D. Warren A. East - Chief Executive Officer, Director
Tim Score - Chief Financial Officer, Director
Terrence Higgins - UBS
Nick Hissop (ph) - RBC
Joel Wamel (ph) - BNP Paribas
Janardan Menon - Liberum Capital
Dede Sam (ph) - RBS
Simon Schafer - Goldman Sachs
Brett Simpson - Arete
Aaron Judge (ph) - Milburn (ph)
Paul Morland - Astaire Securities
ARM Holdings Plc (ARMH) Q4 2009 Earnings Call February 2, 2010 4:30 AM ET
D. Warren A. East
Well good morning, ladies and gentlemen. It's 9:30 so we'll start away. Welcome to our results presentation. I am going to talk for a little while and give you an update on the business and then Tim will give a slightly more detailed update on the numbers and then we'll hand it over to Q&A in the usual way.
So to start with, 2009, we're reporting this morning 2009 and Q4 results. 2009 for the semiconductor industry was a pretty horrible year, particularly in the first half, and so we're very pleased to be standing here this morning talking about a year with our revenues down about 10% versus an industry down about 20%. To see that continued outperformance of the semiconductor industry at large is very encouraging and we're doing that through continuing to gain share in end markets and gain penetration in markets where we're already established and there are trends happening in smart phones, mobile computing, micro controllers, and consumer electronics that are helping us on our way there.
We have had a very successful year for selling licenses, more on that later. Our physical IP strategy is in good shape from both a licensing point of view and some very pleasing royalty numbers coming through from that part of the business. We exit the year with a backlog at record level, up sequentially and also substantially year on year so that is a good position to be moving forward on. And the way in which we've managed the business during 2009 obviously, as I say, it was a difficult year and so with revenues coming down as was inevitable, we had to be fairly disciplined about where we made our investments during the year. We need in our sort of business obviously to continue with investing in major technology developments and so I think 2009's been a pretty good year in terms of managing that financial discipline. That's translated into by the end of the year, Q4 margins at record levels and record cash generation.
The slide is exactly the same format as the last couple of years of results presentations. So very much we are going to be talking about how we are concentrating on executing this strategy which has not changed for some time and will demonstrate how we're achieving growth in our mobile applications, how we're increasing the value in products where we're already present, and how we're extending the IP that we're outsourcing to our customers.
So industry as a whole, probably the worst year for semiconductors getting on for a decade and for ARM, a year of resilient execution closing with several records and we will talk more about those I'm sure.
So now I'll delve into a bit of the detail hopefully. Growth in cell licensing, this is obviously an engine for ARM's future growth. Licensing is all about equipping the ARM semiconductor partners with technology that drives future business, and given the wider economic conditions in the first half it was a very good year for building our licensing base. You can see our licensing base grown to over 660 licenses, second half in particular strong for the number of licenses sold, and our exit rate of $36 million in the quarter in Q4 is an encouraging place to be.
Now, behind those numbers there were some quite meaningful public announcements from some of our licensees. So the Infineon architecture license for security is clearly a meaningful announcement for ARM, breaking into a new area with a leading player in that area.
We saw some encouraging licensing in the quarter as well for the Cortex licenses and for Mali, and this morning we're starting to talk about the fact that one of those Mali licenses was with Samsung who's obviously a leader in the industry and an opinion former in the industry as well. So that's very encouraging for our graphics licensing.
When you look at the pie chart at the bottom, which again we have shown in several of these presentations, the picture is as it has been for some time with licensing driven by both mobile and non-mobile, and we're very encouraged by a continuation of the non-mobile licensing, and within mobile we're increasingly seeing ARM processes used in peripheral controllers in the phone away from the traditional, just the base band mode, and just the applications processor. So we're talking connectivity, cameras, power control, touch screen, and those sorts of applications.
Now I'll just look at the licensing in a slightly different way, and again the left-hand side of this slide is a chart with which I hope you'll be familiar. It's the traditional format where we show the base of licensees that are out there to generate future royalties and the numbers in parentheses are what's been added to those licenses during 2009.
Now I showed a very similar slide this time last year where on the right-hand side of the slide we talked about three new processors that we were intending to launch or deliver during 2009. Well, we did absolutely deliver those products in 2009 and we've sold licenses and so now here's a look ahead to 2010 where we have another three new microprocessors lined up for launch and/or delivery in 2010. And as shown on the slides we've been signing up lead licenses for each one of those processors and you can see the Eagle Processor, this is a high-end applications processor targeted at the top end application processor market. We've got some real core aimed at realtime applications and we've got a new addition to our microcontroller product range as well. So those are all coming to market in 2010 and so that's a bit of a look forward to what's happening there — similar slide to the one we looked at 12 months ago.
One of the things, a little detail as you might notice here down the right-hand side of the slide though is that with the growth, particularly of the microcontroller product portfolio, we have updated some of our market size numbers and in the complete slide pack in the appendices you'll notice that the Ian has updated the charts again, charts which I hope most of you are familiar with, and you'll see that the total market that we're aiming at has grown a little.
So our licensing is all about driving future royalty and it's happening in a variety of ways, but numbers across the bottom are showing our licensing aimed at semiconductor partners aimed very specifically at growing our market going forward in the mobile space, but also in some new areas. And outside in the demonstration area we have ST today. Some of you will have seen the demonstration there. They are showing off their digital set top box and I already mentioned the Infineon design. The numbers down in the bottom indicate where is ARM in these markets today and where is ARM going forward.
If I switch now to royalties, then our Q4 royalties are up sequentially and well ahead of the industry. It was a record number of cores shipped by our partners in the relevant quarter which is of course calendar Q3, and that's driven by serious increases in microcontrollers, serious increases in smartcards, and the traditional areas in mobile phones, consumer products, and enterprise. And the pie chart at the bottom of the slide is again a familiar pie chart. You should see that the proportion of mobile/non-mobile is switched around slightly and I think that's primarily driven by the growth in the embedded area, the little yellow wedge on the slide, and we should also note at this point that the Cortex products are growing very strongly in the last quarter as well. We're now up to 2% of the total being driven by the Cortex products.
So in the year as a whole where did these volumes come from? A thing to note here, I had lots of pictures of electronic products into which ARM has designed over the years. I think a key thing here is the proportion of telephones that are there, and this is just more than telephones and it is adding some color, I hope, to the royalties' picture and noting as well, although in the 3.9 billion microprocessors, but for the year it was a record 1.3 billion in Q4. And this breadth of different products into which the processes are designed is what is enabling the continued outperformance where we're outperforming the industry.
And in the chart on the left-hand side, again this is an update of a familiar chart where you can see the ARM growth line at consistently a gap above the industry growth rate, and obviously over the years that's cumulative, that consistent gap builds up and that's really what's driving our continued outperformance of the industry as ARM grows its market share.
If we look at the numbers for the year as a whole then the units were down arithmetically, but that must be put in the context of industry shipments, including memories industry shipments at large down about 20% and then we've broken down some of the key market sectors where those shipments have happened and how ARM is performing relative to those key market sectors. So ARM's shipments generally are outperforming the industry at large right across the application space.
Now a little bit of detail on that, another metric that we're tracking on an ongoing basis and we've shown charts like this over the years, looking specifically now at the mobile space. The increasing amount of value per ARM powered product. If you'll recall, the first slide that I started off with is strategy about growing market share into new areas, but also increasing the value of ARM per consumer product that gets shipped.
In phones the number of ARM processes per phone average 2.1 for the year up from 1.9 and for the last quarter that we're just reporting it's a record 2.4 ARM microprocessors per handset. And the sort of things that are driving that are the phone market dynamics themselves, more high-end smart phones getting smarter, and the second thing is ARM being designed into a wider range of functions within the phones. So the phones are now coming out with functionality like touch screen, people paying attention to have more sophisticated power control schemes within their handsets, and you need a microprocessor to do those sorts of things and that's where ARM is getting designed in and that's helping drive up this average number of ARM microprocessors per handset.
And those two factors translate, as we move to the bottom half of this slide, into royalty revenue as a proportion of handset ASP. The chart on the bottom shows that that has consistently increased and is actually up about 70% since 2004, so over the five year view. And we believe that that trend is likely to continue and ARM11 A-class processors do get designed into higher value chips and so we'll see that trend continuing. And if we move to the bottom of the slide and think beyond smart phones about mobile computers and mobile computing generally, well we like to think of those, if you like, smart phones on speed where we'll have even higher value chips and even higher proportions.
So that's phones. Now it isn't just phones. It's the wider gamut of consumer electronic products. I was at the Consumer Electronics Show the first week in January and it was very impressive to see the number of exhibitor stands there with ARM technology on their stands. We did a little survey and it produced a very high, high number. And it was encouraging to see ARM being designed into an ever-wider range of consumer products.
Translating that into numbers in some of the sectors which we track, in things like set top boxes and digital TVs, already ARM has sat on a market share of about 30%. That's what we're measuring from what's coming out in 2009. Now, more significantly in 2009 we've had some announcements of design wins, and again I'll refer to the ST demonstration outside. Design wins which, at the moment, is our licensing activity and a design win activity, but which will translate into shipments, royalties, and increased market share in the future. So we're already gaining market share and that market share gain is set to continue. And it isn't just with the generic ARM general purpose microprocessors, it's things like the specialist processors, the MALI graphics cores as well, finding their way into those products. I'd encourage you to go and look at the demonstration to go and see what that MALI graphics processor can really do.
Moving to mobile computing, there has been a lot of noise about mobile computing over the last 12 months in the press, and as I say, there's a huge range of different form factors out there. We've got some of them in the demonstration area outside. There has been a lot of noise. At the moment it's the wage where there's a lot of semiconductor partner design wins that have been announced and some early products. What we're really interested in is not the number of net books sold in 2010, but just how this whole mobile computing market space is expected to grow over the next several years.
And with that I'll set that in context in the wider number of Internet connected devices that we see. This is what is really exciting us about the opportunity facing ARM over the next several years. You have a huge range of different form factors. I cannot tell you what they're going to be. Different companies have different terminology to describe this. Some companies talk about the three-screen strategy. We have rather more than three screens on the bottom of this slide because we can't really tell how big these screens are going to be from here, but what we can tell is that OEMs who are designing these products like a choice of suppliers. The ARM business model gives them the choice of suppliers and the ARM technology is used for a range of different functions within these electronic products. And what we're spending a lot of time and effort on at the moment is developing the software ecosystem and the development environment so that this architecture is ideally placed to gain the lion's share of the market as it develops over the next five years or so. And an indication of that is shown on that slide.
Stepping back a little about market share and market gain, I wanted to summarize, and we had a slide a little earlier in the presentation about how our licensing was targeted at some growth areas, how does that play out once you get from licensing into design wins? I think different sectors were at different stages and that's what this slide is meant to show. On the right-hand side of the slide in the mobile phone space it's all about innovation and this is one of the reasons we're also very keen on mobile computing because it's a place where a lot of innovation is taking place, and we've always said the mobile space is a great place to design your products because it's a very demanding environment technically and economically, and if you can develop things which are successful in this space then they can be applied into other application sectors.
Backing off into the enterprise space, the key factors are a little bit different. It's about multi-sourcing and having common architecture, reuse of software, reuse of tools and so on. And our share is growing there. It's obviously not as developed yet in that space as it is in the mobile space. In the consumer electronics space, represented by digital TVs and set top boxes it's about growing the ecosystem and the development environment, but it's also about market making semiconductor companies switching to ARM generally to reduce their total cost of ownership and cost of dealing with these markets and the share is less well developed today, but it's clearly on that trajectory moving towards the right-hand side of the slide.
And then in 2009 we also saw some activity in microcontrollers and smartcards. We saw very strong growth in volume shipments there, but actually if you look at the total market, our share in this market at the moment in spite of that very strong growth is very small, and what I'm hoping to demonstrate here is the huge potential growth opportunity in this sector of the market as we move from the position where we are on the left-hand side of the slide more towards the position on the right-hand side of the slide. Tremendous opportunity there. We've been bringing new products to market, we've been bringing new tools to market, our semiconductor partners have been launching ranges of products, and so the future for microcontrollers is bright.
Switching gears now and a look at our physical IP business, physical IP business has performed resiliently in 2009. We've sold three new platform licenses in the fourth quarter which if you look at the number of platform licenses we typically sell in a year, this is a very good result for a quarter and there is a span there of technology as well with the GLOBALFOUNDRIES at the very bleeding edge, and that's key for singing up big semiconductor companies and giving them something to work with. But other semiconductor companies and big ones in some parts of their portfolio make use of older technology and it's instructive to note that we're still able to sell platform licenses for new physical IP products on older geometries.
Most of the royalty today comes from older technologies, but since 2007 when we've had advanced nodes you'll now see that nine of the new platforms are delivering royalty. We've pushed the leading edge and I'll talk more about that in a moment or two. We've pushed the leading edge and we're expecting to see royalties at 32 nm before the end of 2010.
Let's switch to royalties for a moment and just talk specifically about that. Q4 was a very pleasing quarter with underlying physical IP royalty. That's excluding catch ups, up significantly year on year. And the chart at the top shows foundry revenues, and if you actually extract from the diagram, which I admit is quite difficult, but basically foundry revenues declined by about 5% and you need to look at the Q3 '08 bar and the Q3 '09 bar and that corresponds to the royalties that we record in Q4 and look at those bars and you will see our royalties up, the foundry spaces is down, and that's indicative of more continued increasing penetration. So it was record underlying royalties and indicative of gaining penetration.
With that gain coming from the future, a bit like our processor business where the licensing activity is equipping semiconductor partners with our technology to drive future business, so it is with our physical IP business where selling physical IP platform is equipping foundries with the appropriate technology so that we can then get design wins and then go on and generate future business.
And the chart is a snapshot of where we are today in terms of the platforms that we have out there spanning the full range from 250 nm right down to 28 nm, and you can see some of the data there. And we're typically growing that by, as we say, about eight different platforms per year.
Now today only about half of the total are delivering royalties so there's royalty potential building up in future and you can see the new royalty payers coming out in 2009. That's a trend that we're going to see continuing. Actually the leading edge that I referred to, the 65 nm and below, that's now contributing about 10% of our physical IP royalties so most of it, 90%, is still driven off of the older technology which again I think illustrates the tremendous potential for future royalty growth in this physical IP part of our business as well as in the processors part of our business. And we also note that an increasing number of the top 20 semi players are users of our physical IP today.
One of the interesting things to note is what's been going on inside the operational aspects of our physical IP business. There we've been, as you know, doing a lot of investment over the last several years and that investment has and continues to transform the business. It has transformed the business into where it is now, a leading supplier of physical IP. We've had investments in technology and more on that in a moment. We've also had investment in the development process and the internal environment that's used to create these products. And what that has done, that leads us to be able to build better microprocessors and we've achieved some processor design wins utilizing our physical IP which delivers better power performance characteristics.
And that investment is also transforming the operations of the business and that helps to improve the productivity internally. Those two factors together mean that as we move forward physical IP profitability, we expect to improve because licensing and royalties are driven by the delivery of new technology into the marketplace and at the same time as that licensing and royalty grows, the costs actually grow more slowly because we're now more efficient at creating the products and so with better productivity we have better profitability.
Snapshot update, a slide that we've shown for the last several, several times, and the three bubbles on the right-hand side of the slide is where the updates have taken place since we last showed this six months ago. Since six months ago the top right-hand bubble there where we're talking about optimized Cortex processor product, we'll be licensing 32 nm during 2010. Back at the very beginning of Q4 we licensed GLOBALFOUNDRIES and we've also delivered 32 nm physical IP products to the marketplace. So progress in terms of deploying the technology, progress in terms of engaging with the marketplace, progress in terms of delivering new products.
One of the key ways in which that physical IP progress translates, and this is the last slide I'm going to show, is building better microprocessors. It's a key result area for us, how our microprocessors perform is important, and our optimized physical IP keeps our microprocessors at the bleeding edge of competitiveness.
So last year I showed a slide ab it like this which was how we intended to improve the performance and power characteristics of the Cortex A9 and the way in which we expected to extract from the physical IP those improvements. This year I'm showing you the results of that and this shows that if you go for high performance we can get the Cortex A9 up on a 40 nm G process to 2 GHz. If you optimize for power consumption instead you can take the Cortex A9 and run it at the same 800 MHz, but actually reduce the power consumption to 500 watts. So that was the activity in 2009.
In 2010 we're taking this activity forward, and what we've done on 40 G we'll be doing on 32 nm and when we go to 32 nm then we expect to improve the power characteristics. When the thing's running we expect to improve the power characteristics meaningfully by between 10% and 20%. We expect to improve the standby power by at least an order of magnitude and therefore improve battery life. And if we optimize for high performance then we expect to be able to deliver the 2 GHz part with about a 30% improvement in power consumption.
So we laid out our plans last year, that's the results for 2009, that's what we're aiming at in 2010, and that's what physical IP is doing to our microprocessors, improving the competitiveness. With that I'll hand it over to Tim to talk about how this has translated into business.
Good morning, everyone. The usual quarterly change of the microphone position. The good news is I've got five slides on the financials. There's a lot of detail in the release. There's a lot of detail in the slide and the detail, I think the format of that most of you will recognize. So I'm just going to spend a little bit of time briefly summarizing, but also talking about some of the things that you'll probably be thinking about as you look at your models for 2010 and that's really going to be the focus.
And a bit similar to last year, we came into last year quite a difficult year to call actually. When we were standing here this time last year I think the general sense was the industry was probably heading down about 30%, certainly 20%-30%, and that's actually quite a difficult context to manage a business. So a year on, as Warren said, we're using the word resilient quite a lot today, but I think it's actually an appropriate term for the way ARM has performed during the year. We have outperformed the industry in line with previous years. Licensing, I think, has held up fairly well. Royalties, gains in market share — and that's enabled us to maintain an operating margin north of 30% for the year and I think that was by no means a given coming in so that's very encouraging.
And we have continued to manage headcount carefully as Warren said. I mean, in fact we end 2009 about 30 people less than we started and in fact the end 2009 headcount is actually below the end 2007 headcount too, so we really have been balancing this quite carefully, but absolutely maintained investment in the key projects and I'll touch on that a little bit more in a minute.
We had a pay freeze in as you know I think which stayed in throughout 2009. I'm going to talk about OpEx and OpEx guidance in a minute, but there has been an annual pay rise effective January '10 and we'll touch on that.
And from a balance sheet capital structure it's been a strong year for cash conversion and cash generation. We started the year just under £80 million, we're ending it just over £140 million. And in Q4, net cash in at over £30 million, a very, very strong performance.
Looking at Q4 in particular, again I'm going to go through this pretty quickly because we've covered most of it — I mean, it was a strong licensing quarter and we saw the pickup in licensing I think in Q3, number of licenses. Q4 strong both in terms of the number of licenses and I think we talked last time about the broad mix of models that we had. We're doing more term and more per use so the number of licenses is higher, but the absolute value of bookings in Q4 was very strong which has obviously driven the backlog at the end of the quarter.
And part of that of course is the substantial long-term deals that you see up there, the renewal of the ST micro subscription, the Infineon deal that Warren referred to, and the GLOBALFOUNDRIES deal. These are not deals that are going to yield step changes in license revenue in the first half of 2010 because the revenue recognition here is longer term, but these are certainly deals that sort of underpin future licensing revenue for a long time which combined with a technology portfolio of more mature cores that yield revenue recognition immediately, that is quite an attractive mix as we come into 2010. And as it says, they are backlogged very high.
Royalties, as we've said for a long time, are gradually increasing as a proportion of total revenue. In most periods royalties are going to grow faster than licensing and of course the impact on that over time is to drive up our margin both at the gross level and at the operating level. I'll touch on that in a second. And I am going to talk about OpEx.
OpEx in Q4, a little bit above consensus. Two things there really; one is gross margin was considerably above consensus so there's a little bit of reclassification there, particularly around the PIPD engineering time between OpEx and cost of sales, but also with a very strong revenue performance in Q4 and a very strong booking and therefore sales commission performance in Q4, the accruals or the charges for those payments in Q4 is higher than run rate and they would drop back down in Q1. So when you look at the combination of that dropping down and inflation for the pay rise coming in you get to a broadly flat picture.
So just to talk a bit about gross margin, because we started — well, in 2008 the gross margin was just over 89%, in 2009 it's just over 92%, and in Q4 it's just over 94%. So the sales side guys are sitting here thinking what's the gross margin going to be in 2010, and I think that's a fair question. So the way to look at this, I mean obviously revenue has a huge impact on this, but let's focus in on the cost pieces, and you can see from that chart that there are really three elements to our cost of sales. One is the engineering — the time of the service guys when they're generating service revenue, there is the product cost of some of the development system products, and then there is the engineering time from the physical IP division that is spent not on generic technology development, but on customization of technology to yield revenue.
And you can see that the first two of those are fairly stable and they would very much track services and development systems revenues. The variable tends to be the physical IP which is why you can see in Q1 of 2009 the gross margin was lower at 90% and that was a quarter where the PIPD engineering charge into cost of sales was higher.
So I think my message is that number will vary depending on what they're doing, but I think it's probably fair to say that a gross margin base now of around 92 is the way to think about it rather than the 89 or what we saw in the first half of the year. I think consensus before today's results for 2010 gross margin was about 91.5, but it does look as though somewhere around 92 is going to be the place looking forward. And as I say, over time that grows as royalties grow as a proportion of the total revenue.
Now we talked a bit about balance, let's just discuss how we are looking at this at the moment. I mean, throughout the last couple of years where we've managed our headcount very carefully we've continued to be completely focused on the technology roadmap development that drives this business forward. What we expect to happen in 2010 is a gradual increase in headcount subject to the market panning out in the way we sort of all currently think it's going to. And therefore for Q1, as I said before, probably going to be fairly flat with Q4, incentive accrual is lower, wage inflation in there. And then if we do increase headcount then obviously there could be a gradual increase in that number. So I would expect consensus OpEx to rise a little bit from where it is. I would expect gross margin to rise a bit. I would expect consensus OpEx to rise a bit.
And yet where it is in the 48-50 range, FX translation is clearly a big issue. Half of our costs are in dollars. What's the translation rate going to be in Q1? We don't know. What's the mark to market FX impact going to be at the end of Q1? That's going to depend on what the rate is at the end of the quarter so there are variables that you can't be specific, but it feels as though it's going to be broadly flat against Q4.
But I'll just reiterate that our sort of longer-term guidance about that we expect to grow our cost base about half the rate of revenues very much stays in place and that is a function again of the royalty growth over time and that's the thing that drives the business to a sustainable margin, the 40% and above which we've been talking about now for some time. And you can see that we've been making some good progress towards that in the last couple of quarters.
So what are we actually saying today? I mean, what we're saying is we see, hear, and read what everyone thinks is going to happen to the semiconductor industry in 2010. We read all the reports. We talk to our customers. I don't think it's fair to say we don't have any information available that says it's not going to happen, but we are saying is we don't have a crystal ball into the second half of the year either and therefore we're not going to be declaring victory on a big up year for semiconductors, but we completely understand why the current estimates are the way they are, but we'll put a qualifier on that.
What we do know, because of our licensing portfolio and because of design wins we know we had, this business comes into 2010 very well placed in a sense to continue to outperform whatever the market throws at us and so we are confident about that. Therefore, what we're saying today is reflecting on the industry improvement expectation that's out there and reflecting on the position that we have in the marketplace, the backlog that we have going into 2010, we would expect our full-year revenues in 2010 to be at least in line with what's currently out there in the market. And just to stem off the question, coming into theses results it's about $555 million consensus revenue for 2010 as of this morning.
Now with that, before we get into Q&A I think Warren is going to come back and make a couple of summary and longer-term comments.
D. Warren A. East
Thanks, Tim. So summary of 2009, I think we've summarized it to death really. Very resilient execution, but a balance there about building the platform for the future. In 2009 we knew it was going to be difficult, but we kept our investment going in new products and as I said a little while ago, that enabled us to deliver new products and enabled us to have new products for release and delivery in 2010. So we did that at the same time as outperforming the industry at large. The new products came in the form of new microprocessors, new tools, new physical IP, and that takes us into new markets. We're positioned with a record number of new licenses in the year as well and they're the basis for future growth going forward.
So we do expect ARM to accelerate its gains in market share, particularly as we're exposed to areas of long-term structural growth. So in mobile phones we're seeing 2.4 microprocessors per phone and we're seeing that continue to increase. In mobile computing we see a new range of products with all sorts of different form factors. We're seeing some significant customer announcements with digital TVs. In disk drives we're seeing increases in market share. We're seeing key announcements of market-making semiconductor companies in areas like smartcards and areas like microcontrollers where we're also growing the portfolio of products and we're gaining traction in things like graphics and outperforming the market in physical IP as well.
Now that was all about where we are and the trajectory, and we did want to step back at this point and just sort of set that in the context of the longer-term picture for ARM. In November 2010 we'll be celebrating our 20th birthday in this business and as we look at the industry for the next decade we can see people struggling with technology and the need for technology which delivers on low power which is key to a huge range of electronic products. We see people struggling with managing costs and we put both of those together and say ARM has a business model which enables outsourcing and ARM has technology which addresses these longer-term technology challenges that our customers face. And we're putting effort into supplying an ecosystem which enable those customers to use our technology, take advantage of it in the growth markets that they see coming forward, mainly today a snapshot of those markets would be characterized by more electronic functionality, more software functionality in products, more of things like connectivity, Internet, clouds, and so on.
So over the next 10 years we think we're pretty well positioned to take advantage of this. As we sit here today we know that our semiconductor partners have shipped over 18 billion microprocessors, so over the next couple of months in our 20th year we'll go through 20 billion microprocessors, and the ecosystem is well established with huge numbers of engineers. So these customers faced with this burgeoning electronic marketplace can use ARM technology whether they're building products for entertainment, productivity, making people better, saving energy — you name it, all of these companies have different drivers of their business. But the common factor is that ARM is uniquely positioned to enable and take advantage of that growth market if you take that slightly longer-term view.
So with that we'll hand it over to Q&A.
Question and Answer Session
Terrence Higgins - UBS
Thanks, this is Terrence Higgins from UBS. A couple of quick ones if I could, just on the headcount increases, Tim, I just wondered if you could talk to where those headcount increases will go to and what the sort of strategic priorities are in terms of R&D over the next two years or so? And then secondly I just wondered whether we can talk about Intel and competition during the course of the latter stages of this year and into next and what your assumptions are in terms of smart phone market share going forwards? Thanks.
Yeah. I mean from a strategic priority standpoint the headcount investment will be largely within the processor division and within the media processor division. I think we've said here recently that our expectation is that the physical IP division, if anything over the next few years, is actually going to have a headcount gradually reduced as we move out of this phase of catch up of technology and we move into more productivity and efficiency within that division. And so but we do see significant opportunities in the processor division for further technology development to grasp the opportunities that Warren described and also of course the multimedia processor is a much younger business that has a very attractive prospect in front of it. So in terms of — those two would be the key in terms of the support functions, the overheads. These will all be very, very carefully controlled and the thrust is going to be in those areas.
D. Warren A. East
Okay. With regard to the question about Intel then, absolutely we see Intel as a very serious competitor for us in this marketplace and clearly the largest semiconductor company in the world has got to be a serious competitor. Actually the ARM world is also of serious scale with approximately the same amount of ARM microprocessor revenue per anum as Intel microprocessor revenue per anum. Our funding goes to development activities.
We absolutely would expect Intel to make some progress at improving how appropriate their products are for embedded applications and things like mobile phones. Similarly, ARM is making progress in its ecosystem development, in its range of standard application software that's out there, and in terms of position within perhaps markets that are a bit more traditionally viewed as Intel markets and we had some pictures in the presentation about smart books and net books and those sorts of things. So ARM is making progress in that area, and one of the messages I hope you got out of the presentations that we both gave is investment in future technology.
So right now ARM technology is a long way ahead of Intel technology in terms of appropriateness for these types of markets. Intel is making progress, ARM is also making progress, and ARM is also investing to open up that gap going forward with some of the improvements that I talked about.
The 9% sequential decline in the average royalties per unit, you had previously I think talked about that you reached about $0.05 royalties per unit when you reached about a billion of microcontrollers. So was there any seasonality in that or was the decline actually slightly stronger than you expected?
I mean, you are right. It is about the product mix. I wouldn't say it was necessarily unexpected. It's what pops out of the arithmetic. The fact is that microcontrollers have just grown on top of a large number by 60% and smart cards have grown by 100%, and modeling is never completely accurate. The result which pops out can move in both directions depending on the product mix so not particularly surprised and not particularly concerned.
We'd point out that actually the royalty in the quarter in revenue in terms exceeded our expectations and that of others by about $3 million, and also the volumes exceeded our expectations by a couple of hundred million units as well. So it's just one of those things. You have higher volume, you typically have a lower price per chip.
And then maybe a similar question on mobile, 2.4 costs per unit, can you give us maybe an idea where you see that at the end of the year, the run rate?
Hard to call the end of the year because really it is about calling both the proportion of smart phones, but also the increased smartness of simpler phones. And the message we tried to get across in the presentation is that is not just about having more smart phones with an applications processor as well as a (inaudible) modem, it's about ARM technology being used in a range of functions within the handset. So we absolutely expect to see that number increase, whether it goes to 2.6 or 2.7 I couldn't tell you.
Nick Hissop - RBC
Thanks. Nick Hissop (ph) from RBC. You mentioned Samsung as a new customer for the MALI graphics processor. I wondered if you could give us an idea of what some of the drivers were that led Samsung to select MALI and if having a tier one manufacturer, whether you would have increased confidence in other Tier 1 manufacturers taking MALI more seriously than they have so far. That's the first question.
D. Warren A. East
Well first of all on Samsung specifically we talked about it as much as we're going to talk about it. Samsung will be having their own product announcements when they announce the products and then they'll be talking all about where it's going and the way forward so that is for them.
In terms of other companies like Samsung, well I would count companies like ST as pretty serious leading-edge companies and we're happy to have them with us today. We're happy to have Mediatech in the fold as was announced before, and so I think it's a matter of time and I'm confident that some other large players will shift in due course.
Nick Hissop - RBC
Okay, one second question. On your guidance of $555 million, if you look at the sort of industry guidance for 2010 semis, it's something say 10%-15%. You're looking at 10%-11% that you're saying you're comfortable with and yet your historical ability has been to outperform that by some margin. What are you basing your comments on your comfort on in terms of what you think the industry will do?
So what we're saying is 555 which is about 13%, right? I mean, I think it's probably lower than most people's expectations for 2010 and what we're saying is if most people's expectations which is for more than that comes to pass, we will expect to do at least as well as 555 because you're right, we're struggling to find a year where we haven't outperformed the overall industry. Our only qualification in our guidance today is we can't call the industry. What we can say is we expect to outperform the industry and therefore if the industry grows in line with consensus then we kind of know what that means for the expectation of ARM revenue. But without that qualification we are here gazing into a crystal ball into the second half that no one knows anything about and there's sufficient uncertainty in the macro world to not be standing here or sitting here in early February high fiving that it's all glorious and it's all 20% up for the industry. That is too early to say, but whatever the number is we will be very disappointed if we're not ahead of it.
A couple of questions. I mean, clearly you've licensed the MALI quite extensively now. How should we look — I mean, have the royalties on the MALI started coming in already? How should we be looking at the MALI royalties versus the processor royalties? I mean, is it a percentage of the chip ASP given that those chips may already have your processor on it for instance? And secondly, I mean there are going to be multiple processors and multiple chips selling out there for the smart phone market which integrates the application process of functionality with the baseband functionality so maybe dual core ARM processors. How will that impact the ASPs in this smart phone cycle?
And then finally, Tim, one question for you. On the overall guidance, I mean when we look at some of these and what you've said in the past, for instance ARM has a model of growing 15% (inaudible) on revenue terms and OpEx growing at much lower rates than half that revenue level, why haven't you stuck to — I mean, clearly you're giving whatever the market is so I mean are you trying to say that you're on model at this point, ahead of model, or you're not willing to say because you're not willing to make a call on the industry?
Thank you for directing your other five questions to Warren, I appreciate that because I was trying to unravel those. But I mean no look, this is a bit like last year. We sit here at the beginning of February, no one really knows what is going to happen in the full year. Consensus was 455, it drifted up through the year. We're usually slightly ahead of it and we ended up at 490.
As regards what does ARM grow in the medium term, I mean yeah, we do see this business as 10%-15% grower through cycles, but we've been through some pretty interesting years. I mean, two years ago our royalties grew 28%. Last year they're down a few points because of what's going so I think it's very — so what we're saying is medium to long term this business outgrows the industry, this business has a business model that's driving operating leverage to margins that we think are sustainably north of 40%.
For this particular year, we can only really comment in reference to what we see around us which is an expectation for the market of around 15%-16% growth, some are higher, few are a little bit lower. That will be a great outcome and we would expect to outperform it.
D. Warren A. East
Okay. The questions about MALI and royalty; typically we've said that from licensing the microprocessor core to generating meaningful royalties is a time delay of about four years. Now, obviously some companies get there faster than that and some take longer, but it's about four years and it's exactly the same with the MALI core as it is with the general purpose ARM core because it’s the same process, designing a system on chip device.
So we started MALI activity around about the end of 2006, effectively the beginning of 2007, and you're right. We've licensed the core extensively over the three kind of years since then, and we have actually seen our very first MALI royalties coming through. But it's not very meaningful at the moment because of that four year lag.
How does it work? It's very similar to a generic general purpose ARM microprocessor core. It's a percentage of the chip price typically and if you integrate a MALI core on the same chip as another ARM microprocessor then it's similar to integrating two ARM microprocessors on the core. Typically both cores come under royalty and the second one has a discount. And perhaps because of the special purpose functionality in the MALI core, the discount is probably not as great as it would be for a second general purpose ARM microprocessor. So you can think of it as an incremental royalty.
And whether there is an applications processor or a baseband and exactly how many cores are integrated, the principle is exactly the same. Multiple cores command multiple royalties with discounts for integrating multiple things on the same. In some instances we win out of that integration and in some instances the semiconductor companies win out of that integration and there are all sorts of schools of thought about which is the best way to integrate.
Joel Wamel - BNP Paribas
Joel Wamel (ph), BNP Paribas. I have got a question concerning your backlog. It seemed to me that it's in the hand of a couple of customers, so I'd like to understand if you are seeing the Tier 2 customer coming back or if (inaudible 00:13:53) Tier 1 so far? And if we could assume the backlog would stay where it is today going forward in Q1? And second question, is there any reason why you haven't gained any market share in the automotive?
Yeah. I think what we're saying today is if you look at — quite a spike in the backlog actually. What we're really saying is that most of that uplift is being generated by these sort of longer-term deals. With regard to sort of the general health of the licensing and the Tier 2s and all that, I mean I don't think they are back up to full speed by any means. I mean, I think we would expect to see that situation sort of further improve. I mean it's quite sort of regional I think in terms of that, but so I think the backlog is a couple of things.
It's pretty healthy licensing generally towards the back end of the year and it's significant long-term deals. And it's that latter point that yeah, I'm saying don't expect the spike in license revenue in Q1 and Q2 just because of that backlog. It's helpful and it contributes, but the opportunity pipeline for Q1 and Q2 deals of more mature processors which generate revenue recognition immediately is, in a sense, more important than the take from backlog. But I mean the backlog contribution to first half revenue is still pretty healthy by historical standards, I just wouldn't want you to map it proportionally against the increase in backlog in the fourth quarter.
D. Warren A. East
And the question on automotive, where are we in automotive at the moment, included in the packs you'll find we've updated the segment slides on Slides 31 through 33, but referring specifically to automotive, we're on about two ARM microprocessors per car at the moment. Automotive is a very slow market to adopt our sort of technology. It's very conservative. There are some established players out there with 32 bit processors in engine management systems. Engine management is a relatively low volume application for us and it's not one that we have historically concentrated on. And as ARM makes progress in general purpose microcontrollers then I think you'll see some more progress in automotive.
As we make progress in the multimedia errors, navigation and those sorts of things, you'll see us making progress in some of those in automotive, and to be honest the analysis on the Slides 31 through 33 is probably not perfect and some of the automotive navigation could well be coming through in some of the other areas as well as in the automotive line, so it might actually be a little bit ahead of the number on the automotive line at the moment.
Janardan Menon - Liberum Capital
Hi, it's Janardan Menon from Liberum Capital. I will try and slip in a few questions here. One is on the commission and bonus payments that you paid in the last quarter, how should we think about that? Is that something that's more likely to pop up in Q4 every year or if you do have a good licensing quarter in any quarter of the year can they pop up here and there?
Second question is you sort of finished the catch up phase of your physical IP business and now you are licensing 32 28 nm physical IP. So does that give you a lift in terms of licensing revenue on the PIPD division from the sort of $9-9.5 million that you're running at, or is this it? I mean, is it pretty much at the run rate that you think is going to be there in the long term?
And three, there has been a lot of talk about the net book market and how ARM is penetrating there. Do you need an additional product there above the sort of A-class processor, especially if in the future Microsoft comes to ARM processors, et cetera, and you have a chance of getting into the laptop market as well? Is the smart phone application's processor oriented product enough for that market or do you need something which can run at 2-3 GHz and compete with the best that Intel can throw at you?
And the a fourth if I can throw it in is on the 32 bit microcontroller market alone, just on the 32 bit part, do you have an estimate of what your market share was at the end of last quarter?
Okay. Can I volunteer for the first one and then you'll have the other three ones?
D. Warren A. East
Where delegation is a good idea. I think you need to look at sales commission and bonus payments separately. When I talk about the bonus this is the corporate wide bonus plan that is tied into sort of annual revenues and profits. And what you're trying to do with that of course is trying to accrue through the year so that when you get to the end you've accrued the full amount and so therefore in a fourth — I mean, it could happen in any quarter you could have a spike, but in the fourth quarter — and you can see sort of from expectations that the outcome was a bit better than we would have accrued for in October sort of thing so there is a little bit of touring up. So you have got the accruals in Q1, Q2, and Q3 which needed to be topped up.
Sales commission is something that is essentially paid following the previous quarter's performance and you're obviously accruing for it based on what's actually happened in the quarter. Therefore, in any quarter where there is a significant bookings quarter you will get a higher commission.
To be honest, quite often in our business it is the fourth quarter. I mean, I think we all know because we talked about it before, in the third quarter that is quite a challenging quarter for licensing because it is quite a short quarter and quite often fourth quarter budgets, people's mindsets about getting deals done by year end, this sort of thing, does. This is not unusual that ARM's had a stronger Q4 for bookings than we have in most other quarters, so two slightly different reasons. But as I say, ordinarily you would expect that number to drop into Q1.
D. Warren A. East
Okay. Let me deal with the other three questions. The first one was about physical IP licensing and the fact that we now have our product available at 32 nm and what we're going to see in terms of how that translates to license revenue. I don't think you're going to see explosive growth in our physical IP license revenue, but I think you will see steady growth, not necessarily every quarter increasing, but we have spent time and effort getting this product portfolio in place. It is in place. It is being adopted by leading foundries and the products are being taken up by leading players. So I wouldn't expect to see it necessarily grow as rapidly as physical IP royalty which is of course cumulative compared with the licensing, and as I said in the presentation, is driven off a whole lot of platforms whereas we tend to be licensing probably more at the leading edge. So you'll see growth, but modest.
Net books, PCs, high end, that question — I don't think we need a product which is specifically different. I think if you look at the top end of our application processor roadmap at the moment we have Cortex A9 which is not really in any end products as they exist today. If you go and buy any state of the art mobile computing product it's probably got Cortex A8 in it, today a single core Cortex A8 running at somewhere between 600 MHz and a 1 GHz. Cortex A9 will comfortably run at those sorts of frequencies and indeed with physical IP optimization as we demonstrated can scale up to 2 GHz today on 40 G processors.
The other Cortex A9 has a lot more headroom to go. It's a multiprocessor design so you can have quad core or up to quad core implementations. And one of the products that I had on the to be delivered/released this year is Eagle which takes our application processor products onto yet another level of high performance and obviously we'll talk about the details on that when we launch the product. Sorry, we're not launching it today.
So I think we are addressing that market with the products we have and with the new products we have in the roadmap.
On 32 bit MCUs, I have to confess we don't break that data out separately. We don't do that separately because we regard the 8 big MCU market as fair game for microcontrollers and very much I don't see winning in the 32 bit microcontrollers and just defining the market as that space as being winning. Winning is winning in microcontrollers where today the bulk of the volume is actually 8 bit microcontrollers and that's why we call that the market.
Dede Sam - RBS
Dede Sam (ph), RBS. I'd just like to ask you a couple of questions if I may. First of all, talking about I think your ambitions, because I'm not sure there were targets for ambitions of market share in things like portable computing, e-readers, tablets, net books — maybe Warren if you could refresh our memory as to where you think you could be this year?
Also, related to portable computing, should we assume that the royalties of PIPD will really take off when this portable computing opportunity really materializes? And I've got just two quick followups on licensing.
D. Warren A. East
Okay. I'll refresh your memories. I think what we said in June 2009 was that we have an internal target for 2010 to be in 20% of the net books. We actually missed that target by a long way. There are no appreciable ARM based real net book products out there that you can go and buy today. There are many design wins, but I think the dynamics in 2010 have been held up by a couple of factors. First of all, the economy and the sort of industry at large was not a good environment for exploring new products. Secondly, there's been a lot of uncertainty over which Linux desktop operating system should be released and we had in particular a couple of sort of switches around of which the favorite would be. So, a number of reasons, but internal target kind of missed, but I think as I said probably a year ago we're talking 5-6 million units here when we're talking about these products so let's put that in the context of the 3.9 billion ARM microprocessors chip when we talk about the opportunity.
Your question actually referred to more than just notebooks however, and you talked about e-readers and tablets and those sorts of things. I think you'll find pretty much all, if not all, of the e-readers are ARM based that are out there. The one that we've got out there on the demo area is an ARM11. Most of the others are ARM11 as well at the moment. So, very strong share in that particular form factor. Tablets, when at CES you can see at few tablets based on our microprocessors and we'll see it depends on things like the operating systems and exactly how those tablets take off.
Dede Sam - RBS
I guess you meant 2009 for your target, not 2010?
D. Warren A. East
Dede Sam - RBS
You meant 2010 or 2009 when you said 20%?
D. Warren A. East
Oh yes, by the end of 2010. It was, you're right.
I think he's declared failure a bit early (laughter).
D. Warren A. East
Yeah. Just to point it out, I got my years messed up.
Dede Sam - RBS
Okay. And then just a quick one on licensing, you announced Cortex A5 as a replacement to the sort of ARM920 64 basebands, can you talk about maybe sort of licensing out for that part, what sort of feedback you get from your clients? And also, if you could give us an idea of (inaudible) roadmap at the ARM V8, so your next processor core. What sort of performance are we looking at in terms of gigahertz/power consumption?
All right, Cortex A5 we've had a little bit of licensing of Cortex A5. We've not particularly separated it out in these numbers. We would expect that part to license quite strongly into the user base for ARM926, ARM1176 type products. This is for people who are building an applications processor, but perhaps one that's going to sell at a slightly lower price point than a Cortex A8 or a Cortex A9 so you will see that going forward.
Dede Sam - RBS
I am also wondering if you could give us a flavor of the ARM V8, what sort of features we're looking at?
Yeah. Well, as I said in answer to that previous question, we're not ready to talk about future products in detail today. Unfortunately that comes at product announcements, not earnings announcements, sorry.
Dede Sam - RBS
No problem, thanks.
Simon Schafer - Goldman Sachs
Thanks, Simon Schafer with Goldman Sachs. I want to ask a followup question on Janardan's point on PIPD. So we take the statement that licensing revenue on PIPD is perhaps growing moderately, and if we look at last year your cost base was almost twice the run rate of your sales level. I know perhaps on a reported basis that's not quite the same as you think about it, but in that context, if license is really growing at a moderate pace as you say, it'll take you a good three years to get that business to break even on a standalone basis unless costs are falling. So should we assume costs are falling very substantially or does it just mean that we're looking at a significant loss run rate for an additional two years?
Don't forget when you look at the segment analysis you have to allocate all of, in a sense, the fixed cost of the business across the revenue earning streams. So incremental PIPD revenue makes a big difference in terms of the contribution it generates on top of this sort of fixed — so if you look at his PIPD revenue, his PIPD's direct costs, that is what we're going to see in improving profitability. So you're right, you need to look at the (inaudible) of the direct contribution level versus covering for fixed costs.
Simon Schafer - Goldman Sachs
Yes sorry, I'm not sure I understand. So that means cost is being reduced for that standalone business or it will just take it three years to be profitable or breakeven on a standalone business? I understand the difference in allocation, but —
Yeah. I don't think we're prepared to sit here and say costs in that division will fall over the coming years, but what we are prepared to say is that costs will rise much more slowly than revenue and I'd point you at the reason that we included on the slides, the productivity increase. That's how we're driving the increased revenue, more products coming out of there. The investment has not just been about developing new products. It's about investing in developing the new product development process which enables us to be efficient and enables us to grow the business without really growing the cost line very much at all.
Brett Simpson - Arete
Thanks. It's Brett Simpson at Arete. I had a couple of questions, maybe first for Warren. A lot of the mobile chipmakers have started to report fourth quarter numbers and suggesting down revenues sequentially, and they're warning about pricing for next year being a much tougher environment. So you put a chart up showing the value of the ARM royalty per handset which continues to go up. It's been a consistent theme over the last four or five years. In the face of chip ASP declines for the chipmakers and smart phone ASP declines, I mean how sustainable is rising royalty value per handset in mobile?
D. Warren A. East
Well, of course the answer that pops out is the result of how quickly does the number of ARM microprocessors per handset increase compared with the ASP of those that are there decline? And if one declines faster than the other one increases, then you won't see that chart continuing to increase. But I would say that price pressure and price erosion is a feature of our industry, has always been a feature of our industry, and applied just as well in 2004 when we started doing those charts as it does today.
And so history doesn't necessarily guarantee the future, but it's a pretty good indicator and so I think you'll see that chart continuing to increase.
Brett Simpson - Arete
And maybe just a followup for Tim, can you maybe just confirm where the order cover for PIPD licensing is for 2010? I mean, run rates are normally about the 50% level as you enter the year, is that still the case this year?
I'd say historically it's been more around the 40 which is what we said and I would say it's at the right side of that coming into 2010.
Aaron Judge - Milburn
Aaron Judge (ph) from Milburn (ph). Just one question for you, Tim. In terms of the long-term operating margin target of 40%, what's your underlying assumptions for the margins for PD and PIPD?
We don't go into that level of details in terms of margin and the reason for that is that this business has many elements of the cost base of this business which are unified and one, and so I think the key point is operating leverage in this business is not just about processor royalties growing faster than the commonplace. So I appreciate that that was not a very coherent answer, but PIPD and the sort of things that Warren was talking about is going to be a very meaningful contributor to the increase in the overall growth margin because the revenue line that's growing modestly on licensing and probably better than modestly on royalty with a cost base that's much flatter than the rest of the group, the increase in contribution from that division to cover the fixed overhead of the wider-arm business is going to be really quite meaningful. But we don't anymore break down sort or sort of artificially recreate for the external audience the margin by division because we don't really manage the business quite like that.
Dede Sam - RBS
I got a followup question. We are seeing some of your customers referring to 32 bit controllers at $0.65 today, do you think it could be a trigger for the industry to accelerate the migration to the 32 bit and to skip the 16 bit actually?
D. Warren A. East
Okay. By the way, I think we've got to have one more question after this if there is one because time is marching on. We'll be here to talk after the event. That's one factor which could contribute. You would expect me to say this, but I'm a firm believer that the market is going to migrate from 8 to 32 bit and 16 bit is going to be a niche area. And I think it isn't just the cost that's enabling that, it's all the advantages of a 32 bit and in particular the tremendous industry standard that ARM has become and the knowledge base that's out there around the ARM architecture just makes it very compelling.
If you can do that with no cost increment compared with an 8 bit controller then you can do it with no cost increment compared with a 16 bit controller, 16 bits just does not win. So I think it is going to happen, the cost is one of the factors.
I think this is the last question.
Paul Morland - Astaire Securities
Hi. It's Paul Morland from Astaire Securities. My question is just on the cash actually. You've just come off your best six month cash conversion in the last six six-month periods, can you just say what's behind that and what it might mean for cash conversion going forward?
Well, I think in absolutely terms what you've actually seen in the second half is really quite a pickup in revenues, and specifically royalties, at a time when the cost base has been managed really quite hard, pay freeze and everything else. And I'd also say that royalties which has been one of the key drivers, you tend to collect quicker because it's one quarter in arrears. When the reports come in the cash is either same time or shortly thereafter. So I think the prognosis for profitability which as you can see translates into cash pretty quickly is looking very positive. I think it's on an upward trend and that's kind of what you're seeing in the six-month conversions.
D. Warren A. East
Okay. With that, thank you everybody for your support and we'll see you in July.