Good day, everybody. Welcome to the show. I think I know quite a few people here. It looks familiar. If you remember last time, last Analyst Day, I said one thing at the end. I said, "Come back in two, three years, and you guys will be surprised." Guess what. Even myself was quite surprised for whatever we have been doing. Somehow I feel like all the engineers working with me in the team, we're just like a group of climbers. We're climbing the high mountains, and we are so focused, and we work hard, and we climbing and climbing, and we didn't really have a chance to look back, see, hey, what's going on? What we have achieved.
I think today is a very good opportunity for all of us to look back and see, hey, how we evolve to today and what we are going to do next. Actually, it all starts with two great innovations from MPS. The first one actually go all the way back, actually even earlier than 2010. I think the team started to work on this Intelli-Phase around 2008, 2009 timeframe. I still remember, we tried to pick the best process available back then. I think it was the BCD3, now we're BCD6, by the way. We also tried to optimize a lot of things, like how we can run faster, how we can integrate more intelligence to the power device, how we can achieve the one nanosecond dead time, things like that. Sorry, I go too far. Never mind.
The thing is, we actually delivered the very first Intelli-Phase in 2010. That was a great achievement. We're under the spotlight for not too long, then we realized that just having the power stage, the Intelli-Phase is just not enough. To grab the market, to win all the big sockets, we got to have our own controllers, and that's where the QSMod was invented. Two years later, we have our first QSMod controller, the digital controller, and it worked so well, so easy to use, and was adopted right away. In 2014, we successfully powered the first Intel processor, well, we powered the Intel Grantley platform in June 2014. That's a big milestone. From that point onward, we achieved a lot of things. The next item, I'm going to tell you what exactly we did. Look at the slides.
This is the slide I actually showed you guys last time, if you still remember that. I want to reuse the slide to refresh the memory that the two technologies, how it impacts the CPU use the power. You look at this big bucket. The bucket has water in it. The bucket is representative of capacitors. The capacitor is providing all the energy to the CPU. This is the Intel processor. As you may all know that processor power is going up like crazy. Nowadays, they're looking for 600 watt, 700 watt. I even got a customer asking for 1,000 watt per silicon. Might not necessarily be Intel, but some ASICs 1,000 watt peak power. The interesting thing is that peak power doesn't stay there all the time. Sometimes this power will go down to a few watts to save the power.
How you do this dynamically and fast enough, accurate enough to make sure the processor can still operate in the most efficient, powerful way? The bucket serves that purpose. You bring the energy to the processor, and you take it back from the processor. That's the idea. As of for today, most of the suppliers, the vendors, they're doing it, the power delivery, in a discrete package format. What does that mean? Well, it means like a fire truck. It's a big fire truck with a hose on the top. How you deliver the power to the bucket? Just turn on the valve. Turn on the valve, the water is shooting to the bucket. Right? Fill it up. The problem is how can you control that valve accurate enough and fast enough to fulfill the processor dynamic needs? That's a problem, and that's a challenge.
Whatever they do, of course, they came up with some fancy digital processor, DSPs, monitoring the water level, monitoring the CPU states, the temperatures, the current, everything. At the end, you still need to pair this big hose to shoot a big water, the analog water. That's most of our competitors are doing. Guess what? We completely changed that story. We don't do it that way. We do it in a very different, accurate way, and this is QSMod. Before we go there, I show you some animations, see how the water flows. It's pretty hard, by the way. This is our QSMod controller. Everybody familiar with QSMod, right? Quantum State Modulation. Quantum State Modulation means every time when you deliver energy, you deliver in a very accurate droplet. The total energy is still the same, but you deliver it drop by drop, right?
We measure each droplet. We know how much energy we deliver each time. We control the amount of the droplets. The process is counting all the numbers. The minute I know I shoot 1,000 droplets, I know where the water level is, and CPU knows how much I need. Everything is in a closed loop. We don't really need a fancy algorithm to do this accurate thing. Actually, just the controller, the QSMod, together with our fully integrated power stage, they work seamlessly to achieve that. Again, I show you some animations. That's the regular loading. Now it's more droplets. If the CPU demand is higher, you see a lot more droplets coming to the bucket. That's the full speed coming to the bucket.
One thing you might notice if you pay attention to my slides, our bucket is much smaller compared to the previous one. As I mentioned earlier, the bucket represents for capacitors. Everyone heard about the cap shortage recently in the industry? That's where we can really make a difference. We need much less capacitor to provide same amount of power to the processor in a much faster and accurate way. That's the story I was trying to tell you. The next slide is to answer a question, why only MPS can do it, right? Why only we can shoot the droplets versus a big hose of water. The concept actually is quite simple. You look at this picture, that's what all the competitors are doing. The first thing you see is a big truck.
The first impression you got from a big truck is starts slowly, and it's very hard to break. It's also very hard to turn, right? It's big. The discrete driver MOS has that characteristic. You see that on the left side, the little picture, with 3 dice inside the package, and those are the MOSFETs and the gate drivers. Each FET only have 1 gate control. It's coming from the driver. Okay? When you turn on the FET, you have to turn on the whole thing. When you turn off it, you have to turn off the whole thing. What we do. Well, I like this picture, 42 nice cars. Look at that picture. Also look at the device right here, Intelli-Phase. We actually have 42 power banks, and each power bank is like a racing car, has their own drivers.
All those 42 blocks are synchronized. They can turn at the same speed, and they can deliver the QSMod energy to the load. We can run faster. Not only that, it can be very accurate. Remember I mentioned earlier, our power stage, we can do 1 nanosecond dead time. That's how we do it. If you have a big truck, you can't make a sharp turn. But if you have a small car, although you have 42, but each one of them can make very sharp turn. We can do a 1-nanosecond dead time. We can control so fast and nimble. That's what we're winning. That's how we're winning. Those are great, right? Great technology. What we actually do. Look at the top line, the CPUs.
Since Grantley, we debuted at 2014, we've been engaging with Intel, and we've been powering the Intel platforms for 3 generations now. 3 generations. If you think the Grantley platform was just born, and we actually went through a lot of growing pain at a Purley platform. But now we're fully grown, we're ready. We actually are powering the Whitley CRBs as we're speaking. The Wilson City, we are the leading power supplier for Wilson City coming out. Another thing, I put a dollar number right there. See, the good thing, good news for us is the dollar amount keep going up, because Intel is asking for more power, and they're splitting all the rails. Now we have 5 rails for 1 processor, and each rail need a power, need a controller. The dollar amount almost like 50% increased.
That's all good story for us. For Whitley, besides the Wilson City, we also be on quite a few other key reference platforms with Intel. You guys will pretty much hear about this in the near future. That's Intel. Besides Intel, because it's a great technology, it just grow by itself, and a lot of SOC vendors came to us asking for partnership. As of today, we're having solutions for gaming SOCs, pretty successful. We have graphic GPU vendors we're powering, FPGAs of course, and we're also working with the Arm-based, the server and the tensor processors. Pretty much everything. Everything need a computing power. We have a solution, and we're engaging with them. Not to mention, we're working with worldwide OEM and ODMs, and they start to realize, hey, MPS is really a value provider.
Not just providing the technology, but also providing the service. Thanks to the worldwide apps team, they respond really fast, just like our device. It's really fast, customer love that. Great technology went its own way, and I'm pretty confident we can, going forward, and we're committed too, in this place. Remember I mentioned Purley growing pain? Here's the story I want to share with you. Look at the screen, there's two footprints. They look exactly the same. Industry call it the common footprint. The common footprint does have merit, because it gives the customers a solving the supply issues, solving the second sourcing issues. But to us, for MPS, it takes away our advantage of high density, because we really don't need to make that big. We don't.
As you can see here, on the left side, three dice, they have to make that big, but we don't. We are much smaller. It's a smaller design. For us to access the market, we have to do it. We did that, and that's what the growing pain is. We learned that, and our result has come out pretty good. The good news, although we make the device bigger, we're not sacrificing the performance. It's still a monolithic die, it still run really fast. Customer look at our device, “Hey, great, it's common footprint.” They look at how come you can run so fast? It's kind of different from the other common footprint. You say, “Yeah, that's exactly what we do.” That's the good news.
Once you mold it, nobody can tell it's a monolithic or, unless you use it, and they will see it's really good. I think the common footprint standard design can last for a while, and eventually, it will have a problem. The problem is, you look at the current. From Grantley to Purley, the current is double, the peak current. When the current double, you look at the processor, the size of the processor is more than doubled, because you're consuming more power, you need more pins to bring the power into the device. Not only that, because current increase, you have to push your VR, your power closer to your processor to minimize the loss on the power delivery. Because of all that, the size of the power supply need to be shrink. The density become a key element of the design. That's what we can do.
We are moving the things further closer to the processor. We're even working with customers, the SOC vendors, to put the power supply right on the package. The red line here is the available space. You see the current goes up, but the available space is coming down. The next question is why MPS can shrink the size, not everyone else? Besides the monolithic, and you look at that curve, the frequency over the years, we're starting from 400, 500 K switching frequency, and we're now doing three megahertz. Three megahertz. That's five, six X within four, five years. You look at the competitors, they just saturate. They clamp there. The reason they clamp there is a big hurdle. I just mentioned, it can't run that fast. Because of all this, you look at the size comparison, the blue bars are the MPS solution.
We continue to shrink the size and the black one, the gray one, is the competitors, because they're clamped at the switching frequency, you don't have enough room to squeeze. They stop there. Guess what? The same available space applied to competitor and to MPS, they're going to hit the wall. That's the challenge. It's real. It's very, very real. Just this morning, I was talking to a big customer. They bring this up and they say, "How can I fit that much of current power solution in my server?" There's no way you can push it. They can squeeze the inductor, they can get rid of the caps, but still the device, if you keep at 5 by 6, there's no room to push in. We provide that solution right away. That's good news. Okay. Here's another.
Well, from this slide going forward, I'm going to tell you a few stories, examples, the team has been working on, and we think that's really the future where MPS computing power is going. The first thing, that's a real picture, you guys probably see that, from a Nvidia GPU 2017 last year, and it was powered by MPS, by QSMod. It's a 20 phase design, very high current, need a lot of dynamics, and QSMod fit perfectly for the application. Perfectly. Last year, this design was powered by a 12 volt bus, because the power need is barely enough using the 12 volt to power this thing. Going forward, 12 volt going to be a limitation. You need a much higher current going into the processor. Industry are talking about 48 volt. 48 volt.
48 volt is 4x the voltage, and it's one-fourth of the current. Do the math. 4x. When the input current goes down and with the 48 volt supply, MPS is also committed to that area. We actually started two years ago for the whole chipset for 48 volt solutions. The next slide is just a very simple architecture. This chipset can do a lot of other things. It doesn't necessarily have to be 2-stage approach. This is just a good example. I think this solution, this architecture will take off eventually. You're taking a 48 volt coming in with the DCX, our own resonant converter, unregulated, and come to a very low voltage, intermediate voltage, 6 volt, which is perfect for our process because we optimize the monolithic process for a lower operating voltage. We can run even faster. 3 megahertz.
You see the 3 megahertz mark right there. That's what we're doing right now. Okay. With the 5 volt, 6 volt input, 3 megahertz, you can easily deliver hundreds of hundreds of current, and also make that orange box super small, so you can put next to the processor. Okay. That's the idea. Very simple. Scalable. Not to mention the transient, because it's running at a high frequency. You can interchange 1st stage and 2nd stage. It doesn't have to be both as a chipset, you have to buy both. You can use vendor A for the 1st stage, vendor B for the 2nd stage. There are a lot of flexibility. The bottom line is, efficiency is great, and we have a size and cost, everything has advantage.
This is the solution we are looking at to power mainly the GPUs and the AI chipset, which needs super high current. Speaking of AI needs data center, needs computers, the racks, every time people are talking about AI, the data center is a big farm, right? A lot of racks. Just like this picture shows. It may not be always like that, right? What if MPS can make that picture on the top, the data center, 10X smaller? We make it much, much smaller, denser. What if you can put that portable data center along your freeway? There are a lot of things it can do. The key is, we got to change the way our current data centers look like eventually.
We need to focus on technology and use our advantage to help power the next-gen processor, GPUs, TPUs, and they eventually become a super nice data center and you can just distribute. That's part of the actual autonomous driving, because with that AI engine distributed, you can communicate with your cars much more easier. Not to mention, we're also looking at all the processes inside the car, right? There are a lot of CPUs in the car, and that's a good place to be. We already have all the technology. We just need to focus execution, and that's our step. At the end, the computing power is not just about computers. The computing power is about the computing ecosystems. If you only have a computer, it doesn't work very nicely.
You need to link everything together, and you need a high-speed highway to connect all the data centers. That's what 5G network come to play, that's why we are also working on the hyper-converged computing. All this, again, let me remind everybody, need density. If you have a huge computer, you can't do that. MPS has been waiting for this, and it's the right time, right moment for us to thrive. I feel I'm very grateful and very lucky to have a team, elite team here, power team, and they work with me very hard for almost 10 years, and I think we achieved a lot after these 10 years. Let me finish my talk. I would say, come back to see us in 2, 3 years. You'll be surprised again. Thank you. Any questions?
Questions now or Yeah.
Yeah.
If you have a question, make sure you bring a microphone over to you.
Great. Okay. Quinn Bolton with Needham. Thanks for the presentation. Two questions. One, first started with 48-volt. When do you see that going to volume production? Will Intel have a version of the Whitley platform that supports 48 volts? Second, with the QSMod solution, did you say the Intelli-Phase elements effectively have 42 individually controllable-
gate drivers that turn off on and on each FET individually, You get that sort of quantum-
Thing. Yeah
ability that you talked about. Thank you.
All right. That was loud. The first question first, 48 volt. Yeah, 48 volt is not just Intel, right? It's pretty much like GPUs, TPUs, and all the other stuff. Intel has their advantage by this fiber building, and they're not in the very front end, has feel the need to see this huge current. Those GPUs probably hit that wall first. We see that 48 volt, in two to three years, it will be in really a good time. We'll see a lot of designs. Our MPS solution, we're building the prototype, all the stuff. We're working with a few key, I would say not OEMs, but key SoC, the silicon vendors to provide the total solution. I will see in two to three years, it will really have an impact in the industry, 48 volt.
Hi.
Hi.
Hi, it's William Stein from SunTrust. I had a question also about 48 volt.
Yes.
It's my understanding that the data centers are going to sort of allow 48 and 12 volts to coexist for some, perhaps extended period of time.
Does your product accommodate an architecture that would operate with both voltage levels?
Good question.
How do you expect that to progress? Eventually, at some time, does the whole data center change to 48, then are you more advantaged, or how does your position relative to competitors change as the architecture moves through this hybrid then maybe moving entirely to 48?
Very good question. For the hybrid, yes. We do see that need, actually a lot of ODMs are asking for, "Hey, do you have a 48 volt to 12 volt?" They can just use existing board. The answer for that is absolutely. We have the solution. Actually, the chipset we developed right now can do both. We can do a 48 volt down to six volt or even four volt, depend on the transformer, or we can use the same chipset for the so-called STC topology, the switching tank capacitor topology. That was very efficient for a low step-down ratio, like 4 to 1. It's perfect. The same chipset, different L and the Cs, the same controller, different L and the Cs. We can configure that as an intermediate solution for those guys looking for the hybrid.
If you ask me, what is our technology eventually land? What is the benefit we're getting? I would say the two-stage approach I just showed you with the lower intermediate bus, that really helps, and to promote our advantage. The lower voltage, high frequency second stage, not just solves the 48-volt problem, it also solves the CPU demand, the current, the power delivery problem. The last inch of the power delivery is the key, and you want to minimize that. When we make the second stage at 3 MHz as today, and maybe 5 MHz two years down the road, it can be so small, and you can see that literally next to the process. That's where those standard hybrid 12-volt solution cannot do.
Thank you.
Quinn has two follow-ups, right?
The QSMod.
Sorry, I'll ask the question, then I'll go back to Quinn. Tore Svanberg from Stifel. The good news about QSMod and Intelli-Phase is that they're very innovative technologies, very unique. The tough thing in the power management space is that new and innovative things, it can be really difficult to convince customers to move in that direction. I know in the server area, for Monolithic Power Systems specifically, there's been a learning curve for your customers to really adapt your technology. Can you maybe talk a little bit about where we are in that learning curve for your customers? Do you feel like now you are at a point where Monolithic Power Systems' technology is just going to spread across all platforms?
Well, I would not say that we're still on the learning curve. I think we're pretty much done learning and our customer learning about our technology. We start with all the SOC, the silicon vendors, right? The first target is we be on their reference design, and that's a really good tool for a newcomer like us to be on all the leading platforms and all the customers start to see it. They don't have to really spend the money and time to develop a board in order to evaluate us, because Intel provide all the tool to them with MPS solution on it, and they can see it. Once they see it, they come to us. They say, "Hey, tell me more about your QSMod. Tell me more about your Intelli-Phase." That's how the conversation starts.
We see a lot more momentum coming out, right? We can't control the step, the speed of Intel, because as you know, Intel push out the Ice Lake a little bit, and then we have to wait for that. Other than that, we really have a good momentum.
I guess just the follow-up question on the Intelli-Phase is it, just to discuss the architecture of the 42 individual controlled streams, if you will. Is that 42 driver MOSFETs driving 42 individual FETs?
The key to Intelli-Phase is intelligence. When we divided all this to 42 or 50, whatever numbers of banks, the gate delay from one gate pad to all the FET eliminated, because the driver is distributed, it's next to the FET. Okay? That allows us to do the transition from low-side FET turn off, high-side turn on. There's a dead time. You have to keep certain amount of time to make sure it don't shoot through. Since we control those gate and FET individually, accurately, we can cut that time to one nanosecond. Just for your reference, most of the discrete guys are doing eight to 10 nanosecond. That's 10x.
There you go.
Yeah.
You had mentioned some of the things that always challenges with the 10-nanometer program. My thesis personally is just going to be more use of all the chip models in server, potentially over some interconnects, like you said, or things that might happen. If you can talk a little bit about how you program a lot more chip models approach taken by some of the preventions that might require how might the player program like that type of target?
That was actually was my dream a long time ago, is why can't we integrate everything power into the process, right? That solve the problem. Of course, the industry is so big and it's not an easy job. If someone were pick a silicon, a power IC vendor to make that move, I think MPS is in the top of the list. Our technology is so suitable to be integrated, co-packaged to the process. You're asking me, it's happening right now? It's still early stage, because it's not an easy task. Okay? Thank you everyone.
Okay. All right. Hi everybody. My name is Chris Sporck. I'm the Battery Management Product Line Director. I've never met any of you. I checked the name tags just to make sure. I'm new here at MPS. I've been here just over a year. I came from Qualcomm. A little bit of background on myself. I was running the PMIC System Design Group for the battery chargers. Quick Charge, these types of technologies, USB-C, USB power delivery. Really from the engineering side. Now at MPS, I guess my title is marketing. That's interesting. It's been very exciting, and I'm very pleased to be here. Thank you all for coming and listening. Let's get started. Oh, I have the clicker. Yeah, I didn't even think about that. Forward. Okay. There's my name.
What I want to do today, it's only a 15-minute presentation, but I want to spend some time and explain to you sort of the journey that we've been on and where we're going, how we're really going to monetize and grow this business, and what different markets we're going to attack, and how we're going to win. First off, on the screen here, I have a variety of different market segments or market verticals, and in the order of, I believe, increasing value. On the left-hand side, these markets that are lower ASP, high volume, are usually a lot easier to enter. If you're comparing a power bank, for example, to an electric vehicle, it's a lot easier to design that chip and to get a design win. You shouldn't discount it either, because there's a lot of value there.
MPS, when we started in battery management, portable power was our first big hit, I would say. What we did there is we looked at the system, and we saw, okay, this device has an input, it's got an output, for charging a cell phone, and they've got separate converters for all these things. What can we do to make this system more efficient, smaller, and lower cost, and higher value to the customer, easier to design? What we did is we introduced a charger IC with an integrated buck for charging the battery and a boost for providing power to the phone. What that did is it allowed us not to only take the revenue for the charger socket, but also the boost socket without having to provide both solutions. We saw that market grow. The revenue grew there a lot.
What we did is, you need to diversify. We have other products, such as linear chargers for growing into wearable devices. Our advantage is we have a very, very small device using our process technology. I believe one of the smallest in the industry that's programmable. We have some very nice design wins at wearable, it's fitness band, smartwatch, this type of thing, Bluetooth headphones, which is another big area. All the cell phones are removing the audio jack, which means you're either doing Type-C or you're doing wireless, which is great for us because wireless headphones means another battery, and we love charging batteries. A couple of the other adjacent markets, the non-USB applications, this is high voltage stuff.
We also have some very nice, high voltage versions of our processes, and we use those to create fully integrated solutions for portable vacuum cleaners, POS machines, this type of thing. Finally, connected devices. I just loosely grouped this category, but what I'm thinking there is sort of like the IoT of stuff. It's like connected cameras, gaming controllers, Bluetooth speakers, this type of thing. Bluetooth speakers is especially interesting to us because this market is growing rapidly right now. A lot of the times, it's very similar to the power bank. You've seen probably some of these devices actually have a charging port for your phone. It was a natural fit for us to attack that market because we already had a mature solution. Anyways, that's what we have now, and we're growing in those markets a lot.
The question is: how are we really going to expand the business and increase the revenue in a big way? That's where these two categories on the right come into play. First of all, mobile computing, okay? Everyone knows this. It's like your cell phones, tablets, computers. I think last year there was $1.5 billion smartphones shipped globally. This is a huge market, but it's also very difficult, because you have system on chip providers like Qualcomm, for example, who sell a reference design. For a company like us who's just selling a charger, we're not doing all the software for them, it's difficult to penetrate unless you have a very key new innovative technology.
The other thing I want to say about mobile computing is that right now we're also seeing sort of a revolution in that area where all these devices are changing over to the Type-C connector. I spent a lot of time working on that specification, doing the definition for parts which integrate some of that functionality, and I really see that as a key point for growth. We have to understand that the consumer electronics industry, wireless devices in particular, are going to move to Type-C because it means universal charging. Anyone can charge anyone's device with any charger. Manufacturers don't even need to ship a wall charger with the product because the customer already has one. We're going to take advantage of that trend. Finally, battery management systems, BMS. What is this?
This term collectively refers to a fuel gauge, cell balancing, cell protection, this type of thing. Portable devices have a minimum of one battery cell, but then the battery pack itself may actually have 16 cells in series, stacked on top of each other to achieve a higher voltage. Every single one of these battery packs needs to have protection. What happens if there's a short circuit? We don't want the battery to blow up, right? This market, we can sit there and say, okay, yes, electric vehicles. That is the pinnacle of this market. How do you get there? What do you do step by step to achieve those high-value design wins, high ASP? Well, you have to look at the other large markets in this category, which, for example, electric bicycles.
I think last year, the quantity of e-bikes shipped globally is over 10 million, and it's growing very quickly. You have power tools. You have cell towers. Just think about the migration from 4G to 5G, right? There's projected to be something like 3X the number of towers in the U.S. because it has to be physical eyesight millimeter wave technology. That's hundreds of thousands of towers that all need a battery backup, because if power goes out, you can't have your communication lines go down if there's a disaster or something like that. Anyways, these are the type of markets that we are actively talking to customers right now and have products in development that are targeting.
What I'm going to do after this is I want to focus on those last two categories and show you a little bit more about what that means and what we're going to do to grow into those markets. Okay. Basically 2017. All right? We have baseline revenue of around $16 million for this product line, okay? I've separated the markets into different categories based on the topology of the chip. BMS, we already discussed, that's that whole separate category. Linear charger, that's the small little things for wearable type, tiny batteries. High volume, lower cost. Boost, buck, and buck boost chargers are all switching chargers, so very familiar with these types of topology. We have all of those solutions right now. We are working on releasing the buck boost solution, and I'll explain to you why that's key.
Basically, the point of this slide is to show you, okay, how are we going to grow our revenue? Well, the first step of that is diversifying our product portfolio and targeting these other markets, these other key markets that I'm talking about, in addition to growing in the existing ones. What you can see here is over the next few years, we are going to release two production, buck boost chargers, BMS ICs, and also expand our existing charger product portfolio such that we can attack other markets such as notebook computing, 2-in-1, cell phones, this type of thing. For us to not be in those markets means that we have a huge growth potential ahead of us. Okay.
I think, looking back at the previous Analyst Day presentation, I saw a similar theme to what I was thinking, with what the battery management product line presented, and that is integration. What is our advantage as MPS? You've heard from the computing product line and DC-to-DC, that we have a really good power process. Yes, we can build a monolithic device that is sometimes twice as small as our competitors for the same RDS(on). Okay, but what does that mean in terms of a charger or a battery management IC? Well, what it means is it leaves us room inside of the package to put other features in there that our competitors normally have discrete ICs for. Typically what happens is the competitor goes to a customer, and they say, "Okay, look, you have a battery management system. How many cells is your battery?
Is it USB-powered? Whatever. They say, "Okay, well look, we've got this product portfolio. We're going to sell you four devices or five devices or six devices." The customer has to pick, okay, I've got a charger, I've got input voltage protection, I've got my USB detection, I've got my battery pack protection, all of this stuff. They have to lay out all those things. They have to do that design. They have to validate it, et cetera. Okay.
What we're saying is, look, we don't want to go there and do this super complicated PMIC, where we've integrated every feature, because when you do that, it usually only works for one customer because everyone else looks at you and is like, "I'm not paying for this." The key is to intelligently look at these market verticals and say, "Okay, this market is USB-powered. It's a two-cell battery. They need cell balancing and a charger. Let's integrate those two features." They need USB detection. They need overvoltage protection. We've already done that. We have a product out in the market right now that's a boost charger, USB compatible, and it integrates the cell balancing, which is very nice for a variety of different markets.
What I'm saying here is, we are using our PowerFET process advantage to take that extra area and intelligently choose which features to integrate in that same package based on not just what one customer wants, but what a large quantity of customers in the same market want. That is going to win. Okay. Remember, there were those two categories, those two different markets that I wanted to focus on, of the growth markets, growing the SAM. BMS. You look at this and you're like, "Wow, there's so many things there." There really is. You think about it, from power tools to electric vehicles, to the power wall for your house, this type of thing, there's really a lot of stuff. One approach could be, let's just cast a wide net and try to attack everything.
Now, you can do that, but the problem is, if you want to get the coup de grace, the EV high value, maybe $8 per BMS chip in the electric vehicle, and it's got maybe 100 of them. If you want to get that socket, those designs take a long time, and those customers want to say, "Okay, well, what's your track record in automotive BMS? What other design wins do you have? Do you have any proven your automotive process?" This blah, blah, blah. We do have an automotive product line. We do have proven devices that are automotive qualified. We need to show them a track record of BMS design wins in order to build the confidence that we can use that in an electric vehicle. That's what we're in the process of doing.
How we're going to do that, is we are attacking the electric bicycle market, the power tool market, the battery backup market, like the cell tower, small cell, this type of thing, and we are already engaged with those customers right now. We do have solutions in development for this. I believe this is definitely a big part of our future, and we do have some experts in this type of technology who are working on it right now. The second thing. Mobile computing. This is on the consumer side. This is what I was talking about, smartphones, laptops, this type of thing. I want to tie that all together and say, I believe the key to taking on this market is doing something different, and that is USB Type-C and PD.
You can look at this chart, and basically what it's saying is, this is the number of connectors shipped, not number of devices. A laptop may have four connectors. Look at that. We're talking, I think, can't really see the screen exactly, but around 2.5 billion units this year, and doubling that in 2021. Pretty much 80%-90% of that is consumer and wireless. Obviously, cell phones kind of led the charge there with adopting that technology, but you can see right now that computers are doing the same thing. Apple has Type-C on all their laptops. You're starting to see business laptops now being shipped with Type-C. The key there is we want to take advantage of this growth, and that's what we're doing.
The products that we are developing right now are taking advantage of this by integrating functionality and making sure that we have the compatible features for the spec. More about that. More about the exact products. These are the type of different devices. Notebooks, smartphone, tablet, power banks, like before, VR headsets, wearables. I touched a little bit on this, but the wireless headphones, that's a big market right now. Apple came out with the AirPods and this type of thing, but none of the phones have a three-and-a-half millimeter jack anymore because it adds cost to the phone, makes waterproofing it more difficult, and it takes up more space that a battery can use. That's great because, that means more headphones with more batteries. We have solutions for that right now.
The very key thing that we're working on that we are sampling a product right now, is a buck-boost charger. What this means is, to support USB, you have to support five volts, okay? That's no problem if you only have one battery cell in your device, because it's one cell, it goes from three volts to 4.5. As soon as you have, let's say, two series cells or three or four, we're talking laptops, speakers, this type of thing that has higher power. The higher current. Sort of like Jinghai was talking about with server, you go to higher voltage, you get lower current. These devices go to a higher voltage battery so they can reduce the current. If you do that, you need a buck boost because you need to take five volts and step it up to the battery.
As soon as you need more power, if someone plugs in a beefy power delivery charger that's 100 watts, that thing can go up to 20 volts, now you need to buck down. That's why you must have a buck-boost charger in your portfolio, preferably many of them. What we're seeing right now, if you look on the market from competitors, the only high-power solutions that are available all have discrete FETs. It's a very similar story to the computing product line. Everyone has this huge solution with four to five MOSFETs, and what we're doing is we're integrating it, of course, right? That's what we like to do. We're monolithic. We integrate things. Customers are very interested in that because we can reduce the solution size in half, at least, because we have all the FETs in one four-by-five package, let's say.
This is the key, and this is something that we see a lot of growth potential in. Finally, like I said, it was a short presentation, but I just want to leave you with some points of what we're going to do to win here. Point number 1, increasing product offering diversity. That curve, increasing the available SAM for us, we have to spread out. We can't just focus on one or two or three markets. We have to have solutions for all these markets. We need the buck-boost charger. We need the BMS IC. Trust me, we are working on that right now. You will see at the next Analyst Day. The growth strategy focused on a high level of integration. I can tell you that, my experience before was all on the smartphone side, where everything is integrated.
Space is absolutely critical. Like I said, we need to intelligently point out those opportunities for which features to integrate in the same chip and do so for a wide variety of customers. That's what we've been doing for the last two years, to be honest, and we already have products that are in production that demonstrate that. The 3rd point, fully monolithic chargers for USB Power Delivery. Like I said, the current solutions are all discrete FETs. We're going to do the same thing with that market that the computing group did with the computing market. We have a proven track record there, so I know we're going to be successful. Finally, for BMS, you might ask, okay, well, why is MPS going to win in BMS? It isn't super high power density. It's not really a power IC.
The thing is, we have a lot of R&D investment going into this. It is all about speed and accuracy circuits, right? You have to have a good design team, a good systems team, and we do. We've absolutely been hiring in that area, and we are developing some very accurate monitoring circuits that will compete with the big guys in that market. Anyways, with that, I'll close and point to the question slide, and please give this gentleman a microphone.
Thank you. One of your peers pretty close up the road here, they're working on a charger technology that takes the charger function from the device onto the adapter itself. I know historically, MPS hasn't really participated much on the actual adapter market. Is that something that you think will happen? Do you have technology that could potentially participate in a technology development like that?
I can speak to that in the highest volume market that that's in, the cell phone market, commonly referred to as direct charge. I guess that's what you're talking about, right? We actually do have an AC/DC product line. Goran's right over there. You should chat with him about this after the meeting. I do have a lot of experience in that area, and without naming particular customers or competitors, basically, I'll tell you what happened. USB Type C and PD, the cable can handle 3 amps, okay? That's a standard cable, the one that you have the charger, and you have your device, and you can unplug it from the charger and unplug it from the device.
If you want to go to a captive cable, which means it's permanently connected to the charger, like a laptop, then it can go to 5 amps. The thing was, is that there were a few Chinese smartphone OEMs that thought, "Oh, this is a great idea. We're just going to do direct charge. We're going to take the regular charger that's there in the PMIC, right? We're going to leave it there because that'll be used for when someone plugs in a regular charger off the shelf, right? The buck will step down and charge the battery.
When someone plugs in our special charger, we're going to identify it, we're going to turn off the PMIC, and we're going to bypass it with this back-to-back MOSFET that's super low RDSon. What they're going to do is they're going to take that wall charger, and they're going to either regulate the output current like a current source, or they're going to have a fine voltage resolution change. What happened is they started doing that, and that worked great. Problem is, for the user, they have to carry that one charger with them, which kind of sucks, right? I can't just use yours. You can't use mine. It's this type of thing. Those companies actually purposely tuned down the current for when someone plugs in anything else to make their charger look so much better.
That's what happened, but anyway, these guys went down that road. Type-C comes out. They're like, "Okay, we're going to do this over Type-C now." Type-C is sort of the unifying standard, right? We came out with the power rules, five, nine, 15, 20 volts. Basically, China said, "Well, we're not going to adopt this unless you give us the option to do this super fast charging." We did. They have that option. It's called programmable power supply, PPS. The fundamental problem with the technology is even if you get around the cable thing, the 5 amps, 5 amps isn't enough. If you want to fast charge some of these newer smartphones, they're like 25 watts, 30 watts even. It sounds ridiculous, but it actually is happening. The question is, what do you do from that timeframe?
Now what people are doing is they're going to a different topology of charger, which basically divides the voltage by two inside the phone. You use that same type of wall charger that can regulate the current, except you double the voltage so that the power can increase, and you can still stay under the 3-amp limit of the cable, and it double the current. 3 amps in, 6 amps out, right? Yes, people are doing that, but I don't think that's the future because none of those converters are interoperable, and they're expensive. You're adding a lot of cost to the AC/DC, and I think Goran can speak to the fact that those are pretty cost-sensitive. You want universal charging. I think the consumer is going to make the choice to buy something that's compatible, right? Helpful. Thank you.
I hope that wasn't too much. No, it was great. Thank you. Yes.
Maybe this one will be shorter. On the BMS side of things, you talked about MPS needs to earn the credibility to get into that market. What are some of the milestones we externally can watch for how you earn that credibility? Given the design cycles on the automotive side of that market, when would you think revenues could actually come once that credibility is earned?
Let's take a step back and say, okay, those first markets that we're talking about going into with BMS, that we're engaged with customers right now, like electric bicycle, power tools, this type of thing. I think the e-bike is a really good example case. We have some customers that we're talking to, and if we execute on our schedules and stuff, I think that end of 2019, early 2020, we can start to get design wins there. Then, grow that business, expand into other customers, this type of thing. In parallel to that, we've also got a few external projects that we're working on where we're basically pairing our BMS IC that we have that's in development, on certain designs. For example, tear down an e-bike or working with some students who develop race cars for a university, this type of thing.
Basically, we're doing the R&D and the proof of concept effort on that side and engaging with the customers in parallel. Once we have the design wins there at the, let's say, the e-bike and the cell tower battery backup, power tools, this type of thing, we are absolutely going to continue knocking on the doors of these automotive guys and showing them, "Hey, look, we've done this, we've done that." Alan has the relationships with these customers, and we're going to leverage that to our advantage. Does that answer your question? I don't know if you're looking for Q3 of 2019, we're going to do X. I can't tell you that.
That's good, thanks.
Just a follow-up on the USB-C.
Sure.
USB, was that just mostly a step down or a buck topology? When you go to USB-C, you change the buck-boost and that's your opportunity to kind of engage. You've got a lot of incumbents, I'm assuming, in USB power, been there for a while. If you could just kind of simplify it a little bit, what's the change it gives you as sort of a new player-
coming into this market, the opportunity to kind of come in and take share? I would think most of the power management guys that are doing USB-C or historically did USB-C probably have buck-boost capabilities in-house. Is there something that gives you an edge as you make that transition from USB to Type-C?
Sure. Yeah. It's really about power levels. The previous USB that everyone's using is micro USB, the little trapezoidal connector. That's limited to 7.5 watts. Okay? It's 7.5 watts, and that's 5 volts, 1.5 amps. If you look at Type-C, excluding power delivery, you can go up to 15 watts from 5 volts, that's 5 volt, 3 amps. Okay. Where the big power comes into play is as soon as you add power delivery, it means you can go up to 20 volts at 5 amps, 100 watts. All those markets that were taking those, or let's say the power semis who were taking those sockets before, those converters were all taking a barrel jack input.
If you look at probably the laptop that you brought to this, except for the person with the Apple over there, it'll have a barrel jack, it'll be 19.5 volts, and it'll be 65 watts. The architecture is a buck because they have a 2-cell or a 3-cell battery, and the adapter voltage is always higher than the battery. They know exactly what the power is available, and they know what the adapter is. With USB, it changes the whole thing, right? Because you don't know what the customer is going to plug in. They could plug in their old 1-amp Apple brick, or they could plug in a 100-watt charger. You have to have a solution that can go up to 100 watts, and you have to have a buck-boost. Before they could have a buck, the buck-boost, that means more FETs.
now if you tear down one of those notebooks, you will see the solution's like this for the charger. The inductor is like a 10 by 10 millimeter. If the system is thinner, it's even worse. If you're doing a solution like a Surface or something that's a two-in-one, you need maybe 1.2-millimeter height on your inductor. yeah, you're doing a 10 by 10 inductor. what you need is, kind of look at what computing did, right? The multiple phases, monolithic devices, bring down the area, increase the frequency, this type of thing. It's the same thing there.
The more flexible you make the FETs and smaller package.
Absolutely. Yeah. I wouldn't have brought that up if this wasn't what I believe truly a consumer electronics revolution because there's been guys playing in the notebook charging market for 10 years. I'm not going to kid you and say, "Oh, we're going to walk in there and just design them all out." There has to be a disruption in technology that happens, and that event is happening now. Other questions? Let me off easy here. Okay. All right. Thank you. Oh, yeah. Oops, I didn't turn it off.
Okay. You guys hear me okay?
Yeah.
Good afternoon, everybody. My name is Alan Chen, and I'm the automotive marketing director here at MPS. I want you guys to know that I consider myself to be very lucky, and I consider my team back there to also be very lucky. As you guys should know, the automotive market is going through one of the biggest, most interesting technological revolutions ever, right? MPS, we feel, is in the right place at the right time right now to really succeed in this space. Over the next 10, 15 minutes, I want to talk to you guys about how we're doing in the automotive market today, how our automotive business is. I want to talk to you guys about the major themes shaping our innovations, our R&D over the next few years.
I want to talk to you guys about how MPS is nicely positioned to really succeed. Okay? First, we're going to start off by talking about how our business is today. Five years of development and focus in this market. From 2013, $8 million to last year, $53 million. 7x growth in five years. Let me tell you this story. In the very beginning, we decided, let's go after this space. What we did is we started by drawing from a catalog of 1,000-plus power products, right? Taking them, converting them to automotive, qualifying them, and selling them into this space. What that experience gave us was an understanding of how to develop IP for the automotive market.
It helped us to understand the customers and really figure out what their care abouts are and really figure out what this very unique ecosystem looks like. Fast-forward to five years, 2017, $53 million. We've come a long way. Today, we are now developing all of our products from day one for the automotive space. When we get the opportunity to, we may downgrade that and sell it into the consumer space. It's a fundamental shift in our philosophy, our approach as into the automotive. The other thing I want to also point out is we're growing so fast, okay? 55% year-over-year growth last year. Take a look at the total semiconductor market in auto. It's about 13%, right? 4x growth versus the market. You can see the TAM that we're going after, $7 billion. $7 billion.
Our slice right there is that little green one, which you might not even be able to see in the back. The cake that we have in front of us is huge, right? We've only now just begun to start growing. I also want you to understand a little bit about what our business looks like. It's very diversified. The first little pie chart there on the left shows you how our revenue is built up. The statistic here is 50% of our automotive revenue comes from small customers individually contributing less than 1% to our top-line revenue. Really, our goal is to grow that entire base of customers, right? The second one there, it's a little bit hard to see, just visualize, but basically, we're very nicely diversified regionally. North America, Europe, Japan, China, Korea.
Our business isn't made up in one region. It's the entire world where automotive development happens today. That's the first part, talking about kind of where we come from. I also want to just refresh you guys, right? The automotive landscape, how does this work? You have the Hondas, the Toyotas, the Mercedes-Benzes that you see on the road. These are the automotive OEMs. These guys, what they'll do is they'll define, okay, I've got an infotainment system, I've got my HVAC, and then they'll define what they need in this new vehicle and actually work with tier 1 automotive suppliers, right? A single OEM, a single Honda or a BMW may work with multiple tier 1s to develop that whole vehicle. The tier 1, in turn, will also develop infotainment systems or radio or cameras for multiple OEMs.
Today, MPS is already shipping into half of the top 50 tier 1s worldwide, and we are either ramping at or deeply engaged with virtually all of the others. Let's talk about some of the big themes, these big tectonic shifts, right, that are really informing what's going to be happening next. Really three major ones. The first one you guys all know, autonomous driving. The goal eventually, and I live in San Francisco, I want to get in my car in the morning, go to sleep, and wake up here. It's an hour-and-a-half drive sometimes. It's really far. I think eventually we're going to get there. I don't think it's going to happen in the next few years. I think that people underestimate just how complex this challenge is.
Certainly, what we're already seeing now is level 2 vehicles where I can kind of let it control itself, but I still have to pay attention. That exists today. Level 3 already exists today, and there's so much investment dollars behind this, it's going to happen eventually, right? Second one, I wake up from my nap and I want to watch "Game of Thrones," right? Connected cars. Cars already today, you can have Wi-Fi hotspots. You can get 4G and eventually get 5G connectivity. Cars are going to be talking to other cars, communicating with street signs and streetlights and all this. The idea of the connected vehicle, it's a logical step once you have this push towards self-driving. The third category is electric mobility, so electrification.
Of course, like Chris was talking about, there's this holy grail of fully EV, fully electric vehicles or hybrid electric vehicles. The reality is it's not going to become the majority of cars over the next couple of years, even over the next 10 years. There actually is this very interesting intermediate stage, 48-volt batteries. The idea here is you still keep that 120-year-old lead-acid battery. It's been around a long time, it's very cheap and reliable. You actually add this intermediate 48-volt battery, and that allows you to electrify some traditionally mechanical components. It gives you slightly higher reliability as a result. It allows you to get a little bit better performance benefit from your car, and it also gives you some fuel economy savings. This is really, for a lot of folks, a good intermediate zone over the next 10, 20 years.
Some projections think maybe 10% of the market may shift to this over the next 10 years. Okay. When we think about these three major themes, what does that mean for us? The obvious thing is more semiconductor content, right. If you take a benchmark, today a normal level 2 vehicle, internal combustion engine, versus the car of the future, fully self-driving, electric vehicle, up to 18x the amount of semiconductor content in the same vehicle. That's really good news for us. Like I said, we feel very lucky. For us, of course, that means that already giant cake of $7 billion is only going to increase, easily getting to $13 billion over the next five, six years. Really nice opportunity for us. From these themes, we think about what do we want to go after inside and around the vehicle.
MPS is not here to do me-too parts. We're not here to do copycat simple parts. We're here to offer differentiated solutions, something that gives the customer a compelling value, motivation to design us in. From this, we have five major ones. The first one, the digital cockpit. Think huge screens displaying content on demand, entertaining you. Think USB charging all over the vehicle because we have so many gadgets now that need to be charged. Think about clusters and heads-up displays. Actually today our business is primarily made up from this. Infotainment is already a thing. The shift towards going from traditional analog radio dials is already happening. First category, digital cockpit. Second one, lighting. The shift to LED lighting for the headlamp is actually still underway. It hasn't really happened in full.
Today, you think about some cars that you see coming down the road, you can actually start identifying them by the headlight design. When I first watched "Iron Man" and I saw the, was it the Audi R8, I fell in love with the brand as a result because it's such a beautiful headlamp design. That's already stuff that exists today. The car of the future is going to get even more intense. That headlamp is now no longer just providing illumination, giving you a, "Oh, what is that car?" It's actually signaling intent as well. Once you go to self-driving, that car needs to be able to tell you, a pedestrian, a cyclist, another vehicle, what is my goal.
Today, maybe you can look at the driver and say, "Okay, well, he's going to slow down for me, so I'm okay," the car's not going to tell you that. We see a lot of very cool developments. We have those relationships with these customers, some very interesting stuff happening in lighting. Third category, body electronics. We talked about that 48 volt. Electrification of some mechanical components. Basically, this means anything that's motorized in the car. Seats, side mirrors, of course, long have been motorized. Think about those door handles that pop out when you get to the car and go back in. A little bit extra aerodynamic quality. Think about lift gates now. If you buy luxury cars or even high-end SUVs, the door, the trunk closes and opens by itself. Fourth category, battery management.
Again, the goal is eventually getting to HEV, EV, stackable, complex 400-volt battery management solutions. It's going to take us investment and time to get there. Today that short-term target of 48-volt battery is achievable. It's achievable with the IP that these guys are working on. Fifth, last but not least, ADAS. How much are we going to need in terms of sensors and computing to get to full level 5 self-driving? I don't think anybody knows yet. Certainly, this means radar, camera, lidar, ultrasonic, thermal maybe, a lot of sensors in around the car, even ones looking at you, the driver, to make sure you're alert, awake, present. We see these five major categories.
You'll see that there are some that are missing here, the reason why is because these are either very low-margin opportunities, or we don't see really nearly the kind of innovation that's going to be demanded of us over the next five years. Five major automotive applications. Why is MPS going to win? It always comes down to great technology. You guys heard this from the other two presenters. Really, we have to walk in and offer a compelling reason why they should shift. There's just a couple different vectors that you can choose from. At the end of the day, I want to offer more integration, better thermals, lower noise, lower EMI, something that's a compelling reason. The first one on packaging. You guys know about our Flip chip power packaging. We've been doing this for quite some time. I'll tell you guys a secret.
One of the top five tier ones in the world We are the very first supplier to actually have that Flip chip power packaging qualified by them. I walked into the meeting last year and I told them about this and they said, "This is crazy. There's no way this is going to work. We've tried this from your competitors in the past and it failed horribly." I said, "Well, give it a shot. Try it out." Because we've gone through right there billions of units, right? We know how to do this and in high quantity. Sure enough, they went and they tested it. Eight months later, it passed. They were shocked. We weren't very surprised. For us, that power packaging really offers nice benefit, in some cases, twice the power density of some of these other solutions. Second one, integration.
Here with that Flip chip packaging, what we can also do is start integrating more stuff into the same equivalent package. It looks maybe just like the same QFNs that our competitors are offering, but there's a ton of stuff inside. In one case, we have this LED driver that's four times smaller than the equivalent solution from one of our competitors. Really nice benefit in integration, reliability there. The last one, a full power tree. We can now walk into customers five years in and offer solutions. For example, think of an infotainment system. You'll often find 15 to 25 different switcher opportunities in there, we can provide a full power tree today. We're one-stop shopping for the customer. In addition to that, we've got some great leadership and expertise in very, very extreme and complex topics like EMI or thermal management.
We can provide all that technical support. Let me give you two quick benchmarks. The one on the left is a competitor solution. This is USB charging. Every car has this now, right? The competitor part right there, one, two, three, four, five, six different ICs needed to realize this solution. MPS can do this in one single chip that's equivalent in size to one of their devices. It almost seems too good to be true if it weren't for the fact that we're shipping millions of these already. The second one was that module that I showed you guys. Here, it's not just a switcher. It's got some inductor capacitors inside, and it fits into that little tiny three by five QFN.
This actually has allowed our customers to create some form factors for their applications that they couldn't otherwise achieve because they couldn't get the board size down enough. For some of our lighting customers, they're able to achieve exotic-looking lamp designs because of this kind of technology. It's a really, really nice differentiation for them. Just to wrap up, really, we're at this moment where the opportunity is incredible. MPS happens to be at the right place at the right time. We've learned from the last five years of innovation, of selling. We understand the customers. We have the great products. We see the growth really continuing over the next five, six years, easily 40%-50% CAGR. You'll notice that the units are missing here. We didn't want to make our competitors jealous.
You guys can throw some math at this and figure this out based on the numbers that I showed you earlier. Okay? Thanks very much.
I had a question about processor in the car.
Especially thinking about the competitive landscape. We know Qualcomm obviously is trying to buy NXP. I think we all understand a little bit why. If you think about that specific opportunity, how does MPS intend to go after that market? Because I'm sure there's going to be a lot of reference work that's going to be required. I also assume that the power management for the infotainment processor or the ADAS processor, whatever's going to be in there, I'm sure it's a pretty big dollar content opportunity.
Good question. Obviously, every processor, GPU, FPGA company wants this space. They want to play, right? In many of these cases, we have those relationships already. I can't comment publicly on this, but suffice it to say, if you know the right folks, you have a much better opportunity to get in. We have the differentiated power solutions. We actually are already on the reference designs. You can go to our website and see for a lot of the folks that are really aggressively going into this space.
Thank you.
No other questions? Okay. Feel free to come find me afterwards. Thank you very much.
Everyone, we're going to take a quick 15-minute break, feel free to help yourself to the refreshments in the lobby, and we'll see everyone back here at about 3:45. If everyone could take their seats, we're going to be starting in a minute or so.
I'm on. Yep. I'm just waiting until the last ones are seated. Good afternoon, everybody. My name is Jens Muttersbach. I joined the company when we were acquired as Sensima Technology four years ago. I already asked you to excuse my Alpine accent. I'm going to talk to you about eMotion, that's something that Michael presented three years ago, basically as a vision. That's a slide from that presentation. Just to recap, Michael showed that we are able to measure really an angle position, yet the width of a hair, finally integrated, the concept, the vision is, to move that into an integrated solution to really revolutionize the motion control market. I want to show you where we are with that today. What's the status of eMotion basically as of today? What did we do over the last years?
The important thing for the beginning is that eMotion is really the concept of having a fully integrated solution for motion control, which is then really a one-stop solution for, at the moment, rather complex things. That's founded on two pillars that we need. The one is we need to know where is the rotor of a motor, really. That's an angular sensor that we need. The other key technology is that we need to drive the motor, that we need to provide power. These two pillars are of key importance to then, in the end, carry that overarching vision of eMotion. As most of you are very familiar with power things, you might not be that familiar with that motion thing, motion sensing and motion control market, just to categorize it a bit.
We have things in your everyday life, your mouse wheel is the angle sensor that everybody has something like 10 hours per day in their hand, it's also gaming wheels, things like that. These are pure angle sensors. On the other hand, you have also in the consumer space, a lot of applications that couple that with a motion, with a motor already. Handheld gimbals, robots, also printers are the normal things. That another thing is in the industrial and automotive market that simply carries on. Motion sensing and motion control is one of these ubiquitous technologies that we take for granted, that we are quite often not even aware of anymore. Now let me explain how we are tackling these challenges in this market. I want to dive into the first pillar, which is the angle sensing.
At MPS, we have that SpinAxis technology, which enables us to really sense the position of a rotor instantaneously and on a very small chip. That's an MPS proprietary solution that's patent-protected, and where others are doing a lot of computation, we simply transfer it into a smart and simple time measurement. Sounds very technical, but in the end, we derive customer benefits out of that on the product. We quite often replace bulky optical encoders, which are sometimes the size of a fist, into a three by three or two by two millimeter package. We are able to sense very fast. We have a very robust magnetic setup and can deliver that with very low power consumption.
All of these are technology features, but these have been integrated in a complete product family that we dub MagAlpha, which are then tweaked into the different application fields, into the different interfaces, speed, and resolution needs that people need. We have considerable design wins across all of these markets already, from consumer, over industrial, automotive, into medical. There is the second pillar, the motor drivers, basically the muscles providing the power to really energize windings in a motor. There we are, of course, being a power company, we leverage the strengths that we have there. It's our unique technologies, it's the packaging, the testing, and also knowing how to support these things out during the design and in the market. Again, we derive customer benefits out of that.
We are smaller, more efficient than our competitors with these motor driver things, which provide thermal advantages and cost advantages across all of these different DC motors from steppers brushed over to brushless DC. Again, there it's converted already to a good extent into really design wins because we are able to address these markets with a much higher integration means. The customers don't only get a power delivery chip, but it has all the protection circuitry, all the first levels of logic already integrated. That enabled us to win designs in also across markets, consumer, industrial, and automotive, maybe drones, maybe e-bikes, robotic arms, and similar things. We have these two pillars where we, in both pillars, have unique technologies, unique advantages for MPS.
The next logic step for MPS is, of course, and that was Michael's vision years ago already, to then combine that into the eMotion thing. We are pretty lucky, so to say, that we are really meeting a market in motion because the sheer number of electric motors is increasing. While that market is expanding, the fraction of that market which is being brushless DC is increasing at the same time. That's due to the advantages of brushless DC motors in efficiency, space, torque ripple, audible noise, things like that. The big challenge in that market is that when you want to really get all of these nice features, the investment in controlling them is ever increasing, and that is, for many applications, prohibitive. Let's have a look how such a motion system usually looks like.
The standard solution is based on a brushless DC motor, then you first have Hall elements to control the commutation. You have an encoder for a precise measurement of the position, then you need signal processing and a drive circuit next to it. All that first comes with a pretty high price tag. Hall elements, if you take the cheapest ones, you're at half a dollar, you're up to $2. Optical encoders can range anywhere between $2 and $100. Then all the processing circuitry, drive circuits then attached to it can range from easily $80 to $50 or even above dollars. The price tag is pretty heavy. At the same time, these things tend to be rather bulky. One of these rules of thumb is that the complete circuitry controlling the motor has two to five times the volume of the motor.
Now we start playing with our unique technologies, we are able to integrate everything controlling the brushless DC motor into a single combined solution. That means we have the position sensing, we have the signal processing, and the motor drive circuitry in a four-by-four millimeter package, fully integrated. That's an existing product from us. Not only have we managed to shrink all these things into that small package, but on top of that, we're able to provide a so-called field-oriented control algorithm inside that. That's a very advanced algorithm enabling you to get the best out of the motor itself, the best features in terms of torque, in terms of power consumption, thermal performance, and similar. Some customers even get more value out of that than just reducing the circuitry. They also get better performance out of the motor or can downsize their motor.
Let's have a look what that has an effect on the market. What we are already seeing, what I showed you, part of that is that we are designed in and have revenue already from, let's say, the industries with the fast design cycles. We are to a good part in handheld gimbals, in drones, in toy things that have a very fast turnaround cycle in their design. We already have secured also the design wins into the next levels of complexity, into industry markets, into automotive solutions, into, for example, fans, where we are enabling to couple two fans into novel server ventilation things where you need to synchronize these fans, things which would have not been able without our solutions.
We are then also seeing that we are naturally growing into the higher complexities and even longer design cycles into the ruggedized industrial and also automotive applications. What we are already seeing is a CAGR of above 60%, we are seeing that also continuing in the trend over the next years, 60%-80%. To wrap up, we are pretty well positioned in both of these separate pillars, in the magnetic position sensing and in the motor drivers and pre-drivers because of our unique technologies. Now we are using that to converge it into the eMotion products to create really that unique bundle, which nobody else is able to deliver, which not only provides a high value to the customer, but also technical advantages. In the end, the user wins, we win, we get more dollars per application in the end.
I hope that explains a bit why we are so passionate about that thing, what we call eMotion. Now I'm happy to take any questions if you have.
Thank you. Could you give us an idea about how big this business can be? Obviously you're not going to replace the entire motion control market. Maybe one day you will, but you talked about some of the growth numbers. Obviously, as investors, we have no reference for how big the business could be. If you could give us a ballpark number of how this business could be by, I don't know, maybe 2020 to 2025, that'd be helpful. Thanks.
On numbers, I defer you to Bernie later on. To give you a sense of the market size, the sheer market size is somewhere in the $2 billion ballpark.
Very helpful. Thank you.
That's today. I can tell you also that both markets have a growth rate of between 6% and 8% per year just from the TAM perspective.
Hi.
Hi.
I have a question about the business model in
This product category. I'm told from folks involved in motion control that typically buyers don't buy just the motor, they buy sort of the whole solution. The problem that I've thought that you might have in this market is that you're either asking your customers to go design motors that the OEMs don't historically do, or you're going to motor companies and saying, "Why don't you go design your motor around my part?" Which they might not want to do either. Is that a dynamic that you see in the business? Is that something that sort of you have to figure out a way to get over to get eMotion to accelerate? What are you doing to overcome it?
Actually, the better thing.
Yeah.
Well, you set me up so well.
Yeah, exactly. I was just hesitant to answer.
Yeah. All I was going to say is,
How did you leave out all the stuff?
If you could hold that question until Michael gives his presentation.
Yep.
I think you're gonna be more than satisfied with the response. Are there any other questions?
He's next.
Right after me.
Right after.
Just sort of a question about how you're looking at eMotion. You talked about the 20+ design wins in position sensing, a similar number in motor control today, You showed a chart where eMotion in terms of revenue, it seems like it is still fairly small-
-as part of the overall, I don't know if that was your revenue forecast or that was the total TAM, It's still fairly small. When you talk about eMotion as a business, are you including the individual position sensors and motor controllers as part of that eMotion business? I've got a follow-up.
Just to give you a reference of it.
Yeah.
Last year, how much was-
$5 million, yeah. It was.
That's position sensors, motor controllers and-
Motor drivers, yeah.
Okay.
Yeah, that's combined, yeah.
Just on those design wins for the motor drivers and the position sensors, I'm sure you think about it today, you've won position sensing in a certain number of designs, but the motor drivers supplied by perhaps an incumbent and vice versa for the position sensor. That ultimately, eMotion could come in and give you both sides of that, but it takes time to win.
Perfectly.
With the integrated solution.
Yeah. Like that. We quite often sell the position sensors separately, the motor sensors, also separately, a lot of customers love to have that coupled thing where we really provide them the first sense of integration, while really with a fully integrated solution, we are bringing them one step further.
Thank you.
You mentioned one thing, something like 20 design wins. We have 20 products being designed into various customers. The base is much larger. I believe there are no more questions, I hand over to Michael.
Well, okay. Well, Stefan stole my thunder. Okay. Well, on the other hand, it's probably easier to communicate the message. You set it up, okay. Let me start with this, okay? It's a different message. We have Today, we talk about battery management. We talk about computing, we also talk about motor driver position sensors and automotive. These are, of course, is growing revenues, that's what you're interested in for the next three to five years. Here, we haven't talked about all the other ones. The development activities, all what you hear today is probably 30%, maybe 40%-50%. All the other areas we are working on, the other ones, AC/DC, the new product lines, the guys here. DC-to-DC is our traditional product. That's 70% our revenues, it's still growing.
We have all the other ones. Other ones, electronic fuse is electronics, actually is electronic circuit breaker. These can be used for cars, for anything that conducts a lot of currents need to shut off. We develop a lot of sensors, too. Today, what you hear is only a fraction of the things we do. We have a lot more product. These are, I would say, conservatively More than 2,000 different parts, probably more than that. I lost count. How do we leverage these things? How do we do that? Also, as I repeatedly said in the earnings call, that in the next two, three years is pretty much set, because it's set by two, three years prior to that. I see we're going to a $1 billion.
The company we set as is, will naturally grow to $1 billion without much of a changes. How do we go back to this page? How do we leverage these to go beyond that? That's my topic today. The $1 billion to $2 billion. I visit all these customers, I made conclusions. That's the way we're going to go. It doesn't cost that much. You always blame on me, I spend too much money. I'm okay. This program, it's a free running. Okay? It's a free running. It doesn't cost too much, okay, and have very little risk. The concept is difficult to tell you. Okay, maybe you walk out of here, say, "What the hell he's talking about?" I'm not a very nice presenter. Maybe in the conversations, we can engage, you can learn more, okay?
Feel free, raise the questions anytime. I think it will help the concepts. What is this? It's E2E business. In the end, I think I find out how you get high margin, at least. Higher margin side, you not only selling the component, you're selling your service. Ease of use. In history of MPS that we do integrate monolithics. That's the name we come from. We integrate a lot more component than our competitors. That's ease of use. Customer design don't have to do a lot of major design effort. I'll go back to this page. All these product, we do monolithic. It's all monolithic. We integrate a lot more than our competitor do. Still require a senior, a very experienced person to implement our product to their application. Or use our MPS FAE, field application engineer, to help them.
It's not an easy task. These 2,000+ product. Now think about the next models. What are we going to do? First things, I'll show you a couple of videos. These are the video during our customers visit, I find out that's a great way to communicate to you. First is this. It's a small warehouse. Underneath it's kind of robots. These robots moves millions of products. I believe this is the future of all warehouse. Okay, the next one is a parcels sort. I visit these customers and when each parcels package at the right place, the correct model speeds it, and on a conveyor belt, is it gets into the right place. Well, it's not in there somehow. Well, the other videos. I think the other one is on the other end, how you get the package sent to the conveyor belt.
How you have the package land on each. This each little is one cart. You land on it in the middle of it, not in between. Sorry, I don't have the videos, and some technical issues. Here is, what's all of these? Actually, next page. This one, a person drop a robot or drop a package, and it shoots right into the middle of this. On top of this red board is a sensor. It check the speed, the size, everything. This one, it moves the conveyor belt. This one, each there's a car. This conveyor belt moves sideways. This company, the owner designed the entire system by themself, and he probably has about 30% or 40%. The market is growing. He's captured more and more in the shipping or warehouses. Mostly is e-commerce company. Guess, all of these, what's inside.
A motor. You have each belt, they have one motor in there. They have to synchronize. Each of these has a motor. There's a motor and need a driver. There's a motor board and an optical sensor. That's a position control. He told me he spent a year and a half to design these things. Why? Because, well, it's not that he can't buy one. You can buy one for a few hundred dollars to control this motor, the brushless motors. If we can use a cheaper one, the motors end up this big. You cannot, brushless motors are the highest energy density motors. Here's the design. If the control, you see the other pages, okay, it's fancy controls. It cost a few hundred dollars. That put him out of the game. He designed his own. He's not expert of this. He knows electronics.
It took him a year and a half to stabilize these things. Guess how much effort he put it in. This thing costs $15. He's consuming every year, somewhere 50 to, now it's growing, is 50,000 units a year. This year, he only shipped almost 100K. Consumes this, 100,000. He's 100%, even in less than six months, he has 100% growth, this one. Does he really want to buy this one? Will said that. This is like a, he doesn't want to buy these things. He doesn't want to do it, even. His difficulty is operate the entire, from the metal to the end. He does the sheet metal too, because he can't buy these things. He has to do implement all the sensors, everything around to make it work.
He couldn't hire enough staff, even. That he's facing the problems. Here is, I'm not talking about only motor. He has to talk about DC to DC. How do you power up these things? This one, he designed his own DC to DC on it. The DC to DC blow up, all that issues, he's complained to me, that kind of things. Okay. Sensor now, just recently since last year, he can't source the component. Optical encoders, he had to buy, they increased the price from $6 to $12. He had to change the encoders, okay, and do all kind of things. He had to do, not only the motor drivers. He had to deal with AC to DC to DC encoders, everything. Which he really doesn't want to do it.
If it's a reasonable cost, he design for him, he will glad to do it. He glad to hand it to you. I didn't make the point very clear. If you buy a couple of hundred dollars of these motor control, it still doesn't serve his purpose. He have a different motor. He have a different needs. Everything has to be custom tweaked. That's his situations. I give you another one. You see the, what's the tear down, what's inside. Actually three motor. One is the lift, left right. What's inside is a battery. You have motor control, DC to DC, and a BMS. Another examples It's less sexy. It's textile machinery. These are the last century things. These are just undergoing, and you don't know, these are undergoing a tremendous change. Full automations, hyperscale. You see this line.
When I see this line, I was amazed. This factory, probably millions of motors in there. Each thread, each tension, all have a motor controlled. You have socks machines, you have knitting machines, you have all kinds of things. It's all become automations, all computerized, all become small quantity. Everything has to be in the computerized, do the very small, smaller quantity, but different kinds. What's inside? Just only one section of it. These are motors. These are control boards. Each thread that you see it inside this one, has a tension management. You manage that each thread's tension. There's two motors in there. Millions of these things. They have solutions, it's working probably in the last 20, 30 years. That they have solutions. If you want to do something new, they have to redesign these, now is undergoing a big change. Another example.
I visit a customer, they do this in this business. Building control automations. They're building actuators. It's a $96 billion business. These are sleeping market. Okay? The electronics portion of it's about 10%. It's really about 10%. You have about $9 billion opportunities. Okay? It's growing. Now, you know the connectivity, everything is undergoing a big change. Everything had to be connected. First example is air flow. All these things, even this, we build it. They have a damp switch, a damp control somewhere. You have to go in, going there. If the rooms gets hot or cold, okay, you have to manually operate these things. These things all can be automated. You have a small thermostat. It's Wi-Fi controlled.
Naturally, you operate the switch, or you operate the valve, or the valve's, okay, call it, is an air damper. Whoops. What is that? No, it does not touch screen. Okay. Okay. These are It doesn't show up here.
It's not the HP.
It's HP. Those kind of things, and the air dampers, that actually is a valve. They call it dampers. Not only in the industrial office buildings, but also in your home. All right, we're back in business. Okay, let me flip a couple. Okay. This is just a typical home. If you have a small thermostat, you naturally want to put a little more. The more valves you can control, you don't have to heat up the entire house or cool down the entire house. This is undergoing a lot of changes. There's a lot of small company doing. Whoops, sorry. Oh, yeah. Okay. I will go right direction. There's a lot of small company doing these kind of thing, particularly is in the box. These are sheet metal guys. Think about sheet metal guys making these dampers. Okay.
They can't hire a team of engineers to do this. They need some kind of solutions. Even you've gone down to different levels. Okay. Well, it's a motor. Motor controls. Also the DC-to-DC, and the supply, the Wi-Fis, all that, the supply, the communications. That's what's inside here. Even the actuator guys, those guys making the actuators, they had to design this electronic control segment. Do they really want to do that? If we give them a reason of not charging them on the legs of solutions. Another example. Well, it's not. Okay, I'm sorry. Okay, this is the water. It's a water. Hot in the cold waters. Flow control valves. Okay. Far, you have a valve a person turn on, turn off. These valves now, you have smart valves, connectivities. What's inside the valve?
Inside this box, you have a motor, you have a motor control board, you have a DC-to-DCs, you sometimes AC-to-DCs. Put it in that control box. The valve guys have to design the electronics. Think about these things. Textile machines, I don't even want to attempt to quantify what's the size of a market. I don't know, okay? I know it's big. This one is something $96 some billion over here in the total, maybe electronic size. Take a very, very single-digit fraction of a few billions. Okay. I'm going on the next ones. Okay? We're talking about billions stuff. These are not a typical your household tools. You have a fastener, you have a electrical lawnmowers. In Europe, you see very often. Here, we still have a mowing people. I think soon, other people were driven out of the country.
They have to think about how you mow your house lawn. All of these, it's also going to experience all these electrifications. Okay? You have to use a bat. Okay, what's inside? Typically, you have a battery, you have a BMS, you have a battery management motor control. All of these. These are really smart tools. They sometimes. Okay, what is this company then outside of the smallest country? Again. Also, besides Switzerland, they're making the tools Hilti. Yeah, Hilti, yeah. The gentleman told me, okay, they're making these tools. They count the usage. They rent those tools. The tools get even smarter. I'm talking about industrial applications. These are all kind of stuff is happening. With another example, massage chair. There's 3 or 4 million massage chairs a year worldwide. Beds, hospital beds.
These beds, you can buy one for $300 or $400. You have a zero weight. You can use a remote control these things. What's inside? Same thing. Exactly same thing. This board and the motors and the optical encoders. Another example. I just put the example there. Just make sure, I'm talking about a motor. Is everything related to motors? I have this one here. Take a guess what is that? It's a thermostat. It's a smart thermostat. If you put a position, this is the display, this thermostat. This is the control board. This is a display. I don't want to name the names who's making these things. Here, look at this. Besides, very, very complicated, these are relayed. This is AC/DC, and a DC-to-DC. Here, we offer something like this.
You almost get rid of the board to be about this size. It make it very, very thick. How many millions of units of a thermostat? Not talking about industrials, just talking about consumer size. Something like a worldwide, like 80, 90 million units. This one. My conclusion here is that there's a lot of underserved market segment. There's underserved market segments. Going back to the first slide and the first tear downs. Those are small-sized customers. They sell 50, 200 thousand units. For IC guys, for IC companies, sell them what? In this case, is there anything related to power? They expected under $2, or $2, you are so expensive. You're making only what, $100,000 to $200,000 a count. For large customers or for even MPS, our salespeople, our FAE, it's not efficient to serve those customers.
There's plenty of them. If the sheet metal guys, if we want do some electronic items, these are actuators, they have to go to a consultant company. The consulting company do it, then he give it to those contract manufacturers to do the manufacturers. There is a lot of these customers. Even for the bigger ones. You're talking about all these well-known companies. These are Honeywells and these, not Schindler and Bucher, Schneiders, and these are large companies. They have a team of people working on these things. They have their own peoples. If we have a project comes in, they have to go start from scratch. What we do is give them a demo board. They're evaluating this. How about this? I give it to them this. Plug-and-play solutions.
You replace this board with a standard model behind it. No optical encoders, no DSP, nothing. Get rid of this board. If you want to, they can even buy a model. Of course, we don't make a model. We don't make all these components either. We're just giving them solutions. These solutions, we're making standard product. These all standard product. Standard mechanical size. These are all standard. We just only a few examples, we're going to have a lot more. Mechanical size are standard size. They can buy one of these and a board along there. There's three wires comes out. Power, ground, communications. They can use the software to adapt whatever they want. We sell them, I will happily sell them for this one, $15. Or plus encoders, $20. This person's doing the converter bills.
He want to do a running change even. This one's. I don't know if I lost everybody or not. This is the new concept of what I'm talking about. We're selling boards now. We're selling a real plug-and-play solutions. These solutions, why other people can never. Lot of module company out there. Lot of module company. MPS, why MPS can do, oh, your inventory will skyrocket. You're going to pile up a different type of modules, these ones. These are costly. We're the first company we can do programmable. We don't have to have a different kind. We don't have the hand-holding to change the component. That's the most difficult part. These are all software. Reconfigurable. We give them a module, we reconfigure for what they need. They can do them. We don't have to do it. We're just giving them a software.
Module is a good examples. Even we want to do it, we're not making the modules. A lot of customers says that, "Give me a module, too. Just like you said. You design everything for it, just give me that." He will hook up with the signal. He'll hook up with the brain of the whatever the systems. He can operate. Whoops, I have a phone coming. Sorry. It's our operation managers. Let me show the last. Here's a call. He's doing. Of course, if we're selling a module, we have to sell him three times as a price as we buy it. The key is here is a total solution. Solution providers. We provide the convenience. In the long run, of course, he not going to buy from us.
He can buy the board, or even if he want to get even lower cost, one of these, he can still buy our chip. This open bill of materials. I don't know if you get that. We're selling board, we're selling components. Different makers, we're selling our own chip. If you're selling this board, we sell you $20. We give you the list of a component, we tell you how you can make this board. These components are MPS qualified. MPS on the MPS list. You can sourcing from our MPS or sourcing from other company. If you do your own, totally do your own, buy MPS chip. It doesn't contradict to existing models. That's my conclusion. We're going to do this. Not only the eMotion, not the sensor, I will do DC-to-DC, our biggest product line.
We do AC/DC. That's why in three, four years ago, or just three years ago, talking about we got to. Because we involved all the component. This instance will probably involve all the cost for a few dollars. We only can sell a chip for less than $1. We created this business. We should capture it, we should provide the convenience for all the users. Recaps, who are our customers? Underserved customers. I believe there's hundreds of thousands of these people around. There's tucking the industrial parks, buy 10,000, 100,000 units a year. Another target is quick time to market. I want some solution. I'm plugging a user, or [somebody schooled]. The solution doesn't work using this. All these picture you see it, we're ready to put on our website.
The more I talk myself into it, I truly believe this is going to happen. Whether we can grow very big, I don't know. The risk is very small. It doesn't cost you a lot. Okay. Maybe a penny a year or something. Maybe a penny. It doesn't cost. Risk is very minimal. The gain is a lot. Okay. The next question is, how are you going to get to all these guys? Okay. I give you a summary. This is our E2, engineer to engineers, okay, through e-commerce. Okay. Here's what we do, is our operations. We design the chip. It's all chip-based. Everything we do, the module is our key, most critical components on the board. It's programmable using BCD5. Somebody already says six. Okay. I don't deny that. Okay. They have an e-commerce.
The e-commerce website. We build a tool, interactive tool. They input what they want, they will see the performance out of this. They order all the programs. Program these modules for them. These are custom modules. We ship. The first target now, I keep saying was 10 days. Now is 10 days. We promise our customers, okay, whoever want to buy these things, custom chip. It doesn't really matter. It's the chips or the modules. 10 days you will have your custom product. By the way, these are all fully automated. Fully automated from the customer selecting the product to customer's input, and we're going there. For cost analysis, not quite automated. Okay. We have a lot of people plugging this one's testing, okay. It will be fully automated. That's my visions. That's our e-commerce model.
Next thing is I let Dean to talk about how we get to these small underserved market segment. Probably Dean going to do a demo. All these are our website. I say we keep fucking it up. Now they're finally is up. Okay. It's the first very crude, the first site. Okay. We are, I think it's up, running, and actually you can program a product. You can buy a product. Okay. That's the first step. All right. I think we should ask the questions later, right?
Yeah.
Do it now? Okay. Dean's doing. Okay.
There we go. Hello? Yep. Great. I'm going the wrong way. There we go. Hi, I'm Dean Gannon, Director of E-commerce, and I'm here to talk to you about the other part of Michael's vision. Just an aside, this was the first time that I've seen Michael's presentation all run together, and it reminded me of why I joined MPS a couple of years ago, because he explained this to me when he was recruiting, and I thought, it really makes sense. You have all these underserved customers. You figure out a product that fits, like he's talking about, and then you figure out a way to reach them. I was brought on board to help reach these customers. There are hundreds of thousands of these underserved customers, and you can't have a sales team go out for these accounts.
They're nice-sized potential accounts, but to support a sales engineer and a salesperson, they just don't work. What are we doing to pull in these customers? Just like any consumer website, we're using all of these different online marketing tactics to allow customers to discover MPS. It's the same thing Amazon does. It's the same thing I did when I was at Amazon. MPS is just starting to use these tactics, but we are seeing some early successes. We've hired an experienced team that's been really successful doing this previously, and we've seen some good things so far. This is an engineer-to-engineer business, and so it's a little different than a consumer business. We have to really understand how our customers are interacting with our website. We've put in place analytics so we can watch their behavior.
We can see, are they interacting in a way that we know is going to lead to a successful outcome? Are they ordering samples? Are they using our engineering tools? Are they downloading data sheets? Are they looking at pages that we want? Now that we've been tracking that, we're looking at how we brought them to the website. The different ways that we showed on the previous slide, like through search engines or ads, we've been following those, and the channels that are working well, we've started to put more emphasis on this. You can see from this graph, prior to 2016, we weren't doing any online marketing. The website served as a source for our existing customers to get information, so it was like a reference website. We started in the middle of 2016 to just do some basic online marketing.
We have a shoestring budget, very limited resources, because our team is really focused on launching the website in Michael's colorful language, how it launched. We did it. Even with our focus on something else, we saw 100% increase in users to the website. We are taking those tactics now. We're able to put more resources and time and effort, and we're pushing out all those tactics to bring in more customers worldwide. The other part of what we're having to deliver, not only do we have to bring in all these underserved customers, but we have to allow them to really self-serve. We can't have an engineer spend time with the textile customer and the conveyor belt customer. We just can't do that.
We've done is we've built self-service tools that allow our customers to do the engineering, our tools provide the solutions. This is what we've built here, and this is what we launched a couple of weeks ago. I'll just give you a quick overview. This is what a customer typically would do. If you think of this underserved customer, they're not a power electronics expert. They're responsible for designing, Michael was showing, they have to design the controller, they have to design all these other interfaces. The power is the last thing on their list that they have to worry about. This is a simulation of a lab bench that a customer would have to use.
If you're thinking about, let's say, that robot example that we looked at a second ago, the customer will put in their power supply, this would be the battery that powers the robot. On the end, they have the load, which would be say the LED lights, the motors, and the little controllers that are in there. This is our product in the middle that takes the battery power and puts it into a usable state for the controllers. Let's see. Okay. What our customers will do is they actually enter what their battery information is, their load information, that's all they need to do. Then they can run some simulations. They can run, for example here, they would run a startup. There's a bunch of simulations that the engineer will run through to make sure things are working correctly.
In this case, he wants to see how quickly a system will start up. Just to step back for everyone to get on the same page about what our customers normally would have to do. To get to this state, normally what a customer like our, say, the robot customer would have to do, they would have to design the physical product, they'd have to get the components, assemble the components physically in their lab, throw it on the bench, make this measurement with the oscilloscope to get this trace. All of that to just get to that first point to see if it actually is working the way they need to might take them two, three weeks just to get to that point.
Then if it comes out that, oh, this is starting up too slowly, it's taking two and a half milliseconds instead of what he needs, he has to go back, he has to contact engineering support. He's a small customer, he's an underserved customer, engineering support is not going to want to pay attention to him. He has to sit there and do some debugging and troubleshooting, which might take him a couple more weeks. He has to reorder. Actually to make the change, he has to physically change the hardware. He has to get parts, he has to get them back in and plug them in, that's just a whole another design cycle, which might take him one week.
What he can do with our system here is he can actually, instead of changing the hardware, right now what I'm doing is I'm changing a software configuration on how quickly the system starts up. I'm not changing anything. I get to log in quick. Luckily, that Remember Me thing works now, Maurice. I can log right back in. I have my design I was working on. You get to see what happens, actually. I turned this on at the beginning of the presentation, so it timed out. I was able to log back in. It has my design saved, so my software design that I've saved for the system. I come back here. I'm going to speed up how quickly this starts up. Again, this is just a software change.
I'm going to run the startup simulation again, see how quickly it starts up. Before, if you remember, it was 2.5 milliseconds. This is about a 1 millisecond startup time. They're able to figure out all quickly like this. Within minutes, they can figure out, does this meet my system's needs? Let's say they do this, they save their design. They're saving. Think about what they're saving here. They're saving a software configuration that we're going to load onto the chip for them. They're able to do these quick iterations and get through a lot of different design cycles, which normally, if it's hardware-based, they're going to spend 3, 4 weeks just getting to this point.
Once they have that saved, at this point, we've made it so the experience from this point on is just like buying anything online. They have their design that they saved. They say, "Okay, I'm going to pick this one." They add to cart and check out, it's exactly the same as any other e-commerce platform. Just as simple as buying a pair of shoes, you can order a customized product and you get it. Our goal is to be able to turn around in days. We're going to be able to do this. Right now, we can do it in 10 days. There's no reason we can't do it in 2 days, that's what we're going to be shooting for. Again, just a quick summary. Our customers design, put in their parameters, we optimize their design, we ship it in days.
As you can imagine, looking at the opportunity for us, it's not just the motion control examples that Michael is talking about. We're putting this capability into all products that make sense. It's across our whole product line portfolio. The pipeline for our product releases, it's pretty full. We'll be rolling products out over the next 12 to 18 months that have this capability. Again, customers that are very underserved, they're not power experts, but they need expert power solutions. It's a $9 billion addressable market. We're going to reach them through these online marketing tactics that I showed the beginning, we're going to deliver them self-service with these online tools that allow the customers to self-service and reach these hundreds of thousands of customers at scale. Thank you. You can help me. I stand here.
Yes. Thank you. Dean, I had a question for you. I'm sure as you got into this industry, you started looking at some of Monolithic Power Systems' peers. TI is kind of the 800-pound gorilla in the space, ti.com sells a heck of a lot of power management chips.
Yep.
I don't know if you have the ability to go in and experience their e-commerce platform.
Yep.
In a very brief explanation, what would you define as the main difference between ti.com and what Monolithic Power Systems is doing with the e-commerce process?
Yep.
Is it programmability, or are there other things that we should be aware of?
I think the key thing that we're offering that's very different from them is the ability to software configure your hardware. That's what we're leading with, because you can't do that anywhere else. There's no other company that's offering that's where we're pushing forward right now. There are other areas where I think we're catching up. Two years ago, we were a reference website, and now we're moving beyond that's really the key thing.
As we add on and we launch all these other product lines with this capability, TI has nice tools that allow engineers to do a lot of the stuff themselves, in the end, to change things, if they have a change in their system. As my engineers have explained to me, it's not uncommon to, for the first time you turn on the board, that it doesn't work exactly the way, because the way you designed it is a little bit different than the way the data sheet being spec'd. It's not uncommon. The way that you have to adjust for that is you have to physically change that. The way we can with our system is our customer can plug that board in that they have in their lab, they can change the software, upload a new configuration, they're done.
They can move on to their next problem. That's really, I think, where we're differentiating and where we'll win.
Maybe say something. The fundamental difference between a Digi-Key and you said it's TI, ours is much bigger. You have another company that allows it from the start to go with me. What they're studying is this board. What they're studying is this board. I don't think we use.
Yeah.
What they're studying is really this board. Okay. It's pretty much fixed. Yes, we can do some surgery, do something to change it for the protocol like this. Basically, it's fixed. This makes it that we contribute programmable. We're making a series of standard mechanical designs. This board, this round thing, okay, these are standard parts. You have a different power, a different size, everything can fit in the back of a model, including heat. If you go to a thermostat, okay, and a different board, this one. All the electronic fuse or electronic circuit breakers, you have to build it. We can reprogram to what they want. The leading of it, we can do that. I can go on and on and on. This, to me, is fundamentally different, and I guess it's a first. We're doing this.
Hi. I'm sorry. It's Chris Caso from Raymond James. I just wanted to dig in a little more on the value proposition for the programmable analog. It sounds like what you're saying is, the fact that you can reprogram it on the fly, if there's something in your design which is not doing right, you have the ability to upload something different. I guess from that, is there a cost disadvantage in doing that? How is that solution, if that's the main benefit, how much of a premium do I pay for that, and what market segments would that be more appropriate for?
The cost actually is part of it. The first thing. Selling this solution to replace this is a value proposition. His time in a year and a half was spent on this to save a few hundred dollars, okay, and per unit. If we use our own solutions, we're not replacing. Our solutions are more appropriate, okay? They're not all MPS products. Sometimes it goes to 1,000 watts of power, okay, or a couple of thousand watts for motor drives. We either have to use a discrete set. They integrate that, it doesn't do the job. Our controller there is MPS. For this, we provide that the value proposition is a turnkey solution for a variety of applications. That's it. His volume goes up. He wants to do everything by himself. That's again, cost. In the end, it's the core of this solution.
We didn't add on more cost. We're doing memory. We're doing many digital memory analog power, all controlled by the MPD5 now. That cost two years from now will be cheaper.
Your view is, if I use the programmable solution and my alternative is go to ti.com, build it with discrete components myself, your view is that your solution will be equivalent to the cost of doing it with discrete solutions?
Oh, absolutely. Yes.
There's no
This is a very good example. It's a very underserved customer. He does everything by himself. I just want to have the value of all volume.
Michael, just sort of a follow-up question. Delivering reconfigurable modules makes a lot of sense because you'll have a set number of programmable SKUs.
That you can deliver once you start getting into board-level products with passives and everything else on the board. That sounds like that gets very custom very quickly. Who's doing the board design as part of this e-commerce delivery? You're going to do the board. Does the software automate the layout? Does the customer say, "I've got a five by seven PCB board, and I've got these components on it?
Okay. Let me cut you off and answer your follow-up first. Okay. In using the DC-to-DC module, we just draw current and they can pick what they want in there. In a standard module, we design our stuff, and we have a lot of applications in here. We design what we call integration. We give it away. Instead of we doing these modules, we doing this module. We have people to do that. To cover the DC-to-DCs in that last angle that I was giving, I think the more way less presentation that we have the slide. We have a few products, these are standard products, to cover the entire range. It's a given, maybe it's not 100% of a market, maybe 70% of the market approximately we cover.
Let's say we have 10 modules in the entire DC-to-DC. We can cover for 50% of the market size.
That's pretty good.
Yeah. This one, we are not making any more. Maybe 10 of these modules, and this is a big one, okay, and this is a small one, and we are making small. We're making 10 more of these ones, 10 different modules. One on the back of the module. If you want a custom design, this is what we do. If we cannot cover, okay, and there is specific needs. That actually is not our business. Used to be? Yes. We can provide a service.
The main focus of the program of line.
Yes.
Really.
That's our focus.
Traditionally, the market for these underserved sort of customers where you can't afford to serve them direct, they're not big enough, volume times ASP is not big enough, that tends to go through distribution. We've talked a lot about Digi-Key and Mouser.
Yeah.
This seems like perhaps it's something that might reduce your need for distribution partners. Am I reading that right? Does it change the relationship with distribution?
Our distribution is really a fulfillment. I would say that 90-some percent of our designing activity, all from the MPS, the EE sales. We are actually uniquely and they do the fulfillment. Okay. Although our volume is at about 20%. Okay. In terms of the main creation, all we.
Okay. You're not using FAEs at Avnet?
No. Very little. Very, very little. I wish they could give us a normal team.
We're an underserved.
They got one hanging fruit.
We're underserved.
Okay.
Thank you.
Okay.
Is Bernie going to present some financials or?
Yeah.
I'm sure we'll all have gross margin questions about some of the models.
The gross margins, okay, I'll show you. I didn't say that. Okay. I forgot to mention. Okay, this one. This one, I think we can sell these size of modules, okay, for replacing this for $15, our margin is enormous. I want to say what, okay, way above MPS. Okay, MPS is today's gross margin.
this is very accretive to your-
Absolutely. Actually there's nothing to lose, maybe we will gain.
Do you have a clicker?
Okay.
Clicker. Oh, thank you. Hello, everybody. I'm Bernie Blegen, the last act. I want to thank you again for joining us. I know that this represents a significant commitment for all of you to pay us a visit, both in terms of your time and resources. I think it's a very meaningful day. Particularly, I think that you're probably pretty used to at this point, interfacing with both Michael and myself, although we did show a little different side of Michael today. That's why it's important to be able to actually meet some of the product managers who are actually driving these lines of business. It was interesting because in a couple instances, it almost came off like we were leading an engineering class here. That's the way we think. Most of these people are PhDs.
They're very senior in their roles, and they're very committed to very passionate, in fact, about the technology. I think that's only something you can fully appreciate when you come on-site here. Again, thank you very much for taking the time to be able to see us. We're running a little bit over, so I'll just kind of give you a preview of what we think we're going to do, and I hope that you can stay for as much of it as you can. I'll finish up with the financials. I'll ask Michael if he has some closing comments. We'll have a short Q&A period at that point. I have some additional things I wanted to cover, just sort of the recap. We have a product demo here.
We're going to have a mixer, and the people who presented as well as some of the other employees that make these things happen will be available to talk to. You can ask them about just about anything you want, except for revenue expectations. You can save that question for me, and I'll just give you a sort of a quizzical look back. Any case, without further ado. One of the things that I do, I've been with the company now almost seven years, and in the role of the controller as well as now the CFO. One of the things that I've found that is really the value creation is our predictability. If we say we're going to hit a number, we hit a number, and in fact, what we try to do is do a little bit better.
Not significantly better, just do a little bit better. As we look at, we're going to update the quarter, we're going to do a little bit of a recap of our history, we're also going to lay out guidance, as far as what we think our financial targets should be. Again, it's going to be pretty boring. It's pretty much what you've already seen before. Any case, back in April, we offered guidance on what the quarter was going to look like. Surprise, we've raised the midpoint for revenue, gross margin, and for operating expenses. The way we did that is we held with the upper end, we raised the lower end, that's what changed the midpoint. It is incrementally positive news here as far as how we're viewing the Q2 guide.
When you look at our history, you have to remember that the formative period of time as far as developing the strategy that's resulted in this performance was really 2010 and 2011. At that period, what we identified is that we needed to go into diverse markets, those markets had to be relevant, large. They had to have long design cycles. They had to have ASPs that did not suffer downward pressure the same way the consumer did, they had to have gross margins that were in excess of our corporate average. If you use this as sort of a scorecard for how we've done, the bars represent our revenue every year with the last year is reflecting in the consensus estimate for Wall Street. The two lines, what you see there, that's the stock price.
What we've done is we have our stock price in the blue, in the green, we have our peer group, the SOX Index, how it's performed over that time. Basically, the results of it is that our revenue last year, we were at 21% growth compared with a market of about 11%, nearly double that. Over the last three years, we've had revenue CAGR of 19%. When you look at the stock price during that period of time, our stock has gone from a little under $20 up to 135-ish, 7x growth over this period of time. That compares with the SOX Index, which has done, mind you, very well at 3x. Our performance is more than double that of the industry. When you look at the diversification, how is that strategy paying off for us so far?
Well, again, the two top areas that we talked about today were automotive and compute and storage. It was very interesting because I was reviewing the presentation materials from three years ago, we actually had relatively concrete plans on how we were going to develop and manage the growth for each of those two. I would probably say that in some of the other categories, it was more visionary. In the concrete area of automotive, obviously 2010, we were at 2% of our total sales related to automotive. That now stands to somewhere in the low mid-teens. Over the last three years, we've experienced a growth rate of 65%. As you'll recall that the CAGR going forward is somewhere annually between 40%-50%, it's a sustainably high rate of growth.
Likewise, in compute and storage, we went from about 10% in 2010, now it's double that at 23. Here again, we have a CAGR of 30%. In industrial, which I got a lot of questions as far as what its long-term prospects and performance were going to be, because we broke it out from automotive and it suffered a little bit in the comparison. Here again, you can see that industrial has grown at 20%, nearly 20% over the last three years and represents about 13% of our business. Many of the initiatives that Michael has talked about in the engineer-to-engineer, in the e-commerce, in the eMotion. I think somebody referred to that as the killer E's, that is really going to probably benefit industrial and other the most. We're really investing in that area.
Consumer, as you can say, is not performing at the same level as our corporate average. What you can't see on this slide is the shift as we've gone from traditional consumer into high-value consumer. High value is like home appliances, and they have characteristics that are much more aligned with what you'd expect to see in industrial. Finally, we have communications. Well, every pound has to have one dog, but don't treat it like that, because in the communications, it's actually one of the areas that we believe is going to be future high growth as we see the convergence of the changeover in technology was going more IC-based and the 5G implementation. That's really 2020, 2021 is when we see it.
Now when we talk about MPS, so much of the conversation seems to be weighted on revenue growth, and I think that is a significant aspect in our value proposition. That's not the whole story. I think that we get as much value from the operating leverage. As you can see here, that compared with 2012, our revenue has increased 120%. That's very significant. Look here, the operating margin has experienced a 70% improvement going from about 17% up to almost 29%. When you take the leverage that we have with the revenue growth, that results in EPS going up almost 3x, 214% increase from $0.93 on a non-GAAP basis to $2.93. How has that affected us as far as free cash flow?
We are generating an awful lot of free cash flow, and we started early, as you can see in this model back in 2008 and 2009, to be able to formulate a capital allocation strategy between stock buybacks and dividends. This is still a work in process because we recognize that other peer companies return a much higher percentage of their free cash flow to shareholders. There's a different value proposition. We are very much a growth company and will continue to be, but we also recognize the need to be able to return cash. Just looking at this slide, which has the green line is our actual cash balance, and the blue line is what we would have been at had we not returned cash to shareholders. You can see that 52% of our cash has in fact been returned to shareholders.
Here I want to look at the growth and diversity. This slide is very important to me because when I look back on the presentation materials that we had three years ago, we had identified most of the major categories on here, computing, automotive, industrial, infrastructure, networking, and consumer. We'd actually identified what were the end markets that we saw that met our criteria being large, long design cycles. It's been interesting because even in the three years since we've presented, we've added so many new sub-markets. For example, we heard about GPUs, which had not been on our radar screen previously. We're looking at artificial intelligence. In automotive, ADAS was very aspirational. Now it's getting to be very real. When we look at industrial, the factory automation and robotics is a new area.
In the infrastructure, again, the 5G opportunity, and on consumer, wireless charging and augmented reality. What does that do for us? It increases the available market that we can go after because we have markets that are growing and markets that are adding new products, and we happen to be in the right place at the right time. When we look at what we've showed you, 2015, again, that's a very healthy picture to be able to go after a target market of $12 billion. Now because of the expansion I've talked about, we actually believe that we're going after a $17 billion opportunity. As you said, in automotive, that's increased to $7 billion, and you saw in the slides, that's actually just the tip of the iceberg. It has the potential to grow significantly in the next couple of years.
Motion control. This is a market that we knew is fundamentally big, and now we're seeing opportunities we didn't even anticipate. The fact is that brushless motors, which is a key driver for this market for us, the adoption from brush motors to brushless is occurring at an accelerated rate. AC/DC, unfortunately, we haven't even done more than just touch on it, yet this is one of our significant growth drivers, particularly in industrial, but also in consumer. The modules we've talked about. Cloud computing is increasing. This might in fact be a modest number. Networking and telecom is something that is going to expand greatly, and we're part of it. You've already heard about our battery management story. When you look at the numbers, you go, "That's terrific." Here again, that's only part of what we're trying to do.
What is important to us above all else is not just maintaining, but accelerating our rate of technological change. For example, we want to have full digital solutions that are augmented by software, so you get what we're referring to as synthetic analog. We want to be able to take, as Michael was describing, I've just pointed out here that we want to take our sensor technology and make it field programmable, with an integrated software. In fact, we want to do that with all of our product lines. We are working on advanced power analog processes. In the current year, what we're doing is we're going from 8-inch to 12-inch in the fab. We're going from 90 nanometer down to 65 and 55 nanometers. We want to push that advantage that we already have.
In the markets where we're strong and we have a commanding Well, not commanding, but we have a good foothold, we want to continue to see the gains in compute and in automotive. We also want to be able to develop with the investments so that we can plan ahead in order to take advantage of future growth in both infrastructure and in industrial. What does this all lead to? We have a target for our growth, and it looks like Oh, that's kind of disappointing. It looks the same as what we showed you 3 years ago. Revenue growing 20% annually, gross margin going into the mid-high 50s, our reinvestment in R&D and SG&A staying at 50%-60%. Now, I should add that these are targets.
When we announced our targets 3 years ago, we proceeded to have revenue growth that was approximately 19%. That was followed the next year by 18%. We got followed that by 17%. Last year, we had 21%. If you take your numbers, that would also yield us a 21% growth. These are targets. We're not so exacting, it's a few points, plus or minus. The reason that I bring up a few points plus or minus is some of the initiatives that we're looking at as far as the process technology, as far as doing the successful launch of the e-commerce initiative, as far as getting sales and marketing resources that are tied into the markets that we've identified that are going to generate revenues 3 years from now.
As far as the design and application engineers we need to develop the next generation of products, we are making a concerted effort in those investments. We're probably going to go a little bit high on the operating expenses. Will we exceed it by a point or two points? Possibly. Might we do that this year and next? Yes. They're very worthwhile investments, and we've demonstrated a history of being able to generate positive returns on those. It's something that I think that you're familiar with Michael. He is very frugal. We are very deliberate in how we make our expenditures, but we see an opportunity so bright right now that we want to pursue that. When you run the numbers, just simple math, and you say, "Okay, if you hit those goals, what does that do?
What does your target become?" We become a billion-dollar company in 2021. Prior to my seeing Michael's presentation, where we're already anticipating getting the $2 billion, I thought this was going to be the big part of my presentation, kind of let the air out of that one. I think what is important to talk about is that with the gross margin improvement, the leverage in operating expenses, that we're going to see our operating margin go up to around 36%, which would represent about a seven and a half% improvement from where it stands last year. Those are meaningful and significant numbers, and we believe it is possible to do that for an extended period of time. Lastly, on this slide, I want to point out that we have added one for capital allocation.
I'm still trying to digest and figure out all of the implications of tax reform. It is still very much a moving target. Until I have a lot of confidence on that, I want to be able to, again, set a target that I know I can meet or exceed, and then have an opportunity to improve on that down the road. For our current target today, we're saying that we're going to return between 30%-40% of our free cash flow in the form of either dividends or buybacks. That compares with what we did last year, which was 34%. That sort of wraps up the quick, the short and abbreviated, but meaningful financial presentation. What I'd like to do is see if, Michael, did you have any additional thoughts before we go to the Q&A?
Yeah, we'll go to Q&A.
Okay.
Well, thanks for coming. We have a lot of people talking about calls. We'll show you something today. We'll show you something today, Rick Fearon.
Yeah.
Yes. Thank you for this very informative analyst day. Bernie, you've talked the last few calls about inventory days moving up because of the business model. Based on what we heard today, any changes? What is the target? Does this mean a particular change in how the inventory days will look like going forward?
Yeah. There's a current question as far as what do we think is going to happen in Q2, and then there's what's the long-term model look like. I think, for those of you who followed us, our historic model has been between 135 days and 155 days. Last quarter, we were at about 177. In the current quarter, when you look at last quarter, really, if you look at the types of new products that we've come out with and the almost strategic nature of the partners or the people that we're selling to, people in automotive, people in compute, and some people in gaming, they're not customers that want to go lines down because their first-time supplier has not delivered on time.
What we did is we made a conscious decision to build ahead in those 3 categories, and that accounted for most of the increase that took us out of our model to the 177. We're very conscious that that is not a sustainable level of inventory for the company. What we're doing is making short-term decisions as far as how to manage that down. Long term, I'd probably say that there will be a bump up in our days of inventory, but no more than five days on the model. If we were at 135-155 days, it would be easy for me to imagine that now is 140-160 days.
Thank you. I also had a follow-up for Michael. Michael, you talked about from $1 billion to $2 billion being a low-risk strategy. Should we assume that you can get to $2 billion organically, or is there any M&A in the picture there? Do you feel like you need to go out and acquire anything or is this going to be an organic $2 billion?
I think it's [Seamers market] . As an example, if we want to acquire some complementary technology, we have cash, okay, we can do it. Now we're expanding our business to outreach to these underserved smaller customers. The key is that absolutely we don't want to acquire revenue. Okay. I think the revenue for us, the MPS is cheaper to grow the revenues. For any complementary technology, any small business size, we're not ruling out. That's our position going forward to $2 billion. Some people do modules, very complementary to MPS. Why not?
Hi. Thanks for taking my question. I have two, actually. First, Bernie, for you. In the past, in addition to the model you've presented, you've talked about gross margin expansion of something on the order of 10-15 basis points a quarter. Is that still the sort of formula that we should think about for gross margin traction over time?
Yes. The model that we've been following, and actually we've demonstrated a very good track record on that over the last 12 quarters, is to have gross margin increasing 10-20 basis points quarter-over-quarter. What that results in is that annually we'll be up between 40 and 50 basis points. When you look at the results of the model that I shared with you, when we get to $1 billion in growth, that's simply doing the math along that.
Yep. Understanding your business is not a spreadsheet. We now have this concept of going from $1 billion to $2 billion. This is out of the modeling horizon for most of us, I think. I wonder what the margin structure looks like as you plan that path to significantly bigger. I think one of the really attractive things about the company is that even going from where you are now to $1 billion, it's still such a small part of the analog market. It's sort of this paved runway almost. The company's doing so great at that, and as we think about $2 billion, that's sort of the dream, what do margins look like in that scenario?
I think it had to be much higher. Although this is at the very beginning.
Right
This is the opportunity that I see it. A bigger motivation is a lot of money to be made. The margin had to be higher. There's a need, there's a lot of money. Customers happier, we're happier. I can't think there will be glad to pay for that. The bottom line is the margin has to be up. What is the financial models? Anything. As long as I see, I believe there's a lot of money to be made, Okay, we're not going after that. The model has to be better than current model.
Thank you.
Yeah.
This can be either for Michael or Bernie. I know you both love OPEX questions so much, so I'll let you share in the fun. Can you hire fast enough to stay within that 50%-60% growth rate as the dollars are getting bigger on the revenue side?
Hiring, it is difficult, particularly in the Bay Area. We can't find enough people, qualified people. You turn around, think about it. Why Bay Area? Why here? They have so many other companies making so much money, and they can hire so many people. Why do we have to compete with them? We don't have to stay here. That's why. Also, in the last 10 years, we grow Is it the last 10 years? Or more than 10 years. The last 13, 14 years, we're growing China from zero to over 1,000 people over there. It's also difficult now, because domestic industrials, and particularly just lately, the semiconductors demands so much. They want to start startups everywhere. It's difficult to hire people there, too. That's why I'm going to Spain. There's a lot of good engineers there. Maybe I shouldn't openly say it.
Forget it.
An unrelated follow-up question. Bernie, you mentioned about the mixing up in the consumer business. How far along is that? I know you've used the consumer segment as a bit of a modulator on the gross margin to keep it consistent.
As all these new products kick in, it seems like it'd be harder and harder to have that slug be big enough to moderate what the rest of the company is doing. I guess the underlying part of both of my questions is, it seems like the margin should have some good tailwinds.
Yeah. I think this is a follow-on to the earlier question that said that when certain of these initiatives really start to contribute, particularly on the e-commerce and the field programmability and serving the underserved market, all of those have margins that are significantly higher than our current run rate. You are right that there is, not a limited pool, but probably a point where we wouldn't be able to continue to accelerate revenue growth through taking on incremental consumer business. At that point, I think we would end up having a margin acceleration that is likely to occur after we get past $1 billion mark.
I think it's the margin-wise. The margin has to go up, because as we're transforming a company, we need a brilliant group of people here. They have to be higher paid. Other people are here and it has to be higher paid, otherwise, why are you working in MPS? As we grow, the money has to be more concentrated. To be smart about it. In the Bay Area, if we necessarily have to hire the best, it has to be above the market rate. It's necessary hire in the Bay Area. You can hire somewhere else. It's much cheaper. You still pay above the market rate, you get the best people. In the end, we need the best people, and we need to pay above the market rate. That's really MPS behind it. We have to do that.
The margins, you said it has a tailwind. Tailwind for airplane is not very good. It's wobbly.
He's a pilot.
I have to turn around to think about it. The tailwind. All these four, not even talking about other one, e-commerce, not talking about. In all the existing product, and the MPS going to those market segments, the price is not determining you're winning or losing it. This strategy we're going up, we thought about eight years ago, and we're now competing these large, I call them mega market segments. Price is very important.
Michael, the question is about what's the prime differentiator for Monolithic Power? What's the secret sauce that allows you guys to do things that the competition can't do? Is it the BCD process, the heart of the competitive advantage? Or is it more than that? Give some discussion on what allows you to do things that TI and the other guys can't do.
I can't comment on the different companies because I've never worked there. I can comment on us. That's a very broad question. I think there's a while we can do this, first things, myself, I'm not a professional manager. I enjoy working on these products. So looking for innovations. I think the innovations start from the CEOs to translate to all people. I think that's very important. Frankly, shareholder return is not the primary, not the issue for me. I'm looking for a better product and a better opportunity, and at the same time, making a boatload of money. That's all I'm looking at. That's why I'm enjoying doing it. I think the company's doing that. To summarize, the differentiation is the culture in MPS. I think is we're fundamentally different. We're not looking for shareholder.
The shareholder return is, of course, important, but not the primary reasons as existing of MPS. It's a byproduct. It's a measurement. Comes out.
Making boatloads of money, I think we're all in favor of that.
I guess two questions for Bernie. You mentioned the 30%-40% capital, or returning 30%-40% of free cash flow to shareholders was partly due to uncertainty around tax, and I noticed you didn't put a tax rate in your longer-term forecast. Today, you're about a 7.5% tax rate. Where do you see that tax rate going, say, in that 2021 model?
Yeah. For the initial implementation of tax reform, there are certain discrete tax items that were grandfathered into the model. Our tax rate will stay low for the 2018, 2019, and 2020, at the 7.5% rate. After that, if we don't take any other steps, we would start to come up to probably another five percentage points after that. What I'm looking to do and what we're studying is what are other opportunities where we can still get tax advantages. That's more of a, you've got the framework, but it's a work in process.
I maybe misstated in some of the public disclosures, but I think in the past, you guys have talked about performance stock units for management and how they've been set at future price levels to incentivize management to obviously drive shareholder returns. I was just wondering if you could give us a sense where the latest round of PSUs were sort of struck as perhaps one metric where the stock price could be headed. If history repeats itself.
What I can do is I can tell you the structure of the program, it's as follows, is that right now we're in the middle of a MPSU program. The way it's set up is that the price targets are established based upon our ending stock price, either at the end of December or the average for December, whichever's higher. What we do is there's a formula which is in our 10-K that graduates what the stock prices are. To date, we have actually met most of the existing price goals. There are five prices for each annual increment, and we've hit three of them so far this year. As far as to your second point, where the stock price is going, those targets again are mechanical in nature, it's based on the closing price of December.
Okay.
It looks like we're about at the end of the questions.
Any other questions?
Okay.
Okay.
Go ahead.
The last things.
Yeah.
Okay. The last thing is we build a car. A lot of people know about it. Today, let me tell you why we build a car. When I start to work on these kind of things, when I work on these things, first, very few customers they have an appetite, they have an interest to work with us on these kind of thing. You don't know what the hell you're talking about. They want solutions. Our board member is a Max on models. He's making models, and we work with them. Even there, they said, "What do you really want?" We're building a learning platform. How do we do this? How do we give it out? For the last year, we've learned a lot. What kind of things we want to do? Make something difficult and potentially very useful.
We're building a car. We're building a car, is a full electrical car, and can be self-driven. That can be used in the suburbans or in the city. Can get on the highway. Can get on the even all-terrain, any fields. All these cannot be used mechanically. Everything had to be electrical controlled. Here's the car. It looks a little bit crouchy. A lot of these are pure, it's straight out of our labs. Come on free to take a look. These ones is all drive by, not by wire, by Wi-Fi. All the control system is we can take it out if we use it. Whoever builds the ADAS, put a model on it will drive itself. Come for you to take a look. Maybe I give you a demo.
Half the function is not working yet. It's not working. Lot of things is, in the end, the software issues. Come over, take a look. Don't let me overrun you.
This obviously concludes the formal part of the presentation. Thank you.