Darin Billerbeck - President and Chief Executive Officer
Joe Bedewi - Corporate Vice President and Chief Financial Officer
Lattice Semiconductor Corporation (LSCC) UBS Global Technology and Services Conference Call November 16, 2011 10:30 AM ET
I welcome you to the Lattice presentation. Presenting today are going to be Darin Billerbeck, who is the CEO, and we have got with us, Joe Bedewi, who is the CFO. Darin?
Okay. Thanks, Doug, and thanks for everybody for joining us here today. Good morning to everybody. So I am going to talk a little bit about Lattice, where we are. Spend a little bit of time talking really about what's different between us and the other two big guys. Because a lot of people are constantly asking us, okay, so you are Lattice, who are you and what makes you believe that you can be successful, especially in an environment where there is really only three players. So we will get into some of those as we talk about, we will talk about the financials and everything.
So let's jump into who we are. But before I do that, a lot of things that I say here today that could be considered as forward-looking, ignore them, right. So for the most part, there is our safe harbor. So as you all are aware Lattice plays in the programmable logic business. That’s a really a huge benefit I think in today's advanced technology, because as technologies become more and more expensive, it’s easier and cheaper actually to programmable logic in a lot of applications. There is always going to be ASIC replacement and things like that.
But we do allow our customers to rapidly implement a lot of their technologies, especially in an environment where cycle time is important. We lead with low power, low cost and what we call affordable innovation. So the innovation that we strive for is not the Intel, Joe Bedewi and myself -- our CFO and myself both came from Intel. We spent 20 odd years each there. We know about hi-tech innovation. The innovation we talk about at Lattice is all for low power. It’s for low cost. And there is a lot of innovative techniques that you use for a low cost in what we call that affordable innovation for application specific areas. And we are going to talk about a couple of those areas today.
We also want to focus our efforts on really growing our FPGA. We do sell a lot of CPLDs and PLDs that are great margin products. We will continue to sell those products through time. But we do want to focus all our effort moving forward on FPGAs. We have a healthy balance sheet with really zero debt. So we have a lot of cash in the bank and a lot of people have talked to us about what you are going to do with that cash. How are you going to use it? Are you going to give it back to the shareholders? We have heard everything. And we are not going to talk a lot about that today but I am sure in the break-out meetings we will.
And then in the last ten years we have shipped about a billion units. And a billion is lot in anything. And I think the first person that ever coined a billion was McDonald, when they did their hamburger and now they are probably in the trillions now. But billions are huge numbers. And we know a little bit of -- something about high cost and low numbers and high ASPs versus what we do which is really the low cost, high volume and the low ASPs. And we will talk more about that also.
We kind of everywhere just like everyone else is. Our primary sales are in Asia and Japan. Which is why you see a lot of our sales force overseas. We did recently open up a new facility in the Philippines. That facility was one which was an old Intel design facility, that we picked up the entire design team all at once. So they do a lot of our math layout and our silicon design. Our main development facilities are headquarters in Oregon and our headquarters -- not headquarters but our silicon development area in San Jose.
We do have core competencies and center of excellences in the United States with low cost geographies. So the center of excellence mirror image for San Jose is China. The mirror image from Oregon is the Philippines. And so for everybody in the room, the Philippines is much lower cost than China. So a lot of people think China is the lowest cost, it is not. The Philippines, Vietnam, are typically lower cost.
So let's talk about our product differentiation and then we will get into growth strategy and then our financial strength. So under the products you could see there is really four segments that people play in. The high end FPGA which is typically X&A, we do not play in. We don’t have million LUT devices, we don’t have 20 gig SERDES and we don’t have thousand dollar ASPs, right. So let's make that clear and we don’t strive for that, that’s not our focus.
What is our focus is what we call the mid-range FPGAs which are really data focused and they are moderate SERDES approaches. So we have a 3 gig SERDES that fits in nicely into 3G systems. We have a 2 gig SERDES, or 1 gig SERDES that fits nicely in the 2.5G. And soon we will have a 6 gig SERDES which will be LTE. So we pride ourselves on having the right performance SERDES for the right interface. And today our volume shipments are in 2G and 3G, not LTE.
People in the room can argue that LTE is future and I agree, there is no real volume shipments today. So we focus where the shipments are. Where low cost and low power and the right performance is in the right target. And I will explain the differentiation between us and X&A as we move on. The low density is really where we say -- we call about between 256 LUTs and 10,000 LUTs. This is where low power, this is where the right features sets, the right innervated and hard IP is really important and this is what we serve with our XO and XO2 product lines. We will talk about that in a little bit. Then we have mixed signal, which most people in the room are probably not aware of. Where we do analog mixed signal devices for what we call Power Manger and Platform Manager. And that’s in networking, primarily communications, and we will talk about that.
So if you look at the product line up versus this, you will see that ECP3, soon to be ECP4 which we launched last week, and XO2, all fit nicely within the data path for the wireless. And we call that the mid-range FPGA. That’s somewhere between 17K LUTs and 250K LUTs. So again not the million LUT devices but fits nicely within the areas that we believe we can be successful on.
On the low density we have got our MachXO and our MachXO2 devices, which are all embedded flash. So they call those instant on devices because they essentially configure themselves without external memory. And then you have the mixed signal which is our power and platform manger series. If we look at the ECP3 mid-range product, again lots of design wins. Probably eight of the ten major OEMs we service. Primarily we do it against our competitors with again lower power and lower total cost. Most of this is the last mile of the wireless and wireline infrastructure. A lot of people call it remote radioheads, pico, (inaudible) you hear everything. That’s really what we target. We have the lowest cost per bandwidth, which is important for us because for customers that are shipping high volume when they look for a cost reduction, they typically come to us.
So the nice thing is we are not going to be there at the very beginning of any interface development but we are there when they want the cost down. And in Asia everything’s about cost down. The lowest power also, again, it reduces the CapEx. A lot of people are focused on the power of the base stations and the monthly energy rates that they are having to burn just to keep these things rolling. And then last it’s feature rich, which is a 3 gig SERDES is DDR3, which some of our competitors do not have. And then cascadable DSP blocks, which are really important in this particular application.
If you look at the mid-range keys for us, again, wireless infrastructure. We spend a lot of our effort focused on that. That’s really the primary driver of the business and the growth of our 3G, FPGA market. We also do some wireline access, even though people will say that’s moved to 10 gig and 20 gig SERDES. There are still a lot of 6 gig and 3 gig SERDES applications for wireline. And then finally we do video processing. The same data path and SERDES that you do for comms is also very usable in the video. Specifically, for video surveillance and also video switching. So we have a lot of applications that we service that way.
If you look at our low power MACHXO2, the picture on the far right, when you talk about small packages, I don’t know if you guys can really even see that from the back of the room. These packages are so tiny that in many cases it’s hard to even probe these products. So the innovation is how do you probe them and then how do you test them, when the packages are 5X5 and you can barely even see them, not even pick them up. So this is a great application for smartphones. It’s a great application or a great product for things like digital still camera. So when you rotate your digital still camera and that image rotates either horizontal or you flip the screen and it rotates screen to screen, that’s all done with an FPGA. And that market is actually pretty large.
And again, you may ask why were FPGAs not used in digital cameras or in smartphones in the past. It’s because the power envelope was way too high. So we have focused heavily on our 65 nanometer low power. It’s actually lower power then you could get on a 28 nanometer device. At least in the density ranges that we have. We integrated the right hard IP, so it adds performance to the products. We don’t have all IP, but we have the right IP for the XO2. And then, again, we priced it correctly. So this price is somewhere between a 1 buck and 10 bucks and sometimes 20 bucks for the application.
In the higher ASP products are typically because they are in more expensive packages. Packages is about 50% of the cost nowadays. So that may be stunning because it used to be silicon was everything. And these particular devices, because they are so small and they are so cheap, the package turns out to be a pretty good driver. If you look at that low density key products, again, we are really focused on the smartphones. Lot of applications in smartphones from some of the LCD screen management to different sensors. A lot of people call it a hub sensor network which enables them to attach a lot of different sensor technologies into the hub based on an FPGA, so they don’t have to service that with I/Os from an application processor.
We also do a lot of things in serves and we do clock generation and some of the bus memory interfaces. And again lot of that is just glue logic, it’s controlled playing glue logic. Things that are pretty cheap and simple but they are easier to be done on an FPGA, and because they are more configurable, they are more reliable once they designed the system. And then we talked about digital still cameras. You see a lot of that in high-def video cameras and the digital still cameras. So lots of features and functions that we do to help them lower cost and then extend their battery life.
If you look at the power management, which many people don’t see or haven’t heard of from Lattice, the value proposition is pretty significant. Basically, the simplistic view of this is we take a lot of discrete functions that are done with capacitors and resistors and diodes and things like that. We integrate those on chip, on a piece of silicon in a package. And then essentially we mate that with an XO2 programmable logic device, which then gives them programmable analog. It’s a very simplistic concept. The value proposition for server networks and comms infrastructure is humongous, because they have the flexibility to add or subtract voltage plan, or add or subtract different sensors and then reprogram it so they have one board design for many applications.
So pretty simplistic. The value proposition is there. It’s about 7% of our revenue today, and it’s our highest margin product. So big focus for us today. And then we will look at the mixed signal. You can see our focus for sales is really the telecom. Why? Because that’s the channel we use for our ECP3 in for products. We have the office automation. Same power savings. Which frustrates me every time I go to a office automation product like a printer and you go ahead and you turn it on. And the thing’s off and it takes five minutes for it to restart. That’s the exact thing power management that you see. It’s not our product that makes it take five minutes but that’s the power management.
And then we do solid state hard drives for things like hot swapping. A lot of the solid state hard drives, when you pull them in and out of some of the applications, we do a lot of those power management techniques. If you look at the growth strategy, you can see we are focused on a pretty big segment of the market, not about total market but a segment that we think grows. We think we can attack it and defend it. And that’s really the mid-range, the low-end and the power management segment.
We expect that we want to grow faster than the market. At least before the macro events of Greece and Italy and everything, the market trends were much greater than they are today. I think the FPGA market was slated to grow about 6.5% per year. I think that’s gone down a little bit in the last couple of months. But we still expect to grow faster than the markets. And our value proposition is really low power, low cost and affordable innovation.
So if you look at the exact market segments, we do a lot in wired comms, about 42% or 45% of our business is wired and wireless comms. And that’s served with many products. We talked about ECP3 being data, we also have control products like XO and XO2. So we spend a lot our effort really in the comms market. It’s a big market for us, but it’s not the only market. And as we move forward, we are going to be spending a lot more time in the industrial video which is an adjacent market for us. Along with the consumer market, both through video consumer and also for just low end price point consumer I/O expansion and glue logic.
So in this low density market, that’s smartphones, that’s laptops, that could be digital cameras, that could be TVs, that could be everything. And then on the consumer it’s really the industrial version of consumer for video processing. So you will see a lot of that that we call consumer. And then on the increased market focus for our mixed signal is just sell more to the comms guys. We have a channel there already, we just need to continue to sell on aggressively to position that product.
Why are we different? If you look at Xilinx and Altera, and I made the comment earlier today in one of the banking meetings that Xilinx and Altera are New York City and we are Portland. And the difference between those two cities is how big the blocks are how many routes you can get through the cities. So if you think of us as Portland and you think of them in New York, we are smaller but we are a lot lower power and the silicon we take up is a lot less. That by definition is simplistic. It’s easy to do. Could they do what we do? Absolutely. It’d just cost them a couple of hundred million dollars to develop and then take them three years to get there.
But when they waterfall their product into our space, that’s like waterfalling New York into Portland. So it doesn’t fit. So the nice thing is we have a really compelling, competitive architecture that we think wins. And we believe it wins and that value proposition resonates with our customers. So if they waterfall there they still have higher power, they still have higher cost. They are not as focused on in innovation to cost, and innovation to the right performance and that’s what we do.
Prime example on XO2. Why we opened some of the consumer markets. When you take a product from, around 30 microwatts and you make it 19 -- or 30 milliwatts and you make it 19 microwatts, that’s pretty significant. So a lot of people are looking at price points, and looking at power. And that’s exactly what we delivered on XO2. And by the way this is 65 nanometer because 28 meter will be twice or three times the expense. So when people talk about cost and when you move to a technology, it’s because you have to have either the right innovation, you have to have enormous dye, which these are not, or you have the right power or the platform that you want to do needs specifics that the technology doesn’t handle. So high-k/metal gate is lower power if you have a giant dye. It’s not lower power if you have a small dye.
The innovations. I am not going to get into this but we do a lot of innovations along the way. Expect to see a lot more in the silicon, in the software and also in the packaging technologies that we are focusing on for cost.
Financial strength. Again this quarter was not so good, at least Q3. And I think the guidance for Q4 like everybody else is going to be down mostly due to the macro events and some of the comms restructuring and may be some inventory that’s going on. But we will make it through and we restructured throughout the year. And that restructuring affords us to really have a lot more variable cost. So in the downturn we just scale down the variable and in upturn we scale up with contractors and the likes of that. So we feel pretty comfortable about the work strength that we have been working on for the last year. So whether the market is up or down, we think we can flex pretty successfully with that.
The outlook, obviously, the decline for a 9%, you already know about that. Gross margins are somewhere between the 50 plus or minus a point. And the operating expenses are pretty much flat, including restructuring. When we look at our business profile you can see, we spend a lot our end markets in consumer and computing with the big ones being communications. Europe and the Americas aren’t significant compared to Asia and Japan. So we spend a lot of our effort doing sales and a lot of support in the Asian countries. And we have no customers that are large than 5% 10%.
So the nice thing is we are not dependent on any one customer. We are probably more dependent on the comms market then we are on a customer. So the market itself is probably the biggest risk that we have. If you look at the quarterly growth for new products, it’s great. It’s been a good trend for us. We expect this to continue especially with ECP4 coming on line and our XO2 product which we released this last year, as it comes into production. So we feel comfortable about our product portfolio, we feel comfortable about our road map. We just have to execute.
And then finally, if you look at the revenue. For Q3 we ended at about $81.7, obviously we made some good cash in the process and we really have no debt. So lots of cash in the bank, no debt. So the company is in solid financial position. So again, we are committed to programmable logic. We really want to focus ourselves on low power, low cost and coin it affordable innovation and we are financially strong.
So that’s it for my presentation, and I think there is some breakout room after this.
Yeah, breakout is in the Carnegie room but maybe we have some time for a couple of questions here, if there are any.
(Inaudible) change the trajectory of the strategy at all as you came on board versus (inaudible).
Yeah, so may be everyone didn’t hear the question completely. But a lot of people have followed since Bruno came on line. And Bruno did a really good job. I mean Bruno came in and really righted the ship to a company that was probably asleep for a decade. In fairness to the company, right, not fighting the people. But the company was started the same size as Xilinx and Altera in 1983 and we sat for 200 million -- or sat at $200 million revenue for ever.
Bruno came along and essentially he righted the ship in a couple of ways. One was cost. He stripped out a lot of the cost. What he didn’t tackle was really set up the R&D structure and streamlining the operations. Right. So he did a great job of getting us back to profitability. And the strategy of him was really, okay, let's do XO2 to extend the XO family, which is very successful. But after that, that was kind of it. And so what we have tried to do is we spent the first 90 of days of my CEOship, if you will, and we dug into every detail of this business. We pooled probably the smartest 40 or 50 people on the company together. Held them there for about ten days until we could crank out at what we call strategic long range plan. And that means going through product models, going through all sorts of models and understand things and saying, okay, if we are going to be in mixed signal, this is who we are going to be. We are not going to do as much clock management, we are going to do power management. Right.
We are going to be in XO. We are going to own everything 10-K and below. How do we do that? What's the process to own everything that’s 10-K and below. And then on ECP3 we actually have to make money on the comms market, versus this constantly chasing Xilinx and Altera. So we set deliberate strategies on who we were, what we wanted to do. And then at that point it was restructuring the company. It took about $6 million to restructure. We closed Pennsylvania, right. We consolidated all the capabilities of Pennsylvania into San Jose and into Oregon. And the two sites, both the Portland or the Hillsborough sites, essentially both have completely different capabilities. Silicon design and development in San Jose, which is where the competency resides for programmable logic. And then we do all the back-end processing and finance and the business in headquarters. And then we mate those two sites.
So we streamlined the organization. So we could flex up and down depending on the environment. We have a clear roadmap. And then we have a process that we go through. Strategic long range planning, product line business plan, annual operating plan, and we are telling you guys everything that we are doing. I think in the past we didn’t share roadmaps with our customers and we didn’t tell everybody what we were doing. And those are two things that are changed. And we sound a lot like Intel, right. Because it kind of is. Joe, is originally from Intel, I am.
Now Intel has some great capabilities and probably more capabilities then anybody and they are a leader for sure. So we kind of do the Intel-like thing at Lattice. So we do just enough to have the right amount of process and discipline, but we don’t overwhelm it with government policies. So we try to simplify because we are small, and we have to be nimble. So I think the big difference, he righted the ship, we are taking the ship somewhere. Any other questions? Or is it so simple.
If there are no additional questions we will....
(Inaudible) empowered management was something that Bruno had talked about as a potential differentiating factor for Lattice. It seems like you guys have taken that to the next level. In terms of the actual product architecture, is there anything that you guys have put in place from a design point of view to facilitate that? Or is that just implicit in the size of the devices that you are making that are intrinsically going to be more power efficient at those sizes and something that’s basically varying or something that was developed for a much larger array? Any process, know-how at your foundry partners that helped you implement that better power characteristics?
Yeah, I mean some of it comes from the fact that when we designed XO2, which was the latest generation of the XO product family, we took that into consideration. So we designed that for power management and we did have an analogue mixed signal power manger group, that then we parlay into what's called platform manager. But I think the trick was actually deciding that we are doing to do that. And taking a team and simplifying the team -- the team for mixed signal resides in Hillsborough. They are standalone team. They are not flexed into any other team, right. They develop all the capabilities and the focus. And we now have a high-end, a low-end and a mid-range platform manager strategy. Before we just had a product and what we did was we say, if you are going to have our product you actually have to have our product line, and it’s needs to be scalable through time. And it needs to align with the other roadmaps we do in the low density. And so we have actually done that. In fact there is some other features on the roadmap that will be coming, that are pretty significant. So it’s really focusing an effort and saying, if you are going to plan mixed signal than actually commit to it. Create a group that’s focused on that. Staff them and then let them go. And that’s what we are doing. So completely different group then the LD and the mid-range group. All quite on the front.
Breakout room is in the Carnegie Hall downstairs. So thank you very much, Darin.
Okay. Thank you.