I think we're going to get started now. My name is Steve Cantor, I am the Vice President of Corporate Relations for Entegris, and I'd like to welcome everybody to the Entegris 2015 Analyst Meeting. Before we begin, I just want to run through a few things. The first thing is our agenda. We have four speakers today, and I think the prepared presentation should take about an hour. First we have Bertrand Loy, our CEO and President. We have Todd Edlund, SVP and Chief Operating Officer, Jim O'Neill, our Chief Technology Officer, and Greg Graves, our CFO. I'd like to ask you all to hold your questions until the end of the presentations. We should have enough time to cover everything. The second thing I'd like to cover too is, we may be making forward-looking statements today.
As such, we encourage you to read our disclosures with the SEC very carefully to understand all the risks and uncertainties regarding those statements. Before I turn it over to Bertrand, I do want to let everyone know in the room that we are having a reception over at the W from 5:00 P.M. to 8:00 P.M. this evening. You should have cards in front of you that will give you the details. If you're interested, we do have some product over there and some refreshments and hope to see you there if you have time. With that, I'd like to turn it over to Bertrand.
Thank you. Good afternoon, everyone. Thank you for joining us here at SEMICON West today. It's been a little over a year since we completed the acquisition of ATMI. We thought that the best way to spend our time together today would be to reflect on our recent performance, the performance of the last 12 months, and to share our views about the future and the potential of the new Entegris platform that we have created. Over the course of our respective presentations, Todd, Jim, Greg, and I will be touching on all of those major themes. The ATMI acquisition has been a resounding success. It is creating significant value for all stakeholders, our investors, our customers, and our employees. We are optimistic about the prospects of the semiconductor industry.
We believe that the Internet of Things and advances in process technology will be good for the industry and will be good for Entegris. As a result, our SAM will expand. We have a very unique value proposition to offer, a value proposition based on a broad technology portfolio, deep application knowledge, and long-standing customer relationships. We will leverage all of those capabilities to gain additional market share. We have a stable business model. We are not an equipment company. Most of what we do is unit driven. Finally, I have the privilege of leading a very talented team. I am proud of our operational and financial discipline. I expect significant expansion of our EPS for the many years to come. Entegris has changed a lot in the recent past. I thought that a few key numbers could be helpful. Those are not estimates.
They are not projections. These numbers reflect what we have accomplished thus far. They are trailing 12 months metrics. As you can see, the top line of Entegris now is exceeding $1 billion. The vast majority of our revenue, 77% to be precise, is unit driven and recurring in nature, which provides stability to our business model and stability to our financial performance. We generate very healthy levels of profits. Our EBITDA margin of 22% compares very favorably against our peers in the semiconductor space, but also against more diversified specialty chemical companies. Finally, the last 12 months have provided you with an opportunity to get a glimpse at the earnings power of the new platform, and Greg will share with you details around our pathway to continue to expand our EPS going forward.
A year ago, early 2014, Greg and I met with many of you, and we shared our excitement about the impending, at the time, acquisition of ATMI. A transaction that we characterized as very compelling, both financially and strategically. I'm here to say that I'm very pleased with how things have turned out. If you look at this very crude scorecard, I would say that the integration is now complete. More specifically, our sales and application teams as well as our technical platforms are fully integrated and fully aligned. Todd and Jim will describe to you why those two platforms are so complementary and how Entegris is now in a position to provide unique value to its customers. $30 million of merger synergies, that was our target. Those $30 million of synergies are in place as of now, and that is six months ahead of schedule.
Furthermore, we spend less than originally planned to realize the synergies, we're feeling good about that as well. The acquisition of ATMI did provide the opportunity to leverage the balance sheet, and in the process, we lower our cost of capital. We are starting to unlock EPS growth, I would expect that to continue as we continue to grow, as we start to capture the full benefits of the merger synergies, and as we delever the balance sheet. Having said all of that, I would say that beyond the short-term financial success, beyond the larger scale that we have created, what is most exciting to this leadership team is the opportunity and the desire to create a better platform. One of the major conditions to our future success will be the ability to maintain our growth momentum.
Let's talk about growth for the next few slides. As you know, 80% of what we do ties to the semiconductor manufacturing process, and more specifically to wafer starts. This chart illustrates wafer start growth, we are using a proxy for that. That's the MSI index or the million of square inches of silicon produced. That's the red line, we compare that to the amount of capital that the semiconductor industry has been spending in wafer fab equipment, that's the blue line. To make it easier to read, we have normalized the data. We are indexing it to the activity level of 2001. Let me make a few comments and observations around all of that. First of all, again, we are not an equipment company.
Most of what we do is unit-driven, our business is really more closely aligned to the red line. Certainly, there is some volatility in our business, as you can see, but it's nowhere near as dramatic as what pure equipment companies are experiencing. You can see that in the difference in amplitude of the cycles. Lastly, and most importantly, we participate, Entegris does participate in the one segment of the semiconductor industry that has been and is expected to grow the fastest. It's also fair to say that the drivers behind IC demand have been fairly hard to predict, diverse, erratic at times in the past 15 years. We all remember the years when it was as simple as trying to estimate the new timing of the Windows product release cycles and try to understand how it would impact the demand for new PCs.
Our attention started shifting to the rate of penetration of new form factors, notebooks first, then tablets, and most recently, smartphones. I'm here to tell you that going forward, I am convinced that the Internet of Things will have a very significant and lasting impact on the semiconductor industry. Cisco estimates that today, 10 billion devices are connected to a network in one way or another. Cisco further estimates that by 2020, 40 billion to 50 billion devices will be connected. We are entering an era that will actually reshape our lives, our daily lives, an era where many more devices will be connected, and that will require better network performance. It will require more powerful computing capabilities, which in turn will be demanding ever larger and more energy-efficient server farms. Yes, it will also most likely require less obtrusive and more user-friendly end devices.
In other words, hyperconnectivity and big data will be driving IC demand, which in turn will be driving wafer starts and ultimately will benefit the unit-driven business model of Entegris. I can already hear some of the skeptics in the back of the room whispering, "Yes, but." Yes, the Internet of Things will have a big impact, positive impact on IC demand. Many of those devices will likely be produced on trailing-edge technology. My answer to that is, first of all, who knows? More importantly, it doesn't matter. For Entegris, it really doesn't matter. For those of you who are following Entegris very closely, and those of you who understand our business model well, you know that two-third of our semiconductor business comes from products and solutions that are used and consumed daily in trailing-edge fabs, where we have very strong market share.
It's true that we like to talk about the leading edge, today will be no different. We'll talk a lot about the leading edge. We don't do that to confuse you. We do that because every node transition is a unique opportunity for Entegris to recast its value proposition. What it means, it's a unique opportunity to increase our SAM and increase our market share. Those windows are narrow, but once we lock our solutions into the fab recipe, our solutions are sticky, and we can enjoy many years of steady recurring revenues and cash flows. As you can tell from this slide, fabs have long lives.
That's why we spend so much time, so many resources, and so much R&D money on the leading edge, is really to make sure that we actually benefit from our successes for many years to come. The punchline of all of that is that I think that the Internet of Things will benefit trailing edge fabs. I also believe it will benefit leading-edge fabs, and in both cases, Entegris will benefit from this trend. Another very important aspect of our business model is the fact that we sell solutions across the ecosystem. We sell a broad array of solutions to fab customers, of course, but we also sell high-value components and subsystems to OEM customers. We also sell filtration and packaging solutions, to name only a few, to materials suppliers.
Jim will tell you why this positions Entegris uniquely to develop defect reduction and yield-enabling solutions for our customers. You see, contamination control, safety, cleanliness, stability of process chemistries are increasingly important considerations for this industry. They present very complex challenges that are very pervasive across the ecosystem and across the supply chain. Let me tell you, nobody is better equipped than Entegris to understand those challenges and to provide solutions. From a financial standpoint, the value of our position in the ecosystem is really around the very well-balanced and diversified customer profile that you can see on this slide. How do we win in the semiconductor ecosystem? We believe that to win, you need three distinct attributes. You need a very broad technology platform. You need deep and real global capabilities.
By that, I mean talent, I mean application knowledge, I mean technical labs. You need a very strong, ongoing commitment to operational excellence. Todd Edlund will dive into all of those considerations in great details, and you will appreciate that, number 1, Entegris has all of those three attributes. That we will, number 2, we will be putting them to good use to create unparalleled value for our customers. In doing so, continue to expand our market share over time. If I want to summarize my previous four or five slides, which were focused on our core semiconductor markets, I would say that for the many years to come, we expect a favorable business environment in which Entegris will be doing very well. In addition to that, we will be supplementing our growth strategy with a sharper focus on adjacent markets.
We are constantly looking for ways to leverage Entegris platform in new areas. We are looking, for example, for new applications that require cleaner, better, more performance materials. We are looking for new process technologies that are more susceptible to complex contamination challenges. Today, those non-semi applications represent about 20% of our top line, and I would expect this ratio to continue to expand over time. In closing, I would like to share with you our short term to midterm priorities. I will do that at very high levels since Greg will be going in more details in this section. Let's start with the financial objectives, which are pretty straightforward and are consistent with what we did show you about a year ago. We want to continue to grow, and we want to grow faster than our underlying industry by about 100 to 200 basis points.
We want to continue to deliver very healthy profit levels in accordance to our target model. We will continue to generate strong cash flows that we will use in the short term to pay down our debt. We will continue to invest to be better at what we do and to earn the right to be viewed by our customers as their indispensable strategic supplier, and that's something we will never take for granted. Lastly, I would say that we are very proud of the quality of our execution, and you have the commitment of this team that we will continue to run our business with rigor and with discipline. With that, thank you, and I'll turn the mic to you, Greg.
Okay. I turn it over to Todd just to really kind of summarize the key things that I take away from Bertrand's presentation. A, is this ATMI acquisition has gone really well. You can index our stock price against almost anything since we announced it, and it shows up in the stock price, and it shows up in the numbers I'll talk about. B, we like the industry the Internet of Things, we think, will continue to drive unit volume, big data, hyperconnectivity. Our underlying industry dynamics are good, and I think specifically, we like the way we're positioned in that industry because, A, we're unit-driven, and B, we sell to everyone and we're beholden to none. What's important about those two points is it makes our model more stable.
When you look at specialty chemical companies or industrial tech companies like filtration companies that command much higher multiples than we do, it's about the stability. We think within semiconductor, we're as stable as one can be. With that, I'm going to introduce Todd Edlund. Todd is our Chief Operating Officer. He's been in that role for about six months. In that role, he's responsible for all of our business units, both the CMH and the electronic materials. With that, Todd.
Thanks, Greg. Good afternoon. Thanks for coming. I'm going to talk a little bit, take a little bit of a different slant on what Bertrand talked about. He talked about the integration of the companies has gone well. It's a great combination. I look at it from a perspective of, does it make us a better partner to our customers? Does it enable us to do things now that we couldn't do before? I'm pleased to say that that's what I'm seeing from my chair, which is really looking at R&D, product development, strategic direction. We're starting to really see the opportunities. I'll describe a little bit of that to you, Jim O'Neill will follow up with some more details about how the things that are going on in the industry affect us. Well, we're not advancing. Oh, there we go. Okay.
Just make sure we're caught up. Bertrand talked about this, three pillars of success. These are the things that we need to do to address our customers' needs. The portfolio is really about, are we able to help customers get to 16, 14 nanometer, 10 nanometer, seven nanometer? That's the portfolio. Do we have the solutions that they need when they need them? Be caught up, be ready, be able to be a partner to them as they advance their state-of-the-art. We can't just do that. We also have to do it because, as you saw, we serve the global ecosystem of semiconductor. We have to do it globally, and what I mean by that is locally. We have to be able to be in Hsinchu with the same level of talent that we have when we're dealing with North America.
We have to be in Korea with the same level of talent that we have in Hsinchu, Japan, same thing. I'll talk a little bit more about that, being able to be local is the only way that you get invited into the fab to work on a daily basis, or that we can get the customers to come work with us in our facilities on a daily basis, and we've been able to do that. Further, there was a time when that was good enough. You got a good product, you got the right price, we'll buy it.
That's no longer the case, because the demands for yield and reproducibility of our customers' processes are driven back to us to make sure that we've got supply chain manufacturing capabilities, quality systems that are going to ensure continuity of supply, safety of supply, and also be able to drive down costs, because we need to do that, and we need to help our customers do that, and we actually do that very well. All three of these are really essential to being an enabled, trusted partner in semiconductor. If we do that, we're going to deliver this value proposition, which is higher yields, improved performance in terms of the chip capabilities, and reduced costs. We want to help our customers achieve all three of those. The breadth of our portfolio.
You've probably seen this slide from us before, I'm not going to dwell on it because Jim's going to talk about it in a little more detail and give you some specifics about how what we do for different inflections happening in the industry spreads out across the company. It's an interesting thing to think about if you look at Entegris, you look at just pick any of these main fab processes, we serve all of these. These are some of the solutions that we have for them to think about what's unique about this company. You look at clean, wet etch and clean processes. We make the chemistries, we make specialties, formulated cleans. We actually make the containers that they go in, so they get to the fab in our packaging.
We do the filtration during the manufacturing, then we do the fluid handling and distribution throughout the fab and through equipment. We actually can touch the liquids from making them all the way through till they're used on the wafer. It's the same thing in litho. We make packaging that actually takes resists and solvents to the fab, keeps them in a pristine state, helps with the dispense to the tool, and then the actual dispense to the wafer. In a lot of cases, filtering those filters, those resists along the way, both in manufacture and in the fab. Doing gas and liquid purification around the whole litho cell, around track and scanner, and then actually handling the reticle pods.
We see so many parts of these processes that we have a unique perspective that it's really of value to our customers to be able to have us come in and help them advance these processes. Implants the same way, from delivering very, very toxic chemicals, gases in a very safe way. Actually providing some of the materials used in the fab, handling the wafers with our chucks, doing gas purification around those processes, and then providing coatings and materials that extend the life and improve the COO of the implant tool. Just three examples, but this kind of depth across the process gives us a unique perspective to help them find the sources of issues, help them achieve their ramps. We're in a lot of collaborations, and we've talked about this before, and it's not something that you can just want to do.
It's something you have to be really invited to do by the customer. How do we do that? You've probably seen this from us as well, and this is something Bertrand actually constructed for us years ago. Help describe how we become that relevant, trusted partner to the companies that we work with. One, of course, listen. You have to be there, you have to be local, you have to be in the fab. They have to be working with you directly. You have to be able to give them solutions to try very quickly, work with them as they test them, and then get to a final solution and ramp that to a reliable supply with good continuity in a very quick manner. Figure out what you learned from that and apply it again quickly.
This whole cycle of learning, we have to do faster and faster. If we do that, we achieve our goal, which is to be the first company that our customers call for help or for solutions or to collaborate. We want to be the first ones, and as you can see here from our JDA agreement count, it's grown very steadily year upon year. Many of these are multi-party JDAs. We might be working with the IDM, their equipment provider, their material provider, their wafer provider, all together to solve a problem, because you can't do these things by yourself anymore. No company can solve a problem alone. Our ability to know all of these partners and to have technical contacts there is key for us being more relevant today than we were in the past. We have a full product pipeline.
As I said, that's one of the important things that we need to be able to serve our customers with. Several years ago, we recognized that we weren't quite caught up to the technology roadmaps of the industry. The industry was going below 45 nanometer, 32, 30, below 28. We needed to catch up a little bit. We made a choice to increase our investment in R&D. You can see we've done that steadily over the last several years. We did it in a focused way because we had a result that we were after. First was half of that increase was really advanced research, science, separation science, chemistry, basic knowledge around contamination control. A lot of PhD-level capabilities here and overseas to make sure that we would be able to help at the leading edge. We plugged that gap.
About a third of those technical resources. Our technical resources in Asia have grown by about a third during this time. We've really tried to focus and make sure we've got enough people near the major customer sites for semi that we can be that relevant partner and work with them daily. If you look at our product pipeline today, most of what we're launching is really intended to serve 10 nanometer and beyond. A lot of what you hear about today is ramping of 16, 14, 12 nanometer kind of things. 10 nanometer is really where the bandwidth is coming to a close. We've had to have the products ready for that right now and working on seven nanometer to follow that. We've had to have a good portion of our increase focused on that leading edge.
As Bertrand said, there's a large tail of business with Entegris. We have a lot of stuff we sell in the fabs all around the world every day. They want help with cost of ownership. They want help with continuous improvement activities. A portion of our increase in R&D has been focused on that as well. Make sure that we take care of that tail business in addition to the leading edge. The most important thing, I think, is that we did that in a very careful way. We did it in a deliberate way so that we could maintain our target model, achieve the financial performance. How did we do that? We did it primarily by moving dollars towards R&D and trying to get efficiency out of other parts of the company to make that happen. We really look to the whole company.
We're able to achieve this increase in R&D. It's there. It's about where we need it to be today. We did it by moving dollars to stay within our target model. I think that we're unique amongst our competitors in our focus of R&D on this industry. If you look at our competitors, many of them do a little bit of what we do. They might serve other industries more than they serve semiconductor. We're highly focused on supporting semiconductor. Our customers are recognizing that. Again, as Bertrand said, we serve the ecosystem of semiconductor. There's a common area that all of these entities are worried about. That's technology performance, process yield, device costs. That's what everybody from equipment manufacturers to materials providers and device manufacturers are worried about.
We're really at the center of that, because we work with all of them. We know the technical contacts at all of these kinds of companies. They're all our customers. We're working with them on a daily basis. When I travel around now as a representative of the new Entegris, I hear things that I never heard before. Quotes from customers, "You guys are at the center of this. We need you to help solve these problems. That's why you see the JDAs. You are the ones who can see each part of this. You know all these different aspects. You can bring it together." When I walk into major device manufacturers now, it's a different dialogue than it was two years ago.
They recognize the breadth of what we do in the fab. They want to work with us in a deliberate way, such as the JDA. Beyond that, we've seen. This is the other things we're excited about, and Jim will talk a little about too, is we've seen the opportunities that the companies coming together provides from a technical perspective. I'll give you one simple example from the wet chemistries and wet surfaces. Before we understood how to filter or purify a chemical, but we didn't always, and in fact, we rarely knew the exact makeup of that chemical. Now we're a chemical provider as well, so we know everything about the chemistry that's going to go into this, and those chemical developers know everything about how it's going to be filtered or purified.
There's a tremendous amount of interplay between a filter and what it's filtering today, far beyond taking out particles. That's one of the things that we saw as an opportunity. It's the same for gases, gas-wetted surfaces, and for solid delivery. We know many aspects of these things. We were obviously excited to start to think about what can we do in terms of positive synergies from bringing these companies together? That's what we're really working on now. Jim and I are working very closely together on, let's look at everything we do and how can we do things better now than we did before. Not just from a cost perspective, but from a product performance capability perspective. The obvious things are the easy kind of close in. I can give the customer data on how this filter and this chemistry work together.
I can give the customer data on what our purification does to this gas. I can give customers data on what we can do in terms of safe storage and what it looks like coming out of the canister. I can tell them more today than I could tell them when we were separate companies. That's quick and easy, and that's stuff we're doing today. Close behind that, and starting actually a couple of months ago, start to work on optimizing those products, make the filters a little bit better matched to those chemistries so they get the best combination, as one example. Beyond that, start actual brand new product development projects. Let's start working on a new chemistry for a new application. Let's do the filtration and purification with it. Let's start working on a new deposition material.
Let's talk about the metrology that we need to provide and we can provide around that, the filtration to make sure there's no concerns about solids making it through to the film. All those things are just getting underway right now. Then longer term, even more fully integrated technologies. We'll bring sensing, purification in, change bath lifetimes. Jim will talk a little bit about some of this, but we're pretty excited that more than just the cost synergies we've already delivered, there are some opportunities to do things better. These are long payoffs, and these are done under JDAs in a lot of cases, so we can't really talk about specifics of them, but it's starting to happen where we see these companies together can move faster than they could move before.
We mentioned that we try to have a lot of these resources near customers, and that's been one of the key things that I've liked about seeing these companies come together. Both companies have made investments, obviously, to be present in the major semiconductor centers of the world. We were pretty pleased with the complementary nature of the resources that ATMI had put on the ground, as well as Entegris. We had a lot of manufacturing capabilities over there. They had some great tech centers. It was really actually fun to bring those together to enhance the speed and capability of what we're doing and do it near the customers. Let me get this advanced. There we go. We've brought together tech centers in North America, working on filtration, working on purification and chemistry, deposition materials, wafer handling, sensing.
Then, in Asia, obviously Korea and Taiwan, we've actually combined our tech centers physically now. The people are working together in the same facility, done some expansion to bring those together, and actually have customers working with us in those facilities on a daily basis to help develop and move faster that cycle of learning on new products. The third leg, you remember the three pillars again, the third leg being operational excellence. One of the things that we also embraced several years ago is the fact that it's not going to be just about delivering something on time at the right price anymore. There's ever-advancing requirements from the leaders in the device manufacturing world to do more green initiatives, to have better roadmap alignment, but focus on contamination control, apply statistical process control in a very consistent way.
Entegris had started this journey, really, I think, first, and one of the real benefits of bringing ATMI in is we were able to quickly start to proliferate that out across their manufacturing as well, which our customers were pleased to see. Now they can get from one resource, a company that understands all of these requirements that it takes to be that relevant, trusted partner. That's only continuing, but we're committed to do it, and really not just because, hey, a customer asked us for it, but we really see that this is going to make us that better company that Bertrand talked about. Again, it's all about building the trust for these semiconductor companies, which are really some of the most demanding customers in the world. Are we succeeding? We've had some evidence of that. One, of course, wins.
The number of engagements that I mentioned that we have at leading-edge developments have increased greatly. We've also started to be recognized from a very tough customer, a very good, important customer of ours, Intel. We were one of 19 recipients out of their global supply chain to receive the PQS award in 2014. As Tim Henry quoted here, it's about all of those things we've talked about, cost, quality, availability, technology. You've got to be cost with a roadmap. You've got to be a reliable supplier. You've got to be meeting all those things on that stack chart I showed you. We're pleased to get it, but we really see it as step one, and we need to continue this journey.
We talk to them all the time about what's next, what do we need to do to get to that next level, get further into controlling of our supply chain ourselves and so forth, and that's all the kind of activities that we're working on today. It's a start. We're very pleased and proud to have gotten this award. It recognizes the work we've done. It's really, again, just the beginning of where we're going to head next. I'll have Greg come up and make a comment, and then Jim will talk about some of the things happening specifically in the technology world and how that impacts Entegris.
If you peel back the eight or 10 slides that Todd went through, really what it's all about is it's our competitive advantage. It's why do we win, and the reasons we win are, A, we're pervasive across that fab environment. The fact we're pervasive across the supply chain puts us in a position where we've got much greater applications expertise than our customers. If you think about it, we come at it horizontally. I said our customers. Most of our competitors come at it from a vertical perspective. Pall's coming at it. They're looking at filtration. Somebody else might be looking at flow control. We're looking at it across the piece. That applications expertise really separates us. The second thing that we're doing is we're investing. Todd uses the term investing.
I think about it as resource allocation, because we're investing in that we're allocating more resources to ER&D, and we're allocating more resources to things like our new product development process, like improving our internet site, like expanding the global platform. We're doing all of that within the target model. It's about making smarter investments and smarter choices, and those are all things that are going to help us advance the ball and continue to become more competitive. With that, I'll introduce Jim O'Neill. Jim is our Chief Technology Officer. He came from the ATMI side of the house. Previous to that, he had a long career in IBM's semiconductor process development area. With that, Jim.
Great. Thank you. Good afternoon. This is a really exciting time for Entegris from a technology perspective. I'm going to continue Todd's story of how we're going to grow the new Entegris by talking about how the major technology changes in the industry are really beginning to play into the strengths of this new Entegris. As you're aware, Moore's Law has driven semiconductor technology development for the last 50 years. Historically, this has really been a story about miniaturization, making things faster, making things denser by making them smaller. Today, the story has changed a little bit. We have begun to reach fundamental limits in patterning and process technology, and we're beginning to experience inflection points in many areas, including things like in lithography, transitioning from single exposure processes to multi-patterning techniques.
In the types of device architectures we're seeing, from planar devices to now three-dimensional devices like FinFET or 3D NAND in memory. The industry has implemented a whole host of new materials that are intended to improve performance and reliability. All this is really in an effort to keep pace with the spirit of Moore's law, but it's come at a tremendous expense in terms of process complexity and increasing yield challenges. Really, this is where Entegris comes into play because as a company, Entegris' strength is really all about helping our customers achieve yield in a timely manner. If you consider the impact of new materials alone, you notice as chips become more advanced, that is higher density, more functionality, the processes used to make them become more complex and more reliant on new materials.
New materials are the dominant contributor to device performance improvements over the last several technology generations. The introduction of new materials brings with it the increase in the number of process steps, and therefore process complexity. This in turn makes our customers' job, the device manufacturer's job, of improving yield for their technology even more difficult. Their already difficult product introduction cycles are even more difficult to achieve. This is really the crux of the problem that Entegris is trying to work with our customers on to overcome, using our strengths in materials discovery, materials development, and critical materials handling capabilities. Materials are increasingly important in advanced semiconductor technology development. They also happen to be part of a very long and complex supply chain.
There's significant opportunity to introduce contamination throughout the supply chain, from the point where the materials are manufactured to where they're delivered on the wafer. The ability to protect materials and deliver them safely throughout the supply chain is absolutely critical. With a fundamental understanding of chemistry and materials interactions and how materials are used in the fab, Entegris is sort of uniquely positioned to enable defect-free material solutions throughout the supply chain, from the point of production all the way to the point of consumption. Let me just remind you of the types of technologies that our customers are struggling to implement in their most advanced technology nodes. These are things like multi-patterning, 3D structures like FinFETs, or the whole host of new materials, all in an effort to keep pace with Moore's law.
The impact of any one of these changes ripples throughout a large portion of the integrated process. It's not just isolated to the sector where that process change has been implemented. That's what's really important here, because an ability to recognize the interplay between materials allows Entegris to leverage the breadth of its product portfolio to help our customers enable any one of these implementations. Let me show you a few examples of how this works and how the breadth of our product portfolio really plays into the strengths of Entegris to be able to help our customers implement these rather difficult changes. You've seen this chart before from Todd depicting the breadth of the Entegris portfolio. Let me explain why this breadth is so important in executing these inflection point technologies.
If you consider the example of multi-patterning, a technique used to improve the ability to achieve ever-decreasing line widths and improve patterning control. The effect of multi-patterning is not just limited to the lithography sector, where the photosensitive resist is dispensed and the actual imaging occurs. The effects ripple throughout large portions of the rest of the process, including deposition, where patterning stacks are deposited, etch sector, where they are patterned, the clean sector, which follows, and then because of the multiple patterning steps which occur, there needs to be an increased attention to overall defect management and control. In effect, the overall impact of multi-patterning is felt throughout the integrated process. Entegris' broad portfolio of products is really what enables us to help our customers implement technologies like multi-patterning. Similarly, you could look at 3D NAND.
This is a technology or an approach that's used in memory to improve memory density and speed. Such 3D structures drive changes throughout the process from how materials are etched to how they're deposited and cleaned. They place increasingly stringent requirements on how devices are doped, and again, increased attention is required on overall defect management and control. In short, the implementation of 3D structures ripples throughout the entire integrated process, and each of the challenges that are posed by this ripple effect can be addressed with an Entegris solution. Again, the breadth of Entegris' product portfolio plays very well into our ability to help our most advanced customers implement difficult changes like this. Last example I'll give is one of depositing a new material, this is cobalt, as a new material that's being explored to improve interconnects performance and reliability.
The use of cobalt requires more than just new deposition precursors and new precursor delivery vessels. It requires new formulated cleans that are compatible with the material, both after etch and after polish. Then those cleans require new filtration and fluid handling capabilities. Again, you see the effect of changing one material in the wiring stack for the interconnect structures, ripples throughout other process modules where Entegris plays quite strongly. Just to dive a little bit deeper into the technical aspect of this. Cobalt is considered to be applicable at the finest wiring levels in the circuit. Those are the levels just above the transistor. At these levels, the resistivity and the reliability of copper, which is today's commonly used material, begins to degrade as the line width gets smaller. New materials are being explored, and cobalt is one of them.
Whether it's being considered as a new replacement for contacts or a barrier layer or a capping layer or a new via layer, each one of these implementations drives the implementation of a new precursor, a new precursor delivery system. It also requires an entirely new infrastructure that's compatible with it, from the formulated cleans that prevent galvanic corrosion, to new CMP formulations that are suitable for removing cobalt. Each of those materials in turn drives new fluid handling capabilities and new filtration and purification capabilities. You see that all of this ripple effect plays into Entegris' strength in being able to provide not only the material, but the material handling capabilities to enable the implementation of new materials like cobalt. It's really not just about having an aggregate portfolio of products that touch a sector that our customers are working on.
Consider the example of a wet clean module in a factory. Entegris has a whole lot of products that are useful in this sector, from the fluid materials, to the purifiers, to the filters, the dispense tubing, the containers, the nozzles, both on the tool and in the subfab. What's really important here is how we take those components and put them together in a way that brings value to our customer. Typically, the unit process engineer working for our customer will take this aggregate of products and assemble them in such a way that hearkens back to giving them a process that performs, one that yields, and one that can be executed at an affordable cost. Today, Entegris can do this. We can provide much more comprehensive material solutions so our customers don't have to invest their time in doing that.
They can focus on more important aspects of the technology, such as the product performance and the device design. I spent a lot of time traveling to many of our advanced customers with this message about how we can put the pieces together, and our customers acknowledge the value that we can bring to the table. Our challenge is really to work with them to identify what are the specific problems that we need to work on for them, and what tailored solutions can we bring to the table to solve their problems. Let me give you one example of a portfolio that we're beginning to assemble to bring product synergies there to our customer. That is, again, an example from our wet etch and clean sector. This is a copper post-etch clean process intended to remove the residue that's left behind after the etching process.
Well, Entegris has long been a company that's provided leading formulations for cleaning advanced metallurgies after etch. Today, these formulations are matched with our best filter technology to provide lower defectivity cleaning solutions. Moving forward, we're working to co-optimize the formulation and the filter to provide the optimized cleaning solution. We're also bringing in our InVue process sensing capability to understand how the formulation bath ages and when it needs to be replenished. We have the goal of developing and providing purification capabilities that remove impurities that build up in the bath during the cleaning process. This is an example of a synergistic combination of the breadth of our portfolio that provides benefits from a performance, yield, and cost perspective. For our most advanced customers. I think the real question is: What does all this really matter to Entegris' business?
Each of the inflection point technologies that I described requires solutions that Entegris is really uniquely positioned to provide. Whether it's precursors for advanced patterning stacks or new metallurgies for advanced wiring, or new formulations and filters to improve yield, or containers or fluid handling capabilities to safely deliver clean chemistry to the wafer, all of these inflection point technologies are really good for Entegris, and our ability to help our customers adopt them is a key element of our growth strategy. Our participation in these challenges and working with our customers on these inflection point technologies, I think is what's really going to enable us to outpace our competitors and outgrow the market by 100 basis points-200 basis points. That's really our goal here.
Thank you, Jim. I think bottom line, materials are becoming more important in the semiconductor manufacturing process. You go back to 90 nanometers-45 nanometers, it was lithography, it was scaling. As we come down below 45, materials become much more important. We're in the unique position of being able to integrate materials with the delivery of those materials, whether it's a high-purity chemical drum, the purification and cleaning of those chemistries, and bring that all to bear on behalf of our customers. I think the last point Jim makes is, as you run around this SEMICON West, you hear people talking about FinFET and cobalt and all of these advanced technologies. In each of these advanced technologies, we play a role, those things are all going to be important to us as we sort of drive toward that goal about growing the industry.
Let's switch over, talk a little bit about Entegris from a financial perspective. This chart is actually, was the last chart in Bertrand's deck. Talk a little bit about the financial objectives, our investment priorities, and the target model. The financial objectives, really, this is what we want to be graded on by you. It's what we grade ourselves on. For us, it's about we want to achieve growth in excess of the market by 100 to 200 basis points. We want to consistently achieve that target model. We're focused on reducing our debt, and ultimately, we want to grow the earnings per share. We do continue to invest heavily in the business, whether it's R&D or some of the other things that Todd talked about.
We have a continuous drive to be among the most profitable companies in the sector. When we compare ourselves, we're comparing ourselves both to small and mid-cap semiconductor companies as well as filtration companies, specialty chem companies, but those beyond kind of that semiconductor universe. We stack up, as Bertrand said, quite well. I wanted to just give you our perspective on the scorecard or our report card. If you think about 2014, on a grade point average, that would give us a, I'd say we had a 3.6. We made three As and a B minus. In other words, we hit the target. We achieved the target model. We paid down debt. We grew EPS nicely, but we didn't grow as rapidly as the industry in 2014. That happened for a number of reasons.
The fact that we had currency headwinds, some of the migrations to the advanced nodes didn't happen as quickly as we expected. We had some capacity constraints in some of our filtration areas. The point is, we made a B minus in 2014. In 2015, we're much closer to that A average. If you think about the first year-over-year in Q1, we were up 5%. Our revenue was up 5%. That was even in light of about a 2% headwind from currency. We had very nice year-over-year growth in Q1. Then on the rest of the metrics, we continued to do well in Q1, and I'll talk about those other metrics now. Starting really with the target model. For us, the target model internally is kind of our holy grail. We've been operating to a model since late 2009.
We've only changed it twice. We changed it once in 2012, and we changed it last year when we did the ATMI acquisition. The way this model works for us, focusing on the top half of this page, is going at different revenue levels, 250, 280, 310, what can you expect us to deliver from an operating profit perspective and an EPS perspective? Like I said, we started out using this as a Wall Street tool, but internally, each of our businesses has a model that rolls up to this. We have business reviews. The talk is all about how'd you do relative to the model. This has really given us great discipline internally. The bottom of the page just shows how have we done against that model. The orange bars are what was our operating margin in that given quarter.
The red hash mark is what did the target model say that our operating margin should be? You can see in the last 21 quarters, we've delivered on that model or delivered on our commitment 20 out of 21 quarters. When you talk about that scorecard. They give us an A there, I think 20 out of 21 says we're doing quite well. Let's shift a little bit to the balance sheet here. Today, I feel very good about our balance sheet, very confident in the balance sheet. We've got about $740 million in debt, $340 million in cash, so a net debt position of $400 million. It's a ratio of 1.7 times EBITDA. Through the first quarter of this year, we had repaid $75 million of the acquisition-related debt.
We said we'd repay $150 million of it within 18 months of the acquisition, which means we've got another $75 million that we're committed to repay by the end of Q3. The balance sheet continues, like I said, continues to be in good shape, and I think you can expect to see that leverage continue to come down. The last thing on that report card was earnings per share growth. The chart up on the screen now shows our earnings progression from 2012 through 2016. We're only going to focus on that period, 2013 to 2015, because that really shows the power of the ATMI acquisition. In 2013, we made $0.59 a share. In 2014, we made $0.69 a share. 2014 to over 2013, we had about 15% earnings per share growth.
Shift to 2015 versus 2014, the street said $0.80 for 2015. $0.11 improvement over where we were in 2014. Another year of sort of a 15% earnings per share growth. Think about where we are in 2015. In Q1, we did $0.18. You take the top end of our guidance in Q2 or within our guidance range, and we're essentially halfway to that $0.80 after Q2. The $0.80, it's the street estimate, but it's certainly not a number that we view as way out there. 2016, we think we can do close to $1.00 per share. When we talk about that earnings per share growth, we've delivered a nice progression, and we think we'll continue to drive good earnings per share. How do we get to something close to $1.00 in 2016?
If you look at 2014, the far left, those are our published results, $962 million in revenue, $0.69 in EPS. Had we owned ATMI for the full year, our revenue would've been $1.076 billion. Our baseline in terms of revenue is $1.076 billion. Industry, we expect to grow about 2.5% a year. That essentially assumes CapEx is flat in 2015 and 2016, and wafer starts are up somewhere 2.5%-3%. That growth, plus 125 basis points of outperformance, would take us to $1.156 billion in revenue by the end of 2016. If you take the target model, which has significant leverage in it, as you move from that $250 million that we showed on the prior slide to $300 million, we're flowing through about $0.40 for each incremental dollar of revenue. You get up to the $1.156 billion number, we expect operating margins of about 19%.
If we hadn't reduced any of the debt, our interest cost would be about $38 million. Those numbers come down to a $0.95 per share number. As we de-leverage, if we're to de-leverage $300 million from where we started, we're halfway there already, to the end of 2016, that adds another essentially $0.05 per share to earnings per share and brings us to that $1.00 number. Our cash flow historically, as a company, those of you who have followed us for a long time, we've been a very good cash flow generator. In 2014, we had cash flow from operations of $126 million. We think that'll bump up to somewhere in the range of $160 million in 2015. That big increase from 2014 to 2015 is really driven by three things. One, we will have owned ATMI for the entire year in 2015.
We only owned it for eight months in 2014. In 2014, we had very significant integration-related costs. The other thing in 2014 is we had not achieved the synergies. As we come into 2015, we have much less in the way of integration costs. We've achieved the synergies. That step-up is a realistic number, and then that moves to $180 million in 2016. Over the course of a three-year period, we'll generate $465 million in cash from operations. The CapEx or the orange area, we're assuming $65 million in 2015 and $65 million in 2016. That's slightly higher than we've talked about before. It relates primarily to two specific areas where we've got meaningful growth opportunities that we're going to invest behind, as well as some incremental investment that we didn't expect related to the specialty gas business at ATMI.
The point to the chart is very strong cash flow generation. The question becomes, what are we going to do with that free cash flow? If you think about the world that we live in, the job of the management team, and Bertrand and I specifically, our number 1 priority is capital allocation. Today, we've been using most of that capital to reduce debt, which is really all about giving us maximum flexibility. As we continue to drive that debt down, we'll start to look at other priorities. For today, I think about it primarily as a debt reduction story. Key takeaways, I think, from the meeting today, from the finance section, A. We're thrilled with how the ATMI acquisition has gone. We believe in the industry. Part of the reason we really like the industry is because of where we're positioned.
We're positioned in a much stabler place than most of the other companies you'll think about in this industry, because we have a diverse customer base, because we have a diverse product line, because we're 80% unit-driven. I think the final point is we're proven executors. We show that with what we've done over the last year on the ATMI integration, how we've delivered on the target model. We think as we continue to execute, the leverage we'll generate will continue to drive growth and earnings per share growth specifically. With that, I'll bring the rest of the group up, and we'll take any questions.
Hi. Thanks for taking my question. It's Jerome Citodi. First on the revenue outperformance of 100-200 basis points for Bertrand, does that include any synergies you would expect going forward, or could that add to that number?
It is inclusive of it, as Jim actually presented in his section, you could see that a lot of the additional growth momentum that we expect will come from the positive synergies that we expect to realize as we combine the capabilities of the legacy Entegris and the legacy ATMI. Jim did present a few examples. Those are only a few examples. I think we're working on more. I hope that there could be more in that. I think we have enough conviction today to actually commit to that 100-200 basis point overage.
Okay. Greg, just one question on CapEx. You mentioned some kind of gas delivery systems. Can you elaborate on that or-
Yes. The specialty gas business or the SDS business at ATMI, we own essentially the cylinders or the delivery systems for that product line. As we introduce new products there, which we're introducing more of them than we initially expected, you can't take a canister that you've used for one gas and shift it and use it for another gas. As we introduce new products, we're having to invest in what we refer to as our fleet of canisters. That number, I would say, has proven to be higher than we initially expected, in part because we're introducing a lot more new mixtures than we initially expected.
On the IoT that you mentioned, Bertrand, you said that it doesn't really matter if it's on leading edge or lagging edge. Is there a benefit on the margin line if it's on lagging edge and on the revenue line if it's on leading edge, or does it kind of balance out between the two?
It's a good question. I'll try to keep my answer simple. We could go into a lot of complexity. At high level, I would say that the size of the opportunity for us at the leading edge remains larger, number one. Having said that, I would say that the margin profile, the market share profile at the trailing edge are just as appealing as what we are seeing at the leading edge. That's what I was trying to say in my comment.
Amanda, think about that as the operating margin. Because essentially what's happening at the trailing edge, you may have slightly lower gross margins, but you have essentially no ER&D. Whereas at the leading edge, you're going to have higher gross margins, but you also have higher investment.
Just one more question. On the net leverage, you said that it was about 1.7 times EBITDA currently.
Correct.
What is a leverage ratio that you feel comfortable with going forward?
As we think about the story, if somebody says, "How does this play out in a perfect picture?" is we de-lever down to kind of 1.0 times, and then we are presented with another nice opportunity like the ATMI situation, and we take it back up to two, 2.5 again and sort of continue that vicious circle. I would say, I think you should look at the $360 million of notes largely as permanent capital. I mean, those are eight-year bullet repayment. That leverage level is likely to exist for a relatively long period of time.
Not to play on words, Amanda, I think we are comfortable with the current leverage ratio. We generate a lot of free cash flow. The reason why we are really spending so much focus on paying down the debt is, first of all, we want to demonstrate to all of you that we can manage within a levered world. We made a commitment to pay down the debt, so we want to deliver on our commitment. Again, we're paying down the debt to regain flexibility so that options can open up down the road.
Christian Schwab. Quick question. We're going to pay $75 million in the next two quarters, Greg?
Right. We paid $75 million through Q1. We said in Q2 and Q3, we'll pay another $75 million total.
Right. What is the total debt reduction for your fiscal year 2016 goal? Is that another $75 million? Does that equal $150 million of the term loan note, or is it more than that?
I would say at this point, I don't want to say exactly, pin ourselves on the capital allocation for 2016. We'd certainly be in a position to make that kind of repayment from what we'll generate and what we'll repatriate.
Quickly remind us, of your $241 million at the end of March, what % of that is offshore?
Approximately order of magnitude, $100 million in the U.S.
Okay. Then if we go to that customer slide that you had on your diverse business mix, it's a little bit different than it was. I don't have the slides with me, but I know it's different than what you had last year. Your largest customer is a little bit bigger, and your next two customers are a little bit smaller than I believe they were last year. W alk me through some of what happened in the marketplace that shook that around a little bit?
Well, first of all, you're looking at trailing 12 months numbers. I don't know exactly what was the basis for what we presented last year, but our number one customer is a large Taiwanese customer. They had a phenomenal year last year, and so did we. I think that's the primary reason for the bump in terms of their importance in our customer mix.
Thank you. Richard Rorick. Greg, on the chart where you had the march towards the dollar in EPS, I don't have the slides in front of me, but I think earlier in the year, you kind of had a bottom-line number of $1.10. I don't know if that was off of a different base to start with, but was there anything changing in those assumptions?
I would say the only thing that might be different in those assumptions from what we've shown in the past is the industry growth rates are probably slightly lower. I don't remember showing a $1.10, though. It doesn't mean it didn't happen. I don't.
Just putting words in your mouth.
Yeah, I know.
On the CapEx, can you talk maybe a little bit more behind where that extra investment's going to be made this year and next?
I would just say we have a significant opportunity. We really have a significant opportunity within our specialty materials business, specifically in the coatings area, would be one. We also have a relatively meaningful opportunity within the graphite area. Those are not foregone conclusions that we're going to make those investments. As we're running the analysis now, they look pretty promising.
It's really growth related, and it's capacity as well as capability related.
Are there levers to push to increase the exposure to the aerospace side of the business?
These actually are both, I'll just say broadly, they're both electronics focused.
We have some exposure to aerospace, it's pretty small in the scope of the company.
Thank you. You were saying the graphite and the coatings all account?
Yes.
That's one of the growth opportunities. The other part of the growth opportunity that will require additional CapEx would be the cylinders for specialty gases. As you know, the way we differentiate ourselves from our competitors would be, one, we believe that we have safer packaging solutions. Also, we continuously introduce better gases, those are new mixtures. Every time you introduce a new mixture, you need to invest into a new fleet of cylinders, that's what we're doing.
If you look at the slide that was Entegris wet etch and clean roadmap example at the end of your presentation.
Yes.
You've got sort of these different years here, 2014, 2015, and 2016, with the escalating impact. I think it'd be helpful maybe just to hear from the different members of the team who are responsible for those different steps. What was sort of the selling process on selling them, we can do A, B, and C? What was the technical implementation process on, we did our internal work to figure out how to do A, B, and C? What's the sort of capital and revenue benefit?
I'll take a first crack at it, and then I'll turn to Todd, you can explain maybe the business model, and then maybe on the technology side, Jim, you can actually provide some more details. Again, remember what Todd said in his presentation. First of all, increasingly, when you deal with the advanced nodes, you work actually more closely with the customers than ever before, and you're really literally invited to collaborate and to co-develop. That gives you a lot of insight into their process challenges And it is really allowing us to really tune the types of solutions that we can develop with them.
I would say that for the last year, all of us and many more within the company, have spent a lot of time on the road telling the story about the value of the new Entegris platform and reminding our customers that we could do a lot more for them than what they've been used to seeing from either Entegris and ATMI. Let me turn to you, Todd.
Well, yeah, you covered it. Really, there's two things we did. When we first came together, Jim and I got the technical leaders of the company together. What do we want to learn from each other? What have we always wondered about what the other could do and how our products interact? Those pretty fertile. It gave us those three blocks of wetted surfaces, gas wetted surfaces. That led to some ideas around that. Then as Bertrand said, really, we visited a lot of customers and talked about the new company. Showed those graphs of the breadth and depth and how we touch all these processes to end. They give us ideas back. "Hey, with all that, you should be able to do this. You should be able to help my cost of ownership. You should be able to enable this process.
If we bring these things together, I'll have to help as a customer because I'll have to do things a little differently myself, but it's got to be a payoff for me." Those kind of validated. These are the right commercial investments. We kind of narrowed down to 8 to 10 things that we're working on right now that we think are going to feed into that. Jim gave you a little bit of a view there of what some of that looks like specifically in terms of path length extension, improving cost of ownership, enabling the next nodes.
Yeah, I think one of the most exciting parts of integrating the two companies was throwing two highly technical teams together and sort of letting them go at coming up with new ideas. Both teams who'd been out in front of the customer, understood what the customer's problems were from their perspective. Then when they got together, they came up with a lot of new ideas. Did we get it right the first time? No. What we brought to the customer sparked a discussion. The discussions really led to much better refinement of the ideas that we're now working on. These have to be continuously validated as we generate data. I think, in doing things like this, bringing the two companies together, getting the two technical teams together, we've opened up a path for increased dialogue with our customer.
If nothing else, that's probably the most valuable thing that we can do as a technical company, which is engage them technically, get a customer's input, and then develop according to their needs as opposed to developing things and trying to push them.
It was really interesting for us, I would say that some customers got it day one. The minute we announced the transaction, I had calls from some customers saying, "This is wonderful. We need to get together now." We said, "The merger is not even closed. We cannot do that. But I'm glad that you're understanding the value that we're going to bring." Other customers, it took a little bit longer, but not very long, and very soon they were saying, "Well," as you said, "Todd, you don't understand. You're really now in the middle of it, and your value has grown tremendously." Some of the customers, it took much longer, and it took several meetings.
We tried to get some easy things to do, demonstrated the value. Now we are at a stage where we said, "Oh, well, we could maybe do a lot more." I think, again, every customer has had a different pace of adoption or understanding of the new value proposition. I would say today that most everybody gets it, and that's really exciting.
I just add one point from a process perspective. That is sort of picking the eight or 10 projects that Todd talks we're working on or evaluating our R&D pipeline and deciding where to invest is a very formal process for us. Next week, we'll have three days of meetings, what's called a portfolio. It's essentially a portfolio review committee where all the business leaders, their technical leaders, their finance people come together. Everybody goes through kind of, here's my pipeline, here's what I'm working on, here's where I can use your help, here's where we need investment. It's a very rigorous process and happens every quarter.
I think that's a good point. We are managing the company differently. I would argue that in the past, a lot of the business units were really managed in a silo way. Todd and Jim are doing a great job at really making sure that there is a lot more cross-fertilization between the BUs. We do that in a very prudent and careful way in terms of what we share and what we don't share, even between the BUs internally. This is happening to a much greater extent than it ever did before.
I got the mic now.
Okay.
Wedbush Pacific Crest. I have a question on the leading edge demand. Last year, you undergrew the model a little bit because there weren't as many, or there wasn't as much in terms of advanced node transitions as you expected. I'm wondering, are you seeing any change in either the cadence of node transition or the number of wafer starts at advanced nodes, and how is that impacting your thinking moving forward?
You're right. I think one of the headwinds last year was this lack of activity at the leading edge. One customer, a Taiwanese customer, had a lot of activity there. They did well. We did very well, and that shows in the customer mix that I presented to you. Besides them, there was really nobody else that was really active at the leading edge. I think that if they had it their way, the cadence would probably intensify. As Jim said, this is really becoming really more complicated. In terms of the number of wafers being run at the leading edge, it was less than we expected last year. This year, most likely, it would be also probably a little less than we expected. Again, as I said at the beginning, I think we are not an equipment company.
The type of opportunities that we want At 20, at 28, all generating very nice steady revenues for us. I think that the leading edge is the hump, if you want. That's the incremental growth that we are after. I would say that this year, we have a lot of hump already in the portfolio. It would be nice if we see more wafers coming from 14 and 16. My guess is that we will do well with or without it.
Okay. I didn't see in the slide deck here, maybe I missed it because I walked in late, a mix of leading edge versus lagging edge like you've done in the past. Did you have that number today?
Yeah, it's really a hard thing for us to guess. If you think about the way we usually approximate that, is we look at the new products that we have been introducing over the last three years, typically those products go into the leading edge. That's somewhere between 25%-30% of our revenue. That's how we come up with the approximation that we make.
Okay. Then finally from me, you have some non-semi revenue. You talked about growth in adjacent markets, except when you talk about your long-term model, you're still talking about outgrowing the semi industry. I'm wondering why you don't factor in some growth in adjacent markets. Maybe if you could just comment on that, what your expectations are.
You want to take that one?
No, go ahead.
It's because we try to keep it simple.
Yeah.
We could build a very complicated index that would be aggregation of multiple index, I think you would all be scratching your heads and saying, "What does that all mean?" We try to keep it simple, and at times it's probably going to be an easier comparison. At other times, probably would be a harder comparison, We decided to go for simplicity.
To be clear, we're definitely we talk about where we're investing, We continue to invest in some of those adjacent markets, some of the newer technologies that ATMI owned around carbon, around resource reclamation or efficiency. eVOLV, as they called it. As well as we've made investments in new business development people for the first time. We've got real professional new business development people helping us look at those new markets.
Hi, could you talk about the relative revenue for a memory wafer versus a logic wafer? Is it much different? If so, could you just share?
I'll pass it on to Todd in a minute. I think it's a hard question, we don't track that this way, as a matter of fact. We are trying to change our systems that we would have a little bit more visibility to that, we don't really have a precise way of tracking. At high level, I would say that the size of the opportunity for what we do is usually a little greater in logic environments. This is changing, this is changing actually pretty rapidly, and that's getting actually interesting, even on the memory side as well. Todd, Jim, anything you want to add?
I'd just say that I think that you think about the complexity of the chip, the complexity of the process, that's generally good for us. There's also differences in how those kinds of customers use our products. Some of them will try to use them longer, shorter, different dilutions, et cetera. There's a lot of factors that play in, it'll depend on the node that the memory's at, as opposed to the node that logic's at. I think that Bertrand generalized it the right way.
Hi, just one quick question on your earnings tolerance with regards to the ups and downs of the fab utilization. For example, whether it is one person or 20 people in this room, we need to keep the lights on anyway. I was just wondering if this would be an adequate analogy for your unit-driven products from the perspective of your clients.
I think the first thing to think about is that if you think of semiconductor units, they've only been down sort of twice in the history of the world, in 2001, and then in the 2009 time frame. We don't really think about it can happen, but it's clearly a much stabler, I mean, if you look at how many times capital's been down, it's been much, much greater than that. If you think about our cost structure, one of the things that the Target model has done for us, though, is it's created much more discipline around variabilizing cost.
When somebody says, "Hey, we want to add this functionality," or, "We're going to invest more here," we're always asking the question, is there a way to make those costs variable so we don't lock ourselves into a situation where we can't flex down if things were to get really ugly.
Thanks. You guys talked a couple of times about not being an equipment business anymore. There are a couple pieces that still look a little bit like an equipment business. Maybe the FOUP business looks like an equipment business. Is there a possibility of perhaps divesting pieces of the business? Not specifically that one, if you don't want to comment on it, but are there pieces that you might cull from the portfolio over the next several years?
Right. I think that the reason why we try to discriminate between unit driven and more capital type of products is really to make it easier for you to follow and understand the trends in our own business. Frankly, we have some really exciting CapEx types of product lines, FOUP being one of them. If you think about the contamination control in terms of the microenvironments around the in-process wafer are great and are very significant, and we add actually a lot of value to our customers by helping them develop new solutions for their advanced nodes. Yes, it follows different buying patterns, but the margins are good. We have an incredibly high market share. It's a good product line, and it belongs in the portfolio.
It's consistent with the overall value proposition that we're trying to provide to our customers, which is being the leaders in contamination control, advanced chemistries, et cetera.
I think I would just add to that we do look at every business. I have a lot of business to look at and decide where to place our bets every day, certainly every quarter, as Greg mentioned, and our portfolio review committee. It's really about where can we do the most for the customer, get the best return on our investment, agnostic of really which product line or business unit it is. That's where we really require our business unit leaders to be very mature and think about the good of the overall company. We'll make those bets in the right places.
We may see parts of the business where we feel like we're not so certain about the growth perspective there, so we're going to make sure we get the funding in the right place, so it's going to provide the most growth for the company on a daily basis. It's really a portfolio from my view, and I try to do the best allocation I can.
To me, the gems of the business are the filtration part of the business and the chemistries part of the business. Now that you've owned ATMI for a year, I guess what I'm a little surprised about is I didn't hear much about the chemistry side. I hear SPS, which is chemistry, I understand. Then you guys are talking about the graphite and the coatings area. One of my concerns about ATMI, I think I've told you this, was that they were getting picked off on certain applications, like first process CMP clean, by little Asian guys. Do you see a roadmap now that you've owned it for a year and change to be able to differentiate the chemistries enough to maintain it or strengthen your position in chemistries? Does that just take longer, and so we're not hearing about it now? I don't know if you could just discuss that.
I'll let Jim answer the second part of your question, but I would just tell you that in general terms, as I stated at the beginning, I'm very pleased with the results of the integration and the acquisition of ATMI, and that includes the quality of the overall portfolio. The reason we didn't talk about. You probably noticed we didn't talk about products at all today. We didn't talk about filters, we didn't talk about chemistry, we didn't talk about products. That was by design. We didn't have a lot of time, and as you know, we have a very broad portfolio. In fact, we didn't want to fall into the trap of singling out one product line and not another one. We kept the message at a much higher level.
Actually, that message is very consistent with the way we're trying to engage with our customers today. We talk about what we call internally our Blue Ocean Strategy, which is really trying to distill it down to quintessential value that we can bring to our customers. A lot of the discussions we're having with the customers is really around the value proposition of contamination control, advanced chemistries, et cetera, and not the individual products. That's the big difference in terms of how we're approaching customers today versus how we were approaching customers in the past. Now, in terms of the ability to differentiate the chemistries or the mixtures, I'll turn to you, Todd, but I like what we see in the portfolio.
Yep. Yeah. As Jim covered, there's a lot of cross effects as they change one part of the process. They bring cobalt in, there's a lot of concerns about formulated cleans and what are the effects of that going to be throughout the other process steps. What I'm seeing, I'm liking the chemistry part of the business because I'm seeing, especially at advanced nodes, 10 nanometer and beyond, there's a whole new set of demands that really weren't there. Not only for the formulation, but also for the purity, consistency, ability to deliver that chemistry, which I think we're going to be uniquely capable to do. We'll stay in that specialty world where we can really bring differentiation through our chemists around the world and do something different and deliver it in a very specific, clean way to the customers.
I think those demands are really growing, if anything. I think maybe in a little bit of a lull in that in the last couple of years, but I'm encouraged from what I'm hearing from the market.
Yeah. On the technical front, every customer's integration scheme, the way in which they build a chip is a little bit different. Increasingly, we're finding that our chemistries need to be tailored off of a backbone kind of chemistry. That's one point. The second point is that with greater exposure to the overall customers' needs through the combination with Entegris, we're identifying areas where our products are working much closer together, not just materials handling and chemistries, but one chemistry with another. Formulated cleans and deposition is an area that's absolutely ripe for significant acceptance by our customer base at this time, and it's an area that we're pushing quite hard. Both preparing surfaces for deposition, cleaning materials that are new to an integration scheme. Because of the breadth of
They call one day, do you have this precursor? Then they say that we need to clean. They get a call the next day, it's for the clean. Well, we got both.
I would tell you, Todd, and maybe I shouldn't go down that path, but when we're doing the due diligence on ATMI, I was very skeptical about the appeal of the deposition materials. I really thought it was mostly a commodity business, and I was dead wrong. I would argue today that we have some really, really exciting opportunities on that side of the business, and I like to be proven wrong that way, believe me. Now we need to put the pieces together. Jim gave you a glimpse at how we're doing that. That's actually a very, very nice and very sound business.
I think we're almost out of time, but if there's one last question we can take. Otherwise, we'll wrap up. Okay, Jerome.
You mentioned on 2014, you mentioned capacity constraints hurt you. Now with the membrane i2M facility being done, is there a way to quantify how much it hurt and how much you could benefit this year?
Do you want to comment on that, Todd?
I don't know if I could. I don't think we lost any market share during that, and the customers hung with us. We tried different combinations of products, tried to keep them satisfied while we brought up i2M Center. Now it's up. It's got almost all of the membranes of volume qualified, and we're late in the PCN process where their customers are starting to convert to supply from there, so we're starting to see that relief. There's been other areas we're being constrained as well. We've been constrained in our chemical drum business, but we've added capacity that we're quickly filling up, actually, in Minnesota. We're going to be adding more capacity in Asia very soon because we're full there. There's a lot of demand on some other parts of the business that are going to start to free up as well.
I don't think we lost any share in it. I think we have some opportunity, actually, to grow. We see some good opportunities for, you said filters was the gem in your eyes. I like it, too. I do see that growing as we unlock that capacity and just because of the demands of these advanced chemistries and resists going forward.
To be clear, I think that the filtration opportunity is probably one of our largest growth opportunity going forward for the next couple of years. It doesn't necessarily only have to do with removing the capacity constraints. It has more to do with the adoption of that particular UPE media into a number of applications. All right. Well, again, thank you all for being with us and for the questions. That was actually very nice. We enjoyed the dialogue. Thank you.
Thank you very much.