Warren East – CEO
Tim Score – CFO
Didier Scemama – RBS
Kai Korschelt – Deutsche Bank
Gareth Evans – Investec
Nick Hyslop – RBC Capital Markets
Nick James – Numis Securities
Sandeep Deshpande – JP Morgan
Andrew Gardiner – Barclays Capital
ARM Holdings PLC (ARMH) Q4 2010 Earnings Conference Call February 1, 2010 4:30 AM ET
Hi good morning everybody, and thank you very much coming to our update on business results for 2010 and for Q4 in particular. It’s nice to see everybody this morning. We, Tim and I are on a very tight timeline to finish this at 10:30, but the good news is that Ian is here. And so if there are any unanswered questions at 10:30, then Ian is right there and Jonathan is at the other end of the room. So we should be able to cover questions. So, sorry that we have to dash-off at 10:30. Now hopefully this is going to go forward and we will get started.
So 2010, it was a record quarter in Q4 to finish a record year, and clearly the numbers are going to speak for themselves. However, one of the things that I’m highlighting on the slide here technology chosen by market leaders. The quality of our customers in 2010 has been something of note, and also the applications that people are now targeting ARM technology at. New areas of growth beyond 2015. So lots of the press releases that you read, lots of the licensing activity that we’re doing now is all about fuelling the royalty growth beyond the period of those market share tables that Ian puts in the pack and you all know and we’ll come those tables in a few moments.
Good news is that in 2010, we could see ARM continuing to gain share in the sort of markets that we want to gain share in, in areas of strong structural growth, smartphones, very strong growth in smartphones, mobile computing 2010 was the year of the tablet and we’re going to see a lot more of that. But we’ll also touch on the growth and thinks like microcontrollers and consumer as well. Physical IP strategy is on track. And in fact in all of those three normal areas that we target for driving growth in outside of mobile, increasing the amount of ARM technology in every consumer – every time a consumer goes into a shop and buys an electronic product, and in indeed one of the ways of getting that by extending the IP that we’re outsourcing.
So a good progress in all of those three areas led to a record year for numbers. And we were particularly pleased with Q4 from a bookings point of view driving the already record backlog still further higher. So 75% higher than it was when we stood here in this room 12 months ago. And translating all that revenue growth into profits and cash, we’ll see about that but it’s not just translating into profit growth, it’s a combination with a fairly sort of balanced approach, we had a combination of more profit but also investment in the future, investments in our product roadmap.
So we grew nearly 180 people during the year, particularly in our processor roadmap and our graphics technology at the same time is increasing the profit margin from 30% to 40% and growing the earnings per share by 70%. So 2011, how has it started? Well, it has been a very busy start to 2011, but of course that is really a reflection of hard work done in 2010. And we started at the Consumer Electronics Show at the beginning of January. There were two announcements and we thought we’d put them on the slide here, everybody has read about it, however in recent weeks but just to set the context, I mean the first announcements, Microsoft choosing to support the ARM architecture with the Windows operating system.
This was a great presentation by Steve Ballmer and his team. I think the really interesting point here was the support for several ARM partners, and also real applications, not just the operating system. So the fact they demonstrated real applications running on ARM, clearly highlighting that of course that is where there is a lot of work still to do. Really Microsoft came to a little bit closer to ARM in 2010, but that’s a result of many years of working with Microsoft. So that was a great presentation. Another great presentation at CES was from NVIDIA. They’ve had fantastic success with their ARM powered Tegra chipset in terms of design wins. And what they were talking about in CES was licensing Cortex-A15, that’s the new Cortex-A processor that we launched in September.
They’re extending their Tegra family around Cortex-A15, but also they’ve become an architecture licensee for new products smoothing out in the computing spectrum. So taking ARM right across the computing spectrum in future, that’s their plans. I think these announcements in enough themselves, very interesting but actually it’s indicative of progress generally with new markets fuelling growth for the years to come. This is obviously all about computing applications, so where is ARM in computing today. Well there is a blurring between computers and smartphones and that’s certainly helping the ARM’s very strong presence in smartphones is helping us get into computing. But at CES there were some great brand names and the whole show was indeed a showcase for lots of ARM technology and showed ARM very strongly positioned as the consumerization of the internet that we talk about takes place over the coming years.
And I was particularly pleased to see the Hand-cranked PC in the corner there, because having taking a lot of stick over the years for our terminology about ARM powered products, this is a genuine ARM powered product of course. So this is all about ARM in computing. And we just thought it would be worth taking a moment or two to talk about the Cortex-A family of processors, and why Cortex-A in computing. And so we’ve prepared this chart and the presentation version and the reference version in your packs, I think probably slightly different one, they’ve eventually come out.
But what we’re trying to indicate is if there is a market size and a market potential in each of these areas, place where we are at the moment. And in some of the areas we have a very strong market share today, but we can see the market developing over the years. And in another areas even as the market develops over the years, we should see even stronger growth from ARM if we’re able to grow our market share. And you can see there has been of talk about some new areas of computing, where we have zero market share today, but where we hope to gain market share in future. And the Cortex-A processors, the box on the bottom of this slide indicates why the Cortex-A processor in this particular application space.
So in smartphones then clearly it’s about performance and it’s about power consumption. And as we move across to mobile computers than people want both thin, light weight products with a long battery life. ARM technology is great for that saving space, saving power and at the same time bringing an ecosystem from the smartphone space. The ecosystem around PCs and smartphones are becoming much more of a single ecosystem. And in each one of these areas as well, the final surprise of the application processor is going to be a bit different, but here in mobile computers, we see mobile computers as really a high-end smartphone. So the price of the applications processor probably being above $20 today may be going down a little bit in future.
And in DTVs it’s a similar story. It’s about high performance required as the TVs becoming more connected to the internet and as the TV becomes a hub in the home. ARM’s market share is around the 30%. It has grown in 2010. We grew about 60% in this space in 2010 versus the market growth of 30%. So that wound our market share up a little during the year, but again you see the Cortex-A8, Cortex-A9 processors today probably A-15s in future. And as we move across into service, you can see the Cortex-A processors scaling to deliver higher levels of performance. Its more use of the internet, drives more server usage. ARM is very attractive in terms of power savings and potentially cost savings going forward.
The ARM Cortex-A processor support multiprocessing and that delivers the high levels of performance required for server applications. And in this space we have zero market share today, and we have an ecosystem that we need to develop. So this is very much a potential opportunity and its market growth in the out years. And similarly, I would say in desktops where there is a very established capable and competent incumbent architecture. And so ARM presence in that space is probably some way out but certainly as our performance – as the performance of ARM processors increases over the years, and as the ecosystem develops with companies like Microsoft supporting the ARM architecture, then that’s clearly a market opportunity for the future. So a little bit of aggression [ph] there on computing.
Back to the business, and let's look at the business in a bit more detail. So processor licensing has a slide which is a format that you should recognize. And 2010 was a great year for licensing in terms of growing our existing base of licenses out there to generate future royalties. So that’s now grown to over seven – or about 750, over 740 licenses with active licenses out there with potential to generate royalty. And we added 90 odd licenses to that during 2010. Some of the quality of the licensing there was very good as well with the customers involved as I mentioned. And the quarter that we’re just reporting was an excellent quarter for licensing, with the great increase in the backlog at the same time as record level of license revenue.
And really the licensing obviously is broken down in the earnings release properly but across the range of our product portfolio. And the bottom bullet there, eight licenses for Mali graphics as well. We shouldn’t forget the Mali product line that graphics processors, also very pleasing end to the year to be able to report eight licenses signed. And that’s important in establishing an incremental royalty stream for ARM going forward. The sorts of applications that people are licensing ARM for in the pie chart is again an update on the pie chart that we show most quarters, but pleasing to see, people are still pressing ahead with mobile but also a lot of licenses for ARM in non-mobile, in the embedded space, in the home space and in enterprise for wireless networking.
And you saw a couple of announcements during 2010 where people are licensing ARM for server applications as well. So what’s that done to the licensing base? We’ve updated the chart here. Its 743 licenses and the breakdown is shown there, the numbers in brackets, so the numbers that we added in 2010. And so fairly consistent licensing across the portfolio, but particularly pleasing to see the number of Cortex-A licenses that we sold. The Cortex-M, the microcontroller, processor that’s been a great success over the last few years and microcontroller royalty revenue will come onto in a few moments is starting to pick up, but 2010 was a great year for licensing the Cortex-M series to generate even more royalty further out in the future.
Now for the last several years, I’ve stood here with a chart a bit like this and the arrows on the right hand side of the chart, I’ve been talking about the new processors that we’re launching in the year in question. And our marketing folks got a little bit of ahead of me this year. They decided they were sick of Warren standing up at the earnings presentations and announcing products. And so we actually launched the Cortex-R5 and R7 yesterday, or was it the day before? Yes, yesterday. So they got in ahead of me, otherwise we would have some code names there as well in the usual fashion.
But launched a couple of real-time processors already for this year. The Kingfisher is aimed at lower end smartphones and feature phones and cost-sensitive digital TV applications. It’s an A class processor, a small A class processor. Mali, I’ve got a slide a bit later on to talk about Mali. And at the microcontroller end, we’ve got a system design product to work with a Cortex-M processors that we plan to launch later this year as well. So the product portfolio is rich at the moment, and 2011 will be no exception to recent years in terms of bringing new products to the market. And that enables us to continue to grow our licensing to generate royalties in future.
And here is a little bit of a new slide, the chart that we’ve created on the bottom left of the slide is in the back of your packs in proper size. But what this is showing is how adding to the license base in numerical terms is necessary but not sufficient. What’s important is not just the number, but who has the license and what they’re actually doing with it? And so the dots in each market segment represent important semiconductor companies for that segment. And behind each dot there are of course important must have OEM customers that we need in order to gain market share in each of those sectors. And so that’s a chart you’ll see developing over the years as we try to show you how we’re – what we’re doing behind the scenes to grow our market share.
That leads on of course to volumes of real products. And if we make it work, then the 4 billion to 6 billion growth that we’ve seen in 2010 will continue as we go into 2011 and beyond. And here are some examples of the products that came to market in 2010 based on ARM. And you can see a huge range there of products. Again a little illusion looking at forward to the appendix, we’ve updated the charts with the different market sectors, the volume of devices, the volume of course per device in the total available market for ARM. I am not going to go through those charts and I’ll ensure to spend time going through them over the coming months, but we’ve updated them for this year.
And we’ve actually changed the chart in a couple of ways. We’ve added some new lines showing new market growth potential. We’ve done a little bit of reclassification between feature phones and smartphones. And we’ve noted that because of adding some new lines that the total available market has actually grown. So we’ve slightly restated our market share for 2009, because we’ve said actually the market is bigger than we stated a year ago. So with the same volume of course our market share is a little bit smaller.
The good news is of course, so this shows that we’re targeting an even bigger market going forward. And you can see how our share has grown in 2010 when you look at those slides in detail and you can see the new enlarged definition of the total market available for ARM. And of course if we get all those applications and design wins correct, then that translates into market share. And you can see on the right hand side of this chart, how our market share is growing in 2010. And in each of the key market segments we target you can see ARM’s growth outstripping the growth of the underlying market. And in revenue terms, on the left hand side of this slide, you can see ARM’s royalty revenues continuing to trend ahead of the underlying relevant semiconductor market.
And that’s been consistent as you can see over several years, consistent out performance continued in 2010. Now switching to Physical IP. Physical IP, it was actually a good year for our Physical IP business and it was a good fourth quarter as well. We passed another milestone in the fourth quarter with our first Physical IP subscription license. We see more progress in terms of developing platform licenses to generate Physical IP royalty going out in future. And in Q4 itself, was very strong underlying royalty for our Physical IP business. It was 12% growth in our Physical IP royalty year-on-year. By the way 30% of that Physical IP royalty is now coming from 65-nanometers and newer. And that’s a trend which is set to continue.
When we look at those royalty numbers and we analyze it in a bit detail we can identify approximately 1.9 billion chips now in 2010 having ARM Physical IP. So we’re starting to develop a metric there which is starting to get comparable with microprocessors. And you can see the incremental royalty that we’re adding to a chip that goes out with ARM Physical IP as well as microprocessor royalty. As far as the technology roadmap is concerned in our Physical IP, that’s also on track. We announced with IBM just the other day early this year, a collaboration right down to 14-nanometers back at the 20-nanometer node, 22 nanometer node.
We’ve actually take our first 20-nanometer structures. We’ve done three 20-nanometer tape outs now and so the technology program is well on track. If I go onto how we productize some of that technology then 2010 certainly been a great year for using our Physical IP to build better microprocessor. The products that we’ve refer to have Processors Optimization Packages where we develop Physical IP to enable our partners to make better implementations of our microprocessors, delivering in some instances a 30% to 40% uplift in performance and/or power consumption.
And it’s all about delivering better performance, better power consumption and better time to market for our semiconductor companies, making them more and more competitive. So fourth quarter, last year we did four processor optimization packages. We sold a total of 10 in 2010. And of course that will translate into incremental royalty per chip as we go forward. And typically, a Physical IP uplift is about a 50% of an ARM microprocessor core royalty. So that’s quite encouraging as we look forward for incremental royalty from Physical IP. Another way of generating incremental royalty is with graphics. This is further incremental per chip. It was a good year for a Mali graphics. Samsung came clean with the usage of Mali 400. And Mali 400, they’re set telling the world is giving them about a five-fold performance improvement over their previous product. Meanwhile ST talked about licensing the next generation Mali graphics. And they’re talking about great success with their existing Mali graphics processor having 10 design wins so far for their Mali 400 version.
But the next generation giving them another five-fold increment in performance. So very much performance is where we’re driving the graphics roadmap. But at the same time being ARM and being Mali technology, the emphasis is on performance with power efficiency. And Mali is designed into many mobile products now. We’ve started licensing this after the acquisition at the end of 2006. And Mali is expected to be shipping in meaningful volumes in 2011.
All of those things give us incremental royalty per chip. There are really three ways of generating incremental royalty per chip that we’ve identified here. The Cortex-A processors themselves typically come onto slightly higher percentage royalty. The adding multiple processors per chip, we talked about at the Analyst Day back in May and have that 10s to add increment to our royalty. And then additional IP in the form of either Physical IP or of Mali or in some instances indeed of both. We have examples of chips that show the very things that I’ve been talking about for the last minute or two, where both Mali and physical IP is present alongside an ARM microprocessor.
And so the average percentage that we receive per chip is starting to creep upwards. So this is very, very early days for that phenomenon at the moment. So where does that take us? Tim will summarize for these results in a moment, but I’m just going to take a slightly longer term view. We’ve just done our 20th anniversary and so looking forward to our 30th anniversary in 2020. It’s really a story of growing beyond the sweet spot where we are. This is a spectrum of computing across the bottom of the chart there. And we’ve already started serious volumes in Baseband.
And over the last several years, we’ve grown out from Baseband to a very strong position with a developing ecosystem in the middle of this computing spectrum. And what we’re doing in terms of developing our new product portfolio is developing the sort of products both at the microcontroller, the high volume, low performance end of the spectrum pushing out microcontroller sensors and so on. That’s all about generating huge incremental volumes, and also at the high-end of our roadmap developing the processing performance. So that’s our development, our licensing activity which will lead to royalty growth as we look beyond the charts in the appendix.
So the ARM really does become the computing architecture across the full spectrum. And we think that we should see at least 100 billion ARM microprocessors by 2020. And with that I’ll hand over to Tim for some further information on Q4.
Thanks Warren. Good morning everyone. Good to see such a big crowd on a busy day for (inaudible) technology company reporting. Warren has talked on quite a lot of financial information already. And there is considerable detail in your packs and in the release itself. So as usual, you’ll probably be pleased to know that I’m not going to fully buster our way through to 10:30 and leave all the questions to Ian. I’m just going to pull out some of the salient points that I think might help as we try and interpret these results and look at our forecasts.
So, on the royalty front, as has been strong, strong contribution to growth from mobile and non-mobile. 1.8 billion units in the fourth quarter. And I remember at the Analyst Day of 2003 when we were heading for 700 million to 800 million units in the year when Bruce Beckloff [ph] and his predecessor talked about 2 billion units down the road in a year. And the Analyst community was in shock and said Tim is that a financial forecast, you must be kidding. And now we’re pretty close to doing that in a quarter. So that’s quite interesting. And it’s a 60/40 contribution. We’ve been saying for years that non-mobile is on a trajectory to grow faster than mobile.
But the whole mobile surge and smartphones has kept mobile ahead but keep your eye on non-mobile, that’s happening. I think in terms of looking forward, we’ve seen some fairly – some mixed reporting in the fourth quarter, which of course makes up our Q1. A number of our bigger shippers are reporting sequential growth. Some of them are reporting sequential decline, I mean our reading of the overall industry looks to be around about flat to may be slightly down overall for the industry. So obviously we need to sort of think about that as we look at, take this 82 billion royalty number we just reported, look at in the context of Q1.
In terms of licensing, I mean most of you have been following ARM for a number of years, will recognize this as an exceptional quarter in pretty much every dimension. First time we’ve reported over $50 million of revenue. First time, we’ve signed more than 30 licenses. And at the same time as high revenue, we’ve also grown the backlog sequentially as it says there by about a third. And interestingly the contribution to license revenue in the quarter which tends to be in the spectrum to 40% to 60% from backlog existing at the beginning of the quarter, it’s quite high up with 60%.
And so the way I would encourage you to think about this is in the last year or so, as we came to the downturn and we were reporting license revenues of around $30 million, I was saying that more typically you should expect us to be in the $35 million to $40 million revenue range in a sort of normal quarter. And of course if we had a high number of units or varying contribution backlog, it could be higher or lower than that. Now I would encourage you to think about $40 million to $45 million as if you like the base case for licensing as opposed to the $35 million to $40 million. That doesn’t mean to say we can't have license revenue numbers below or above that but I would encourage you to think about that as the base case as we look forward. So that’s licensing.
In terms of costs, Q4 costs little bit higher than consensus and higher than run rate largely because the very strong out performance in revenue and the exceptionally strong performance in bookings has driven some higher, what’s referred to there as performance related remuneration which basically means the corporate bonus and sales commissions. We’re at a higher level than normal run rate. So normalized OpEx was £61 million in the quarter, and you can see from there that we are guiding a return into the high 50s for Q1.
Now these revenue uplift is clearly yielding margin expansion. A 10% margin improvement over the year. You will notice I’m sure that the earnings in Q4 benefited from affectively a one-off tax adjustment related to the recognition of a deferred tax asset in the US. Thinking forward, I would retain the guidance that we’ve given you for a number of years, recent years now of about 27% as you look into 2011 and beyond. If you strip out the unusually low tax rate in Q4, you get normalized earnings about $2.56 as opposed to the $2.90 we reported.
Looking at the full-year, I’ll say we’re touched [ph] on this. It’s a very strong year for the industry, 2010 a bounce year from 2009. The good news is ARM continued to outperform. We performed better than the industry through the downturn. And we performed better than the industry in the bounce. And I think when we look long-term, that second point is very important, over the last three years, royalties are gradually edging out as a percentage of total revenue. And this is of course what is really driving the operating leverage and the margin expansion. And it’s interesting to see that the operating margin is actually 7% up over the period, pre-going to the downturn and now coming out.
And I think we’d all expect it to be seriously higher than 2009, which was a down year for the industry in terms of revenue, but it’s pretty encouraging to us that its up 7% compared to 2008 which for ARM because of the royalty shift one quarter, it was a pretty clean year in profitability terms. So that’s pretty encouraging. I think we touched on the investment in people, led out by almost 180 people in the year, over 90% of them in R&D. So the finance and supporting functions to continue to work very hard to keep this business in good shape as the R&D engineers get invested in. So it’s all good news, we’re controlling our overheads and back office pretty well.
We expect to have more investment in people in 2011 but at a slower rate than we’ve seen in 2010. Very strong cash flow. £180 million net generated in the year, that is more than double but just under £90 million we reported last year. And you’d have seen consistent with the interim, we’re increasing the dividend by 20% which you can interpret as the Board’s confidence in the medium and long-term earnings prospects for this business. You will recall that we grew the dividend through the downturn at 10% per annum and now we’ve increased that to 20% this year.
So in summary then on the outlook, I think most people believe that the semiconductor industry is going to grow at more typical growth rates. The ones that we see and have in our mind at the moment is somewhere between sort of 5% and 8%. And I think in that context, we think we’re very well positioned to outperform that as we have done for the last few years. I mean the royalty momentum I think is there clear to see. The high order backlog whilst a significant proportion of it relates to revenues going to be recognized in future years beyond 2011. It’s also clearly helpful for 2011. And we go into each quarter next year with a higher opening backlog position than we would have had going into the equivalent quarter in 2010. So that is encouraging.
And reflecting on that, we expect the dollar revenues for the full-year to be at least in line with current expectations which are in the mid-690s, currently 695, 696. So we expect to be at least that number as we’re look into 2011. But it’s very early in the year. We are in the first day of February. As I often mention over the last four or five quarters, there is still a lot of macroeconomic uncertainty in the world. And whilst ARM is very broadly exposed, we’re not immune to cycles and if consumers slowdown. It does have an impact.
So it’s very early on in the year but we think we’re very well placed to outperform. So very quickly before we go to Q&A, in summary, what we’ve seen in 2010, I think is probably an acceleration of influential market leaders choosing our technology. Many of the big players now adopting our technology in a very broadly across their groups and their divisions and across multiple product lines. That's going to continue to drive the royalty momentum that we’ve seen. And as Warren said, looking further out, we have some very exciting opportunities in you know – if you like untraditional markets for ARM where our technology is now stretching into those zones. That brings exciting challenges. It obviously brings increased competition but it ultimately adds to our overall royalty growth opportunity.
So with that I’ll open the floor to questions.
Didier Scemama – RBS
Good morning. Didier Scemama, RBS Group. A couple of questions if I may, first of all, you started to talk about computing since the Microsoft endorsement. I was just wondering, if you could may be quantify the sort of blue sky royalty rates you would get from a – let's call it a notebook processor that would have ARM CPUs, ARM GPUs, ARM video codes as well as the Physical IP. And if you could maybe outline some of the potential ASPs we’re looking for those end products and then I have a quick follow-up.
Well, really we – it’s a bit early to start for us to start speculating too much on exactly where ARM is going to get to in the computing space. I mean we’re saying there is a potential there and what we are highlighting on the chart was actually the potential for the Cortex-A processor. Already in lot of the other functions, you see ARM starting to be presence. So we’re kind of already there. The other thing, there is a huge uncertainty of, you know exactly how big this marketplace is, what’s a computer, what’s a mobile computer and so on. So a lot of caveat. But broadly we’re talking about the application processor socket, the main brain of the unit. And typically today, applications processes in high-end smartphones are of anywhere between $20 and $40. And the royalty for ARM is we’ve been through the sort to standard ARM core, the fact that Cortex-A processors come under slightly higher royalty percentage.
The fact that if you add a graphics processors it’s similar to an ARM processor. The fact that if you add a Physical IP, it’s about 50% of an ARM processor. So the big question is okay, so what sort of price as in future of that chip. Does it stay in the $20 to $40 range or does it sort of increase or decrease. And for that I’d sort of point at history and say for the last several years, our experience in smartphones is that new chips get introduced. The get introduced at a relatively high price. And over a period of two to three years, they decline in price, even as the functionality of these things increases, the average price after that decline stays more or less flat.
And so I think the two of the advantages for ARM in these sorts of applications are performance and cost, billing materials costs for the product. I can't really see the price of those things changing much. And it is a very high-end smartphone that we’re talking about from an architectural point of view. So that’s a quick run around how you make up the answer without actually giving you the answer in pounds, shillings, and pence.
Didier Scemama – RBS
Okay, I appreciate that. In terms of physical entity, one of your American competitor basically is banging the drives at, its main advantage versus the ARM architecture is his processor technology 22-nanometer high-k/metal gate? And basically the companies that are manufacturing ARM based processors are clearly creeping up their CapEx to meet demand from their customers but potentially to catch up. What I was wondering is if you could give us a sense of where you think the ARM ecosystems and partners will be at the juncture of 2013 in terms of processor technology and whether the Physical IP division is going to see a step change effectively at those nodes as you enter the computing market with very high-end processors?
I think we will see the ARM Physical IP making headway in that area. But I don’t think it will relate to or I don’t think it will manifest itself as a step change because, we are already seeing change in our Physical IP business. We’re already seeing our growth in underlying royalties. We’ve seen the licensing activity and we can see that activity coming into royalties as we go. So I’m not sure that is going to be a step change. On the slide we talked about 20-nanometer and 22-nanometer, that particular node 20 or 22 translating into royalty in 2012 and beyond. So that’s going to start in 2012.
The 14-nanometer collaboration that we announced last week with IBM, that’s probably going to be starting to generate royalty in about 2015. I know that the American competitors talk about being ahead of the ARM world. The reality is that the likes of TSMC and IBM are making massive investments as well in the technology. And let's face that all these things have built with the same equipment. So the notion of some companies being a very, very long way ahead in terms of having massively different technology is just not really very true. And I point you at TSMC and I point you at IBM to get much better answers to that question.
Didier Scemama – RBS
I think next one, yes.
Kai Korschelt – Deutsche Bank
Thank you. It’s Kai Korschelt at Deutsche Bank. I had a couple of questions, and in terms of true analyst fashion I wanted to test the assumptions on the mobile computer estimate for the market which is through up there which I think you said would grow to 750 million devices in a couple of years, I mean that would be a roughly doubling of the PC market from here. Can I just ask what – whether you assume any sort of cannibalization from smartphones because that was highlighted as a separate market or just kind of what the key driver of that pretty rapid increase is? And my second question would be on just on the ASPs for application processors, it looks like – my understanding is high-end application processors are more in the $15 to $20 ASP range at the moment but they do seem to be higher for tablets for example versus smartphones, although there is no architectural functional sort of reason for that. Do you think that is sustainable or do you think as that category grows in volumes, prices will come down and sort of even out with the smartphone royalty? Thank you.
Yes. We tend to answer the first question about the market for mobile computers. We’re certainly here at the end of 2010. And we’re talking about a potential market size for 2015 to roughly three fold growth from 230 million today to 750 million. The 230 million today we’ve put in the footnote to the slide, it includes laptops, tablets, e-books and all those sorts of things and that’s fairly clarified number. Now we’re getting our market numbers from a mixture of industry analysts but particularly people like Gartner. And if you look at what people like Gartner are saying for 2015 that’s what they’re saying. Frankly, I have no better intelligence than you about whether what they’re saying for the market in 2015 is right or not, or what it’s actually going to be.
But we’re simply highlighting that as an opportunity because that’s what market pundits who were paid to make these things come up with. In terms of the ASP, then as I said chips tend to be introduced at a sort of high level and they tend to decline with time to a lower level. And we would say sort of high-end apps processor, that gives probably average of about $20. The sort of apps processor that you find and by the way we reclassified some of the smartphones and feature phones. So ARM’s classification has been a bit different from some of the industry analysts over the recent years.
We’ve aligned with industry analysts. And so you’ll find high-end feature phones which we had previously referred to as a smartphone. They do include an apps processor but they don’t include an operating system which permits applications to be downloaded. And that seems to be the crucial difference. In those sorts of products probably you’re talking about a cheaper apps processor, at the lower end of the range maybe it’s sort of $15 and below. At the high-end, state of the art Tegra 2 based apps processors. They’re at the high-end of the range. And so on average, we say it’s about $20. It is going to fluctuate, but generally when we look at markets like that the process on average stay pretty solid as individual chip prices move within that. I think Gareth and Nick.
Gareth Evans – Investec
Thanks. When you typically talk about quite long lead times between licensing and then the eventual royalty, I just wondered it feels like I was seeing an acceleration in development cycles both in the end market products and also on the chipset side, I am thinking about (inaudible) chip and some of the new entrants coming through. Is it fair to say that we’re seeing an acceleration in development maybe a shortening of the lead time between a license and royalty? And then sort of almost related the spread over SII [ph] data. So you’re kind of spread over the semiconductor industry, obviously continue to outpace it. I just wondered whether you feel that that again can accelerate from here. You had quite a wide spread over industry growth in Q4. Whether you feel that given your access to higher growth markets that might controls etcetera, you’ll actually see a widening out of that difference? Thank you.
On acceleration on the time scales, it’s true. That we have seen a couple of examples in the last 12 months. People have been able to go from licensing through to sort of launching products, possibly a little bit quicker than we had seen. I don’t think those examples are particularly indicative of a trend that people who are using things like our processor optimization packages that I highlighted. And I highlighted – people buy those for reasons of performance, reasons of power, and time to market. And undoubtedly utilization of a more complete IP offering does enable somebody to get a product out sooner rather than later, couple of examples. I am not sure that really the times scales are sort of compressing. So we’ll wait and see what happens in 2011. If we get many more examples in 2011, then obviously more of those examples will indeed be a trend.
And I do apologize, Gareth, I’ve forgotten your second question.
Gareth Evans – Investec
I had a question with spread. How does ARM’s growth compared with the overall industry?
I think you’ll be quite tempting to note that cross into the net growth. I think we need to be quite capable, I mean in this situation you often hear us about talking about growing in broadly 2X and indeed, I think last year. If you adjust the industry, take our analog and memory and compare it with the stuff that directly relates to our business, we did indeed grow it about 2X. We were 29%, the industry was 14%. (inaudible) with ARM’s potential rapid penetration into microcontrollers. We could overtime see a little bit of movement on that but we would remain sort of thinking it in about the 2X really.
Nick Hyslop – RBC Capital Markets
Nick Hyslop from RBC. Can I go actually to the other side of the equation on ASPs where you have historically said 1% of the chip price is the right number. You guys started talking about 1.2% for some latest prices and you’ve even mentioned 1.5%. Could you just update us on your thinking on what you’re charging for the latest licenses for the new contracts that you’re taking out? And whether that’s a trend that continues particularly as we start to look at these higher value markets that you’re addressing? What level of royalty should we think of, if we didn’t know what the ASP was?
Yes, I think what we’ve started saying is sort of 1% to 2%. I am particularly at the sort of higher functionality, higher performance end, the Cortex-A type processors we’d be talking at the sort of higher end of that range. Things like the Cortex-M processor, then we’re certainly talking the low-end of that range. And what you have to remember is that still approximately 90% of our royalty revenue is coming from the licenses sold prior to 2006 before we really got into licensing our Cortex-A processors.
And so any uplift from those higher royalty rates is still having a very, very marginal effect on what we’re actually seeing in terms of average royalty today. And which is why in the presentation I said it’s very, very early days, but we are starting to see that effect come through into the numbers.
Nick Hyslop – RBC Capital Markets
As we look at your addressing of the computing market, is it fair to say you’re not in at the moment from a royalties perspective. It might be more reasonable to think towards the higher end of that percentage for these new applications are very high value processes.
That is a very reasonable assumption.
Nick Hyslop – RBC Capital Markets
Thank you. And then the second question that I had is that, and it has been asked several times before in these meetings but I think is more, more relevant. You had £290 million of cash now. You’ve increased the dividend 20%. It’s very obvious that the free cash flow is very consistent now across the quarters and growing, and with your increasing revenue from royalties that is not going to change. Is the structure now right, balance sheet structure from given the confidence that you should increasingly have in the cash flows from the business?
Well I don’t think balance sheet structure is a science, a pure science, I mean certainly what right now the Board is comfortable with broadly £300 million of cash in the context of our opportunity. Our plan of record our there around capital structure is that we have a progressive dividend policy, that means that overtime it has a relationship learnings. You’ve heard me say in these types of sessions that we expect to increase the payout ratio overtime. It’s currently at 30%. We think it’s appropriate as this model develops for that to increase gradually. So I think that’s one plank. Now if overtime we need to do things other than in managing our capital structure. We review it on an ongoing basis.
We obviously communicate that ahead of doing it but the plan of record at the moment is progressive dividend with that trajectory.
Nick Hyslop – RBC Capital Markets
There are some more towards the back.
Nick James – Numis Securities
Good morning. It’s Nick James from Numis. I just had a question on Mali, which you said you expect a meaningful volumes to come through in 2011. Can I ask you is that coming from the smartphone segment or from the digital TV segment? And if there isn’t meaningful volumes in the smartphone segments in 2011, when do you expect meaningful volumes from that segment?
Smartphones will be the first volumes, but we would expect it by the end of 2011 to see some DTVs as well.
Nick James – Numis Securities
Sandeep Deshpande – JP Morgan
Good morning. Sandeep Deshpande from JP Morgan catching up [ph]. Couple of questions, firstly, I mean reverting back to one of the earlier questions on the royalty rates and the potential ASPs in a PC market, I mean there are investors who are thinking at this point that well in the PC market, ARM could potentially have royalty rates of 4% and 5%. Would there be any basis to those assumptions, can we talk about octet core A9s for instance. What would be the royalty rate on that for Tegra 5 processor for instance? Secondly, can we talk a little bit about the PIPD and the penetration of PIPD in these compute cores, given that I mean that could be an addition to the royalty rate. Some of the leading at least publicly announced on the three licensees of the Windows processors which is TI, Qualcomm and NVIDIA. None of them are known licensees of your platforms in PIPD. So if it has that changed?
OK, I’ll answer the last part of the question first. There has been no change to what’s known about whether those companies have licensed our Physical IP or not. However, I would say that all of those companies are well aware of our Physical IP offerings and these companies with whom we have a consistent and ongoing engineering dialog. And so I certainly don’t rule out the prospect of some usage of our Physical IP, from those and those sorts of companies in future in this space. And Physical IP by into the Processor Optimization Packages absolutely, I would expect to find being used as a platform for some of the apps processors in computing over the coming years.
So that’s certainly a very credible royalty increment on the apps processor. Now obviously, I am not going to talk about octet core Tegra chips in particular because that individual customers. But what’s alluded to in your question is adding cores in a multi-core system, can that increment royalty? Yes, it can. And in some instances it can, in some instances, it doesn’t because royalty rates negotiated is slightly higher level in the first place. And what you deal with is an average across the different chips that a different of the – certain manufacturer is producing.
Whether the 4% to 5% is the right sort of level to be modeling? Now we sat here at the end of 2010, beginning with 2011. It sounds a bit racy but it’s not all together out of the question. If you think about multiple microprocessors on a chip, if you think about adding graphics, and if you think about building the whole thing on ARM Physical IP, all of these things contribute individual royalty. And at the high-end of the Cortex-A spectrum, then I said in the 1% to 2% range you’re still adding some multiple processors, yes you take off discount but it does add up as we illustrated at our Analyst Day in May. And the Physical IP also adds an increment.
So it does sound a little racy, but the notion of north of 3% is not entirely out of the question.
Sandeep Deshpande – JP Morgan
Are any of these lead licensees for the Windows OS using your graphics IP? Thank you.
They’re not amongst our public graphics licensees. None of those three individual companies. One of them, I wouldn’t expect to be. They’re specialism is graphics.
Andrew Gardiner – Barclays Capital
Hi good morning, Andrew Gardiner from Barclays Capital. Just I would be interested in hearing some thoughts on the other side of the computing equation, away from the royalty side but more on the units. You’ve given us a bit more detail in the slide pack this morning about how you see the total market for chips. But I’m wondering if you can give us a bit more insight into how you see your addressable market changing over the next three, four, five years based now on what we now both about the chip side in terms of A9 to A15 and even looking further beyond that, what NVIDIA has now announced with Project Denver in the next generation of the ARM core, but also on the operating system side with Microsoft. So how do we – in terms of the actual addressable market within the total CPU market for ARM over the next few years, how is that going to step-up overtime?
I think we’re illustrating this on slide 32. And I’d encourage you to look at slide 32 and may be talk with Ian afterwards or with Jonathan who puts slide 32 together. That’s where we see as the market opportunity. As I said in answer to an earlier question, it’s based on industry analysts data. Things like the Microsoft announcements are part of enabling the ecosystem to allow ARM to call this an addressable market opportunity. So undoubtedly the announcement made by Microsoft at the beginning of this year. It’s a tremendous boost. It opens tremendous market potential that would otherwise be much harder to unlock. I would still be potential but to achieve it without Microsoft would be making an assumption that Microsoft’s competitors would take a lot of share away from Microsoft which frankly, that’s up to Microsoft and their competitors and nothing to do with us.
So it’s a great indicator. And it’s an indicator that shows that the what we’re laying out on slide 32 is an incredible possible addressable market.
I think it’s probably also worth adding that if you look on slide 31, which analyses what actually happened in 2010, you’ll see that our estimate there, this is the market share of 28% and on the restated 2009 is 23%. And what we have observed over the last three or four years is that our share is incrementing at somewhere between 3% to 5% percent per year. And we don’t particularly see anything in terms of the licensing opportunity and the announcements that are coming out from our partners to that type of growth should change as we look forward.
I think we better make this the last question for us. Well as I said earlier, the others are in the room to continue with informal questions.
Yes, thanks for that [ph]. On the server market, I apologize being late. So perhaps you already addressed that point but on the server question I think you put on the timing 2015 something like 70 million units, if – 110 million units I apologize. Could you talk about that actually, because what I am trying to understand is first of all, do you need 64-bit support to penetrate the server market whether that’s the SOHO [ph] opportunity or the datacenter opportunity? Second, what sort of again of royalty rates would we get on datacenter capable ARM based processors or on the SOHO capable ARM based processor, that would – just so that where we can understand in model what used to be 0% market share business for ARM?
Well, the 64-bit question, its true a lot of service today, use 64-bit and some applications which run on service rely on 64 bit. Some rely on 40-bit addressing and that’s built into Cortex-A15. And so whilst Cortex-A15 is a 32-bit processor, it has the extended addressing. And this is a sort of ongoing thing. There are certainly server applications today which – for which 64-bit or lack of 64-bit is not a barrier, nor a 32-bit processors perfectly adequate to address that in typically multi-core configurations and Blades with multiple mutli-core chips on the Blade.
That’s where we are today. And out in 2015, then we’re not in the business of announcing our 64-bit products today by the way. But it’s logical to suppose that at some stage in the future, ARM will extend its architecture in that direction. And it would certainly be helpful as and when we have those sorts of products. But right now, we’re not ready to talk about those sorts of products and then large chunk of the server market is available with the product roadmap that we have.
And just a final question I’ll leave for the team. I think what explicit about your R&D direction for 2011. What I am trying to – what I am worried personally is that because of the opportunities that lie ahead for ARM in computing, you need to step-up your R&D may be in 2012 and 2014 so that’s the ARM partners have got the course they need in time to, if you want to start to fight with our friends in California. Is that risk, is that something that you would completely eliminate in terms of possibilities?
I mean Warren talked on his first summary slide about sort of financial discipline and balance. And yes, this is the way we look at it, I mean obviously, we have a world of opportunity ahead of us. But I think we see – R&D goes up in absolute terms. It goes down as a percentage, the margin goes up. We do not see sitting here today the need for a step change in our approach to investing in R&D from the trajectory we’ve been on, and talk about going forward. So you should still see, expect to see steady margin expansion in this business as you have seen over
Well, thanks very much everybody. We’ve got to formal Q&A at this point. Thanks for coming. We’ll be back presenting in July.