Connected world.
All right. Morning, everybody. Welcome. Thanks for being with us here today. We really appreciate it. Also in a great location, Nasdaq, here in Times Square. Hey, on behalf of myself and the new management team at Lattice, we're really excited to share with you some of the opportunity that we see for Lattice moving forward, and very happy to share some of our long-term plans for the company. We have a good crowd here. I really appreciate those that showed up in person. There's a number of people here who have known me actually for many years. I appreciate that, too. Actually, if you know me well, you know that one of my hobbies is home improvement. I like to do plumbing, electrical, I'll do tile work, hardwood floors, all sorts of stuff. It's kind of the way I relax on the weekend.
A number of years ago, my wife and I bought the house that we're currently in, and we bought it for a couple of reasons. The first one was, it's in a great location. It's in the Santa Cruz Mountains, has beautiful views of redwood trees, that sort of thing in Santa Cruz Mountains. Beautiful location. Second thing was, it's got nice architecture, nice layout. Some people would say it's kind of got nice bones. The one issue with the house was, it needed to be remodeled. It was a bit out of date. No problem. I like to remodel, so I was happy to do that. Now after the remodel project, it's a great house in a great location. The house reminds me a little bit of Lattice when I started. Lattice is a company that's, first of all, in a great location.
It's in a great part of the semiconductor market. If you look at the semi market overall, there's commodity parts of the semi market, and then there's high-value parts of the semi market. Parts of the market where there's high complexity, high degree of software content, high barriers to entry. That's FPGAs. FPGAs are part of that industry. Lattice is certainly part of that part of the semi industry. It's in a good neighborhood. Secondly is it's got a really strong core franchise. It's got good bones. When I joined Lattice, this was pretty clear just within the first few months. I've had the opportunity in my career to work in a number of different types of businesses, high-end CPUs, high-end graphics processors, DSPs, et cetera. I can tell you after a few months at Lattice that this business is incredibly resilient, very strong core franchise.
Lot of opportunity in front of us. It does need a bit of a remodel. We started our remodeling project about nine months ago. We're already showing some improvement, but we've got some more work to do and happy to share the blueprints of that plan with you here today. Okay? As I kick off here, I thought I'd start by answering the most common question that I got asked when I joined Lattice, which was, "Hey, why did you join? What's the opportunity in front of the company?" I thought I'd start with that. First of all, and I mentioned this a little bit already, Lattice is positioned really well, not just within the semiconductor market, but positioned really well in terms of end markets. This is very important to the foundation of the company.
It's positioned in great end markets that are large, they're growing, they have secular growth trends that will drive FPGA consumption or semiconductor consumption for years to come. If you look at the size of our market opportunity versus our revenue base, plenty of room for us to grow in the coming years. Next is, we have very differentiated product line. Our products are very different than what others in the industry, FPGA industry are doing. We're focused on low power efficient FPGAs. That's very important. Small size, easy to program. That's very important to a growing number of applications in the industry. Applications that require that power efficiency, that small size. Things like edge computing, artificial intelligence at the edge, IoT, industrial IoT, big secular growth trends that are underpinning some of those end markets I talked about.
As well as 5G infrastructure and servers. We'll talk about these applications today. One of the other strengths of Lattice is a very strong customer base, very diversified customer base. Global distribution channel, but also a position with world's top OEMs. The OEMs that really drive the market, drive an incredible amount of semiconductor spend. Lattice has, in a lot of cases, multi-generational position with these OEMs, and opportunity to grow more. When you start to put that all together, what you'll see today is a company that's got great opportunity to expand profit moving forward. Not just top line revenue growth, but gross margin expansion and EPS growth. We'll talk more about that today. We have been making a lot of changes over the last, say, six to nine months to Lattice, and you'll hear about these changes throughout the day.
What I thought I'd do is just kind of summarize some of those key changes right up here at the front. Number one, and this is probably the most important change to understand about Lattice, is we are 100% focused on FPGA now. That's 100% focused on our core franchise. That means all of our R&D, all of our sales, all of our marketing focused on FPGA. That was not the case a couple of years ago. We're also trying to support other ASSPs from our Silicon Image acquisition. That's behind us. All of our effort is focused on our core profitable franchise. Next is we've shifted our focus in terms of where we're sourcing our revenue moving forward. Our sales team is really focused on driving sustainable multi-year revenue streams.
In the past, we were subject to some volatile revenue streams, handsets, et cetera. It's not the case anymore. We're very focused on building multi-year revenue streams. We've done a lot of work on the product roadmap. You'll definitely see that today as management team presents. We've improved the cadence of products that'll be coming out. We're speeding up the cadence of new products, investing more in system-level software, system-level solutions. Some significant changes on the product roadmap. Underneath the hood, we're revitalizing the culture. This maybe isn't easy to see from the outside looking in, but it certainly affects the performance of the company. Cultural changes like just speeding up the beat rate of the company, having the company operate at a much faster pace, driving accountability, driving a performance-driven culture. This certainly affects our results. That feeds well into financial discipline.
I think you'll also see a greater focus on financial discipline, consistent profitability, consistent cash flow generation, cash flow expansion. Certainly a big change for the company has been the new leadership team, which you'll meet today. Re-energized leadership team that's kind of re-energized the whole company. This is a leadership team that has very deep industry expertise, not just semiconductor expertise, but very deep FPGA expertise. We're pretty excited about these changes. We'll talk more about this throughout the day. Our mission moving forward, very clear, very straightforward. It's to be the low power programmable leader. What that means is being the leader in FPGAs that are very power efficient, small size, easy to use, very power efficient.
The reason that mission is critical to the industry is because there's a growing number of applications that require not just programmability, but power efficient, small size, easy to program. Again, things like artificial intelligence at the edge, IoT, industrial IoT. This is a mission that look, we're great at this. We're the world's best at power efficient FPGAs. We've been doing this for 35 years. We know how to do this, so this is where we're focused moving forward. As we take that focus to the market, the five main markets that we're focused on are shown here. Collectively, those represent about a $3 billion market opportunity for us moving forward. These markets, very healthy, underpinned with a lot of new secular growth trends. Growth trends that'll drive consumption of semiconductor and FPGA content for years to come.
Examples of that would be in the communication space, 5G wireless infrastructure. This will certainly be a growth vector for the industry for years to come, but also for Lattice, and we will talk specifically about that today. Industrial automation, so the automation of factories, more intelligence within the factories. Automotive electronics growing very fast, both in gas and hybrid electric vehicles. Lattice's small, low power focus, very important to those type of applications. We've got good, healthy end markets that we're positioned in, and relative to our revenue base today, plenty of room to grow. We're bringing to the market a very well-rounded portfolio. First of all, we have a general purpose product line that allows us to address a number of different applications across many different markets and applications within those markets.
Beyond that general purpose product line, we have specialized, customized product lines that are really specialized around specific functions, specific tasks within those markets. Things like video connectivity, control, and security. A nice, well-rounded product line. If you look underneath the hood at where are we investing, where are we investing our R&D is, it's very different than our competitors. Other FPGA companies are investing in very large FPGAs. This is where their R&D is focused. Large FPGAs, very high power, for all sorts of applications like, for instance, data center compute applications, accelerators and data center. We're focused on the other end of the spectrum. This is our specialty, is small, power efficient, 10 millimeter on a side down to one millimeter, 1.4 millimeter on a side, milliwatts of power. That's very important for a number of applications.
Again, applications where power efficiency, programmability are important. Some of those growing applications are shown here. These are a lot of the edge computing applications that we see growth in today. Artificial intelligence, inferencing at the edge, embedded vision, robotic motor control, precision motor control, and industrial automation. These require not just programmable solutions, but power efficient, which is where we specialize. If you look at, as we approach these markets, these applications, how does our customer base look? If you look at our customer base, we have a very strong, very diversified sales channel. This is something, part of this I knew coming into the company, which is that Lattice has a very strong global distribution channel. It actually lets us support 9,000 customers. Tremendous diversity of revenue stream, diversity of customer base.
That really provides a lot of ballast in the boat for Lattice and makes the business very resilient. The thing that I was surprised about when I joined Lattice was the extent of the OEM relationship. If you look at the largest OEMs that are really the market makers that drive a tremendous amount of semiconductor spend in the industry, and a tremendous amount of FPGA consumption, Lattice has great relationships with those top OEMs. We've been in business for 30 years. We've built really strong relationships with those customers. I believe we're under-penetrated at those customers. There's a lot more opportunity to grow, to gain more share of wallet at those top OEMs. You'll hear more about that later today as well.
With these OEM customers, one of the most important things is the product roadmap, because anybody that's worked in semiconductors knows that OEMs don't buy point products, they buy roadmaps of products. We have been doing a lot of work on the product roadmap, really tuning up the product roadmap. I wanted to share a couple examples of that here with you. If you look at where are we focusing our R&D, there's been a couple big changes over the last year or so. The first change is just which projects are we investing in. If you look at where were we spending our R&D, say, beginning of 2018 versus the beginning of this year, a couple of big changes. First of all, we canceled and shut down any investment in non-FPGA projects.
We can focus 100% of our R&D on FPGA. Within FPGA, we're optimizing the portfolio within FPGA. We shut down low ROI products, projects, and funneled that investment into high ROI FPGA programs and driving faster time to market, a faster product cadence. That's one important change. You'll hear more about that later today. The other change is we're investing more in the software that sits above our FPGAs, software system-level solutions. What this does is this allows a customer to take our products to market quicker, faster. They can rely on our reference platforms, our software libraries to get to market much quicker. A couple of big changes that we've been making in the product roadmap. If you look at the roadmap, moving forward, it's very healthy.
I want to share with you some of the products that we're bringing out just over the next 12 months. We've got a good, solid pipeline in front of us. The team is excited about this, the rollout of these products that are coming over the next 12 months. First one, which we're launching today, is our second generation of sensAI. sensAI is our software stack for artificial intelligence inferencing at the edge of the network. It's a software library that customers can use to design our products into applications at the edge of the network that bring intelligence to edge devices. We've expanded the capabilities. We've significantly improved the performance versus our first generation. Steve will talk more about this later today, but this is a great example of where we're investing more in software moving forward.
Another product that we're launching today, this is a really important one, MachXO3D. We took one of our most popular FPGAs, and we added to that new security technology that allows these devices to be used as a platform root of trust in a hardware platform. Providing very foundation level of security for hardware platforms. That could be for all sorts of different platforms. It could be for servers, for client devices, for networking platforms. We started sampling this product to OEM server customers late last year. We're really happy with the progress on this product and anxious to get this ramped into production. This, we're also launching today. In the second half of this year, we're going to introduce another new product, CrossLinkPlus. This is an enhanced video bridging solution.
This is used by our customers that have to support many different types of displays. Displays for industrial applications, for automotive applications, consumer, et cetera. A pretty innovative new product. Again, Steve will talk more about this later. Probably the one that's got the most excitement internally, both the R&D team is really excited about this, but also the sales and marketing team, and that's our next-generation FPGA platform. Our next-generation platform, this is a complete grounds-up rebuild of our architecture, as well as it's on a new process technology. We taped out the first in a family of devices we taped out in Q1. We have the silicon back now, silicon looks very healthy. We'll build from this new platform generation, a series of a family of new products that'll extend our capabilities, that'll extend our features, et cetera, and our competitive advantage.
It's not just added new architectural features, but it's built on FD-SOI technology, which gives us a big performance per watt advantage, so it's a big power saving. Really competitive new platform. Really excited about this. You'll hear a lot more about this today. If you look at the product pipeline that we've got in front of us, the market position, the position we have with our customers, you start to put that all together in terms of what does that mean for the business moving forward. First of all, in terms of revenue growth. Revenue growth, we're expecting the next couple of years to be in kind of the single-digits range, and that's really kind of 2019 and 2020. Beyond that, in 3- to 4-year range, accelerating to double-digit growth. What drives that double-digit growth?
It's a number of different growth vectors. 5G infrastructure. We're very well positioned in 5G control plane applications. That'll drive growth for years to come. Positioned well in terms of servers, the server generation, and the products that are used in both server platforms as well as client platforms. In a number of industrial automation applications, automotive electronics. We'll walk you through a number of these specific growth vectors throughout the day and give you more details on each one of those. What really kind of drives the inflection point from single to double-digit growth over the coming years is those new products that I just talked about on the prior chart.
Those new products that we'll introduce over the next 12 months really drive the inflection point, as well as some new sales initiatives that Mark will talk about later that we put in place to both expand our design win pipeline, but also drive faster time to market in our design wins. Good, healthy growth over the coming years that we're anticipating. Beyond revenue growth, gross margin expansion. We believe there's significant opportunity for gross margin expansion. I've been pretty consistent about this since first joining Lattice, that I believe the company should be driving higher gross margins. If you look at from 2017 to 2018, gross margin improved by about 1 percentage point, so a little bit of improvement year-over-year in 2018. In our most recent quarter, Q1 of 2019, we were operating at gross margins above 58%.
I believe that there's no reason the company can't operate and sustainably above 62%. That's going to take some work to get there, but we believe that there's, again, there's no reason this business shouldn't be operating at a much higher gross margin. A number of strategies we've already put in place to drive gross margin expansion. Number 1 is pricing optimization. This is really about ensuring we're getting the right value for our product in the marketplace, that we're getting the fair value for the type of product we bring to the marketplace. This is a strategy we built out in Q4 of last year. We started implementing that in Q1 of this year, started seeing some initial benefits in Q1. We expect that to continue to benefit us moving forward. Another one is product cost improvements.
We built the plan in Q4, started implementing that in Q1, started seeing the benefits of that in Q1. We believe we'll benefit from that moving forward. As you saw in the last chart, there's a mix shift that we're anticipating in the business as well. We're expecting most of our growth to come from comms and compute, industrial, automotive, and as our mix shifts towards those segments, that helps drive gross margin expansion as well. Okay. If you put that all together in terms of a long-term target model, revenue growth, low double digits, gross margin above 62%. Our target is to try to get OPEX right around 35%. We think that's the right sort of steady state model for the business. That puts our operating profit in the 25%-30% range.
If you look at 2018 as a reference point, 2018, we were at about 17% operating profit. This is a significant expansion in our operating profit over the years to come, and that'll obviously drive all sorts of EPS growth. We're really excited about this long-term model. This is our focus moving forward. With that, I want to just kind of summarize a little bit, maybe go back to my remodeling analogy. We started the remodeling project about 9 months ago. We've made some good progress already. We've got a lot more work to do, but what you'll see today is really the blueprint of where we're headed over the coming years in terms of that remodel. We do believe we've got all the right ingredients within the company.
We're positioned in great markets with healthy growth trends, got very differentiated products, very differentiated technology, well-positioned with customers, and I think we've got a lot of opportunity for expansion of profit moving forward. Okay. With that, I want to introduce the management team that you'll hear from today. First, you'll hear from Esam Elashmawi. He's the head of marketing and strategy for us. Before Lattice, he worked for Microsemi. He ran a number of different businesses at Microsemi, including the FPGA business. I think he's been in FPGAs for over 30 years. I think he started as an intern in FPGAs, actually. We can ask him about that later. He'll talk about where we're focused in terms of what markets we're focused on.
More importantly, he'll walk you through the big applications that are driving some of our growth over the years to come and explain to you why and how we're winning. You'll hear from Steve Douglass. He's our head of R&D. Most recently, he was at Xilinx. He has over 35 years in the industry, decades and decades of FPGA experience. He heads our R&D. You'll hear about the changes that he's making to our product development strategy, he'll also walk you through those new products that I showed you. He'll show you some more detail underneath each one of those and why those are competitively differentiated. Mark Nelson is our head of sales. He joined us at the very beginning of this year from Intel PSG, where he was head of Altera sales before that.
He'll talk about our customer engagement strategy and talk about the design win pipeline that we're building and how we're accelerating that. Sherri Luther is our new CFO. She started at the beginning of the year as well. Decades of finance experience, a CPA. She was most recently at Coherent, where she spent over 15 years, part of the finance team that drove a pretty extensive expansion in the business at Coherent over those years. Glenn O'Rourke will join us for Q&A. He's our head of operations. He joined us from Xilinx in December. He drives global operations for us. Again, decades of experience in FPGAs. Before Xilinx, he worked for Lattice. He's a returning Lattice employee, so we're happy to have him back on the team. With that, I think I'll hand it off to Esam. Esam?
Thank you. Thank you, Jim. I also want to take this opportunity and thank all of you for being here with us in New York. Jim talked about my first job in the industry actually was an internship at an FPGA startup. That's what it was. In the late '80s, we were startup companies at that point. The industry, or FPGAs, evolved quite a bit since then, and it is in a good part of the semiconductor market for all the reasons that Jim talked about. Now, in those 30 years of being associated with FPGAs, if you go back and ask some of my old team about the amount of amazing talent that I've been able to work with, they'll tell you, if you ask them, that the best hires I've ever done in this industry were employees that came from Lattice. They'll tell you that.
They'll also tell you the reason why I used to say that is because the Lattice employees, not only were they hungry and they did not take incumbency for granted, they also understood one of the basic fundamental principles about success in the FPGA market. What was that? They understood that it wasn't about the biggest, largest, fastest FPGA. They understood that in every end market, and the vast majority of applications had to have the small FPGA. They had to have the small FPGA. Our industry wouldn't survive without those small FPGAs. They also understood that those small FPGAs required innovation or modernization to the features and attributes and capabilities that today's systems require. They understood that, and that's really, really important. When I started off in the industry 30 years ago, we couldn't imagine today how important low power was.
We didn't design for low power. We didn't design thinking security when we started building our FPGAs. We didn't realize how many different variations of memory devices are going to be out there, and you had to evolve on the interfaces, whether it be DDR2, DDR3, LPDDR3, DDR4. You had to innovate on the small FPGAs, and the industry required it. That was something key that I always found that existed with the Lattice employees, and it's true. Now, if you think about over the past 30 years, the value proposition of an FPGA, it hasn't really changed. The value proposition, the fundamental reasons why people choose FPGAs has not changed. What has changed, if you ask me personally, I tell you that the demand for these value propositions has never been stronger than it is today. What fueled the FPGA market was the comms and industrial.
We didn't start selling FPGAs into the automotive industry until over the last decade, and now it's just increasing because of the value propositions that FPGAs bring. I remember visiting a premier automotive company a little over a decade and a half ago saying, "You can use FPGAs." They're questioning, "Well, can we use FPGAs?" "You can." Today, it's a norm in the automotive industry. In the compute and data center market today, it's a norm to have an FPGA there. The value propositions have never been stronger and in all the end markets as well. You know the push in the comms for 5G. You know what's happening in industrial with factory automation. The compute market is adopting it. The automotive market has adopted. Every end market is now adopting FPGAs.
You take something like an AI, which is an application that goes into every single one of these end markets, and you see the value proposition of an FPGA being even stronger. In something like AI, where it's fragmented, it's a race to market. That's the beauty of an FPGA. If you're doing an AI type application, your secret sauce, how you develop your inferencing, that's your secret sauce. You want to have an FPGA for that. You want to differentiate. You don't want to be doing the same thing that everybody else is doing. You may want to even deploy your system and upgrade that in the future again. The value proposition of an FPGA does that for you. The supply chain, knowing that you're buying a part that's right off the shelf that many others are using, that's key. That's a value of the FPGA.
The parallel processing brings you performance as well. Never have I felt over the last 30 years the importance and the need for these value propositions as I do today with all of the end markets that we've got. Now, with that said, what does that mean for the FPGA market and the growth in the future? Clearly, the FPGA market is expected to outgrow the semi portion of the market, and it's not a surprise for all the reasons that we just talked about. According to IHS, it'll outgrow it by 2X, and we'd expect that as well. The next question you should ask me is, okay, that's the FPGA market. It will outgrow the semi growth by about 2X. Where does Lattice play within that? If you take the FPGA market in 2018, it's a $5.6 billion market.
For simplicity, I'm going to segment it into two groups. The first group is where Lattice plays, which is the $1.6 billion of that $5.6 billion. That's the portion of the market that requires that small FPGA with innovation. $1.6 billion of the $5.6 billion total TAM of the FPGA. That's our SAM. The rest of it, the $4 billion, think of that as the larger high-power FPGA, a simplistic approach. $1.6 billion TAM in 2018, what does that look like moving forward? If you go forward, what you find is that our SAM almost doubles from $1.6 to $3 billion. Why is it doubling? You've got the traditional FPGA market that's growing, and we talked about the CAGR of that. Also over the last year or so, Lattice has introduced capabilities that allow us to have SAM expansion outside of the traditional FPGA market.
When you think about AI, we just talked about all the value propositions of AI. We're not selling and taking sockets away from other FPGA companies. We're taking sockets away from microcontrollers, some ASSPs, some SoCs. That's SAM expansion for us. When you think about, and we'll talk more about security, the capabilities that we're putting in our chip for security, it's not just about securing your design in our FPGA, it's about implementing security capabilities for your system. We're not going after the regular FPGA socket. We're actually creating new sockets that FPGAs couldn't do before. That's SAM expansion. You think about the technology we have around embedded vision, video connectivity, that every system now, whether it be in a factory, in an automobile, in prosumer type goods, there's a screen you need to connect to it. We're not going after the traditional FPGA designer.
We're going after new sockets. We're replacing other ASSPs that are doing video connectivity. You could use a larger FPGA with modernized features to do that, but it's at a very high cost. You need that in the small FPGA. That's what we're doing. You end up with about a doubling of our SAM over the next few years. The next question you should ask me is, well, how does that play into your end markets? Let's take a look at the end markets. The largest one that's going to benefit from that is what I refer to as the value-based end market, is the industrial and automotive market. Very sticky, long annuity to it. It's a beautiful market to be in. Good margins in that market. Followed by comms and compute, which is a good, healthy market.
Consumer, which we've already shared with you that we're very selective of what we pick in consumer. We want it to be sticky, multi-generational, and there's a lot of applications like that in the consumer market that value those attributes in a small FPGA. I'm going to dig into each one of these end markets in more detail. Conventional wisdom says you all, we should start with the industrial and automotive, it's the largest, but I recognize I'm not going to get away without starting with the comms and compute and 5G. This is the one that everybody tends to talk about and ask about. We all know what's happening in the comms side of our comms and compute around 5G, massive build-outs, refresh of equipment. The question is: Where does Lattice play within the 5G infrastructure? Let's take a look at that.
In 5G, we're going to start off with a base station system. And in a base station, you have a remote radio unit, and you have the baseband unit. Let's take the remote radio unit as an example first. Within a remote radio unit, there are two major functionalities, and I'm going to simplify this again. What are the two major functionalities? There's what we call the control plane of a system, and there's the data path. The data path, think about that as that's the fastest signal that has to go through there. That's what's transmitting the signal. That's what's going to your phone. It's going to the back offices. It's doing the radio processing. You need really fast performance. You can't have that fast performance without a control plane that's managing the entire system.
What turns on, what turns off, what features should be done at the right time? You need a control plane around that. Lattice plays in the control plane. The other FPGAs you hear about play in the data path. Two distinct functions. In the control plane, we do things like power management. We do bridging. We do control functions. That's where we are in 5G, and that's a critical aspect of 5G. You need the small FPGAs, the low-power FPGAs in every end market and the vast majority of applications. That same application that I just described in remote radio unit applies to the baseband unit, applies to the micro cells, small cells. You'll find us across the entire infrastructure for 5G. The next question we typically get, well, how does your 4G footprint compare to 5G footprint? Let's take a look at that.
In 4G, we also focused on the control plane. Between 4G and 5G, we raised our value proposition to our customers, and we have really good coverage with the top OEMs, where we're doing more than just a control function in the control plane, but we're doing power management and bridging functions as well. If you take a look at that with the existing design wins we have, with really healthy coverage across the OEMs, that translates to at least about a 1.3x increase in $ content. Let's move on to the other portion of the communications and compute end market, and let's talk about data center or our compute segment. We all know that the data center build-outs are being driven by the need of big data, the cloud that's happening there, and we play in that as well.
Let's talk about where does Lattice play within that, and I'm going to take an example of the server market. In a server market, there are platforms that get released. It's important to understand that we are processor agnostic. Doesn't matter which processor they're using in a server, we're processor agnostic because we do a lot of stuff around the control plane. In a typical server platform, let's take, for example, current server platform. You can refer to that as the Purley platform. We all hear about Purley platform and Intel. The prior server platform, you can think of that as Grantley. If you take a server, you've got key functionalities that happen in a server. We're on the control side again, and you might hear that there's other FPGAs in servers. Those are the accelerators. You can have FPGAs, GPUs. We don't play there.
We're on the control side of the server because every application requires small FPGAs as well. If you look at what we did in the prior platform, we did simple control stuff around power management and controls of IOs. You'll also notice our attach rate, which refers to all the companies that build servers. Our attach rate was 25%. If I normalize that from a $ perspective, let's call that $1. If you come to today's platform that's in production, what you'll see is we've added more value to the customers. We started putting system status monitoring functions there. Our attach rate went from 25% to over 80%. Because of the higher value, the ASP went from 1x to roughly about 2x. What do we see for the next platform that's going to come out?
We're working with our customers. We've talked about we're in an EAP program, early access program, with key OEMs on that. Everybody's preparing for the next platform. We've also added more value proposition to that market or to the servers as well. How? We're solving now new security requirements that are coming up in the industry with our MachXO3D that Jim said we're announcing today. That allows us to get more value from the customers. We're going to continue to strive for a similar or higher attach rate. If you take a look at what that means from an ASP perspective, we've almost tripled our value to the customer compared to where we were in the prior platform. That's an example of how we're growing, what's fueling our growth in the future.
Let's move on to the value base or the portion of the end market that I really enjoy, which is the industrial and automotive. Very sticky, lasts for a very long time. We all know what's happening in factories. The economics makes sense for factories to automate as much as possible. There's a lot of AI capabilities going into factories. It just makes sense for them economically. It's also an end market that we've had lots of good growth and design win momentum. We talked about our 2018, our industrial grew by about 18%. It's a healthy market to be in for us. We've got a lot of momentum in this market. There's also the automotive portion of this end market for us as well.
Let's dig into these in a bit more detail and see what's fueling the growth for us today and what's driving it for tomorrow. If you take a typical factory, there's lots of applications that FPGA solve for a factory. I could fill this screen with stuff that's fueling the revenue for us today. I decided to take some of the key ones that are relevant and that matter, and that we're all familiar with. Motor control. Every factory requires a motor. FPGAs, the value proposition of FPGA, it gives you precision because you can do things in parallel. Motors consume a lot of power. You want low power FPGAs to do motor control. Motor control is something that FPGAs play into. They've been playing into for a while, and it fuels our revenue that we've got today. Think about industrial network. You've got old equipment, new equipment.
How do you interface old with new? They're different interfaces. FPGAs are ideal for that. Sensor bridging, lots of aggregation of data. How do you pre-process that? What do you do with all the data that you collect out of a sensor before you send it to an application processor? FPGAs do that. There's a lot of things that are driving our growth today. What do we see driving our growth tomorrow? How are we going to get to the targets that we talked about? There's a lot of new applications coming in factories today, and these are examples of applications, and there's more, that actually we're starting to see ramp up, ramping up, contributing some of the revenue, and we expect these to continue to grow. Package detection, collision avoidance.
You've got a lot of autonomous robots moving in factories. We've been open in the past to talk about if you've seen the Amazon robots, you'll see Lattice inside, and that's replicated for other types of robots as well. Predictive maintenance. There's lots of gears and belts are going on. Do you wait for an equipment to break or you use technology to do predictive maintenance on when things start to wear out? FPGAs are being used today for that. Lattice FPGAs are being used to do that today. There's a lot of stuff, including the embedded vision that we talked about, which are new markets we're going after that are fueling the growth of our industrial end market. Let's look at the automotive market. Again, I remember a while back where we had to convince this market to use FPGAs. Today, it's a norm.
If you've got an FPGA, let's see how I can use it. With our value propositions of small and low power, and you think about the security that we're adding as well for our FPGAs, this is a market that's going to be healthy for us. It's also one of our fastest-growing markets. If you look at a vehicle today, we solve things around infotainment, things around display. We all talk about ADAS. Lattice is in ADAS as well. In fact, on TI's website, it's public. They've got a reference design, and they reference Lattice in that reference design. As they win on those, we win as well. There's a lot of things that we do from an automotive perspective, but there's a lot of content, electronic content being added into vehicles today. Why is that the case? Because automakers want to differentiate.
It's about the electronics that they put in vehicles, not about how fast the car can go now. It's about what features can I bring into my vehicle to differentiate. There's a lot of other applications that are driving our growth in the future. EVs, electric vehicles. One of the largest or the largest electric vehicle company in the world that make everybody else look very tiny actually exists in China. We don't see those vehicles here from a A consumer can't buy them. You can see them at Google campuses and other places. It's BYD. In BYD, there's not a single vehicle that coming out of the production line does not include at least one Lattice device on it. If you look about what's happening in technology with e-mirrors, one of the large OEMs in Germany, the side view mirror isn't a dumb mirror anymore.
It's actually a display that's showing you an image. Lattice is inside that. The rear view mirrors are no longer dumb anymore. I believe there's one here on the display as well. That's Lattice inside of that. It's not just display, it's about cleaning the display. If it's nighttime, do you want to see bright lights or do you want to see a clear image? If it's foggy, we can defog it. We can do pre-processing on the image so that the display gives the user an enhanced experience. Driver monitoring. If you think about in the shipping industry, commercial vehicles, people want to monitor the drivers. Are they dozing off, not dozing off? You'll find Lattice inside those applications as well. A lot of stuff that we see that's driving the growth for us today and in the future. Let's move to the consumer market.
This is a really good market to be in if you're selective about what you choose to play in. We are selective. We look for things that value an FPGA value proposition that are sticky and multi-generational, there's a lot of that out there's some on display here today. What are the type of things? Well, high-end projectors, soundbars. There's a lot of stuff that actually fuel our revenue that we've got today. When you think about what's happening in the homes or other consumer type goods around smart home, security cameras, object detection, voice detection, the dots, the spots that are out there. I heard a voice, I heard a command. Smart thermostats, smart appliances. That's where you're going to find small FPGAs because they are needed in our industry, that's what's fueling our growth in this segment as well.
When you take all of what I just went through, what does that translate to? Well, we're excited of the fact that we've got an opportunity to double our SAM. That wasn't the case a few years ago. That's the case today because we're adding capabilities into our product that we're not just going after the traditional FPGA design anymore, you're going to see examples of that when Steve presents. We're excited about the applications we're in and the applications that we're building for the future. Personally, I'm really excited the fact that FPGAs today are more of a necessity than they have been in the past. With that, I'm going to invite Steve to come up and talk about our products. A really quick one. Steve's been in the FPGA, this industry like me, 30 plus years.
In those 30 years, by the way, we call it a small community. We know of people. We meet people. I've never actually met Steve prior to joining Lattice, but I've heard his name quite a bit. I'm glad to have him on our team. One of the nice things about Steve is he did a stint within marketing. I've done stints within engineering, that makes for excellent collaboration because there's a feeling of the challenges that we both have. It's always good to have somebody that knows what we got to deal with, and I know what he's got to deal with. I'm really pleased to have Steve on our team. Steve?
Okay. Thanks, Esam. It is great to have a guy in marketing that understands what goes on in R&D, I do have a few notions of what his challenges are on the business side. Good morning, everyone. I'm Steve Douglass. I'm the new R&D leader here at Lattice, I have been in the semiconductor industry for 35 years. The very first project I worked on was a chip called the 80386 microprocessor. You may have read about that in history books. You might have even seen it in museums. In fact, it was actually featured here a few blocks away in the Museum of Modern Art. They had a full wall-sized part of our chip hanging on display, I saw it for myself in 1988. That was a while ago, but it's been a great journey. I really enjoyed a few years at Intel.
I heard about this really exciting new technology called programmable logic, where customers could define and redefine the functionality and make a chip do anything they wanted it to. As a young new engineer, I was really captivated by this technology and the possibilities that it could bring to the marketplace. From that point on, I was hooked. I spent the next 32 years in programmable logic. I spent 12 years at Cypress trying valiantly to compete with Lattice, not always too successfully. I spent 20 years at Xilinx, and at both companies, I had both R&D roles and business leadership roles. I got to tell you, it's been truly amazing to see the perpetual transformation and revolution of not just the technology of this sector of the market, but also the business side, too.
The possibilities of FPGAs and what they can do for customer system is greater now than it ever has been before. I got to tell you, I'm really excited to be a part of the Lattice team to capitalize on these opportunities and drive growth for the company going forward. When I first joined Lattice last fall, there were a lot of things I saw at the company that I really liked. We had a talented, dedicated engineering team that had a really strong core of FPGA design experience. I saw a product roadmap that was very customer driven. It was really focused on solving key customer issues in certain applications that other FPGA suppliers weren't really addressing. The products themselves were highly differentiated, and it was driven by innovation from the R&D team.
I thought this was a very strong base for the business and really excited of what the capabilities of our core engineering team were. However, I did see several challenges. We had way too many projects going on at the same time. The team was spread way too thin to be successful in delivering these products on time. There were multiple product architectures that were being developed independent of one another and very little sharing of design information from product to product. I also saw a lack of system-level solutions that could help our customers implement their designs more efficiently and effectively. I worked with the team, and we came up with a strategy that would address all three of these challenges. As Jim mentioned, we are now solely focused on FPGAs. It's what we've been doing for over 30 years.
It's what we're really great at. It's what we'll continue to do. We're now doing platform-based design, I'll explain that in a little bit. It's a very efficient way for us to implement multiple product families off of a common product architecture. We're also expanding our investment in software and soft IP to create more complete solutions to allow our customers to implement their designs more efficiently in our products. Let me touch on each one of these elements of the strategy in a little bit more detail. First of all, the most important element of the strategy is focusing on FPGAs. As Jim mentioned, when the company was designing products a year ago, they were working on way too many products on way too many diverse technologies, required different skill sets, different design methodologies. It was very inefficient.
We decided to cancel all of the non-FPGA products and focus solely on FPGAs. Within the FPGA projects, there were a lot of projects that were really very narrowly focused. They were spot opportunities in the market. They were rifle shot products that really didn't have an opportunity to influence our broader market capability. We canceled those projects as well because they weren't really going to deliver the ROI that we needed to grow the business. We focused the team on a handful of highly impactful FPGA design projects that could really help drive the business going forward. We also decided to accelerate our investment in software, after we had made all of these changes, we are still within our financial model, which is to have an R&D spend of about 20% of sales.
We think that this is really the right level of investment that we should be making to serve the markets that we're trying to address with the product portfolio that we're trying to build. The second element of the strategy is to do platform-based design. Platform-based design is when you have a single product architecture, from that architecture, you can build multiple product families that are optimized for different applications in the marketplace. Jim showed our product portfolio, we have a general purpose family, we have multiple FPGA families that are targeting specific types of applications and functions. This is the perfect strategy. I know from personal experience, it's the most efficient strategy as well for developing multiple families, targeting different areas of the market, and getting to market more quickly.
It takes a little bit longer to develop the platform and get that first product out, once you get that first product design done, the rest of the products and the other families come very quickly. Our overall investment is actually smaller, our time to market is faster, and we can deliver our products to the market with a much more predictable cadence of innovation. The third key area of our strategy was expanding our system-level solutions. What I mean by system solutions, it's really software and soft IP, which are pre-engineered solutions that help our customers implement their designs more effectively.
In the early days of FPGAs, it was sufficient to just give them an FPGA design tool that allowed them to compile their designs, a few soft IP building blocks, like a memory interface controller and a couple other FIFOs and other building blocks. They could build a design pretty quickly and get to market. Today, customer designs are far more complex, they really need more help from us to implement their designs more effectively and efficiently. We are building solution stacks, which are a combination of system software and soft IP that is targeting certain types of applications. These are the applications that Esam already identified as areas where we want to be targeting growth. Artificial intelligence, security. We are building more complete solutions in these areas so customers can build their designs more quickly and get to market sooner.
It will accelerate the adoption of our products, it gives us a certain stickiness, because when they made a commitment not only to our silicon products but our software tools and our IP, that's a stronger commitment from their engineering team. It's a stronger dependence for us, it really allows us to proliferate future business with those customers. This is a great strategy that's working very well for us, we will continue that strategy going forward as we build more solution stacks in other targeted application areas. We've also, by the way, revamped our FPGA design tools. We have introduced very recently our Radiant Design Suite, this was a grounds-up redesign of our FPGA tools that allow customers to implement their designs a lot more efficiently and effectively, they're going to be getting better system-level results with this new tool.
We feel that these investments are very important to our strategy going forward, these will accelerate the adoption of our silicon in the marketplace. Okay? This is our R&D strategy. We've already deployed this strategy, we're accelerating the execution of it, we're already seeing some excellent results to date. In fact, Jim mentioned that we have several new products that we're introducing over the coming 12 months, I'm going to talk about each one of these in a little bit more detail. The first one is the second generation of our sensAI solution. That is one of those system software solution stacks that I talked about that we're adding extra investment in. This is a very targeted area to help customers build power-efficient inferencing at the edge of the network. Here's a brief video to introduce you to our sensAI solution.
To build intelligent, autonomous devices, support for AI at the edge is necessary to address issues like data latency, privacy, bandwidth constraints, and cloud computing costs. AI at the edge requires smart devices to process data and take action in real time using onboard neural networks. Lattice Semiconductor's award-winning sensAI solution stack has the tools needed to implement always-on AI and FPGAs that meet the small size and low power consumption requirements of edge computing. Lattice sensAI, accelerating AI at the edge.
You just saw, we've all heard, that there's been this rapid expansion of artificial intelligence across all the markets. In virtually every market space, we're seeing smart applications, smart systems, across every possible application domain. We've got smart homes, we've got smart cars, we have smart factories. Everything around us is getting smarter. Hopefully, we are getting smarter too, to keep up with all of this intelligence that's building around us. The artificial intelligence explosion really had its watershed moment with rapid advancements in an area of machine learning called deep neural networks, or deep learning with neural networks, and the ability to train these networks with very large data sets. Now, machine learning really has two major components.
There is machine learning training, which is teaching a neural network to recognize certain patterns in data, whether it's a video, audio, or database information in a particular system. Then inferencing is the deployment of that network in an application to do real-time processing of the data and to do critical decision-making that's part of the operation of the system. With our sensAI solution stack, our customers can actually build inferencing engines that will allow them to implement artificial intelligence in their edge applications and target our very power-efficient, low-power FPGAs to get very power-efficient implementations of their algorithms. In this latest release of sensAI, we've increased our performance by a factor of 10 on things like real-time image processing.
The networks themselves are designed in a different software environment, software frameworks like TensorFlow and Caffe, We are expanding the number of software networks that we support. We have a very seamless interface to these software tools that allow our customers to design their network in these frameworks Then compile their designs with our sensAI design flow. We also have a lot of predefined reference designs that allow customers to see firsthand how to implement some of these inferencing algorithms, things like presence detection and object counting. It really helps them get a running head start on implementing their designs in those systems. I actually have an example here of using sensAI in a particular application on the edge. This particular application is a security camera that's embedded in a doorbell. You guys may actually have some of these at your homes today.
You want to keep track of who's stealing those UPS packages off of your porch, making sure that there are no intruders coming in. In this application, it's critical for the camera to be on all the time and for the algorithm implementing the human detection to be operating nonstop, twenty-four seven. You don't want to have any lapse in your security. The traditional way to implement this human detection algorithm is with a microcontroller. The microcontroller does not have the parallelism of an FPGA, It cannot do the computations as fast, and it can only produce about one to two frames per second of image processing. That's not very good for detecting fast-moving objects in view.
If you use sensAI, though, to target our Lattice iCE40 UltraPlus device, you can actually get up to five frames per second, which is much better detection capability, and it only burns seven milliwatts compared to the 100 milliwatts that are spent by the microcontroller. If you implement this human detection algorithm in our low-power, small form factor FPGA, you get a 70x factor improvement in the power efficiency. We're very excited about the possibilities of sensAI. We're seeing a lot of customer interest, and they're using this in a whole host of different applications. We're very excited to introduce that second generation sensAI today. The second product that we're introducing today is MachXO3D, which is a controlled PLD that has embedded hardware security features in it that allow it to be established as a hardware root of trust in any embedded system.
Here's a quick video to introduce you to the MachXO3D.
Securing critical infrastructure and smart devices against unauthorized access has never been more important, as the cost of compromised hardware can be disastrous. Every device on the network, from an industrial robot or autonomous car operating at the network edge to an enterprise-class server, uses firmware to control internal components. This firmware is an increasingly popular attack vector, and the desire to protect against those attacks inspired us to develop the MachXO3D family of secure FPGAs. They enable a robust, hardware-based root of trust that detects malicious firmware at boot, protects firmware during operation, and recovers compromised firmware to a previous safe state to keep devices running securely. Accelerate development of secure systems from the edge to the cloud with Lattice MachXO3D from Lattice Semiconductor, the low power programmable leader.
In this day and age of perpetual cybersecurity attacks and constant news stories about data breaches, security is top of mind for every one of our customers. With MachXO3D, we're able to help them solve their security problems. The MachXO3D is the first controlled PLD that has all the security features embedded in it to allow it to be established as the root of trust for the entire system. Since it's a controlled PLD, it's usually the first device turned on the board and it's the last device turned off, so you have secure protection the entire operation of the system. The MachXO3D has been designed to be compliant to the new NIST standard for the platform firmware resiliency.
This basically requires that you not only be able to protect against an attack, but you can detect an intrusion when it happens, and most importantly, you can recover the firmware back to a known good state. Many of the solutions out there today can do the protection and detection, but they cannot do the recovery. Not only can we recover the firmware back to a known good state, we can monitor and recover multiple firmware simultaneously. We've had a lot of interest with many customers in a lot of different markets, and as Esam and Jim both mentioned, we have sampled several OEM customers with this product, both server OEMs as well as hyperscaler data center customers. In fact, I have an example design of a hyperscaler customer who's actually building their own server.
You can see from the diagram on the left that there are multiple processing elements in this system. There is a processor that's running the base application software, there is an FPGA accelerator, and there's a board management controller. Each one of these devices has its own firmware. Each one of these is a target for attack. In their original design, they were using a control PLD for not only control functions, but the protect functions. They were using a microcontroller for the root of trust. The problem with using the microcontroller is it cannot do that rollback to a known good firmware. You cannot recover from an attack with this particular implementation.
By using a MachXO3D, our customer can not only combine the PLD and the microcontroller into a single device, reducing the overall power and cost of the system, they can also add this much-needed protection of a recovery to a known good state. This is just one example of how a particular system can be protected with MachXO3D, and it's not just limited to servers and client devices. It can be used in virtually any system that has an embedded processor with firmware. We are seeing a lot of interest across all markets in this particular product and very excited to be launching this product today. Another device that we're working on today is CrossLinkPlus. This is the next member of our CrossLink product family. These devices are optimized for video and display, they have the ability to aggregate multiple image streams from several different video sensors.
We can do up to five sensors aggregated at once, we have built-in hardware for very high-performance video interfacing. We can support not only standard display but high-definition displays as well. It has instant-on capability, it can boot up in less than 10 milliseconds. We're seeing a lot of interest in customers, not only in traditional consumer display applications, but infotainment in the auto area, as well as industrial, where there are a lot of displays on the factory floor. Again, I have a customer example of how CrossLinkPlus can be used to help customers with their implementation for high-speed video connectivity. This particular customer has a display application. They're building multiple different displays of different sizes and different resolutions.
The traditional way for designing these systems are to have an ASIC that translates the data coming out of the application processor and presents it in a form recognizable by the display. This is very costly, though, because you have to have a very high NRE to develop each one of the ASICs, it adds complexity to the manufacturing flow to have multiple devices. This customer realized that they could use one single CrossLinkPlus device. Because it's an FPGA, it's programmable, it can be configured to support all the different variations of screen size, display size, and resolution. This is just one example of the benefits of using CrossLinkPlus. Because we have that high-performance embedded video interfacing, we can support both standard and high-definition display systems. This product is in development right now.
We will be sampling it in the second half of this year. We will provide more information as we get closer to the availability of that product. Okay. Now we get to the fourth product, and probably the one I'm most excited about. This is our next-generation FPGA platform. I mentioned earlier the importance of doing platform-based design. It really is important to deliver our product portfolio in a very efficient, time-effective manner. By using a common single architecture to build multiple product families, we can rapidly deploy our roadmap, which has not only a general purpose FPGA family but multiple tailored FPGAs targeting different application spaces all off of a singular architecture. We have optimized this platform to be successful in all the target applications that we are targeting for growth.
Isam mentioned that we are looking at embedded vision, artificial intelligence, and security as application areas that can help drive growth for the company. Well, in this particular platform, we have embedded systems that allow us to have higher video connectivity performance. We have up to double the bandwidth from prior products and generations. We also have optimized this platform for AI inferencing at the edge. We've included even more advanced features in our DSP system blocks, we've included up to five times more memory on chip. It's very important to have a lot of on-chip memory, because if you can store the entire neural network and the computational results on chip, it saves a lot of time and power to go externally to outside off-chip memory to store data and retrieve it.
We feel that we're going to have much more higher performance and power-efficient implementations in the AI inferencing space. We've carried forward all the security features from our MachXO3D device onto this platform, so it too can be a hardware root of trust. Of course, our mantra, our mission, is to be the low-power programmable leader, we have optimized the architecture of this platform to be low power as well. We continue to use a very power-efficient, area-efficient implementation of our FPGA fabric. Our competitors have been using a more complex coarse-grain architecture of implementation, which may give them a little bit higher performance for their high-end, high complex, high-power FPGAs, that is not well suited for the applications we're targeting.
We are staying with a conventional four-input lookup table architecture, which is half the power and half the area of those competing solutions. Not only have we architected the platform to be power efficient, we are building it on a very power-efficient process technology. We have partnered with Samsung on a very innovative technology called fully depleted silicon-on-insulator. This technology is very similar to standard bulk CMOS. In fact, it has almost all the same processing steps. There's one very subtle difference. There is a very thin oxide layer that separates the transistor from the bulk of the wafer. This dramatically reduces the leakage of the transistors, it lowers the power by 50%. This power savings is additive to what we have architected for low power in the device itself.
In summary, this new platform not only has embedded system features to allow us to have optimized implementation of applications that we're targeting for growth, it also has been architected for low power and has been implemented on an innovative low-power technology. As Jim mentioned, we've already taped out the first product off of this platform, and we have seen silicon, and the initial results are very positive. I'm very confident we'll be able to deliver this product and sample it early next year. As we get closer to that timeframe, we'd be more than happy to share more information with you about that product. In summary, we're very excited about all the new products that we're introducing over the next 12 months.
It's a great rolling start to our product pipeline. We're very excited about the strategies that we're using to round out our portfolio and accelerate the execution of it. We are solely focused on building up FPGAs. It's what we've been doing for 35 years. It's what we're really great at, what we will continue to do, and excel at. We are doing platform-based design for a more efficient approach for building multiple product families, targeting different application areas from a common architecture. We will continue to build system-level solutions with software and IP that help customers build their designs more quickly and efficiently in our silicon hardware. I'm really excited about the potential of the company going forward, and I'm really proud to be leading the R&D team.
I think we're going to be an innovation engine that can really help drive growth for the company going forward. Okay. With that, I would like to introduce Mark Nelson, who is our sales leader. Mark and I never met each other until we worked together at Lattice, but we competed vigorously against one another when I was at Xilinx and he was at Altera. I got to tell you, he is an awesome salesman. I have been on the wrong side of a couple of really big deals that Mark demonstrated his sales prowess with. I also have an inside track in that Mark started his career on the engineering side as well, so he understands a little bit what goes on in the R&D world. I did a stint in sales too, by the way.
I spent the last five years at Xilinx, leading technical sales and support. Mark and I have a great collaboration because we really understand the challenges that each of us are facing and allows us to collaborate more effectively on winning customer designs. Mark.
Thanks, Steve. I'm also very excited to be here today and at Lattice. I thought the best way for me to demonstrate why I joined Lattice was to give you a glimpse into our customers. The few months since I've been at Lattice, I've been traveling pretty much nonstop to customers in every geography and customers in just about every end market segment, both large and small. Most of these customers I've had relationships for many, many years, therefore, much of their feedback and input was what I was expecting to hear. There was also a couple of pieces of their input and feedback that were positive upsides for me that I wasn't expecting. First of all, there were two themes that were common across all of my customer visits.
The very first one was that our FPGA competitors are focused on the very high power, the very complex devices that are targeted mostly at the data centers. That wasn't necessarily meeting their product needs. For example, one of the customers I visited was a very large, one of the largest industrial customers, and what they did was they laid out all of their FPGA usages that they use in the company. The reason they did that is that they wanted to see how many of those would intersect our current capabilities and the roadmaps that we were pushing forward. They wanted and needed a company that was going to continue to invest and innovate for the industrial market, similar to products that we have tailored for those industrial applications. They were very happy to see that we were targeting what they needed going forward.
That was very common across customers. The second theme was around power. It's pretty well known that those very complex large FPGAs have been battling with a power challenge for many years. The customers were very happy to hear that we were going to retain that Lattice heritage around low power architecture in our current and the future products like Steve just showed you. There was one upside that I heard that I wasn't expecting, and that was around our customer relationships. Our customers told me that they like doing business with Lattice, and that many of them have had multiyear relationships with the field applications engineer and the salesperson, and that they built a kind of bond, a trusted bond with them. The second area that was positive upside for me was around security and AI.
As I went to many different customers, one of the things that became apparent really quickly was that security and AI are top of mind to just about every customer in all different segments. The solutions and products and capabilities that you've just heard about are actually elevating the conversation with those customers, elevating it to be more strategic, and also elevating it earlier in the process. With that, overall, the reception has been very positive. With all that customer enthusiasm, it's our job in sales to turn that into revenue and design wins.
Today, I want to talk to you a little bit about what are we doing differently with our sales model that will help us to generate more business opportunities for Lattice and continuing to reach those thousands of customers in the broad market, while at the same time increasing the value that we're providing them. It all starts with structure. When I first met Jim, and many people thought that Jim and I knew each other from the past, our paths never quite crossed. One of the first things that we talked about that really interested me was his mission to take the company back to its roots, to simplify the company around FPGA. And I quickly found out after I joined Lattice that I needed to do the same thing with the sales organization.
We were structured as two different business units with both independent sales force. For a company our size, that creates a lot of inefficiency and a lot of complexity. We've now unified the sales force under me, and we've moved to a very traditional three-region structure. We also had a very complex compensation structure. Our field applications and salespeople were paid with as many as 5 to 10 different variables in their incentive structure. We've now moved to two, revenue and design wins. As you know, design wins are what fuel our future revenue growth.
Also under this new unified sales model, it also helps us to be able to leverage the solutions that we're generating in our applications teams and the skill sets that we may have in our organization, such that we can deploy those on the most important opportunities, no matter where they might sit in the world. The second thing was to align around our customers. Now, you heard from Esam that one of the great strengths of the FPGA business is that FPGAs are used in just about every end customer segment. They're also used by just about every size customer from large to small. As Jim said, each one of those groups of customers is very important and vital to our business model. We need to make sure that we have sales coverage models that is aligned with each one of those groups, both efficiently and effectively.
Let's start with the big customers. As you can see, that $3 billion SAM that Esam talked about, half of that is centered in as little as 20 customers, and we have good positions and good relationships at those large OEMs. It also presents a huge upside for us. Because of that upside at those customers, we're going to increase our focus, increase our direct resources to capture the business with those large OEMs. In that next group of customers, we're going to continue to leverage both internal and external partners to make sure that we have the best access, influence, and that we continue to win with those customers. Then that long tail, those thousands of customers that we talked about.
We're going to strengthen our partnerships with our global and regional distributors to make sure that we continue to capture those customers, but more importantly, even expand it even further. How do we manage thousands of customers? We do that through a very traditional pipeline methodology. Starting on the left-hand side, we uncover and we fill that pipeline with new business opportunities for Lattice. Then we methodically track and move those opportunities through that funnel to the point where the customer commits to Lattice, and then we move that on into production. The good news is we have a very strong and healthy pipeline today, and we measure a healthy pipeline by the size of the pipeline, the quality of the pipeline, and the mix of the pipeline.
As you can see from the table, the majority of our pipeline today is centered around those multi-generation, multi-year revenue streams that Jim talked about earlier. In fact, if you take those top two, the industrial auto and the communication and compute segments, they have production life cycles anywhere from two years to 20 years. Very good. When we look at our increased focus at those large OEMs that we talked about, that's going to enable us to increase the overall size of that pipeline. As we win at those large customers, that's going to also have an effect for us, a positive effect on that broad base of customers. That's because those large customers tend to be the leaders in their particular market segment, and the rest of the customers in that segment tend to follow the leaders.
Another way that we're going to be able to take this healthy pipeline and strengthen it even further is by some of the things that you heard today that are being launched and rolled out today and through the second half of this year. Things around capabilities in our security, AI, embedded vision. Those are allowing us to create more new opportunities that's going to further strengthen that pipeline. As we expand that SAM and we start to reach even more customers, many of those customers may not be familiar with FPGA technology. They may not fully understand what FPGAs can actually do, what problems they solve, or what opportunities can be created for them. Many of them have never designed to an FPGA before. A couple of ways that we help those customers is through demonstration boards and reference designs.
First, we take some common applications, some common challenges that customers are trying to solve today, and we couple that with a demonstration board, and we put it in the hands of customers to be able to show them the capabilities that our products can bring and the opportunities that it can create for them. The next thing we do is we use our application team to create some of the solutions that are particular to our target market segments, and we provide customers those reference designs. You can think of that as a starting point for the customer. The customer takes that reference design and then starts to customize and differentiate to their particular system on top. We take that bundle of reference designs and solutions and demonstration boards, and we bring that to the mass market by working with our distribution partners through some marketing programs.
A great example was last month, one of our distributors worked with us to create a seminar to be able to showcase sensAI and its capabilities. They brought that out to one city, and it exceeded all of our expectations by almost immediately selling out. It brought in over 30 different customers in that one city to spend a whole day in a room learning about Lattice and our sensAI capabilities. Because of that success, that particular distributor now is replicating that seminar in 10 different cities over the next 30 days, and that's one distributor. We take that same recipe or those same sets of programs and then replicate those across all of our global and regional distributors. That's what gives us the multiplying effect across the thousands of customers.
The value that those solutions and bundles provide is it helps to get customers to market faster. Another way that we create value is pretty well-known and standard, that the best way to win and the best way to increase value with customers is to engage very early, to engage while they're architecting their systems. I want to give you another example that we've had over this last month. A couple of customer meetings that I've been involved with. One was one of the largest wireless communications company, and one was another very large compute customer. The result of that meeting was the same in both cases. They proceeded to roll out and lay out for us their complete block diagram of their system. And what they enabled was a discussion with us about how best to partition their system to take maximum value of our security solution.
Those are the kinds of discussions that create value and that customers value. The way that we make that value and turn that into multi-generation value that Jim talked about is by adding more of a solution. Software in addition to hardware. Steve talked about the increase he's making in R&D around software for this reason. We're also increasing our investment in training for our field application engineers and our sales so that we can become more proficient at that system-level approach, that solution selling. We've also built a lot of tools and analytics so that we can align the value that we're bringing customers across to the thousands of customers we have. Those tools and analytics now are providing our sales team the ability to appropriately price our products to the value that we're bringing a particular customer in a particular market segment.
Hopefully with that, you can see why I'm excited to be here at Lattice. We have a very strong and expanding pipeline. The reception from our customers, the pull from our customers has been both strong and strategic. They need a company that's going to invest and innovate in this space. We simplified the structure, which is going to enable us to move faster and be more effective. And it's a combination of these things that's really going to fuel our growth. The next speaker and I joined at the exact same time. In fact, we were the last two to join the executive team. Being the last two gave us a big advantage because we could firsthand see what Jim was building and the team he was assembling.
With that, I want to introduce the person that I know all of you really came to see today, our CFO, Sherri Luther.
Thank you, Mark. Appreciate that warm welcome, I want to give all of you a very warm welcome to our Financial Analyst Day. Very excited to be here to talk to you about our financial priorities and our long-term model. First, I'd like to give you a little bit more about my background and why I decided to join Lattice. Most recently, I came from Coherent, a roughly one and a half billion in revenue company, where I spent the past 16 years. A company with a very solid balance sheet and a very strong culture of managing a business through financial metrics and establishing accountability. My philosophy is that what gets measured gets done. When I really looked at Lattice, I looked at a company that was similar in size to Coherent when I joined Coherent about 16 years ago.
I saw the opportunity to put in place key processes and metrics that I could use to manage the business and grow. When I also looked at Lattice, I saw the opportunity to increase gross margin, increase profitability, to reduce our SG&A, and the opportunity to really saw a lot of untapped potential to increase our cash generation. I'm just going to switch out here. This one's not working. I saw a lot of very expensive debt that needed to be taken care of. When I joined Lattice four months ago, I put in place key financial metrics and processes that we use to manage the business. Metrics focused on improving profitability, increasing our cash generation, and de-levering our balance sheet. I'd like to take you through each of these key areas in a little bit more detail.
I think I have a clicker that doesn't like me today. Let's talk first about revenue. When I look at 2018's accomplishments for Lattice, I see a lot of opportunity that we can increase our gross margin, I'm going to take you through some of those areas of expansion opportunities. When I look at SG&A, I see a very high level of SG&A, I'm going to take you through our plans to reduce that over time. From a cash generation perspective, in 2018, we were able to pay down our debt in making discretionary debt payments of $40 million. In 2019, the first half alone, we've made $50 million in discretionary debt payments. These are just examples of how we've been able to use the metrics that I've been able to put in place to improve DSO and thereby generating additional cash.
I'd like to take you through each of these areas in a little bit more detail. From a revenue perspective, this is a slide that Jim shared, which breaks down our revenue in each of our key four market segments. You can see the growth that Esam talked about in 5G, the expansion of our position in servers, the growth in industrial and automotive, fueled by factory automation and an increase of electrical components in vehicles. We heard from Mark about the sales design win pipeline. We heard about our $3 billion SAM and expansion opportunities due to the proliferation of FPGAs across multiple applications. All of those growth vectors are driving growth in our two key market segments, comms and compute, and industrial and automotive.
Both of these key market segments typically have higher gross margins, and you'll see a bit of that benefit when I talk you through our gross margin expansion strategies. From a gross margin perspective, this is a slide that Jim put up as well that shows our targeted gross margin of over 62%. We talked about our three main areas of how we expect to achieve that. Pricing cost optimization, our product cost reduction, and mix improvement. I'd like to take you on a little bit of a gross margin walk to explain to you what our strategies are in a little bit more detail as to how we plan to achieve these. Let's start with our pricing optimization.
We really did a deep dive, took a lot of time and did a deep dive to understand how the company was pricing our product and to understand the value proposition to our customers. This was a project that Esam kicked off shortly after he joined. It was very insightful. We used a lot of advanced analytics to really understand, across the board, the pricing dynamics of the company and our products. We found some interesting insights. One was that some of our products were quoted to customers based on volume pricing, but yet the customers were not held accountable to the volumes, yet they were able to enjoy the benefit of the lower price. Another example was that we had products that were priced the same no matter who they were sold to, irrespective of what market they were sold into. Our FPGAs are not commodities.
We need to get the value add, fair value for our customers, for our products. Through these advanced analytics, we put together pricing management models that we can use to manage the business and quote to consumption, so we can get that fair value for our products. From a product cost reduction perspective, I mentioned that what gets measured gets done. It applies to our suppliers as well. We partnered with our suppliers to put together yield and productivity improvements as well as efficiency improvement plans that benefit us at the gross margin line. We've also put in place metrics and a very disciplined approach to managing our inventory, so that over time, we reduce our inventory and see that benefit at the gross margin as well. The third area, mix.
I mentioned to you how our comms and compute and industrial and automotive segments are growing over time, and we talked about the growth drivers there. Those segments typically have a higher gross margin, and we expect to see some benefit to that as well in our gross margin. The two key drivers of our increasing gross margin to our target are pricing and cost. I'm going to switch over to share with you our OPEX strategy. Our strategy for SG&A is to reduce it to 15% over time, and we have very clear plans as to how we're going to accomplish that. In Q4 of last year, we consolidated our San Jose facility from two floors down to one floor, and we saw the benefits of that starting in Q1.
In Q2, or in Q1 rather, we announced the consolidation of our Portland facility into our Hillsboro facility. Two sites that were 15 miles apart that we consolidated. It didn't make any sense to have two sites so close together that we were maintaining, we consolidated those. In Q2, we expect to see the benefits from that consolidation. There are other areas as well that we're looking at to reduce our SG&A. Another example is in the area of IT. We are leveraging our third-party suppliers more effectively so that we can get additional cost reductions in SG&A. We are also leveraging our lower cost geographies to get additional benefit and cost reductions in SG&A. Our demand generation is very important to us, and we don't want to harm that. Frankly, the plans I've just laid out to you just make good business sense.
Now I'd like to jump over to our balance sheet and give you a little bit of insight on our capital allocation strategy. I mentioned to you that in 2018, we made $40 million in discretionary debt payments, de-levering our balance sheet down to 3X. In 2019, year to date, in each of Q1 and Q2 so far, we have made a total of $50 million in discretionary debt payments. Our goal is to de-lever our balance sheet down to below 2X. Once we get to that level, then we will determine the best use of our cash. I mentioned in the beginning that one of the things I noticed right away was our very expensive debt that we have. I'm really thrilled to announce to you today that as soon as I started, I undertook a process to drive for looking at refinancing that debt.
Today, we are announcing at the close of market that we'll be filing an 8-K with the details of our debt refinancing, something that was sorely needed and will drive a lot of benefit to the bottom line. Through that debt refinancing, we were able to reduce our interest rate by 250 basis points. That's significant. We also extended our maturity to 2024 by three years, the old debt was due in 2021, giving us more flexibility in terms of the timing that we pay down our debt. We also have greater flexibility with respect to stock buybacks. Again, our goal is to de-lever down to below two. Once we get below two, then we can determine on an ongoing basis what the best use of our cash is. I'd just like to recap our long-term model.
We talked about revenue growth in the low double digits. We talked about the drivers in comms & compute and industrial and automotive. We talked about increasing our gross margin to over 62%. I walked you through our plans as to how we expect to do that. Within the envelope of OPEX of 35%, we expect to reduce our SG&A down to 15%. I walked you through some of our plans as to how we intend to do that, all with achieving a profitability of 25%-30%. A few key takeaways that I'd like for you to have here is that I mentioned that what gets measured gets done. We have detailed plans that I've talked to you through in terms of our gross margin expansion, detailed plans as to how we're going to reduce our SG&A.
We have key metrics in place and processes to drive improving cash generation. We have a new debt refinancing that gives us significantly lower debt service costs, gives us greater flexibility to determine the best use of our cash over time. We feel very confident that we have a solid plan in place that we can achieve. With that, I'd like to turn the mic over to David Pasquale for Q&A.
Thank you, Sherri. We're going to take a minute and bring the entire management team back up for the Q&A panel. When you're asking a question, if you can just state your name and the company that you're with, please. There'll be mics that will be filtering out into the audience. Given all the interest in Huawei, it's been in the news, Jim, we're going to start with that. If you could just give us a perspective on what Lattice's exposure is, that'd be helpful.
Yeah, sure. Certainly, we're complying with the order that was issued last week that Huawei suppliers were impacted by. Just to calibrate you a little bit on our size of business with Huawei, if you look at 2018, it was mid-single digits, think around 5%. We were expecting roughly similar contribution from Huawei this year as well. That gives you a little bit of calibration. I also want to take the opportunity to mention what we said before, which is a reminder that we don't have any customers that are over 20 or over 10%, excuse me. We are very well diversified. If you look at the total active customers, 9,000 active customers, large distribution channels. Our revenue streams are very well diversified.
Chris.
Christopher Rolland , Susquehanna. Thank you so much for hosting the day. It is amazing the contrast between the last one and this one. Yeah, you guys have put an amazing management team together. I guess the biggest pushback that I get in the name is really around FPGAs and ASIC replacement
It seems like there is always this treadmill that you guys are on. What are you guys doing specifically to help combat this, fight against it, create sticky and long design cycles?
Thanks, Chris. I think I will ask Esam to answer that, because actually you touched on that in your section in 5G, right?
What people are talking about, and our competitors are very open about it, that in 5G, in the data path, remember the data path, and there is the control plane. In the data path side of it, there is a replacement of those FPGAs to ASICs, and that is driven by the economics. Everything is driven by economics. If you have got an FPGA that is very expensive, very high power, their goal is once the standards are known or they know what they want to do, they will replace that with an ASIC as soon as possible. That does not apply to the control plane. When you look at the control plane, the economics do not make sense to go build all those different types of ASICs to replace the FPGA. It did not happen in 3G. It did not happen in 4G. We work very closely with our customers.
It's not going to happen in 5G either. That's with 5G. When you take a look at the applications that we're going after today, for what Lattice does on the small side of the FPGA, we're actually seeing that we're replacing a lot of ASICs in the market today ourselves. The economics start to make sense because of the value proposition of an FPGA. When you take a look at the examples we talked around AI, if they're using an ASSP or about video connectivity, there are ASSPs there that we compete with, the value proposition of what we bring to the market, we're actually taking ASICs off the board and replacing them with FPGAs today.
Okay. Tristan?
Tristan here with Baird. You've done a great job articulating the old and the new Lattice. Could you expand a little bit on what the new management team is bringing in terms of new relationship and how that's benefiting Lattice?
Yeah. I think you're referring to, Tristan, for instance, customer relationships. If you look at across this management team, obviously decades of experience, we brought our own relationships with customers across every one of those end markets that we talked about today. I'll ask Mark to comment a little bit on this as well because he's on the front lines with customers. Certainly, that's helped open some new doors, at a number of, what I would say, large OEM customers, where maybe we had a position with that OEM customer, but the relationships that we brought into the company, has really elevated the discussion and elevated the discussion of where Lattice can fit into that OEM's plans over the long term. Do you want to add a little bit to that?
Yeah. I think as Jim said, we have a good position with Lattice at those customers, so we're a known, trusted supplier to them. Through some of the relationships on the team, we've been able, as he said, to elevate some of those discussions. Some great examples, Esam at a large communications company that, we're engaged with, but now we're getting RFQs for other types of applications that they may not have typically given to Lattice. Jim, through some of his relationships, opened the doors very quickly for us as a sales team to talk with client type devices versus just server devices to expand some of our security solutions. Certainly, Steve and I were a part of another tier 1 wireless infrastructure customer, where really we hadn't been that close to for a while from a strategic standpoint.
What I mean by a strategic standpoint is them actually influencing and being access to our long-term roadmap needs. We literally had high-level management and 20 different people from this customer be able to come visit so that they could understand what we were doing long term with our roadmap.
Hans?
Yeah, thanks. Hans Mosesmann with Rosenblatt Securities. Sherri, in terms of the model going forward, what are the puts and takes of upside to that model in terms of top-line gross margins, operating margins? Thanks.
Sure, Hans. Thank you for the question. The way that we approach putting together our model was a really balanced approach, looking at all of the opportunities that we saw in growth margin and in OPEX. As you can see, there are a lot of levers that we have. It's not just any one particular area. We feel very confident that we can achieve those goals. Of course, we're never satisfied. I think that's true of the whole management team, and we'll continually strive for improvement. We feel very confident that we can achieve those goals with the detailed plans that we have in place, and pretty much our blueprint, if you will, as Jim kind of alluded to at the beginning of his presentation.
Okay. Mark?
Hi. Thank you very much for the presentations. Two questions. I think the first one for Mark, Steve, and Esam. Altera and Xilinx have big R&D budgets. What would prevent them from coming into the markets that you're going after, which seem really attractive? You painted a very nice picture of the markets. For Jim and Sherri, to what extent do you need revenue growth to get to your target model?
I'll start.
Yeah, I'll start off with the first one. If you take a look at the other competitors, they've been public about this as well, their focus is where they're saying, which is the large FPGAs, compute data center comms. For those who have been tracking this market for quite a while, and I've been in it for 30 years, I refer to it as kind of the arms race. It's you have to get there. It's whoever gets there first wins the majority of the share. When you have the dynamics that's happening between those two companies that you refer to, it's about who gets there first. If you're distracted with something else, you lose that window, you lose a dominant share of the market potential.
If you're in Xilinx and you take a look at Altera, which is Intel, and the might they have, they're doing the right things they need to do to focus on that market. Anything else for them is a distraction. That's really from a market perspective, how they're making their decisions. If I were there, it's the right decisions to go do.
Go ahead, Steve?
Okay. On the second part of the question, I think it was how much revenue growth to hit model probably related to the OpEx target. We talked about a 35% OpEx, 20% R&D, 15% SG&A. Most of the work there is to get the SG&A down to mid-teens, 15%. We are assuming a combination of direct cost-cutting, and Sherri shared a few of those examples, but there is some assumption in there of revenue growth as well, we're getting better leverage on the model over time. It's a combination of both direct cost-cutting and revenue growth.
Okay. Ruben in the back.
Thanks. It is Ruben Roy from MKM. Jim, just to follow up on that thought, I have a quick follow-up after that. On the revenue growth targets for the next couple of years, single-digit near term, double-digit, over three to four years, let's call it. Is there a portion of revenue related to some of the products that you've discontinued? Is there a legacy portion of revenue that's in that calculus for revenue growth? That's the first question.
Okay. The first question, certainly our remaining revenue streams from Silicon Image are now essentially in harvest mode, right? We stopped investing in any non-FPGA product lines. You can think about the size of that. If you look at just product revenue, remaining Silicon Image revenue is about 10% of product revenue, roughly. We expect that revenue to be on a slow multi-year decline over time. I think that all falls within the consumer segment, our consumer segment. That part of our revenue sits within consumer. The only other piece of legacy Silicon Image revenue would be the IP revenue, which is HDMI royalties primarily. Again, that's just on a very steady, slight long-term decline as well.
Okay, thanks for that. Following up to Tristan's question from earlier, when you go out and talk to customers, you guys talked about maybe being on the same page roadmap-wise. What are some of the other pushback or concerns you're getting from customers? Lattice had a lot of stops and starts with various product lines in the past. I'm just wondering, your top customers, what are you hearing from them and what you need to do to move forward with them? Thanks.
Honestly, it's really quite positive. I'm trying to think of any pushback examples. Mark, you might have to help me on that. It's honestly, they're very pleased to have Lattice 100% focused back on FPGA and investing in a part of the market that they view our competitors not investing in, right? Industrial automation, power-efficient FPGA, small size, easy to program. These are things that our competitors are not really investing in. They're happy to see us investing in areas that they know they're going to need those capabilities for their business. Do you want to add Mark, or?
I think you pretty much covered it, yeah.
Okay.
Charlie?
Yep. Hey, Charlie Anderson from Dougherty & Company. Thanks for the great analyst day and the presentations. I want to ask about the new FPGA platform. It sounds like it's a very high-performance platform. I'm curious, what are the opportunities for content increases from the new platform? What is the feedback you're hearing from the customers on the new platform? Security was sort of a common thread in some of the presentations today. It sounds like there's an appetite on the server side, I'm also curious, some of the other end markets, where's the appetite for security as well? Thanks.
Yeah, let me start on the new platform, Steve, you should add in as well. On the new platform, I think it's a combination of it certainly bringing new capabilities to our product roadmap. We do expect that new platform with its new capabilities, higher performance, better power efficiency, et cetera, to allow us to expand into some applications that we haven't traditionally serviced in the past. If you remember from Esam's section, the SAM expansion he talked about, part of that's enabled through that new platform and through the new capabilities that it'll bring. It'll drive some revenue growth, SAM expansion. It'll also over time, it'll replace some of the existing product. It's a combination of replacing existing products and driving expansion of capabilities and expansion of SAM above that. Do you want to add a little bit?
Yeah, maybe just following up on that, a lot of the system blocks I mentioned are really helping us to optimize different product families for different areas. The beauty of the platform-based design is you don't have to put every feature in every product. We can be very selective in how we architect the product families going forward, and we're really going to be focusing those products on the customer requirements in those application areas that we're targeting growth for. The embedded vision, artificial intelligence, and of course security as well.
Just want to add, Charlie, your second, you also asked about security into other markets.
Oh, yeah.
You're absolutely right. When you think about in all the other end markets where our processors are located, they're located everywhere. Just like we're securely booting it on a server, that same application applies to every other processor, whether it be in a factory, an autonomous vehicle. You name that application, there's always a processor there. It is applicable to other markets. When we first developed this product, it was joint development between us and our customers focused on that portion of the market. The applicability of what we're doing goes into multiple end markets.
Ian Conn, Shannon River. My question is regarding Huawei, or if you prefer a generic customer, when they're cut off from your product, what their alternatives might be in the short, medium, and long term.
There are not pin-compatible replacements for Lattice products, right? One of the things about Lattice products is we're not a commodity. You can't just pop our product out and put somebody else's in. Again, a strength of Lattice's business and its business model. I would say there's no immediate obvious replacement for a Lattice part.
Medium and long term?
Medium or long term, I guess would be more longer term. This could apply to any semiconductor device. You could redesign the platform to try to design in a different type of device. That would be the option, that would be extensive amount of effort. That would be a significant amount of time.
Jessica.
Thank you. Jessica Adams with Assurance. I listened to [Mike Hurczyk] present last week at the JP Morgan TMC conference. They were talking about Microsemi acquisition having FPGAs, which weren't really sort of talked about a lot before that I noticed in their end markets. Are you seeing them as a competitor at all?
Yeah, they kind of decide whether to let Esam answer that question since he came from Microsemi. Let me answer that question, though. Yeah, we really aren't. The Microsemi FPGA business, which is now part of Microchip, is I would say, mostly focused in military aerospace defense. We generally don't run into them a lot. We run in more to as Xilinx and Altera are sort of more traditional competitors. Do you want to add anything?
Thanks. Just one more if I could. Can you tell us the new principal amount of the debt and whether or not the coupon is fixed or floating? You made some reference to the difference versus the previous spread, anything else you can tell us?
Sure. We took on a $175 million commercial A loan, term A loan with a $75 million revolver. I mentioned the interest rate, it's tied to our leverage ratio. It starts 250 basis points lower than our previous debt. It steps down as we de-lever over time. You'll see lots more details in the 8-K and the details that were filed.
Thank you. Matt.
Hey, thanks. It's Matthew Ramsay from Cowen. Two topics from me, I think. One on 5G and one on what I'm sort of excited about with the story, which is inference at the edge and the possibility for that over the longer term. I guess in 5G, Esam, you made some reference in the slide to maybe a 30% content boost for you guys. I was wondering if that was 30% per box, and then there's a multiplier for number of endpoints on top of that, or if it's just a blanket 30% statement. On the inference at an edge, you talk about microcontrollers being an area where you guys are trying to gain market share.
Maybe you could talk a little bit about competition in that space, Project Trillium at Arm, and what the folks on that side are doing in the microcontroller space, or is that a market that you think you can gain share in pretty quickly? Thanks.
All right. On the 5G one, if you recall the slides, what I said was at least 30%. There are different ways to quantify the 5G. There's more remote radio units that are going to be deployed, more components that are going to be leveraged. Our footprint did increase. If you look at the number of sockets we're on, it's higher on 5G than it is on 4G, in addition to the value that we're bringing. It's hard to predict how fast and how much of 5G is going to be deployed. When we take a look at the actual designs we have today with the OEM coverage that we have today, which is strong, we see at least a 30% increase in dollar content, and that's across the build outs that will be there. If the build outs surpass people's expectations, that will go higher.
We chose to pick, again, a balanced approach on how we came up with that number. On the microcontrollers around AI, this is an emerging market, and any time you have an emerging market, the users themselves aren't sure what the best solution is out there. You have multiple people going out to the customer base. "Here's a solution, use my MCU. Here's a solution, use my low-end GPU or SOC. Here's a solution, use our low power parallel processing FPGA." Because it's emerging and because it's new, people are looking for what is the ideal solution that's out there. You can go with a microcontroller if it satisfies your need and you're happy with it. You can go with an FPGA that has parallel processing and lower power. We've seen that when the customer bases that we've gone to.
You can take a GPU and do the application. A GPU will cost more and drive more power. For us, the challenge is about getting to those customers and getting the word out that it's an FPGA, we've been having really good success. For the customers we engage with, they see the value proposition of the FPGA. Working with Mark on the sales team, we've got a lot of programs in place of how do we now proliferate that with our channel partners, we've had some good success.
Tristan.
In the past, you've had some lumpiness associated with the ramp of prior geometry nodes, in terms of OpEx, with tapeouts. How should we look at the fluctuations in OpEx as you're ramping 28 nanometer? How much investment or how aggressive are you going, at 28 nanometer versus prior nodes? Finally, maybe a sense of how much of your total revenue could be 28 nanometer based, let's say two, three years from now?
On the first part of the question, what I would say is, right now we're at about 20% of sales, R&D as a percentage of sales. First of all, we think that's about the right model. You heard Steve talk about that as well. We're at about that if you look at Q1 most recent quarter. We anticipate staying roughly within that envelope. Yes, there could be some sort of quarter-to-quarter variations or fluctuations based on non-labor expense based on expenses that, boards or things like that. It'll generally stay within about that 20% is what we expect. What I would say is most of our work has been trying to fully optimize the R&D spend within that envelope.
Steve did a lot of work on this along with Esam in the second half of last year to really optimize the portfolio within that 20%, because we did find a number of projects that we just didn't believe were good ROI for the company. These are FPGA products, we shut those down, funneled the R&D into the big ROI programs like the 28 nanometer devices and really tried to drive an acceleration in terms of time to market and a faster cadence. On the second part of your question, 28 nanometer revenue, it's probably too early for us to start to give a sense of when that would start to generate revenue and to what extent. As we get closer to the product launch, we'll probably provide more color around how much 28 nanometer revenue would be over what period of time.
In the back. Richard?
Thanks. Hi, Richard Shannon, Craig-Hallum. A couple questions. First of all, you talked about single-digit growth in the near term and then getting to double-digit. Wondering how much contribution you need from the new markets like AI as an example to get to that double-digit number or how much of it comes from the base business. The second question on sensAI. Can you talk about number of customers, pipeline engagement, competitors, et cetera, just to give us a better feel for what level of maturity we're in terms of bringing that up?
Sure. I'll let Esam maybe answer the second part of the question. The first part of the question, I would say not a lot will be driven by AI-related. There isn't a lot of AI revenue, in the near term. It will help contribute to the double-digit further out. Most of what's driving our growth in the nearer term is around 5G infrastructure deployments, that server ramp that Esam walked through, industrial automation, things like that. We are expecting AI-related applications to be a good growth driver out in time. I would say, we're trying to be pretty balanced about our forecast around that and realizing that, look, it takes time for new applications and things like that, new adoption. Do you want to talk a little bit about the second part of the question?
Sure. I gave examples in my slides about applications in end markets, some of which we're leveraging AI. These are all customers that we've been engaged with or multiple customers for a similar type of application. The revenue ramps are there, but it's still ramping up. We're seeing revenue, but this stuff is still ramping up. It's going to take time. One of the things that excites me the most, however, is the sensAI 2.0 release. That enables us to accelerate this because it provides more performance, it provides more ease of use for our customers. This is part of the things that Steve talked about, which we invest in our solutions, and what Mark alluded to, we can scale the business to more users. Where are we on the total of this? We're still in the very beginning stages.
We expect it's going to continue to ramp up, every investment we do to make it easier for our customers to adopt only helps accelerate that more.
Okay, we probably have time for two more. Will?
Hi, thanks. William Stein from SunTrust. Thanks for taking my question and for hosting this informative analyst day. I wanted to talk about industry structure and Lattice's position in that. You've been very clear that capital allocation is going to be directed to debt paydown, there was at least one acquisition I understand done by a prior management team that sounds like you don't view so favorably. On the other hand, Lattice is relatively small, size tends to be an advantage in developing more strategic relationships with bigger customers. How does Lattice envision its position in the industry going forward from a consolidation perspective?
Yeah. What I would say is that, right now, most of our management focus is on the organic business plan. We're really heavily focused on making sure that the organic business plan is as efficient and productive as possible. I kind of talked about that remodeling, right? We're still early in our remodel of the organic business, so we've got a lot more work to do. Now, that said, we are in parallel looking at what might be inorganic moves that we could make, potential acquisitions. We certainly are always monitoring the environment for that. I would say our approach to that is we would look at inorganic opportunities that would really be additive to our organic strategy, that would somehow accelerate the results of our organic strategy.
I don't think we'd be looking at things just purely for a consolidation, just purely to get larger, but we'd be looking at things that are well-aligned and right down the fairway of our existing strategy.
Okay, Cristiano?
Yeah. Cristiano Amoruso, Lion Point. Quick clarification on the comments on year one and year two revenue growth. Does that include what's going on with Huawei?
Yeah. What I would say is more generally, not just year one, year two, but three and four is, look, we haven't built a business plan that's