the ecosystem, a big part of what AMH does. In the fab, the fluid handling, the measurement and control, especially with our acquisitions of PSS and GMTI recently, certainly has enhanced our capability around that. Wafer handling, where we're the world leader. The growth drivers, high purity needs, protecting those substrates, protecting those chemicals. Then, of course, the increase in materials consumption is driving the drum business, and I'll talk about that on the next slide. If you look at the key growth opportunities, chemical packaging, our biggest challenge right now is just adding capacity. Over the next couple of years, we will double our capacity for manufacturing of high purity drums. That growth will come in Asia because that's where we need it.
That's where the chemicals ecosystem is ramping up to serve the growth in Asia that the industry is expected to continue to have. We're very excited about that, and we're making real-time, right now, additions to our chemical packaging capacity to serve that. EUV lithography, as the leader in technology around EUV reticle handling. We're very excited about EUV and the proliferation of that as the advance notes, a good thing for the AMH business. Safety. There's a lot of focus on safety in the fabs, as Bertrand mentioned. They care about how our safety is in our factories, but also very importantly, the safety of our products as they're used in the fabs. Again, this is another area where the business can see opportunities outside of semi, and some of those acquisitions I just mentioned give us an additional view into that.
You can see our expectation here, this business grows a little more in line with the market. It has some high share in a couple of positions here. We expect it to continue to outperform the market about 100- 200 basis points. The next slide. The other thing about AMH that you've heard me talk about before is that it's also a supplier to the other two divisions. It's one of the things that I think is most exciting about Entegris and why we're so happy to have the three divisions within our portfolio. They help each other. If you start on the upper right there, when AMH is actually a state-of-the-art molder of advanced polymers, the largest consumer of high purity grades of fluoropolymers in the world. That we can apply that to make components for our filters.
The AMH division is a component provider, design and provider to the MC division. In the bottom there, you can see we're also, not surprisingly, a supplier of things like chemical drums and packaging to the SCEM division. Unique opportunity here is that we're right now bringing up our capability in Korea to actually blow-mold the drums, the FluoroPure drums, and actually fill them in the same site with some of our chemistries from the SCEM division. That's a very unique opportunity for Entegris. I think most interesting too is if you look at the intersection between SCEM and MC. Matching filters to chemicals is absolutely essential these days. We work with the chemical manufacturers, and the IDMs, and the OEMs, a lot of cases, to design filtration and purification solutions specifically for their chemistry.
There's a lot of that unique need today in how you design a filter solution. We can do the same with the SCEM division. We can move quickly to develop filters that are specific to emerging chemistries that the SCEM division is working on, and we can get to a solution that works well together, and get there fast. That's the main goal we have with the combination there. It all leads to yield in the end. All of these things help our customers yield faster and ramp their manufacturing more quickly if we do well at coordinating between these divisions. Next slide. I won't spend time on this one. This is kind of a takeaway. It gives you kind of a view of what we do around the ecosystem. The next slide. To summarize, it's all about speed to yield.
For Entegris, we have the material solutions. We provide the materials protection, and that gets us faster to that yield that the customers are after. We win by the three main areas I just talked about, the technology portfolio and our applications knowledge makes us first at the table as a solutions partner. Our global infrastructure lets us execute quickly on development and ramp those solutions. Our operational excellence provides that consistent performance that our customers demand. With all of that, we've established ourselves as a smart choice for our customers. These capabilities together and our proven performance build that trust to be that trusted partner that makes us a winner. With that, I'll turn it to Jim O'Neill, our CTO, to talk some more specifics about these applications.
Great. Thank you, Todd. Good afternoon, everyone. Today, I want to explain how Entegris is innovating for growth. I'll highlight several technology trends that our customers are wrestling with and explain why Entegris is uniquely positioned to address them. Next chart. First, let's start with our innovation process. At Entegris, we innovate within a strong product management culture, and we have significant discipline about how we manage the portfolio. We have a rigorous stage gate process based on an Accolade platform, and we have a lot in that portfolio. Ideas typically enter the process through our own R&D and from external work we do with universities and consortia, as well as with our customers. We make choices about where we want to invest our R&D dollars. Clearly, the main focus is to execute projects through the portfolio, but in my view, a healthy portfolio is dynamic.
We work to actively cull projects that aren't headed in the direction that we need them to be going. We've chosen to invest at the leading edge. This is primarily driven by the long development and adoption lead times for leading edge technology, and if you don't invest early, you'll never intersect the customer's qualification dates. Over the last two years, we've increased our R&D spend so that now more than half of our project spending is directed towards innovation for advanced node technologies. The choices we've made have enabled a portfolio that we think is adequate to support our ambitions for growth. Our future sales growth is supported by real opportunities in the R&D pipeline. Next chart, please. A few words about the technology roadmap, which I think is getting quite interesting and is moving in a way that's very good for Entegris.
The struggle that our customers face is to come up with the innovations that keep them on a Moore's law trajectory. They've had to develop new device architectures and new patterning schemes to enable the aggressive density and performance gains that are demanded by their customers. However, underlying all this innovation is a foundation of new materials. These new materials have been required just to enable advanced devices and new patterning schemes. These new device structures are really small. If you consider that the critical dimension of a transistor on today's most advanced logic technology is about five nanometers across, that's about 40- 50 atomic diameters. If you compare that with a strand of DNA that has a diameter of about two nanometers, you can see that we're truly engineering materials on atomic scale dimensions.
Clearly, this is not only challenging, but it requires extreme control of defects, control of materials purity, and it puts unique requirements on the materials that are used to make these structures, things like high chemical selectivity for etches and high conformality for deposition precursors. Next chart. Our customers are literally facing tall challenges. If you consider the structure of any of the main device types in both memory and logic, they're all heading in the same direction. They're all getting smaller and taller. Making these structures requires new materials. It requires engineering these materials with atomic scale precision and control at part per quadrillion level purity and zero defectivity, all at the same time. Fabricating these devices is increasingly complicated, and this makes yield learning for our customers even more challenging.
This is actually a perfect situation for Entegris, more materials and more yield challenges. From my perspective, that's really the combination that Entegris does best. Our customers are increasingly recognizing this fact. We know this because they come to our tech centers, they bring their wafers. They work with our engineers on their challenging problems. To me, that's really the true value of our tech centers. Frankly, that's also how we learn. This type of interaction has happened over and over again in our labs in Taiwan and in Korea, and it's beginning to get traction with our new lab in China. Next chart. The industry is truly moving toward Entegris. Advanced node technologies use more materials, and that's good for Entegris. The number of process steps required to fabricate these advanced devices is also increasing.
This means that there are more opportunities to introduce yield-limiting defects. That's good for Entegris, too. This increase in materials usage and process complexity is especially pronounced for 3D NAND technology because of the increasing number of layers in the device. Consistent with this trend, we've seen for 3D NAND, say, going from 64 layers to 128 layers, we see CVD steps are increasing by nearly 50% and wet steps by 36%. For logic devices, the number of ALD steps is up 33%, and wet processes have also increased, even though some thought they would decrease due to the implementation of extreme ultraviolet lithography. Entegris is increasingly relevant as technology nodes advance. Where are we focusing our attention today? Well, I'm going to show you four industry inflection points and illustrate why Entegris is uniquely positioned to address them.
The next section is going to get a little technical, but I think it's important to understand the trends that are going on in the industry. Next chart. The first inflection point is the trend towards vertical integration. As lateral scaling, and by that I mean pattern shrinking or device miniaturization, as this becomes increasingly difficult to achieve, our customers are beginning to integrate into the vertical dimension. This is especially apparent with 3D NAND structures, where a taller vertical stack translates into increased memory density. These vertically integrated devices are difficult to make. The structures are taller and increasingly narrow, this brings the challenge of engineering materials with atomic scale precision on high aspect ratio structures. There are a greater number of materials in these stacks, this calls for high selectivity of both etch and deposition processes.
The resistivity of conductive layers becomes a challenge at these reduced dimensions, new metallurgies are required to meet the resistance and reliability requirements. Defects control is critical or you'll never yield these devices. For this particular 3D NAND example, Entegris brings to the table things like a highly selective nitride etch chemistry to form the metal channels in the structure, new atomic layer deposition metal precursors to fill them. Materials like molybdenum is a candidate, but there are other metals that are being examined as well. We have the ability to deliver solid forms of these precursors, as well as metal specific cleans and the matched filters that go with them. You see that Entegris brings a complete set of materials and yield solutions to this rapidly growing memory market, and 3D NAND has been a very good market for Entegris. Next chart.
The second inflection point has to do with the wiring that interconnects all the transistors on a chip. Historically, tungsten and copper are the metals of choice for this application. Copper in particular has a problem, that it requires a barrier layer to prevent it from diffusing throughout the device and causing reliability problems. As the wiring shrinks, copper lines become more resistive and less reliable, a new metal is needed. Options include materials like cobalt or ruthenium or molybdenum, all of which are in Entegris' portfolio. Changing the metal causes a cascade of downstream effects. A new metal means new electrochemistry, subsequent wet processes like CMP chemistries and cleans need to be reformulated. New formulations require new filtration schemes, since today's filters are made from chemically modified polymer membranes.
You need to be sure that when a complex chemistry comes in contact with a chemically modified membrane, that they're compatible. Otherwise, neither the filter nor the formulation will work. They need to be developed together, and as the provider of both the chemistry and the filter, this is where Entegris clearly has an advantage. Next chart. The next big inflection point is in the area of patterning. Continued device miniaturization is finally being enabled by the introduction of extreme ultraviolet lithography. I'm a little biased on this one because before coming to Entegris, I spent time leading process development at IBM's Albany operation, which was an early EUV adopter. From that experience, I understand some of the challenges that are posed by EUV. There are new resist materials, new mask materials, new pellicle materials, all of which require their own defect learning.
Due to the statistical considerations based on the energy of an EUV photon, something we call stochastic noise, EUV lithography has a built-in level of process variability. What this means is that all other forms of variability have to be controlled, and so random defect learning is even more important at these small dimensions. To address these opportunities, Entegris brings a number of solutions to the table, including defects control in the form of resist filtration or resist handling and delivery, and reticle pods that are used to protect the mask. We also bring new materials that can be used to make pellicle membranes, as well as novel films that can be used as absorber layers on the EUV mask. Overall, EUV adoption is great for Entegris. It drives new materials and defect control opportunities that we're actively engaged in.
Just consider the state of the industry today. The 7 nanometer node is in manufacturing now, and that was where we saw the first introduction of EUV. The 5 nm node is ramping, and there you see more extensive implementation of EUV. The 3 nm node is in development, and you'll see further expansion of EUV, as well as the possible introduction of a new device architecture called Gate-All-Around transistors. 2 nm node structures have been demonstrated using EUV double patterning techniques. EUV is a rapidly expanding opportunity for Entegris, with a tremendous extendibility. Next chart. The last major inflection that I'm going to talk about is the drive towards zero defects. Industries like the automotive industry, particularly with the rise of autonomous vehicles, this places new requirements on device reliability, and this includes both mainstream and advanced node devices.
We already know that small defects can kill device yield, and even smaller defects can cause reliability fails in the field. This puts increasing pressure on continued defect control. With increasing process complexity, yield ramps are even more difficult to achieve. The real challenge, however, is that for advanced node technologies, we can't even see the defects that we need to remove. A 7 nm node technology can be killed by a 3 nm- 4 nm defect, and latent reliability defects are even smaller. In-line defect detection equipment that you see in the fab can only detect defects down to about 15 nm or so. Speed to yield and reliability learning are a real challenge. This isn't just for advanced nodes, this is for mainstream node technologies, which have their own increased reliability requirements. Again, this creates enormous opportunities for Entegris.
We provide the pure performance materials that are used to make these advanced devices. Frankly, that alone is no longer sufficient. A complete solution requires the ability to filter and further purify materials throughout the supply chain. It requires the handling and delivery solutions that are needed to ensure the integrity of the material from the point where we make it, to the point where our customers actually use it on the wafer. These increased reliability requirements that are driven by industries like the automotive industry drive SAM growth for defects control solutions for both mainstream and leading-edge fabs. Next chart. Why is Entegris uniquely positioned to take advantage of these inflection points? Well, I'd argue it's because of the breadth of our portfolio. Let's consider a specific example that leverages the whole Entegris portfolio. It involves the metallization of a 3D NAND memory device.
Historically, the conductor for each cell in these tall memory structures has been tungsten. As more layers are added to the structures, say something greater than 200 layers, the resistance of the overall device increases. We need a new metal, and molybdenum is one candidate material, but there are others. This seemingly simple change sets up a cascade of effects downstream. New clean formulations, new filters, and you need to ensure the integrity of the formulation that we've made all the way to the wafer. This plays into Entegris' strengths in materials, in defects control, and in materials delivery. Frankly, there's no other single supplier in the industry that can do all of that at the same time. Next chart. What does all this mean for Entegris? Well, clearly we are currently experiencing an unprecedented number of process innovations in the semiconductor industry.
The increased demand for semiconductors means more silicon in both memory and logic. More silicon means more materials. These advanced devices require pattern miniaturization and atomic scale processing. The industry inflections drive the need for advanced performance materials and improved defect control. These changes create opportunities which Entegris is uniquely positioned to address. I think all of this plays into what we've tried to build at Entegris, which is a unique portfolio of materials and defect control solutions that our customers need if they have any hope of yielding their most advanced memory and logic devices. When I look forward, I think it's a really good time to be Entegris. Thank you. At this point, I'd like to introduce Greg Graves, who's the Entegris' Chief Financial Officer.
All right. Thank you, Bertrand , Todd, and Jim for the great setup. Really, my presentation today, there are two components to it. I am going to provide a little bit of historical perspective, and then we will look forward a bit. Next slide, please. Thinking about the historical perspective, this is our historical revenue, EBITDA, and EPS. I want to unpack it a little bit, but starting on the left-hand side of the page, revenue has gone from essentially $1.1 billion to a little over $1.8 billion. So $750 million of growth over the last five years. And I would note that we had a full year of ATMI in 2015, so much of this growth has been organic. EBITDA has gone from $230 million to $536 million, or an increase of a little over $300 million. Why do I highlight those numbers?
If you take $300 million and divide it by $750 million, the growth in revenue, you come up with about 41%. Which has been our commitment in terms of flow-through to the EBITDA line. The other point that I'd like to make is just on the revenue side, we've achieved our commitment of industry outgrowth of 200- 300 basis points. I mean, that 11% CAGR, or 9% of that, is organic. The market has grown about 5%, so well exceeded that, and then obviously, have exceeded it on the EPS line. Let me unpack this a little bit on the next slide. This is what we refer to as our multiplier model. I could've started with one more bar to the left, and that there's a multiplier element to the market, and then our market grows at about twice the rate of GDP.
You'll note, our revenue has grown at 2.2x the market. Our EBITDA has grown at 1.6x . These simple algorithms that we talk about, sometimes I say our model is very simple. It's outgrow the market by a few hundred basis points, drop 40% of it through to the bottom line, and you get an EBITDA growth rate that is, like I said, 1.6x the revenue growth rate. Effectively allocate your capital, manage the balance sheet well, manage the tax rate effectively, and our EPS growth is 1.3x the EBITDA growth.
This is what we'd like to refer to as our multiplier model, and Bertrand said at the beginning, "Can we do this again?" We think when you unpack it and think of it as just outgrowing the industry by a few hundred basis points, generating that flow-through, so we do think we can do it again. Next slide, please. Let me talk a little bit about our capital allocation principles. There's really nothing new here, but I think what's important to note is, first of all, we've been very consistent in how we've talked about capital allocation, and I believe we've been disciplined in terms of the capital allocation. Three components. It's investing in the business, it's investing in value-accretive acquisitions, and it's returning capital to shareholders. First of all, when we think about ER&D, we've historically invested 7%- 8%.
Our goal is to invest 9% of sales in ER&D. I use the term invest, and Jim alluded to it at the beginning of his presentation. We very much take an investment approach to how we allocate our R&D dollars, and this is something we frankly, under Jim and Todd's leadership, have gotten much better at. CapEx, 7%-8% of sales. Might not be that way consistently every year. There'll be years where it's more and years where it's less, but over time, expect to see us invest in the business. You don't get the capabilities that Todd described without investing in capital. Value-accretive acquisitions are a big part of what we're going to do. We intend to be a consolidator. We think to some degree, our execution has earned us the right to do that. Most of what we're targeting is within that core semi market.
A few of the transactions we've done, we've gotten some adjacent markets as well, the primary focus will be the core semi market. We talk about returning cash to shareholders. We talk about a 60% payout over time. That might not happen every year, but over a long period of time, we expect to return 60% of our free cash flow to shareholders through dividends and buybacks. Next slide, please. What have we done with the capital allocation over the last few years? We've spent $1.6 billion on acquisitions, $800 million invested in ER&D, $650 million invested in CapEx. Like I said, I just want to remind everybody, those are very much investments that have allowed us to drive that top line. $340 million on buybacks and $120 million on dividends, so about $460 million returned to shareholders.
I'd note that return to shareholders has primarily been in the last three years. We didn't implement the dividend till 2017, nor did we become consistent buyers of our stock till late 2017. If you were to look back over the last three years, that $460 million we've returned to shareholders amounts to about 60% of our free cash flow. Next slide, please. Just a little bit of a summary on our acquisitions. I mean, we've done nine transactions since 2014. Two of those transactions have accounted for 75% of that $1.6 billion spend, and that's ATMI, which was about half of it, and SAES, which represented about 25% of it. The rest of the transactions were relatively small tuck-ins. They've all had some exposure to our core semi market, and about half of them have given us exposure to other markets. Strategically, they've all been very important.
I'll just highlight a couple. The SAES Pure Gas transaction filled a significant hole in our portfolio around gas purification, and specifically high flow gas purification. The DSC and MPD transactions gave us molecule synthesis capabilities and have been a very significant boost to our deposition business. Anow increased our exposure in China, and Sinmat gave us an entry into the CMP slurry market. The other thing they've obviously done is they've been a huge contributor to value creation. If we had not done these transactions, our sales in 2020 would be $700 million less, and our EPS would be about $0.60 less. So they have been very value accretive from an earnings perspective. Next slide, please. Let's shift our focus a little bit and talk about why I feel so good about where we stand today and our future. First of all, we've got a great capital structure.
I sleep really well with this balance sheet. It's both, it's conservative, and it gives us a fair amount of flexibility. It's conservative in that we consistently maintain at least a couple hundred million dollars in cash. Our EBITDA target on an ongoing basis, our leverage to EBITDA target is 2.0. We're very focused on maintaining that strong BB rating, which will allow us to access to the credit markets in good times and bad. Looking at the lower right, we don't have any significant maturities till 2025. The other thing this balance sheet gives us is optionality. I talk about a gross leverage target of 2x , but we've talked about our willingness to take that up to 3.75x for significant M&A opportunities. That alone gives us $1 billion of firepower with the existing balance sheet and the existing EBITDA.
Anything we buy will bring additional EBITDA. We've got the flexibility to do meaningfully sized transactions. Next slide, please. I want to talk a little bit about why we're so confident in terms of our ability to continue to expand our EBITDA margin. Historically, we've obviously focused on productivity and efficiency. We wouldn't be able to deliver the returns that we have without it. In the last 18 to 24 months, we've made it much more of a central theme in the organization. About 18 months ago, the finance organization, in conjunction with the rest of the team said, "Let's make productivity in the functional areas cool," so to speak. I mean, it's cool to be innovative in the ER&D part of the world. Let's make it cool to be productive in other parts of the organization. What are we doing?
I mean, we're standardizing processes, we're simplifying, we're centralizing, automating, and we're making greater use of shared services. We've just established a shared services facility in Penang, Malaysia, where we're moving a lot of our regional finance functions. We're moving a number of our IT roles there as well. Those will all make us more efficient. Where are we doing it? We're obviously, we're not doing it in ER&D. We're doing it so that we can continue to invest in ER&D, continue to invest in our core business. We're doing it primarily in the SG&A functions, the supply chain functions, a lot of the high transaction count type functions. There's benefit just beyond the productivity and efficiency. As we do these things, we also increase the scalability of our business model and the ability to add things to the business.
A number of the acquisitions that we've added recently, we've almost immediately taken the accounting function to the shared services operation. This scalability is an important factor as well. Next slide. Why is the productivity important? It's really important as we want to continue to drive this 40% incremental flow-through that we've talked about. Doesn't sound like a big number, but when you look at that multiplier model over a period of five years, it makes a big difference in terms of our ability to grow the profits significantly faster than the revenue. This slide, we've operated to a target model for a long period of time. The left-hand side shows at $1.8 billion in revenue. It roughly looks like this year's P&L will be a little bit ahead of that on the EPS line.
The idea being at $1.8 billion in revenue, our commitment's 24% operating margin, 29% EBITDA. As you move out and you work through essentially with that 40% flow-through, you'll see you get out to $2.4 billion, that EBITDA margin has expanded to 32%. EPS has moved from $2.35 to $3.55. That's assuming the same levels of debt that we have, essentially the same share count, and essentially the same tax rate. This commitment to the 40% flow-through is an important concept, and while trees don't grow to the sky, we're comfortable that we can continue to deliver on it. The important way to look at the flow-through, however, as I showed it over a five-year horizon. There are going to be quarters where we don't achieve it. There might even be years where we don't achieve it.
For instance, in 2018, we didn't achieve it because we made an acquisition of a significant business that had EBITDA margins of 30%. It's hard to get a 40% flow-through on revenue from an acquisition that has less than a 40% EBITDA margin. Point is, from an organic perspective, committed to this flow-through as we go forward. What does that mean when you take what we've talked about the growth rate? Let's roll to the next slide, please. What does that mean when we talk about where we're starting from in 2020 and where we think we'll be after three years? Bertrand laid the foundation for that 10% top line growth. If we grow 10% annually over the next three years, that'll take us to $2.4 billion in revenue.
You reflect back to the target model at $2.4, we'll have a 32% EBITDA margin, 300 basis point improvement over where we are, and that's again based on that 40% flow-through. We'll have non-GAAP EPS at $3.55. The other thing I want to highlight on this page, we're introducing new targets around our return on invested capital. Our return on invested capital is about 16%. Believe we can take that to 20% over a three-year time horizon. The return on invested capital is an important concept for us. Everything we do is driven by essentially returns. Jim talked about it at the beginning, our R&D portfolio, there's an NPV and an ROIC on each major initiative within the R&D portfolio. Obviously, our CapEx is driven by that, and as Bertrand showed, when we evaluate M&A, we're looking for an ROIC in the high single digits by year three.
ROIC is an important component to everything we do, so we wanted to lay a milestone goal out there for that as well. Next slide, please. To get from that $3.55 that I just showed you to the $3.75- $4 that Bertrand talked about, you obviously have to take capital allocation into account. This is just some for instance scenarios. The first one is a share repurchase. If we can keep our leverage at the existing level of 2x EBITDA, and we benefit from an expanding EBITDA and the cash flow that we generate, if we were to use that cash to repurchase shares at our current multiple, which is about 25x forward earnings, we would generate about additional $0.10 of accretion by using that capital in that way.
If we were to use it or able to effectively find M&A at reasonable valuations, which we used here is 14x EBITDA, deploying that same capital on M&A would provide $0.25 of accretion. Again, we talked about being willing to take that leverage level up for transformational M&A. If we were able to find a transformational transaction levered at 3.5x , we believe we can generate an incremental $0.50 per share. These are hypothetical scenarios. At one level, they're just math, but what we want people to understand is obviously the balance sheet has optionality for us, and if we're effectively using our cash flow and our balance sheet, we're going to generate returns above and beyond the organic.
The $3.55 I showed you organically, plus kind of the midpoint of these numbers takes you to that $3.75-$4 that Bertrand Loy talked about at the beginning. Next slide, please. I'd just like to summarize the six reasons to own Entegris. One, is we have an exciting industry environment with really good secular growth trends. I think of where this industry has come over my career at Entegris, the breadth of where semiconductors are used, the breadth of the applications, has really driven significant growth. The transition to solid state memory, again, a big growth driver, and Jim touched on some of how that will impact in Entegris. Finally, the industry has become a better industry as there's been consolidation. There's more industrial logic. It's a much more rational industry, I believe. Jim outlined our exposure to the key technology inflection points.
Those are all what are going to allow us to grow in excess of the industry. We've got really strong competitive moats. We've got a great IP portfolio, both in the form of our patent portfolio, our knowhow, as well as the significant applications expertise that we have. We've got a resilient business model. Our products are sticky. We get qualified in, and we've got products that are 20 years old that are still chugging along. We've got a diverse customer base. We've got a diverse mix of products. All of those things make us a very resilient company. We're a big cash generator. We've always generated significant cash flow, and we think we'll continue to generate significant cash flow. Finally, we're disciplined allocators of capital.
We believe I showed you some of the options that we have with regard to that cash flow and the capital that we generate, but those are essentially icing on the cake of the strong execution that we intend to continue to deliver and that we've delivered historically. In the end, we think we've got an opportunity to continue to compound value. With that, I'll turn it back over to Bill for the Q&A session.
All right. Thank you very much. I have been getting questions in over email and from the chat function, so I will get started here. If you have more, please send some over. We'll start with Sidney Ho from Deutsche. This one will go to Bertrand, and it'll be a two-part one. Bertrand, in your slide when you showed the business mix and the customer mix, when you think about that customer mix, and you think about the mix between foundry and memory, how do you see that evolving over the next three to five years?
I think, again, one of the attributes of Entegris that is actually so unique is this very broad and diverse customer base. We're fortunate enough to sell to most participants in the industry and most segments of the industry. Having said that, historically, the industry roadmap of our logic and foundry customers has been a lot more challenging than the roadmap of the memory makers. Historically, as a result of that, we've had more opportunities to contribute t o the logic architectures. This is changing very rapidly, as you've heard Jim describe, and I have actually provided some quantification of that statement. If I want to summarize all of this, I would say that today, memory represents about 30%, roughly, of the revenue we generate with fab customers. I would expect that number to migrate to close to 50% over the next three to four years, depending on the pace at which we see the migration to 96 layers and higher layer count structures.
Okay, another follow-up from Sidney. We mentioned in a few places potential growth opportunities outside of semi. What are some of these potential areas?
We have historically always tried to be opportunistic in adjacent applications and markets. Typically, customers from industries outside of semi would come to us, looking for solutions to emerging problems that they are facing. In some cases, it actually makes sense for us to put R&D monies to work and to develop some derivatives of product platforms that we have developed for semiconductor applications. The industries where we see the most opportunities for us going forward would be medical, life science, and to some extent, aerospace applications as well. Again, I think that given the rate of growth that we expect in our core semi markets, I don't expect a non-semi portion of our business to grow meaningfully beyond the 10% that it represents today.
Expect that portion of the business to be within 10%, 12%, 13% of our total revenues in the next four or five years.
Okay. Got a question from David Silver, C.L. King. Jim, and I'm going to paraphrase here a little, David, but Jim, in terms of the R&D projects you laid out, what are the one or two key ones that a year from now or so, you will judge whether we're successful or not? The ones you're most excited about.
I think the ones that we're most excited about are the ones that are targeting the greatest areas for growth. I would look at things like the work that we're doing in our SCEM division on deposition precursors. Deposition precursors of the form that we call atomic layer deposition materials. These are the precursors that are used to fill those high aspect ratio structures, both in memory and logic. We're already beginning to see some traction there. We think that with some further work that we're doing in 3D NAND, that should be something that we can be proud of. I think we'll be successful and I think we can be proud of it. Another major area of innovation is in selective etch chemistries.
It's no longer just providing chemistries that can remove one material on the wafer, but being able to remove a material selectively in a very complicated stack of materials, similar to what you find in advanced devices. We're seeing many more demands or calls for it from our customers. I think we're beginning to get some traction there. The other major area is in innovation surrounding the fundamental performance of membranes and filters. I think as you get to defects that are sub-10 nm , sub-5 nm , you need something more than a better sieve. You need a membrane that is chemically modified or functionalized to selectively remove contaminants. That takes a lot of innovation and a lot of know-how, and I think we're very good at that, and it's in huge demand right now.
I think if you just look at the success of our Torrento 2 filter, things that follow that are other areas that are very exciting.
All right. I got a follow-up from David here for Greg. Based on your history of steady incremental M&A, how do you rate your capacity for taking on more M&A and all the related integration challenges?
What I would say is I'm thankful that at one level for the COVID year this year, because we did not do as much M&A after doing five transactions in a little over a year in 2019 and early 2020. We've had an opportunity to get ourselves positioned in terms of integrating those businesses. I think I do have capacity for additional transactions as we move into next year.
Okay. I'll go back to Jim on here. This is from Toshiya Hari from Goldman. You have explicit financial criteria for your R&D investments. If so, what are they? How are business unit leaders evaluated, incentivized for their follow-through in hitting these targets?
Yeah, I think that one of the primary things we look at is the percentage of our total revenue that new products generate. We have a loose goal of a vitality rate of something in the range of 30%-35%, something in that range. Clearly, we want new products to carry a higher margin than the products that they're replacing. In order for that to happen, they have to be differentiated. We really need to be able to ensure that the new things that we're coming out with are substantially differentiated in the industry relative to our competitors. We have metrics around that as well.
Yeah. Bill, would you like me to add a little bit to that?
Sure.
Just a couple things. We put every project up against financial criteria for return on investment, as Greg mentioned. They are measured equally against other opportunities for investment. We set business goals for every project, and we actually do, we call them lessons learned or rearview mirror reviews of these projects after they have been out in the market for a while to see if we met those or not. If not, why not, and apply those lessons learned to future projects. In terms of motivating and incentivizing the business leaders with regard to that, one, they all have goals every year, in some case compensable goals, around the margin performance of the business, the growth of the business, winning of new applications and new nodes, which require those new products.
Very specifically on the performance of some of their top projects, the kind of projects that Jim described recently. There's a very tight tie of performance and accountability to be able to delivering those R&D dollars.
Great. Bertrand. This is from Mike Harrison from Seaport Global. You mentioned looking at filtration options outside of semi. How similar is the science? Does it mean you can do it organically or inorganically to get into these non-semi markets and filtration in particular?
Right. There are some similarities, but there are also a lot of differences. This is actually one of the reasons that led us to acquiring Anow in Q4 of last year. This company that is based in China actually brings a lot of the capabilities that we were lacking at Entegris. I think that the combination of the two platforms will actually be a very powerful combination, bringing some of the best membrane scientists to work together with the right application knowledge with the medical and life science applications in mind. It's going to take a few years for us to develop the right platforms. I believe that there are opportunities for us to unlock in the field.
Okay. Another one, Bertrand. Patrick Ho, Stifel. When you're thinking about R&D development, how do you decide between building it organically or inorganically, in other words, using M&A to supplement what you have? What are the kind of key metrics that you look at when you think about that?
The theoretical answer would go along the lines of what Todd and Greg were mentioning. It's really about assessing the risk, assessing the return on the investment. Remember what Todd said in his presentation, which is actually something very important. Our customers not only expect the best solutions, but they expect that within a specific and usually very narrow timeframe. Time to solution is very important. All of those criteria are really something that we think about in the context of two major decisions. The first one is looking at existing applications, existing markets that we serve today. Then the other series of opportunities will be the white space, so areas where we don't play today. We need to make those decisions of internal development versus acquisitions for those two areas. I'll give you a couple of examples.
In the case of deposition materials, six years ago, we realized that this was an area where there would be some really exciting growth potential. Entegris at the time didn't really have any capabilities, so we decided to acquire ATMI. After that, we made a lot of internal investments to grow the capabilities, grow the global footprint of this particular business. A couple of years ago, we realized that we were not going fast enough. We had built some very credible capabilities in lab scale synthesis capabilities, pilot scale and applications. We didn't have high volume manufacturing capabilities. We were still relying on a network of suppliers, and it was a limiting factor. We had to decide to either invest internally in the high volume manufacturing or acquire those synthesis capabilities.
We decided to go and acquire DSC and MPD simply because we didn't have the time to invest internally. Again, I think that we are constantly having those discussions as a management team. I think we are a team that is very pragmatic. There is really no pride of not invented here, and I think we have demonstrated that we can be very flexible in terms of what we choose to develop internally versus what we decide to acquire and bring into the portfolio.
Okay. I got one for Greg from Paritosh Misra from Berenberg. We talked about the 40% flow through as a total company, how do you think about incremental margins in the three separate segments?
Let me think about that for a minute. I would say as we go forward, and Todd showed it in his presentation, if you think about the SCEM business, the incremental margins are probably higher than they are in the MC business. We've made significant investments in SCEM over the past few years, and so much of that is behind us. Whereas in the MC business, we've got some significant investment ahead in terms of capacity. We'll see higher margins than we have today, but we won't see as significant of a flow through. Then we'll continue to see improvement in the AMH business. Last quarter, we saw 23%. I told people not to hold on to that number, but I'm hopeful that we're out of the days where we see numbers in the teens for that business.
All right. This from Chris Kapsch from Loop. Bertrand, Chris asks, "What sort of MSI expectations do we have over the period to 2023? Do you see if memory and foundry logic balance contributors?" Now, of course, we're talking about the market. "Is there one of those that will provide sort of unbalanced contributions?
Chris, I would say that we have stayed at very high level with those planning assumptions, and our view is that the market overall will be growing at about 5%-6% annually over the period. That's a blend of what we expect to see for MSI and CapEx. That's as far, frankly, as we've been thinking about it. I think the key takeaway for you when you hear us talk about the growth is really what we commit to deliver on top of the industry, whatever the industry ends up doing. The way we will do that is every year, at the beginning of the year, we will share with you with a little bit more specificity what we expect to see in the various segments of the industry, and therefore update the rate of excess performance that you can expect.
Again, for today, please keep it at high level. I'm sure, as we know, that this industry is hard enough to predict, we won't go with more details for today.
All right. Jim, this is one I know I can't answer. This is from Max Licht from, how is Entegris preparing for the eventual introduction of quantum computing and hybrid chips that include gallium nitride and silicon carbide?
Okay. Let me answer the second part first. That's the easiest. There are a number of new materials that are coming into play in semiconductors, gallium nitride, silicon carbide being two of them as a way to extend into the power device range. There's a lot of demand there for components for the automotive industry. That's actually an area that our recent Sinmat acquisition is quite strong in, the polishing of hard materials. Through that acquisition, we're not only able to participate in the market, but learn more about other applications and other ways in which Entegris can play in that market. The power device market, gallium nitride, silicon carbide is an area that we are actively playing in. As far as quantum compute is concerned, we have our ear to the ground through consortia, through the work that's being done, spearheaded by Yale.
That's some of the work that's being done by some people associated with IBM. At this point in time, we don't have specific materials programs, but we are listening. Right now, the unique demands for materials is not something that we have found a path into it at this point. It's not clear that the materials are specifically unique to quantum. It's more of the design and the operating temperature, which are the real challenges there. As far as the gallium nitride and silicon carbide, we are actively playing there.
Okay, this is for Bertrand. It's actually a combined I'm going to combine a couple questions from Patrick and Toshiya. What is our competitive edge when targeting potential acquisitions and the integration process that allows us to hit these financial targets that we provide?
Well, I think that there was an earlier question around the bandwidth that we currently have. I think that's actually a very important question because it's not about just closing on a deal. The value is only unlocked if you can really properly integrate the companies that you acquire. The best way to do that is really to start thinking about your integration plans as early as you can. I think that at Entegris, we do that very well. Usually, when we start screening seriously potential acquisition targets, we not only have a team looking at the target from a valuation standpoint, but we already have, very early on, a team starting to think about integration opportunities and risks.
That plan actually continues to evolve as we get closer to closing, so that by the time we close a deal, we are really ready to execute very quickly on the integration, and we don't waste any time. Again, I think it's that focus that we have in most everything we do at Entegris, the intensity that we have, and frankly, the great teams that we have that are able to do those complex integration projects very, very effectively. That comes with experience, and that comes with practice.
Okay. Next question, another one for Bertrand. Charles Shi from Needham. It's quite a long question, but I'll sum it up. When we talk about our sales to customers, we don't talk about the packaging end of the market. Any plans to address that? Anything in the portfolio that exists today?
Yeah. I would certainly invite Jim to jump in if he wants to add. It's true that wafer level packaging is evolving very, very quickly. The processes are becoming a lot more complex, requiring more automation, more precision, cleaner processes as well. We are starting to identify the number of areas for us to contribute around microenvironment solutions, but also in terms of chemistries, filtration, and all of the traditional solution sets that we have developed for the front-end processing. Again, Jim, I don't know if there's anything specific that you want to call out, but again, this is a new trend, certainly, and I think that there are areas for us to contribute.
Yeah. I think if we look at it, and we have looked at it from the roadmap perspective, that particularly 3DIC and certain aspects of wafer level packaging really represent an extension of wafer processing farther into the back end, as you need to prepare the wafer for through-silicon vias, wafer level bonding, chip to wafer bonding, things like that. That represents, I think, an expansion of the field in which we play. We have been working very closely with some of the leaders in the industry, some of the foundry players in that regard, and are beginning to identify opportunities where Entegris can contribute our materials and application knowledge.
Okay. A question from Chris Strom. How standardized are SCEM and MC products that customers use as they advance nodes? It's probably one for Bertrand, at least to start. If a customer is moving from five to seven or seven to five, will it use the same products, or do you customize these solutions?
First of all, at seven, a lot of the solutions are already customized, and our customers, again, are trying to migrate to the new node as quickly as they can, and they're going to try to extend the materials and the solutions as much as they can to minimize the number of changes, and therefore, the risk of the transition. Having said that, every time they see a good reason to change the material because it would improve the performance of the chip or the cost of the chip, they would actually do that, and we will aggressively work with them. It depends. Again, all of the changes that are being made are increasingly very unique to the integration scheme of a customer.
I think that's why the incumbent advantage is so important in this industry because that allows you to have very unique insight, and therefore, allows you to shorten your time to solution for the next generation chip as well.
Yeah, if I could just jump in here for one. There's a figure of merit that I think when looking at it from the customer's perspective, a node-to-node transition would likely leave 70%-75% of the processes the same. That accounts for the stickiness of materials in that space. It's the 25%-30% that changes that represents the opportunities for innovation, and that's really an area where Entegris is focused.
Great. I got probably just a few more questions here. If you haven't sent anything in, send it to me via email. Greg, how would you lay out the puts and takes of the operating margin targets for the three different divisions?
How would I Say that again, Bill, the puts and takes?
They're all looking for increases in operating margins. At a really high level, what are the drivers for the three divisions to hit those increased operating margin targets?
I would say in the MC division, it's going to be the continuation of higher volumes, the continuation of, Jim talked about advances in membrane technology that allow us to continue to capture more value from the customer. It's going to be effectively adding new capacity in an efficient fashion that allows us to continue to expand the margin. In the SCEM division, I think I would say we're in pretty good shape from a capacity perspective. It's going to be leveraging the existing capacity, continuing to develop, again, products that are differentiated, leveraging the capabilities that we've gotten from the two acquisitions that we did last year.
I would say the AMH division, where we don't expect quite as much growth, that's a division where, to me, it's more about solid execution with the existing footprint that we have, as well as some growth within a couple of key areas, liquid packaging, for instance, some opportunities within life sciences. Those will all play into the AMH division.
A little bit of piggybacking on that. Bertrand, another question from Sidney Ho. The growth targets for AMH are a little bit higher in terms of above-market growth than they were before. What's fundamentally changed in that business that gives you/us more optimism about that business growth opportunities? AMH.
Right. I think Greg touched a little bit on that in his previous answer. I think there are a couple of product lines that we believe will benefit from very favorable trends, especially around liquid packaging solutions, both in semi applications as well as in adjacent markets. That's going to be a nice driver for AMH. We also expect some of the recent acquisitions, PSS and GMTI, to contribute nicely to the overall top line of AMH. The other additional good news is that both of those product areas, so PSS, GMTI on the one hand, then liquid packaging solutions on the other hand, have usually better margin attributes than the overall AMH division traditionally. They should have a nice contribution both to top line and to the margin profile of the business.
Excellent. Well, that wraps it up for today. Thank you very much for attending the Entegris Investor and Analyst Day in 2020, the Virtual Analyst Day. We sure hope we can get back together live as soon as possible, especially next time. If you have any follow-up or any additional questions, please reach out to me directly via phone or via email. Thank you very much and have a great rest of the week.