Good morning, and welcome to Entegris' 2021 Virtual Investor and Analyst Update Meeting. I hope everyone is staying healthy and had a good Thanksgiving holiday. Before we get started, there's a few housekeeping items I'd like to cover. For the best performance, and we sent this out earlier, please use Google Chrome or Firefox and disable VPN if it's possible, and any other programs you might have in the background. We posted a PDF of the slides on the IR site. You can also find those on the tool here. You can adjust the boxes in the tool, same as last year, and if you've made too many adjustments and need to kind of revert back, there's that counter-clockwise widget that's at the bottom of the screen that you can use to reset.
If you have any tech questions, you can use the Q&A box to send that to the engineer, and he'll get back to you. After the presentations, we will have a Q&A session. You can send Q&A questions to me via the Q&A box, but easier, you can send them directly to my email, bill.seymour@entegris.com. And I can. Please, if it's not in there already or put your name and your firm name in there, please. Next slide, Anna. We expect the presentation portion of today will last approximately 40 minutes, and we've allotted about 30 minutes for Q&A after both the presentations. Speakers today are Bertrand Loy, President and CEO, and Greg Graves, our CFO. Next slide, please.
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, and actual results could 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 here in the PDF and also on our website. Next slide. With that, I'll hand over to Bertrand. Bertrand.
Thank you, Bill. Good morning, everyone. Thank you for joining us. It's been approximately a year since our last Analyst Day, and we have certainly covered a lot of ground since that time, in spite of a very challenging operating environment, as you know. Much so that we expect to nearly reach performance levels this year in 2021, we originally targeted for 2023. In addition to all of that, our opportunity pipeline has also strengthened in many different ways. All of these various considerations led us to conclude that we needed to update you on our growth targets, our strategies, and on our increased investment plans that we need to accelerate our growth momentum, increase our competitive advantage and ultimately increase our market share. Anna, next slide, please.
Let's start with a few key facts on Entegris. For those of you not familiar with our platform, and I think it would be a good refresher for everybody else. The company was created over 50 years ago. Its original product was the very first wafer carrier solution for the then nascent semiconductor industry. As a matter of fact, if you go to the Smithsonian Museum in Washington, D.C., you can see some of the early Entegris products on display. Obviously, since these early days, the company has evolved very significantly, both in size and in scope. Our mission, which you can see on the upper right here, remains essentially the same. Our mission is to provide unique process solutions to help our customers improve both the performance of their products as well as their manufacturing yields.
To accomplish this, we have built a platform that is very unique in our industry. A platform that is comprised of three divisions corresponding to the three major types of product solutions at Entegris, and you can see that on the left side of the slide here. These divisions develop obviously the discrete products and technologies required for them to compete effectively in their served applications. But they also collaborate very closely together in order to co-optimize chemistries, and filtrations, and packaging, and dispense solutions to bring more complete solutions to our customers and to market faster. As you can see on this slide as well in the lower right, we provide our solutions across the semiconductor ecosystem.
Fab customers represent close to 45% of our revenue, of which logic, both mainstream and leading-edge, logic account for 69%, and memory DRAM and NAND account for the remaining 31%. The rest of our revenue is split between different clusters of the semiconductor ecosystem, including equipment makers and specialty chemical manufacturers. You will also note on this pie chart that about 10% of our revenue come from applications outside of the semiconductor industry. Next slide, Anna, please. One key attribute of our platform is that it is inherently very resilient. If you look at our historical performance, you will notice that Entegris has been more stable than most industry participants. What is so unique about Entegris? What makes our business model so resilient?
I think there are a few different ways you can answer these questions. The first way to approach that question is to remind you all that the vast majority of our revenue, approximately 70%, 70, is recurring in nature, generated by consumable products like chemistries, materials, filters that are used daily in production cycles by our customers. The second way to think about this is that most of our products are customized to very specific process recipes of our customers. Our solutions are very sticky and have very long tails.
Once we are specced in, once those solutions are specced into a new technology node, they are very rarely taken out, and they will be used essentially for as long as the fabs remain in operation, and in some cases, especially in logic, that could be about, you know, over 10 years plus. The third way to think about resilience is that we have a very broad product offering. High mix, low volume with very few product platforms reaching annual revenues in excess of $50 million. Then finally, as you recall from the previous slide, our customer base is broad, is diverse.
To provide some context, additional context to that statement, you can see here at the bottom of this slide, that we only have one customer representing more than 10% of our revenue, and our top 10 customers combined amount to approximately 45% of our sales. The punchline of this slide is that our success at Entegris is not dependent on any one product or any one customer or any one technology inflection, and that translates into resilience. Another final point, an important point that I want to make, on this slide is that in our case, this added resilience does not come at the expense of growth. Let's talk about that. Anna, if you could maybe flip to the next slide. Let's talk about growth. I think this slide is the perfect place to do that.
Before I dive into more details, I want to make two high-level observations here. The first one is that we have ambitious growth aspirations for the next few years, as you can see in that red circle. The other big observation is that we have optionality in our growth plans. With that as a backdrop, let me help you decipher this slide together. We'll start with the main driver on the left, which are for sure the many opportunities we see in our core market, the semiconductor ecosystem, where we expect to benefit from a number of very powerful tailwinds. We expect the industry to grow at approximately twice the rate of GDP.
Because of the growing importance of what we do, we expect to outperform the industry by 3%-5%, which is a bit higher than what we said last year. I will describe all of these trends in a bit more detail in my next slide. Let's move now, for now to the next growth driver, which will be small to mid-size acquisitions, which are expected to contribute 1%-3% of growth a year. For context here, remember that M&A has added about 2% of annual growth in the past five years. What you see on this slide is very consistent with what we have accomplished with our programmatic M&A strategy in the past few years. Last but not least, we are increasingly focused on finding new emerging sources of growth.
I will provide here again a little bit more details in a minute, but given our current opportunity pipeline, we expect another one-two points of growth to come from these emerging growth factors, including in applications like life sciences, where we have already obviously seen some traction. If you add this all up, we expect to deliver a top-line CAGR of approximately 15% over the next three years. Let me now go into a little bit more details on the various cylinders, if I can call them that way, of our growth engine. Anna, if you could go to the next slide, please. This is a slide, I mean, that you've seen many permutations of, and this is still obviously a very important backdrop for us.
The punchline of this slide is that the semiconductor industry is entering a new era. Sensors are proliferating everywhere around us, collecting, generating huge amounts of data. Data that will need to be stored, indexed, correlated, to generate new insight and drive decisions, actions by humans and increasingly by machines. As the world becomes more interconnected, we believe that demand for semiconductors will grow at approximately twice the rate of GDP, as I was mentioning earlier. Another important point to make on this slide is that the demand drivers for semiconductors are now a lot more diverse than at any other point in time in the history of the semiconductor industry. I believe that this multiplicity of demand drivers will be the source of greater stability for our industry.
Whether that is coming from the adoption of AI, EV, 5G or beyond the needs of high-performance computing, to name just a few of those drivers. More broadly, I would argue that the semiconductor industry will likely be one of the most exciting industries to be in for the next decade. On the next slide, this is really a summary of what Jim O'Neill, our Chief Technology Officer, presented last year. You remember that Jim spent a lot of time last year describing the logic and memory roadmaps, highlighting emerging process challenges and connecting those with our solution set. Today I'm not gonna go into that same level of detail. In fact, I want to do the exact opposite.
What I want to do is encourage us to take a few steps back, and from that vantage point, a few things will become clear. The first one is that while of course, the logic and memory technology roadmaps are different, the big themes, the big drivers are in fact very similar. These roadmaps are all converging in their own ways towards new and novel materials, more complex architectures. You can see here a reference to 3D NAND, Gate-All-Around, all the while continuing to push for more miniaturization. That's observation number one. The second observation, which is also very important, is that the roadmaps are becoming more challenging, more ambitious, and our customers, all our customers, are also expecting the pace of innovation to accelerate. In other words, what our customers are trying to do is increasingly more challenging, and yet they are hoping to shorten the development cycle.
All of this is very favorable to Entegris because we operate at the crossroad of material science and material security, and it means that our solution set is increasingly important to our customers and their ecosystem. Simply put, the market is really coming towards us, towards Entegris. And that's gonna bring a lot of new opportunities, exciting opportunities, but also new expectations and new responsibilities for Entegris as well. Let's start with the opportunities on the next slide. I'm gonna focus on trends that will be very favorable to our SCEM division. And as I was mentioning, material science is increasingly driving the industry roadmap, and we expect our SCEM division to benefit greatly from the compounding challenges of miniaturization, high aspect ratio device architectures, whether it's 3D NAND or Gate-All-Around, and the introduction of new materials.
In a simpler way, maybe to say that new materials will be required as the industry advances from node to node. As a result of our close collaboration with industry leaders, we are developing, for example, a new generation of materials with better electrical properties for thin film deposition. We are also engaged in the development of new etching chemistries with greater selectivity that will become critical as 3D NAND devices grow in height and as Gate-All-Around structures are adopted in logic architectures. These charts show the material spend per wafer in the leading edge nodes. This is based on third-party industry data. As you can see, our served market is expected to grow very rapidly, both in logic and memory, as our opportunity per wafer increases in new nodes.
In this chart, just to orient you, the middle bar indicates approximately where the leading edge, high volume leading edge is today, and on the right is where the leading edge high volume production is expected to be in four years from now, roughly four to five years from now. Again, on the left, four to five years ago, the middle is today, and on the right, four to five years from now. The takeaway is that Entegris will be uniquely enabling greater bit density, faster compute, and higher energy efficiency in the next generations of logic and memory chips. Moving on to the next slide, another major area of customer engagement for us is defect management and contamination control. Here I am mostly talking about our MC division, microcontamination division.
By extension, it applies at some level to our AMH division as well. As you know, purity is essential for our customers for two main reasons. The first one is that process purity very directly correlates with wafer-level defectivity and yields. The second reason is that process purity also correlates with long-term performance and reliability of semiconductors. In other words, our contamination control franchise makes our customers' products better and cheaper to manufacture. The one thing I also want to share with you is that, as you know, the desired purity requirements are a moving target, right? Device feature sizes are continuing to shrink. We were talking about a number of new materials being adopted. These new materials are susceptible to new classes of contaminants.
All of that means that the number of species of contaminants is increasing. It means that the permissible size of the contaminants continue to shrink, and it means that the concentration levels for those contaminants are also tightening at every node transition. For us, all of this translates into three things. The first is that we will see more points of filtration across the ecosystem. The second thing is that we expect to see greater usage of more advanced filters. Then the third is that we expect, you know, increased frequency of filter change out. The graph on the top of this page illustrates how we expect our content per wafer to evolve over time in leading-edge logic and memory fabs, as our customers continue to transition to more challenging architectures. Next slide, Anna, please.
Here I want to introduce a couple of additional considerations. Before I do, I want to be very clear here. We're not taking credit for anything on this slide. These two factors are entirely outside of our control, and yet these factors will be very favorable to Entegris. The first one is the graph on the top of the page. It is a simplified version of what we presented last year. What it shows is that we expect more wafers to be produced at the leading edge over time. Where, of course, again, we expect to enjoy greater Entegris content per wafer for both materials and filtration. As you can see on this graph, for simplicity's sake, we are considering all memory wafers to be the leading edge, and the cutoff for logic is 14% .
Again, a very favorable trend as more production migrates to wafers where we have greater content. The graph below, I want to touch on something different, and something that is another very positive catalyst, especially for our filtration business. As you know, the desire for more computing power for applications such as AI and smartphones has also been satisfied by an increase in the physical dimensions of individual chips or larger die sizes, to use the industry lingo. That is particularly true for GPUs. With this increasing die size, you can see how much greater the pressure has become on defectivity and how much greater is the need for microcontamination.
In this illustration, what you can see is that if you have the same defect density on your wafer, your yield will go from 91% down to 63%, depending on whether you're producing smaller dies, you know, like something like a processor for a lower-tier cell phone or larger dies, say, you know, for a GPU. As you can imagine, this pressure on yield will continue to increase as further miniaturization makes wafers susceptible to even more, you know, even smaller defects. That's gonna drive the need for greater contamination control, and in turn, that's gonna drive greater filter sales and drive greater opportunities for our microcontamination division.
One point maybe that I probably want to make also on this slide, which is a reminder, but one, you know, a point of yield loss can be very costly for our customers. For reference, here on this slide, you know, we estimate that one point of yield would equal about $200 million in annual lost sales for a single logic fab. So real money at stake, and that's why, you know, yield optimization is such a focus for our customers. What I'm trying to convey here with this slide is that there are many external forces that will be playing in our favor. The other point I'm trying to convey is that what we do at Entegris is exponentially growing in importance for our customers. On the next slide, I want to touch on our customer engagement model.
On the left, you can see the journey we want our customers to experience when they engage with Entegris. I am sure that you've read Jim Collins' book Good to Great, so you will recognize the flywheel concept. A concept which is meant to call out the major steps the organization needs to go through to create customer value. It's meant to be simple, and it is, but it's also meant to create shared purpose and drive strong alignment across functions within the organization.
To do that from the very early roadmap discussions, in many cases, years ahead of the node transitions, when we start learning about the new process challenges our customers are expecting to encounter and expecting us to solve, from the ideation stage, early sample development work all the way to the high volume manufacturing ramp. As we go through these various steps, our overarching objective at Entegris is to be viewed as the most trusted technology partner by the semiconductor leaders. This flywheel is very important for us. It's more than a model for us. It's a mindset. I like to refer to it as our growth mindset. As an organization, we are always thinking about it.
We are always thinking about how to improve the customer experience at every crank of the flywheel. That means that after every major business win or business loss, we regroup as a team, and we discuss. We discuss what we have learned and what needs to be true to make the next crank easier, faster, better. You get the picture. As we are having those discussions, we are reminding ourselves that what got us to win the last time will likely not be sufficient for us to win the next time, simply because the needs and expectations of our customers will continue to increase, and we know that our customers will test the limits of our capabilities at every crank of the wheel.
We are having all of those discussions in the context of what you see on the right side of this slide, which are the three major elements of our value proposition: technology, global infrastructure, operational excellence. On the technology front, we have a very strong foothold, but we are being asked by our customers to work on more complex challenges requiring earlier involvement. We are spending more in fundamental research, and in that context, we are very focused on new investments we need to make in talent, in metrology capabilities, and more generally, new levels of funding in order to shorten our time to solution. On the global infrastructure front, our customers are increasingly expecting us to engage with their engineering teams in their home markets.
As a result, we have invested over the last few years a lot in local tech centers in Korea, in Taiwan, in Japan, and in China. We have ramped as well, to some extent, our investments in local manufacturing capacity. These choices have served us well, actually very well. We want to do more of these investments because they are proving to be a great source of differentiation for us since most of our competitors will not be able or willing to make similar investments. Last ring on this graph is the operational excellence, which is also another very important area where real differentiation can be achieved.
Our customers will continue to expect tighter process windows, more process control, greater cleanliness, greater control of our supply lines, and as a result, we will increase our investments to meet these ever-growing expectations. Remember that, you know, at Entegris, we do not sell products. We really sell the promise of a timely solution to emerging process challenges. We recognize that to live up to those expectations, we need to raise the bar and increase our investments to continue to win, to continue to increase the differentiation, and ultimately to take additional market share. Greg will provide more details and quantify all of that for you in this section. Next slide, please, Anna. A few years ago, we also came to realize that our core capabilities in material science and material security could have unique value to the biotech industry.
After a period of careful market assessment during which we validated the unique attributes of our Aramus bags, we decided in 2017 to fully fund this initiative, and we invested in a dedicated team with strong preexisting life science background. I will tell you that a few biotech industry leaders guided our development efforts, and they also encouraged us to supplement our internal capabilities with partnerships and acquisitions. That's in part what led us to the acquisition of Flex Concepts in 2018. The various initiatives, organic and inorganic, gave us a complete industry-leading solution for the cold storage and transportation of biologics. Last year, we broke through the first $10 million of revenue for the Aramus platform.
This year, we expect to reach about $50 million in revenue, and now we expect 2022 revenue to approach $100 million for that platform. As you know, today Aramus is used almost exclusively for COVID-19 vaccines, but the unique characteristics of the bag that you can see here listed on the slide make it perfectly suited for other biologics, and we're currently seeing a lot of interest across a wide range of therapies. Going forward, we expect to extend our offering with ancillary solutions around our high purity bags. In addition to this, we are in the process of expanding our filter offering for bioprocessing applications as we fully leverage the Anow acquisition that we made in 2018.
If you sum up all of these various, you know, development initiatives, we expect our life science revenue will be exceeding $300 million by 2030. On the next slide, Anna, please, you can see that, you know, beyond life sciences and more generally, we are increasing our focus on new emerging mega trends. Like in the case of biologics, we are targeting the intersection of high-value markets with unmet needs, and we are intersecting that with our core capabilities and Entegris, be it material science, contamination control, as well as our operational capabilities. To be clear, you know, we will not be wasting time on well-served applications in mature industries.
Instead, we will be focusing on promising high growth applications, and we will do that organically, but in some cases inorganically as well. For obvious reasons, I will not disclose today any specifics until we reach certain meaningful milestones, as we did with life science. What I can share with you is that the chart in the middle here is an estimate of some of our potential opportunities in various emerging applications. We expect this opportunity pipeline to continue to fill up with Aramus and other life science opportunities. As we discussed in the previous slide, we also expect new emerging growth vectors to add up to it and combine. We expect all of that to contribute about one to two points of annual growth over the next few years. Next slide, Anna, please. A slide just briefly on M&A.
You can see that, you know, generally, as a general principle, we see our M&A efforts as a way to supplement and augment our value proposition to our customers in various industries. Generally speaking, at Entegris, we believe that there needs to be a strong, and actually there is a strong correlation between new unique customer value creation and long-term shareholder value creation. This belief, frankly applies to both our organic development efforts as well as our M&A initiatives. At Entegris we believe in being patient, we believe in being disciplined, and this slide highlights on the left the areas we are targeting, and on the right, how we are thinking broadly about shareholder value creation.
Given the principles that I was mentioning, our programmatic M&A strategy is expected to contribute 1-3 points of annual growth for the years to come. My final slide, Anna, please. I think that overall my presentation would not be complete without a word on our CSR program. Last year, we introduced our four areas of focus, which were innovation, safety, personal development and inclusion, and sustainability. I explained to you at the time how we are not only trying to do the right thing, but also how we are connecting our CSR program very closely to our value proposition, our strategy, and who we are as an organization.
I also shared with you last year why we chose a 10-year horizon when casting our CSR goals, and how we expect this longer timeframe will give our teams an opportunity to break the mold and set very aggressive objectives as opposed to just settle for modest incremental change. Having said that, I am pleased to report that we've made some meaningful progress early in this journey. You can see some of that on this slide. In October, we published our first CSR report with data following the SASB framework. As displayed in the bubbles on the right, you can see some meaningful progress in a number of critical areas, in particular in our safety and inclusion goals.
These early results reinforce our belief that it is possible to have a positive impact on our world while improving our business performance and creating very real, very tangible economic value. I know you've all been sharing a lot of very important feedback to us. Please continue to do that. We'd love to hear from you as we continue our journey. This slide concludes my section, so I will now turn the stage to Greg.
Saying hello to all of you, but here we are again, one more virtual meeting. I've been looking forward to it. My presentation, I'm gonna talk a little bit about our history. We'll talk about our investment levels. We'll talk a little bit about our investment levels going forward, and we'll talk about our financial targets. First of all, on the next slide, please. Just a little bit about historical perspective. Over the last five years, we've grown the top line at a 14% CAGR. Let me unpack that a little bit. 12 percentage points of that are organic versus a market growth rate of about 8%. We've outgrown the market by about 400 basis points over the last five years. Shifting to the right-hand side of the slide and the EBITDA.
EBITDA has grown at 21%. Our operating EBITDA margin has moved from about 22% of revenue to it'll be 30% of revenue in 2021. Significant operating leverage. If you look at the flow-through from 2016 to 2021, we've been just short of that 40% target that we've talked about consistently. Earnings per share cumulative growth rate over the last several years of approximately 29%. The point of this slide, the key message here, though, is that we've delivered strong growth, we've demonstrated the leverage in the model, and that sets us up, we believe, to continue to do that going forward. Next slide, please. You know, we've really been true to our guiding principles around capital allocation.
If you look at this from top to bottom, M&A, which we've talked about a lot over the last several years, has been our number one use of capital over the last five years. At the same time, we've also made significant investments in ER&D, as our ER&D as a percentage of revenue has crept up over time, and we've continued to reinvest in the business. We've also talked a lot about our commitment to returning cash to shareholders, and over the last several years, we've returned close to $600 million in cash to shareholders. That represents about 42% of free cash flow over the last five years. In Q3, we took our programmatic buybacks up from $15 million-$20 million per quarter. That'll be our number on an ongoing basis.
Today, we're announcing that we're gonna increase our dividend 25% from $0.08 per share per quarter to $0.10 per share per quarter. Overall, consistently been balanced in that approach and stayed with the discipline that we've talked about, many times in terms of that capital allocation philosophy. Next slide, please. With that, the growth setup that we've talked about, I want to talk about our investments going forward. We are gonna invest significantly more in the business over the next three years. If you think about the growth, we're gonna grow about 20% or over 20% in 2021. We've had that 14 nanometer growth rate over the last five years.
Given the growth that we've had and the growth that we expect going forward, we're gonna increase our ER&D spending from approximately 7% of revenue to 8%-10% of revenue. Our capital spending over the next three years will increase from about 7% of revenue to 15% of revenue. That's driven largely by our investment in the Taiwan facility, where we'll be investing about $500 million over the next three years. Excluding that investment, our capital spending will be about 9% of revenue. As we come out of this significant investment cycle, we'd expect the CapEx to be in that high single digits once we get past the next three-year phase. As we do this, our depreciation will increase from about 4% of sales to 5%-6% of sales.
Our ROIC over the period will be relatively consistent with the 20% we'll achieve in 2021. It'll dip a little bit in the next couple of years, but by the time we get to year three, we'd expect it to be back to that 20% level. Just a little bit on the investment in Taiwan. That $500 million investment will generate about $500 million in revenue. The IRR on that investment is in the mid-20s. The main tenant in that KSP facility will be our micro contamination business, which happens to be our highest margin business. The returns on this investment are excellent and, you know, we're very pleased to be making this significant investment in close proximity to many of the technological advances in the industry and near our largest customer.
In addition to that investment, over the next several years, we'll be making investments in our liquid filtration, both from a capacity and a capability standpoint. We're investing significantly in our global applications labs, particularly in Taiwan and Korea. We'll be investing in a meaningful way in our SCEM business. Much of that growth capacity driven. A lot of the investment in that business is in what we call the fleet, which is or the packaging that we use to ship the product to our customers and that our customers use while our product is on site. All of this is gonna allow us to better serve our customers as well as widen the gap relative to our competition. Next slide, please. From a capital structure perspective, I love our balance sheet right now.
It's both conservative and it's flexible. It gives us significant optionality. Today, we've got less than $1 billion of debt. We're operating below our gross leverage targets and our net leverage targets. 85% of the debt that we have is unsecured, which gives us significant secured borrowing capacity. We have no meaningful maturities for seven years, so very flexible balance sheet. We believe this gives us significant firepower, both for tuck-in M&A as well as transformational M&A. Next slide, please. Turning to the income statement, I've got a couple of slides here before I talk about the consolidated outlook where, I'll set this all up. First of all, when you think about the divisions, our fastest-growing division has been and will continue to be our MC business. Over the last five years, that division has grown 19% on a CAGR basis.
Going forward, like I said, we expect significant growth from this division. That's gonna come in more from the point of view of liquid filtration and purification, bulk gas purification and filtration, as well as new markets with the acquisition that we made in China, about two years ago now. From an operating margin perspective, we expect 34%-36% operating margins from this division. That's in spite of the higher depreciation levels within this division. Recall, they will bear much of the load for the KSP facility from an operating margin perspective. I've talked about the MC business as the crown jewel of the portfolio, really since I became affiliated with it 15 years ago with the Mykrolis merger. It continues to be a very strong business with a bright outlook. The SCEM business won't be far behind in terms of growth.
They've got significant opportunity in deposition materials, advanced coatings, selective etch chemistries. We've talked about the recent acquisitions that give them capability within hard substrates, which equates to power electronics, which is a rapidly growing part of the semiconductor industry. Again, the prospects for growth within SCEM are significant. We expect the operating margin in that business to be 25%-27%. They're operating well below that today, but as they increase the volumes in this business, and also some of their newer products are among their newer products and their highest growth products are among their highest margin products. We expect to see a benefit in that business from improving mix over time.
Finally, you know, last but not least, the AMH business, which has grown very rapidly in 2021 as a function of both the capital exposure of that division, but also the Aramus bag that Bertrand talked about previously. They've got growth opportunities in chemical packaging as sort of the purity of chemicals that our customers are buying both from us as well as our competitors. You know, continues to march up in terms of the expectation. The chemical packaging side of that division will see meaningful growth. They've also got opportunities in EUV. From an operating margin perspective, we expect them to be in that 21%-23% range. They've operated just below that the last several quarters.
Overall, this is while it's our most mature business, we feel very good about the prospects of the AMH business going forward. The next slide, please. Let's talk about the target model. The target model has been a core part of our operating philosophy really since we came out of the economic downturn in 2009. What I want to talk about here first is what's not changed about the model. What hasn't changed is our commitment to deliver on that 40% incremental flow through as we continue to grow the business. This has been consistent with our principle for probably the last half a dozen years.
If you look at where we're operating today, starting on the left-hand side of the model, we'll be a little bit over $2.2 billion this year, and we'll deliver on an operating margin somewhere around 30%. As you move to the right, you get all the way out to the far right, $3.4 billion. We'd expect operating EBITDA margins in that 34% range. What's not changed is the flow-through. There's a little bit less leverage in the EPS than we've had historically. Part of that is we've got a slightly higher tax rate assumption in this model. It's built at 18.5%, where we've operated lower than that the last several years.
We've also had the advantage of lowering our interest costs over the last several years, and we think we're probably about as far as we can go on that for right now. The point I want you to take away from this is to expect the same financial discipline that we've demonstrated as a team over a long period of time. When I say a team, I don't just mean the senior team, but this is part of the fabric of the company. The divisions operate to their own target models, and it will continue to be a tool that we use to drive financial discipline in the company. We've talked about how the divisions are gonna grow. We've talked about the target model.
Let's move forward and talk about when you roll it all up, what does that mean from a growth perspective as well as an earnings perspective. This is our illustrative growth model. We'll end 2021 at about a little over $2.2 billion in revenue with an EBITDA margin of about 30% and EPS in excess of $3.30, ROIC of approximately 20%. We talked about that commitment to top line growth of 15% over the next three years. That brings us to a revenue level of $3.4 billion by 2024.
The leverage in the model, that 40% flow-through, takes us to a 400 basis point improvement in that EBITDA margin to 34% and gives us an earnings per share number of just over, of in excess of $5.30 in, by the year 2024. Strong growth characteristics, strong operating leverage, consistent with what we've demonstrated in the past, results in a very strong financial profile over the next three years. Let me talk a little bit about what that means in the near term. Next slide, please, and talk about our 2022 preliminary expectations. As a backdrop, I mean, we're obviously optimistic about the market as well as our own opportunities to compete in that market, and, you know, believe that the trends in the industry continue to benefit our key thesis around advanced materials and contamination control.
Strong market, well-positioned in that market. We currently expect sales growth of greater than 15% for 2022. I just want to highlight that so that nobody walks away saying, "Well, they talked about 15% over a three-year horizon." There's no hockey stick to that. I mean, we expect to be above that number in 2022. You know, we continue to monitor the supply chain and manage proactively against that. Our fundamental demand, as Bertrand talked about, is probably, is in excess of that 15%, but we're talking about something in that 15% range or slightly more in 2022. You can expect that 2022, you could take that 15% growth and look at the target model, and that, and that's how we would like you to think about earnings in 2022.
Our plan is to provide more detailed guidance around 2022 when we announce our Q4 results at the end of January, beginning of February next year. With that, let's move to the final slide. Really, I just wanted to close with, you know, the seven reasons to own Entegris. I mean, first of all, we've demonstrated that we're a value compounder and that we've, you know, got significant growth, able to generate significant growth and able to generate significant operating leverage. We feel very good about the industry trends. You've got continued advances in technology and all that equates to broader demand for the semiconductor industry. The industry today feels very good to us. We've got accelerating exposure to many of the key inflection points. Bertrand talked about that.
As you move to finer and finer feature sizes, it benefits the filtration business. It benefits the advanced materials parts of our business. We've got very strong competitive moats. That's both from an IP portfolio perspective in the form of patents as well as know-how, but it's also our applications expertise. I think one of what separates us from many of our competitors is the breadth of our product lines and our knowledge across the fab, which makes us a very strong partner for our customers in terms of providing that applications expertise. Now, clearly, the 70% unit-driven piece of the business provides some resilience to our model. We've generated strong operating cash flow. We believe we'll continue to generate strong operating cash flow.
That cash flow will allow us to continue to do M&A, make significant investments in the business at the same time, and return cash to shareholders. We've demonstrated our confidence in that by the increases in our ongoing buyback program, as well as the increase in our dividend. Then, you know, Bertrand talked a little bit about our life sciences business and what we're doing in other sectors, and that gives us some optionality. What's different today is, A, I mean, we've arrived in life sciences. I mean, we've put up some strong numbers in Aramus, and we feel so good about the growth prospects of that product line going forward. We're also investing in the process of identifying additional markets, and that's something new.
I mean, we're gonna have a team that is focused on those opportunities where there are significant technical shifts or significant shifts in end market exposure where we can play. Finally, I mean, we're disciplined capital allocators. At one level, I like to think of what we do as a senior team is very similar to what you do. We're managing a portfolio of investment opportunities on both the ER&D side as well as the capital side. We're weighing that against M&A and returning cash to shareholders. I'd like to think that over the last several years, we've done a very good job of, I'll call it, investing and allocating capital. I think with our discipline, you can expect more of that same approach as we move forward.
With that, I'd like to wrap up, and we'll open it up for your questions. Thanks for your time today.
All right. Thank you for sending questions both to the Q&A tool on the webcast, plus I got a bunch over email, so we'll get started here. The first question is from Sidney Ho from Deutsche." 2020 was a strong year for node transitions in both foundry and logic and memory. In the next few years, do you expect that cadence to continue?' And then, he adds, "Do you expect the content per wafer to increase?" I think we sort of answered that question, but Bertrand, go ahead.
Sure. Yes. I think we've been answering that question multiple times over the last, or the most recent earnings call. I think that the industry backdrop continues to be very sound, very exciting, both in terms of the level of production in the fabs, but also the pace at which our customers are migrating to new nodes. That's true in logic. That's increasingly true in 3D NAND. And as your the question implied, we are indeed expecting greater content per wafer, as our customers progress from one node to the next. It's a very compelling combination of favorable factors for us.
All right. Sidney might not know it, but he has a follow-up question. This is probably Greg. "Can you talk about the financial impact of your recently closed acquisition of BASF's Precision Microchemicals business and the strategic rationale of the investment?"
Yeah. That business, which we closed on actually yesterday, from a financial perspective, it in order of magnitude, think of it as about $15 million in revenue. It will be accretive to earnings. It's got operating margins that are consistent with or slightly above our current corporate average. From a financial perspective, it fits in nicely. From a strategic perspective, I think I mentioned when I was talking briefly about the SCEM business, the opportunities that they have in polishing for hard substrates, which equates to power electronics being the most significant market for that. Strategically, it's a materials platform. It adds to the capabilities that we had when we bought the Sinmat ac or did the Sinmat acquisition in early 2020.
We feel the combination of those two businesses position us very well in polishing for hard substrates.
All right. Next question, Toshiya Hari from Goldman Sachs. Bertrand," the 15% revenue CAGR you were guiding to is significantly above what you presented last year. Can you break down the reasons for that? You know, A, overall semiconductor industry dynamics, B, semiconductor dynamics that are specific to us, Entegris, and three, contributions from non-semi."
Right. I think the overall industry dynamic remains the same. I mean, we've been consistently saying that, we are using a 2x GDP growth rate as the underlying growth assumption for the industry. What is changing is the level of outperformance we expect to be able to deliver, and that's directly a function of the opportunity pipeline, and that's a function of the greater content per wafer that numbers that we've been sharing with you today. It's the adoption of new materials becoming increasingly critical in the next generation of chips, and that's the further miniaturization of those architectures which make them more prone to contaminants, and in turn translates into the greater importance of our micro contamination solution set.
If you combine all of that, you know, we saw that we are increasing the level of outperformance to a range of 3%-5% now. We also have growing conviction in our emerging growth vectors, life science, obviously. We are aggressively working on supplementing that pipeline as well with a number of other opportunities that are still a bit too early for us to go too deep into that, but also very, very promising. As I said at the beginning of my presentation, the growth objective we are laying out here for you is really a function of the optionality that we have in our growth strategy and the quality of the pipeline.
As Greg said, it's really about managing a portfolio of opportunity, be it organic opportunities in semi, organic opportunities in adjacent emerging applications, or a programmatic M&A pipeline.
Okay. As a follow-on to that, "as you pursue multiple emerging growth vectors, should we expect to see an increase in OpEx as a percentage of sales as you build out your R&D marketing and distribution capabilities? And what financial criteria are you looking for when considering potential investment areas, and how will you balance the need to invest for growth versus near-term profitability?"
Right. All of the investments required to appropriately support these new emerging growth vectors are included into the models that Greg presented, both, you know, the target models, so the P&L, as well as, you know, the CapEx investments. You should know that in most cases, those emerging opportunities start as fairly modest investments. They start as derivative development efforts of existing platforms. As we learn more, and usually it's under strong sponsorship from a customer, we reach greater levels of conviction, and then we decide, you know, to potentially invest more. Again, very disciplined approach, fully reflected into the models that Greg presented to you today.
Okay, next question, Patrick Ho from Stifel. He says, "I clearly see the growth opportunities in semis and markets like life sciences and the need for investments. At the same time, Entegris has shown a very good track record of M&A, but for both small and large. How does the company determine internal investments versus potential M&A to accelerate growth in your various businesses? Is semis business naturally more internal investments and things like life science, does it require, more M&A?
Maybe I can take that and maybe use that in fact to complete my answer. I think I missed a component of your previous question, which is how are we thinking about return. We're thinking about that in the context of we want those emerging opportunities to be additive from a growth standpoint, to be additive from a bottom-line standpoint. Usually that's the reason why we are looking for emerging applications, industries that we would expect will be growing much faster than the semiconductor industry, and where the you know the promises of return could be as good or greater than what would we be able to deliver in our own core semiconductor applications.
That's why we have set up a team going through very careful stage gates, asking very simple questions about, "Is the opportunity real? Can we win? And if we win, is it worth it?" We can commit to you that we will have the same discipline looking at those opportunities as we would have looking at the core semiconductor opportunities.
Okay, a follow-up from Patrick. "Given your increasing collaborative efforts with your semiconductor customers and their needs for both materials engineering and purity control, how in-depth are you now in their future roadmaps, and how does this provide Entegris with growing lead times and visibility? And then as part of that, are your customers giving you multiple node outlooks that could provide more stickiness going forward?"
We have been collaborating very closely for many years now, and it is true that that level of close partnership gives us great insight, which is helpful during the technology development phase. As those solutions become increasingly critical to the manufacturing processes that we are serving, we are getting greater visibility into you know the forecast of those customers. So I think that you know visibility, having said that, is never perfect. I think we all know that it is difficult to forecast future demand in this industry. But I would say that on balance, we have indeed probably better visibility today than we had five or ten years ago. I think that's a reflection of the closer collaboration as well as the growing criticality of what we do for our customers.
Okay, next question from Mike Harrison from Seaport." In your slide on emerging growth opportunities, you refer to informed bets. How do you think about risk versus return in these emerging opportunities, and how will you pursue something that would require a large internal capital outlay or prefer opportunities with low fixed costs?"
As I said, you know, the returns usually it's about being growth enabling. It's being also additive from a bottom line standpoint. In other words, having a margin profile that is as good or better than what we would deliver in our core semiconductor market. When it comes to the level of CapEx, I would say it depends. You know, different products in our portfolio require different level of CapEx investments. It's a hard question to answer. I think, again, once we cross that threshold, and when those new opportunities do require meaningful CapEx, we're gonna start calling them out. You have full visibility around the choices that we're making, like what we did with life sciences when we decided to invest $30 million this year.
Again, I think as we have greater conviction, as we make bigger investments in those emerging opportunities, we will give you more insight and more visibility into that.
Okay. Follow up from Mike." Can you give more detail on the Taiwan expansion, including what gaps you're filling and your current MC offering in Asia, as well as the timing of the ramp towards the $500 million annual revenue amount for the facility?"
Think about 2022 being really buildup phase, with the first tools coming in at the end, second half of 2022. Customer qualification starting at the end of 2022, going into 2023. I would expect the first article sold sometime in Q2 of 2023 when it comes to liquid filtration in particular. We're not really filling gaps per se, we are just building for the future. We know that, you know, 3 nanometer, 2 nanometer nodes and beyond will require much tighter filtration solutions, which inherently will have to be using much more capable manufacturing processes. That's what we're investing into. We're investing into capacity for sure, but we are investing into better capabilities, manufacturing process capabilities to make those filters.
That's why you've heard Greg and I talk about those investments as an opportunity to put some more distance between us and our competitors as we advance the technology and we advance our manufacturing process technology as well.
All right. A couple questions. More modeling questions from Josh Silverstein from Wolfe Research. "Greg, the illustrative 2024 outlook assumes margins getting to 34% versus 30%. Does that mean the bulk of the revenue growth comes from MC versus SCEM and AMH?"
No. I mean, I think MC will be our fastest growing division. As I said in my prepared remarks, SCEM won't be far behind. I also said, I mean, you have an improving profitability trend in SCEM. I mean, obviously our most profitable business is that MC business. So the fact that those two businesses are our fastest growers ultimately helps from a mix perspective, but I don't want to lead you to believe that the whole thing is driven by mix. It's really, I mean, if you think about it at a fundamental level, we're gonna increase our ER&D spending as a percentage of revenue. We expect some gross margin expansion as we move through the period. We're laser focused on managing SG&A.
The leverage is gonna come from, like I said, through SG&A and in improving gross margins. I just want to comment, I mean, the gross margins that we're operating at today are not the natural margin of the business. I mean, they reflect, you know, costs related to COVID. They reflect the supply chain challenges that we face. Those would be my comments on the P&L and the model.
Okay. Another sort of modeling question. On the EPS outlook, are you assuming any share buybacks versus the $20 million per quarter that we started in Q3 2021?
We're assuming the natural level of buybacks, the programmatic buybacks, but we're not assuming any opportunistic buybacks. If you look over the last five years, we did do a meaningful amount of opportunistic buybacks when we saw weakness in the stock.
Next question comes from Chris Kapsch from Loop Capital. You frame potential growth beyond core semiconductor industry focus and pipeline M&A as potentially delivering 100-200 basis points of top line growth in compelling adjacent markets. Bertrand mentioned hiring dedicated life science team with relevant experience, but can you elaborate how much organizational capability and commercial bandwidth do you have in these target areas, i.e. life sciences, and can you broaden this capability via solely organic investment?
I think I kind of touched a little bit on that, but more specifically in terms of the life science organization, we have made the investment organically to support the high purity Aramus bag. It's in place, and it's actually delivering great results and performance. We acquired the capability of Anow with their, you know, medical and bioprocess filtration platforms. They have some commercial capabilities. It won't be sufficient to fully unlock the value that we see in that platform, and we will add to it, and as I said, it's included into the numbers that Greg mentioned. Right now we are in the process of stabilizing and improving the manufacturing processes in China. Anow is based in Hangzhou, China.
We are also very focused on going through the regulatory process for many of their products, and the time will come, you know, soon for us to start commercializing some of those products at a global level. Steady pace and a lot of discipline around stage gate decisions, both in terms of technology development and commercial strategies. We always consider, you know, options between direct sales versus partnering. We're looking at all of that fairly holistically and fairly broadly.
Hey, Bill.
Yes.
Just two comments. One is your question on buybacks and share count in the target model. We will continue that $80 million annualized amount that we're talking about. A portion of that is accounted for in the target model because it offsets dilution, but the target model is predicated on 137 million shares flat over time. At an $80 million clip, that share count should come down a bit over time, and that's not accounted for in the target model. Then a question that I'm seeing on my screen from Christian Schwab that I'll read is, "does the more wafers at the leading-edge graph account for the fact that leading edge changes over time?
As in, how much of that chart is a result of natural maturation of processes versus a structural shift toward more leading-edge mix overall? " Do you want me to take that, Bertrand?
Okay. I thought you wanted to take that one.
Oh, you can go ahead and answer it. Or, I can.
I mean, I think, look, it's a function of you know, the natural push towards the leading edge. This is something that has been happening very much you know, on an ongoing basis, and we're just trying to put a fine number to it. I would expect, frankly, the transition to the leading edge to accelerate in logic in particular. We know that the capacity in mainstream fabs is limited. There is certainly some level of new investment, but it's gonna be limited. I would expect you know, that to be a driver for more production to be moved to the leading edge.
Another driver is that if you look at the very nature of some of the new demand drivers for chips, most of those applications, especially when you start thinking about EV, when you start thinking about autonomous driving, I would expect, you know, more wafers to benefit from greater performance, to benefit from the greater energy efficiency that you get at the leading edge. So again, I think it's a natural transition, but I think that natural transition will actually accelerate.
That line, though, on the graph, Christian, that's 14 nanometer and below chips for logic.
Yeah, that was just for simplicity's sake. I mean, we recognize.
Yeah, right.
That, you know, what is leading edge today will become trailing edge. We wanted to try to give you—I mean, since you have the wafer park count, I mean, the content per wafer number at 14 nanometer, we wanted you to have that as a yardstick, if you want, and see how much, you know, the production will be moving at 14 or higher levels.
Okay, follow-up from Chris Kapsch. From an inorganic, I know we've answered this in some ways, but in an inorganic target standpoint, what will be the criteria for acquiring greater scope and scale and additional skill sets to build out this broader presence and relevance in these contemplated adjacent markets for trial?
If I understand the question well, is there an inorganic element to your emerging growth strategy? The answer is yes, but, and I think that the way I would qualify that is, you know, generally speaking, at Entegris, we like to fire calibrated bullets to validate our assumptions, to validate that, as I was mentioning, the opportunity is real, that we can win, that it's gonna be worth it. Don't expect us to make any meaningful, you know, sizable M&A in an area that we don't know or don't understand. Expect us to learn first, and we're gonna learn through either internal development work with close collaboration with customers in those emerging applications or and/or making small acquisitions to just validate our hypothesis.
I mean, this is what we did with Flex Concepts. For life science, this is what we did with Anow for filtration in bioprocess applications. This is what we are doing closer to home with Sinmat and Precision Microchemicals for hard substrates. Again, it's really trying to validate that there's something there, and once it's validated and once we have conviction, then maybe we can get to that next phase. Expect us to be disciplined about that.
All right. Next question, Kieran de Brun from Mizuho." How do you think about your current manufacturing and R&D footprint, and how that positions you to benefit from new customer investments in Asia, Europe and North America?"
As I was mentioning in my presentation, we are trying to move the whole flywheel concept closer to the industry technology leaders. We believe that the best way to accelerate the learning cycles is to collaborate as early as you can with those customers through the investment in tech centers that we've made, to be able to go to pilot manufacturing and ultimately high volume manufacturing close to those increasingly important customers. As I was mentioning in a different question earlier, our customers are not, I mean, only trying to get into those more challenging nodes, they're trying to do that faster than before.
This whole learning cycle, this whole execution cycle has to be as effective and as short as possible, and you best accomplish that when you can do actually both the development, early stage pilot manufacturing and high volume manufacturing closer to the customer. The good news is that we are capacity constrained. Well, it's not a good news, but I mean, you've heard us talk about the limiting aspect of it in 2021. But as we were thinking and as we are thinking about adding new capacity, we have the luxury of being able to almost think about that on a white sheet of paper. That's why we're investing in Taiwan.
That's why we are, you know, making, and we've been making investment in in Japan and in Korea as well, to be able to to fully engage with our customers. As we were saying, both Greg and I, we believe that it's gonna be the source of great differentiation because few of our competitors will be able and willing to make the same types of investments.
Okay. Next question from Timothy Arcuri from UBS." These company-specific drivers are clear, but what is the threshold for semiconductor revenue and/or CapEx and wafer starts, aka, MSI, for you to still be able to grow year over year?"
Well, as I said, again, we're not providing a very precise guidance. What we're saying is that over the next several years, we expect the industry to be growing at about twice the rate of GDP. That's a core element of that growth formula. That's again, as I said, it's a core assumption in the model. If the industry is growing meaningfully below that, you know, it's gonna be hard for us to offset it, even though, you know, I could see a scenario where some of those other growth engines or cylinders could actually contribute maybe a little bit more than what we have on the side. I think that's why we like talking about it as having optionality.
That's why we believe that we have great resilience in our model. Don't get me wrong, we're not immune to the industry, the semiconductor industry. You know, 90% of what we do ties to the semiconductor industry. While we expect to be more resilient, we would certainly be impacted if the industry doesn't grow at twice the rate of GDP.
All right. Next question from David Silver from C.L. King." Based on our breakout of logic versus memory to fab customers, does that need to change in our growth forecast? As a part of that, obviously, does the memory, the share of memory, need to change, or will it change over time?"
I think it will change over time. I think what we're facing is an interesting conundrum is that memory has been growing very, very fast, and you've seen that expansion on a relative basis. In the next few years, as the logic roadmap transitions to 2 nanometer Gate-All-Around, I would expect a new inflection in our growth in logic. I think that, yes, memory will continue to expand, but I think it will be probably if you fast-forward five years from now, it probably could be 40%+ of our revenue. I think it may not be 50% as we thought originally.
The reason for that is simply that we are seeing actually new opportunities in logic, and more opportunities in logic than we thought there would be, you know, three years ago. It's a good problem to have.
Okay, one final question. Aaron Wasserman. He's quoting or referencing what we said. "Most of our competitors will not be asked or willing to make these types of investments." That's really kind of, I think, CapEx and R&D. Can he shed more light on what he means here? Is he referring to the investment capacity of Entegris competitors and what he means by that? This will be the final question.
Could you reframe the question, Bill?
Yeah. I think that the comment we made during the presentation, or you made, was most of our competitors will not be asked or willing to make these types of investments. I think we're talking about, you know, the broader CapEx and R&D investments. And he's asking if you can shed more light on what you mean by that.
Right. I think many of you are familiar with the competitive landscape. In broad terms, we compete with small regional companies that are not well capitalized and don't have the global footprint, or we compete with divisions of large chemical companies that do not have the same focus and the same intensity on the semiconductor industry as we do, and may or may not have access to the, you know, the capital that would be required to have as aggressive a strategy as what we are proposing here. I think in both cases, we believe that we're gonna be able to do better. I think, you know, we have intense purpose and intense focus and understanding of the semiconductor industry.
I think we have the scale to be able to make these investments that ultimately will translate into greater differentiation. So that's what I meant, and hopefully that answers the question.
Excellent. That was our final question. Thank you for joining today. I know there was a lot of questions or several we didn't get to, so please reach out to me with those, and I'll get back to you right away. This ends our webcast. Thank you for joining us, and have a great rest of the day. Thank you.
Thank you very much. Bye-bye.
Thank you all.