Arm Holdings - Earnings Call - Q4 2011

Transcript

Warren East (CEO)

Okay, good morning everybody, and thank you for coming to our results presentation. I have an obligatory slide to flash up in front of you. In the usual manner, I'm going to do a business update, and Tim will summarize the effects, the financial effects of our business update for the last quarter before we do questions and answers. I've been enjoying myself so far this morning with some of the broadcast media because we've been talking about another record quarter for Arm, which brings to the close an excellent year. I think the highlights shown on this slide, really that excellent year and record quarter and increase in backlog and so on, it's been driven by continued activity from the major semiconductor companies investing in Arm technology. In the last quarter, we're pleased to announce our lead partners signed for our Armv8 architecture products that we announced in October.

Actually, we haven't announced these products. What we announced in October was our Armv8 architecture. We will in due course be announcing some products. What you can see here is that some of the lead partners, our first lead partners, have signed up for those products. This is indicative of the strength of the Armv8 architecture for the future. We're very pleased with that. There'll be a little bit more on that as we go through the presentation. We have another subscription licensee in LSI, and I'm pleased with that. When we come to talk about that, LSI has been an Arm partner for a very long period of time. They use Arm in a handful of different products, and they have a product line around infrastructure and networking centered around the PowerPC architecture.

In their press release just a week or so ago, they talked about adopting Arm technology in those products as well. It's very good to see LSI joining the subscription licensees, particularly for that application. If you look at our business performance, we're continuing to grow our market share in all our target markets. I really struggle to single out one area and talk enthusiastically about it at the expense of the other areas because, you know, right across the piece, we're very pleased with the progress there and overall increase in market share that we'll come to. In particular, microcontrollers are driving the volume, and mobile is driving the value. We've talked on, when we put up this slide, we talk about growing Arm, and we talk about the different growth trajectories there in the graphic. Obviously, one of those is extending the outsourcing model to other IP.

I'm very pleased with our progress in our physical IP business, consistently building the processor optimization packages alongside the Cortex A products, continuing to work at the leading edge in physical IP. We're highlighting here our five Mali licenses in the quarter, bringing the total to nearly 60 Mali licenses now. I'll talk again in a little while about Mali's success, particularly in digital TVs. All in all, an excellent quarter, generating record revenues, generating record levels of profitability, and that's coming through into profit. There's been a lot of column inches during 2011 on Arm's progress in mobile computing. I was at CES a couple of weeks ago, and at CES, the interesting thing to note was that Arm was everywhere. That was the really interesting thing to note.

I had an interview really early on, having been there for about two hours, and somebody said, "What's the most exciting thing you've seen so far?" I hadn't really seen anything. What I had already picked up on was that you could find Arm technology in just about everything you looked at, from home automation through to the biggest digital TVs. We were also seeing Arm a lot in mobile computing. There's a huge range of form factors on show at CES, and certainly most of the mobile computers on show at CES were Arm-powered. Those Arm-powered apps processors incorporating our Cortex A processors, but the good news was that some of them were also incorporating Arm graphics and Arm physical IP. Of course, the opening keynote at CES was from Microsoft, and Microsoft led off their keynote with their Arm-based mobile products.

They went on to talk about Windows 8, and they went on to kick off their demonstration of Windows 8 with an Arm-based product. After 12 months since the initial announcement, we're now getting very close to Microsoft's beta launch. At CES, they talked about the beta being available in Q1. After 12 months, we're getting close, and we're looking forward to the Windows activity during 2012. Obviously, it was a consumer electronics, a consumer product show. I mentioned digital TVs. Digital TVs for us, it's an exciting area. These TVs are becoming smarter, and as they're becoming smarter, they're becoming connected. As those two trends are happening, they're using more and more Arm technology. Over the last few years, we've been gradually increasing our share in digital TVs as they've been getting smarter and getting more connected. We think our market share in 2011 was probably about 40%.

The good news is that in 2012, we expect that number to be over 50%. It is driven by the need for these TVs to become connected to the internet. These TVs are also including graphics processors, and they're starting to ship. Our Mali graphics processor is very well positioned in the TV space. We expect that of the TVs that ship in 2012, certainly more than 50 million of them will have embedded graphics. We would expect to have about 70% of the market share in graphics in digital TVs in 2012. A footnote on the bottom of the slide there, we're pleased to see our physical IP evident in some of these processors as well. Computing has been a topic, not just sort of mobile computing, but computing in general has been a topic of discussion around the Arm architecture during 2011.

It's not going to go away in 2012. We're looking forward to some great progress there. The summary is that the mobile computers, still a lot of those products that will be launched this year will be around Cortex A9. There'll be dual-core A9s, there'll be quad-core A9s. That's good for us from a business point of view because obviously these chips are higher priced chips, and therefore translating into higher levels of royalty for Arm. We're also expecting to see the first devices based on Cortex A15 this year in the mobile computing arena. You're going to see some Cortex A15 products just in some of the initial server activity too. Just before the end of last year, in the fourth quarter, we saw an announcement from Hewlett-Packard about their activity, which is based around actually Cortex A9 in servers.

This is pointing the way for servers based on low-power architecture from Arm. Around the same time in October, we also had the version eight architecture announcement. The good thing about version eight is that it includes 64-bit. It includes some other features which are applicable for computing and server applications. It speaks a little to the need for increased levels of security within built hardware support for formal cryptography. As with all Arm architectures, it's totally backward compatible. Code you wrote on Arm 7 in 1994 will run. Of course, it does a lot more than that. We're going to see these Armv8 products out in silicon. We have some announced architecture licenses. Two of them are announced, but we have some other version eight architecture licenses as well.

As I mentioned at the beginning, we've signed up our lead licenses for the Arm implementations, which will be announced later in the year. Armv8 products are expected to ramp into silicon in 2014. We have a little while to wait before the silicon reaches significant volume. That's enough of an introduction. Switching to the business summary now and looking at licensing to start with, the numbers in parenthesis on the chart, which should by now be familiar to you. The numbers in parenthesis are the numbers of licenses that we sold during the last quarter. The 25 licenses that we sold bring us to a conclusion, an excellent year for licensing and a total of over 120 licenses sold. There was a good mix of licensing across the Cortex product family in Q4.

Interestingly, this is the first quarter where we haven't actually sold any licenses for those older microprocessors. On the far left-hand side of the slide, we just sold licenses to the newer products. Another nine Cortex-A licenses during the quarter, another eight Cortex-M licenses. As we look forward for the last several years, when presenting a slide that looks very like this, we've had the arrows and we've alluded to project names for some of the new products which are coming out of Arm Holdings during this year. At the top of the slide there, you'll see our new graphics products aimed at higher performance and also energy-efficient graphics processing. We're going to launch those later in the year. Atlas and Apollo are the code names for the first Arm version eight microprocessors that will be launching later in the year.

The aim there is to target the high-end of the computing spectrum, servers, and high-end smartphones. At the bottom of this slide, we talk about Flycatcher. Flycatcher is the microcontroller end of the spectrum. This is going to be our smallest, most low-power product coming out of Arm yet. This is really aimed at all those intelligent sensors that I stand here and talk about. We're going to see some of those products later this year. This is another chart which some of you are familiar with. It is summarizing the effect of the last quarter's licensing and looking at it over a longer timeframe and seeing how the picture in terms of cumulative licenses is building up. We did continue to sell in Q4 licenses across this broad range of products. Good to see mobile, good to see TVs, set-top boxes, and good to see a subscription license.

The chart on the right-hand side of the slide shows the effect of our licenses on the royalties that we're earning and reminds people of the loyalty of the Arm business model and the fact that royalty continues to be generated for many years. It also highlights the fact that it takes a long time for licenses to contribute significant portions of royalty. We have now entered another five-year period. We have now got sort of four levels of licensing when we split it out according to time. You can see that the 120 licenses that are sold during 2011, of course, not yielding any royalty yet. In fact, the 350 licenses sold over the previous five years, now they are starting to ramp into volume, but you can see this tremendous potential for growth to come from those licenses.

If you look at the progress for the whole year, as well as adding 121 licenses into the cumulative pool of future royalty-generating licenses, the number of partners who are engaged has also grown during the year from 256 this time last year to 290 now. Good news is that we're exposed to more companies, more companies buying into the Arm technology. Now, switching to royalties, again, a familiar chart here. The pie chart on the bottom left is showing the split of where all the volume is going. Mobile now accounting for just, well, 55%, just over half the overall volume. The pie that's growing really significantly at the moment is the embedded pile. This is part of the pie. This is microcontrollers, primarily, microcontrollers and smart cards.

In spite of huge volume there, over 500 million units during the quarter, that is still growing at a very, very significant rate, up 80% yearyear. We have seen very strong growth in this sector for the last several years. We now have over 130 Cortex M licenses out there ready to generate future microcontroller royalty. You are going to continue to see this number grow. As we look at what our microcontroller partners are doing, we're seeing every quarter launches of new Arm-based microcontrollers. These partners bring their products to market, and it isn't normally just one or two. It's line cards of 50 microcontrollers at a time from Arm partners. Good news happening there. One of the other things to bring out on this slide is the value of the mobile royalties.

Obviously, we talk in fora like this and in investor one-on-one meetings, and we talk about the high value of application processors, the higher levels of royalty which the Cortex A processors attract, adding extra Arm IP onto those application processors. People say, where is that going to come through in terms of incremental royalty? It is coming through in terms of incremental royalty. We can see volumes up 10%, value up 20% in the Cortex A and Mali area, and we're bringing that out on the slide. Apart from that, the other point on this slide is continued outperformance in each market sector, outperformance versus the market as a whole.

When we look at that over a longer period of time, that's what the chart on the right-hand side of the slide is showing us, that even as the industry cycles happen over multiple years, Arm royalty growth continues to outperform industry growth. That's a trend which has been consistent. You can see in the five-year CAGR at the bottom of the chart there, it's been consistent. We expect by the end of the licensing and the design and activity that's happening, that trend is going to continue to happen as we look out into the future. Where did all those products go? They didn't just go into smartphones. I always ask Ian, can we have a slide that shows some pictures that aren't just gadgets and smartphones? Here, this is the broadest range of products that we've ever shown you, I think, on these slides.

You can look at the pictures and see where the Arm technology is going. The other point to just highlight on this slide, in the bottom right-hand corner, you'll see we've got miniature versions of slides that you will find in the back of your pack in the appendix. We have these slides should be familiar to many investors. They summarize how Arm's market share is earned across the different product categories. For the last few years, we have been in a position where after 12 months, the analysts like Gartner tend to go back and count a few extra microprocessors that the industry shipped, and so the total for the previous year sort of increases. What we found this year is that the 2010 numbers, which last year we had a total of 22 billion units shipped in 2010, that's what we told you at the beginning of 2011.

Now at the beginning of 2012, when we turn around and ask those analysts, they tell us it was more. The size of the market, backward looking, increases. Arm's volume, of course, is exactly the same as it was. We told you the correct numbers in the first place. The net effect is that our apparent market share, when we look back, shrinks back. Instead of having 28% share in 2010, we think now we had about a 25% share in 2010. Notwithstanding that, the numbers for 2011, the best guess numbers for 2011 at the moment are shown in the appendix in the 2011 slide. We had a 30% share in 2011. If you look back over several years, those market shares have consistently grown from mid-teens to early 20s to mid-20s last year to 30% in 2011. That's a little bit about royalty.

Continuing with the theme of where the Arm technology goes and the fact that I do get a bit frustrated when people think that we supply the technology into gadgets and gizmos. The good news is that all those gadgets, all those smartphones, all those mobile computers, they do use incredible quantities of Arm technology. We are getting our technology used in products that make the world a better place as well. That's what it says in the Arm vision. We want our technology to be invisibly making the world a better place. It is. We're seeing Arm Holdings deployed, microcontrollers deployed in healthcare. We're seeing Arm Holdings in education. The theme at the bottom of the slide is about energy. As you know, energy is very important to Arm. Power efficiency is very important to Arm.

When you blow that up and look on a larger scale, the rate at which we all use energy is important to Arm. The fact that we all are getting more concerned about how well we use energy is a great driver for Arm's business. We're seeing Arm Holdings used increasingly in electric motors. We're seeing Arm Holdings used in smart meters. You will see Arm increasingly used in servers where energy efficiency is really very important indeed. Switching to physical IP, other parts of our business, physical IP. The news here is that over the last two years, state of the art has really moved from 40 nm-20 nm. That's actually quicker than Moore's law. The good news is that Arm's activity at 20nm is absolutely on track. We are engaged with all the major foundries at 20nm.

We are continuing to grow the number of platforms out in the field licensed on older technologies. Some of those older technologies are now starting to generate royalty. We received in the last quarter our first 32nm royalty, and that was booked last quarter. The first shipments were in calendar Q3. As we look down the right-hand side of the slide there, you see not only are we involved at 20nm, but leading edge to us is 14nm. Right now, we're engaged with the leading edge foundries on creating physical IP for 14nm technology, which will obviously be several years in the works before it actually starts shipping in any appreciable volume. It's good to see Arm physical IP there.

It's good to see that because when we deliver our physical IP to semiconductor partners in the form of things like processor optimization packages, it means that our processors can be implemented on the advanced silicon geometries. The chart on the bottom right there shows the consistency of licensing of processor optimization packages. We continued to do that in the last quarter of the year with further processor optimization packages. The summary really is it's an exceptionally good quarter to an excellent year, very strong licensing and continued growth in the backlog. I'm not going to steal Tim's thunder on the numbers. He can cover that in a moment. I think on the royalty side of the business, we're continuing to see Arm gain in market share as these new Arm-based chips come to market.

In terms of additional Arm IP, physical IP, and Mali graphics now generating royalty, we said 2011 was going to be the year of tens of millions of units of Mali graphics. It was indeed the year of tens of millions of units of Mali graphics. We're going to see that trend continue in 2012, particularly with the digital TVs. All the while, we continue to invest in new technology so we can bring things like version eight to market. Over the last year, we continued hiring. We hired an additional nearly 230 people last year. It's a balancing act because we're doing that at the same time as increasing margins and driving profitability to record levels as well. A little bit of an advert on the bottom of the slide for Mobile World Congress coming up at the end of the month. I will be there, as will Lawrence Bryant, our mobile specialist. We will be hosting a session for investors at Mobile World Congress in a couple of weeks' time. With that, I will hand over to Tim.

Tim Score (CFO)

Thanks, East.

Good morning, everyone. Warren is right. He hasn't stolen all of my thunder, but the good news is he's stolen most of it. I only have three slides. I think there's a lot of detail on the financials in the release. There's a lot of detail in the slide deck. What I'm going to do is just try and share with you a little bit of color about how we should think of the numbers from here based on what we've read and talked about this morning. As I think you've all seen, a very strong licensing quarter, $68 million, and a backlog up 20% sequentially. What that, of course, means is that both from a revenue-bearing standpoint on licensing and from a backlog building, very strong and leads us with a healthy pipeline into 2012. The backlog is up over 2x what it was this time two years ago.

That's good news because the license expectations for 2012 are running about twice the level that they were going into 2010. In a sense, the relationship of backlog to target license revenue is actually consistent. From here, I would say that $68 million is a fairly exceptional quarter. That doesn't mean to say it's not repeatable in the medium term, but I think you should view it today as fairly exceptional. The guidance that we're giving is that $60 million ± per quarter is what we should expect going through 2012. Expect us to be over that in some quarters and potentially under it in others. I think that's around about the level that you should think about when thinking about licenses going forward.

On the royalty front, clearly a much stronger outcome than I think we were all thinking about when we stood here or were on the phone actually at the end of October. Certainly, the first two months' data for the third quarter, which is the context, of course, for our Q4 royalties, was fairly flat. There was a pretty appreciable uptick in September, which helped industry numbers. The industry was nowhere near up to the extent of 19%. I think you're seeing their market share gains, which consistent. I think we talked last time about Arm being impacted by Japan and a lot of that recovery to closer to fuller production. I think we've also seen about a quarter of the royalty reports for Q4 shipments, which are, of course, our Q1 royalties. I think there are some signs of some fairly significant orders in the third quarter that I wouldn't expect to be repeated at the same level in the first quarter. With the 10%, it looks like.

Quarter, I would take somewhere plus or minus $90 million as the context for thinking about Arm Q1 royalty. Broadly consistent with what happened in the overall industry, yes, we will grow market share. As I say, in the detail, we have seen some signs of some big orders in Q3, which look as though they won't be repeated in Q4. Round about $90 million is what we're expecting. Margin expansion has 5% points in the full year, up from 40%-45%. We'll come onto that a little bit on the next slide.

For the full year. For the full year, moving away from Q4. Year growth up 24%. Now we've had Arm performing very resiliently in 2009 in the downturn, outgrowing the industry in the bounce-back year of 2010, and outgrowing the industry again in a steadier state year of 2011. That is sort of consistent performance through the cycles. As Warren East said, a good year for licensing. We used to talk about 75-100 processor licenses a year as the run rate. We did 121 last year. I think the key for that, of course, is it's the underpin, it's the installed base for the royalty generation going forward. Behind all this, of course, is investment in innovative and relevant technology for the markets that we are targeting.

You can see that in the last two years, we've been growing our headcount at around about a 10% growth, up from just over 1,700 at the beginning of 2009 to just over 2,100 at the end of last year. Our expectation going into 2012 is that we will continue to invest in our R&D capability. Most of our hiring is in the Research and Development area. You can see it there, 80%. Expect further investment through 2012. The margin expansion that I talked about, 40% going to 45%, driving strong cash generation and consistent increase in the dividend, which again has grown through the cycles and up another 20% this time. Going into 2012, as I say, we're well positioned. The backlog is twice as high as it was a couple of years ago.

Warren East touched on some of the new products that are going to be available for licensing and the new markets that we are developing technology to access. Our traditional target markets, we're continuing to drive up our market share. Warren East showed we went from 25%-30%. Ian Thornton's restatement of the overall market size has left us with more room for growth than we thought we had under the old population. That is good news too. Q1 specifically, as I say, thinking $60 million plus or minus for licensing, $90 million plus or minus for royalty, including the other revenue streams, that's consistent with about $200 million for Q1, which is where the market currently is. Our OpEx guidance, mid-60s.

I think when we look at the models for 2012, we obviously need to take into account the full-year effect of the hiring we've done in 2011. We need to take into account the wage inflation that kicked in on the 1st of January. Obviously, we need to take into account the 2012 impact of the hires that we intend to make through this year. If the overall market holds up and we do invest in the way that I'm outlining, you would expect an ascending trajectory for our OpEx through the year from this base. Looking for the full year, it's obviously very early days. Usually, at the end of January, you will hear us be fairly circumspect about how precise the crystal ball can be looking at the balance of the year.

We're on the assumption that the macroeconomic outlook stays about where it is now, i.e., some growth in the industry, not overly exciting, but some growth in the industry and continued outperformance from Arm Holdings. We would expect to be, as it says there, at least in line with the current market expectation, which is just over $860 million. With that, I think we'll throw it open to the floor.

Warren East (CEO)

Fine.

Can we start at the front here?

Sure.

Okay, you've got it.

Yeah. I won't like to go with the people during machines. The first question would be for Tim Score and the guidance for Q1 specifically. Looking at $90 million there for royalties in Q1, as you just said, if we start from $100 million in Q4 and semiconductor revenues went down to say something like 10% Q4 versus Q3, it looks like it's very conservative. You're not taking any market share. There is no positive impact from the new iPhone 4S, etc. Is your guidance for Q1 overly conservative? That's the first question. The second question would be for Warren on the milestones for Windows on Arm. What's the schedule basically for this year? When shall we expect to see the first tablets? When shall we expect to see the first laptops?

Have you seen any excitement on the software side, basically, to develop or to port more software to Windows on Arm in terms of the software ecosystem outside of Office and Windows?

Tim Score (CFO)

Thank you, François. I mean, obviously, I can understand where that question comes from because we've just beaten expectations for royalties in Q4 by about $10 million. My answer to the question is precisely what I set up there, which is the industry looks as though it's down about 10%. That would obviously imply about $90 million without any market share gain. We expect to continue to make market share gain. Having seen about a quarter of our reports and looking at the detail of that, there are signs that there were, if you like, some overshooting in the third quarter shipments. The trajectory of Q3 and Q4 in the detail is not necessarily completely in line with what you would expect in the overall. Now, it's very early days.

It's very early days. We've only seen about a quarter, and we may well do more than $90 million. What we're guiding to you today based on what we see in the industry and what we've so far seen in a small amount of the detail is that it feels as though it's somewhere in that area.

Thanks.

The second question was about milestones for Windows on Arm. I'm afraid we still have to really refer you to Microsoft for their detailed answers on this. It is their product. They are in control of the rollout of their product. It's fair to say that the project appears to be on track. They talked about beta release being available in Q1. I think it's going to be available certainly well before the end of Q1. As for Windows on Arm in particular and tablets and laptops, I don't think from where we sit, we don't see any difference between these two form factors. The laptop form factor or the tablet form factor, we're going to see both during this year. When the beta is released, there will be, we're led to believe, there will be beta products.

It's really up to the OEMs as to which form factors they emphasize, whether they be tablets or whether they be laptops. As far as the software porting is concerned, it's still a little early. The whole point of the beta program is to start accelerating that application development. The application environment was demonstrated by Microsoft at CES. I'm sure when they bring out their beta release, they'll do a more active promotion of that application environment.

Thank you.

All right, we're on to the second question.

Didier Scemama (Analyst)

Hi, it's Didier from RBS. A couple of questions. First of all, on your revenue growth, not guidance, but sort of your revenue growth expectation for 2012. When I look at capital spending by foundries, it feels as though the wafers out exposed to smartphone and tablets are going to be up well in excess of 50%, typically because the die size of application processors is increasing dramatically. My first question is, do you think this year we could see a firmer decorrelation of your revenues in smartphones from the underlying market because of ASP increases? My second question is really regarding competition. Obviously, there's been a major milestone from Intel recently with some major design wins. It remains obviously to be seen whether those phones are going to sell.

How do you see Intel competition in particular, the comments they made about potential workaround on supporting Android applications on x86? Many thanks.

Warren East (CEO)

Okay. First question on the ASP. I think you're already seeing, as I'm trying to pull out on the slide, we're already seeing some effect of pricing and higher value chips in the mobile space. It seems to be a fairly well-recognized phenomenon that these chips are getting larger. I mentioned we'll see dual cores, we'll see quad cores. We're seeing embedded graphics and that does blow up the size. No doubt about that. It pushes up the underlying cost. However, it continues to be a very competitive marketplace. It's really not for Arm Holdings. It's for those Arm Holdings licensees who are competing in that space to determine the pricing that they're going to deploy for those apps processors. Undoubtedly, the chip areas are going to be bigger. You're absolutely correct. We can't really comment on the pricing. Intel phones and Android.

We've been talking for some time about Intel being able to announce some smartphone design wins. They've announced a couple at CES. I'm sure they're going to announce a couple more at Mobile World Congress, probably. That seems to be a pretty good forum to announce them. Our answer hasn't really changed. We have a lot of respect for Intel's technology, which is excellent, Intel's investment capability, which is very strong. They're making great progress to take their devices and make them more power efficient and get them into the sort of zone which makes them suitable for smartphone designs. They'll inevitably get some. One of the factors that one of the hurdles that they have to overcome is the fact that every day there's 700,000 Android phones activated around the Arm architecture. The application developers are working on creating applications that run on Arm.

Intel are going to have some products that are in the zone. They're going to have to compete with the 20 or so Arm licensees who are very actively supplying apps processors into that zone and into the ecosystem, which is Arm-flavored. This is really a question for Intel as to how well they think they're going to be able to overcome those hurdles.

Thanks. It's Jan Arden from Liberum. Two questions. One is, what is your licensing backlog and your licensing outlook into the current year and for the next couple of years? You've doubled your backlog over the last two years. Is there anything which you see on the horizon which prevents you from, say, doubling it in the next two years? Why would the backlog not continue to rise? Do you see less potential in terms of your negotiations to keep rising that and therefore for your licensing revenues to keep rising over the medium term, even if there are one or two quarters here and there? The second question is back to Intel again, to be honest. They have achieved quite a big breakthrough in terms of power consumption and with their latest Medfield processor. Power consumption has been the key platform for Arm over the years.

From an R&D perspective, what can you do or what are you doing to try and ensure that you stay well ahead of the game going forward so that you can try and keep up a significant gap with them on power consumption terms?

Tim Score (CFO)

In relation to the first question, let's remind ourselves of the context of Arm's licensing. We have said consistently over the last five, six, seven years that we see license revenue as a sort of mid to high single-digit growth stream with royalties growing at a faster rate over time. Therefore, that operating leverage drives increasing margins and increasing earnings. We've seen obviously a much stronger two-year period of growth for licensing, 30% and 40% respectively. As that's been happening, the main questions we've been getting are, that's so far above your mid to high single-digit guidance, is there a lot of, is this a one-off peak or is there something brought forward, is this pent-up demand out of the recession? The answer to all those things is no.

This is a function of Arm's addressable market broadening away from mobile phones further up and further down, and that introduces lots of new technology and lots of new customers able to use Arm Holdings for the first time. What we're saying going forward is still think about licensing as a mid to high single-digit growth stream, but on top of a much higher platform that has been created in 2010 and 2011. Exactly how the backlog develops in that context, I think we'll have to wait and see. Backlog is obviously quite a lumpy concept. In general, we would see it growing consistent with that type of guidance. That's kind of how we see it.

Warren East (CEO)

Your second question was about Intel and their progress in smartphones and what Arm is doing about that. The answer is that, of course, as Intel are making great progress here, Arm is not standing still. One of the key drivers of Intel's progress is their ability to ramp new semiconductor processors fast and to be able to move to later generations of semiconductor technology. We see Arm working, as I was describing, with all the leading foundries at leading edge. In fact, even as Intel are bringing their 32-nanometer Medfield product to market, we're seeing Arm's 32 nm products starting to ship. We're seeing Arm 28 nm products starting to ship. We're seeing Arm engaged at 20nm, and we're seeing Arm engaged at 14nm. In addition to that, there's another whole dimension to this, which is delivering real solutions in a phone environment.

That's what the Arm partnership has been doing for the last 15 years or so. That's what Intel is getting to grips with right now. Dan's building on that lead, taking that system expertise that we've built up over the last 15 years or so, and turning it into things like Big.LITTLE, which we announced back in October, where we have the benefits of high performance with a big processor married in a seamless way to the power efficiency of a small processor. That sort of system-level solution is moving the goalposts significantly. You end up with something that has the performance of a Cortex A8 smartphone of two or three years ago, delivered in a power envelope that uses about 20% of the power. We're continually moving those competitive goalposts. I don't know. You probably saw who got that first. There's so many.

Kai Niklas Walter (Senior Analyst)

Yeah, thanks. It's Kai Niklas Walter in Deutsche Bank. The first one was just on Windows on Arm tablets. Just wondering, from your perspective, what do you think will drive consumer interest in those sort of devices, given the probably, at the end of the day, disappointing Android tablet sales, which from an OS and user experience perspective probably weren't too shabby? That's my first question. Why would that change with Windows on Arm from Q4 onwards versus Apple? My other one is just on share buybacks. Clearly, dividends are the main mechanism to distribute cash to shareholders, but you keep piling on cash in the balance sheet. Is there any level from an absolute or relative perspective in cash terms which would trigger you potentially thinking about a buyback or anything similar? Thank you.

Warren East (CEO)

Okay. The first one on Windows on Arm and Android. A few factors to consider, I think. Android tablets shipping in 2011 has been described by some as you described it. However, I'd say actually when Android phones were introduced, they took off. There was a lot of hype. Actually, they didn't take off in the sort of way that reflected that hype. A few years later, two years later, we're seeing Android phones shipping at half a million units a day, 700,000 units a day, and continuing to ramp and really be a very successful product. I think we should give the Android tablets a little bit more time. The second point to note is, of course, that consumers are very familiar with Microsoft and very familiar with Windows, and they are less familiar with an Android environment.

I think Microsoft have an awareness advantage with consumers that the Android folks didn't have. Now, it's up to Microsoft exactly how well they're going to exploit that advantage, but I think that's a fundamental difference.

Tim Score (CFO)

Yeah, I mean, generally on the sort of management of the balance sheet and the capital structure, we've got GBP 424 million of net cash at the end of 2011. We think that is a reasonable amount of net cash to hold in the context of our investment opportunity. We, I think, demonstrated between 2005 and 2008 when we bought back 16% of the shares of the company that buyback is a mechanism for returning cash that we are quite happy to deploy if we think the time is right. We don't plan to build a cash pile. We do plan to grow our dividend. I think in the medium and longer term, we plan to grow our dividend progressively, by which we mean expect a gradually increasing payout ratio.

Depending on our investment needs and opportunities as that develops, we'll define whether we need to deploy some other supplementary mechanism to avoid building a cash pile. Buyback could well be one of those mechanisms.

Andrew Dunn (Senior Portfolio Manager and Investment Advisor)

Thank you. It's Andrew Dunn from RBC. You've mentioned Mali today, I think, more times than I've heard you mention Mali for a long time. Can you just talk about, and you're obviously gaining market share there against your major competitor, can you just talk about what's driving that, how you're able to gain that market share? You've also talked a lot about digital TV today. How should we think about royalty rates in digital TV? Obviously, the chip price is probably higher than you'd see in handsets. How should we think about royalty rate there? If I could finally ask a question, you've signed a license with Star announced just over the weekend. MStar obviously is a MIPS licensee. I understand that will sit alongside the MIPS processor. The Mali will sit alongside MIPS. Is there an opportunity there to get the full Arm CPU into MIPS into MStar? Thank you.

Warren East (CEO)

Yeah. With the way you introduced the question, I think, oh yeah, that's an interesting observation. Perhaps we did mention Mali quite a lot. I don't think you should read anything into the fact that we're mentioning Mali a lot other than we're at a slightly different stage now in terms of Mali penetration than we were at 12 months ago. 12 months ago, we were still quite early. We are coming from behind. We're undoubtedly the number two player in outsourced graphics IP, and it's sort of almost inevitable that if somebody has a very, very high share and you have a number two player that's coming after them with a competitive solution, that number two player is going to gain some market share. We have sold more Mali licenses during 2011. The total is now nearly 60 Mali licenses.

Inevitably, over time, these Mali licenses, it's exactly the same as our standard microprocessors. It takes about four years to go from licensing to royalty. That licensing has been happening. It's not surprising that the chips will start to ship and the fact that we're continuing to sell more licenses. If you look at the licensing trajectory over the last few years, that's also been growing. The number of licenses per annum has been growing. You're going to see the volumes continue to ramp. I highlighted digital TVs because that is an area where we're particularly pleased to see Mali's success. We've got two factors. We've got digital TVs adopting graphics processing, and the proportion of digital TVs that includes a graphics processor is increasing.

At the same time, the licensing that we've done suggests to us that we're well positioned to have the majority of the market share there, probably as much as 70%. That's the reason we're mentioning it because it's a significant milestone in the progress of Mali. As far as MStar is concerned, yes, they had an announcement at the beginning of the week or at the weekend for the U.K. anyway. We'd love to see Arm processors deployed more widely in MStar. Obviously, we're in the business of selling microprocessors. They're in the business of using them. I think you'd expect us to be talking. Should we work sort of from the front back?

Sam Hudson (Financial Analyst)

Yeah, hi, it's Sam from Redburn Atlantic. Just a very quick question on your physical IP division. You have been very kind to give out the breakdown on the profit side. Given that the catch-up phase of the physical IP division is now somewhat over, but we still see somewhat of a profit loss in this particular division in the last half as well, when do you see this particular division become profit-making or is it still going to remain somewhat of a drag to the sort of the higher, bigger, brother processor division valuation? Secondly, what % of licenses in 2011 were on the non-mobile side and where do you see the non-mobile royalty value to grow in the medium term?

Tim Score (CFO)

Okay, June the 13th. Yeah. I think on the physical IP division, if you sort of track the segmental analysis that we do on a twice-a-year basis, you'll see that the fully allocated loss or the loss after all of the sort of central Arm costs have been allocated to each of the divisions is reducing. I think having established the leading position in technology-wise in the advanced processors, the physical IP division is well positioned to grow. Having said that, the division is going to continue to invest in securing the leadership position in the advanced processors. I think also, when you think about the physical IP division, don't just think about what the division or P&L looks like.

Think also about the value the physical IP is bringing to the processor division, which is significant because these processor optimization packs and the like are making our processors much more attractive. You have to look at, you called it, I think, a drag or a lag or something. We absolutely do not view it like that. We view physical IP as a very significant enabler and enhancer of the processor division, as well as a steadily growing revenue stream in its own right.

Warren East (CEO)

A question about licensing. I think in the roadshow slides, we'll have a pie chart. That's a trend which has been sort of in, or certainly two-thirds thereabout has been in place for the last two and a half, three years. Bearing in mind the time lag that I mentioned between licensing and royalty, we would expect that licensing to come through. Indeed, we are seeing non-mobile continue to grow and slightly outpace mobile in terms of growth, which is why this year, we're now at sort of 45% of the total volume being non-mobile. I think that trend is going to continue. Meanwhile, don't forget mobile is a huge potential as well.

As we look at the market for 2016, and if you look at your appendix, you'll see that Ian's broken it out for 2016, 40 billion units, somewhere just under three, just over rather three-quarters of those would be in what we would call non-mobile. Obviously, we have a very high share in the mobile area, in the 25% area, and we have a much lower share in the much bigger market, which is about three-quarters. We have been doing that licensing. We're going to continue to see an upward trend in non-mobile.

Sam Hudson (Financial Analyst)

In value, though?

Warren East (CEO)

In value, I think in the mobile space, you have high-value devices, apps, processors. You have lower-priced chips, modems, connectivity devices, and so on. Similarly, in non-mobile, you have vast volumes of lower-priced chips and microcontrollers and so on. Similarly, you have high-value devices in networking and servers and in some of the higher-end consumer products as well. I think the non-mobile-mobile split won't really have much of an effect on the average chip price. Let's keep working backward. Oh, sorry.

Nick James (Equity Research Analyst)

No, that's fine. Morning, it's Nick James from NIMUS. Just a couple of follow-ups on licensing. Just in terms of what you described as abnormal strength in Q4, I wonder if you could give us any more details that centered around specific customers or any more granularity there. The other thing was on the subscription license with LSI for kind of higher-end wireless infrastructure or network infrastructure type applications. That would seem to be a new segment for Arm Holdings. If there's scope for further types of licensing in that area over the following year.

Tim Score (CFO)

I think on licensing, it's just a question of if you look at the detail, if you look at the aggregate of revenue-bearing deals that were signed in the quarter, and we project forward as to the likely run rate, it looks higher. It's really just a function of the specific deals that happen to be signed in that quarter. Abnormal in the context of I'm here telling you you shouldn't think of that as the base. It's not abnormal in the sense of it's unrepeatable.

With regard to the networking infrastructure and mobile networking infrastructure, the LSI agreement is a very significant one for us. We're very pleased to see that, as I said. It is one of several pieces in the jigsaw for Arm Holdings to gain share in that space. That space is dominated by different architectures, particularly PowerPC is strong.

A couple of years ago, we had some activity with Xilinx, and they've been deploying PowerPC. Those sorts of devices get deployed in those sorts of products. They've moved to have some Arm-based products as well. LSI moving to have some Arm-based products. You'll hopefully see some others. It's a multi-year journey, but we do expect to gain share in that space. Ian's telling us we must move on. We'll try and pick up the pace.

Sandeep Deshpande (Stock Analyst)

I see there's a question from JPMorgan Chase, guys. Just a couple of quick questions. Firstly, on the Arm processor itself, with Windows having been seen in the mobile market here in Europe, to not great success at the moment, at least, what kind of applications are we going to be running on Arm, on Windows, when it comes later this year or early next year? Are you working with developers in developing Arm-based current Intel applications onto Arm, or is Arm only going to appear in the Metro version of Windows? Secondly, on the Mali graphics processor, another follow-on question. You've clearly done quite well in the digital TV part of the graphics market, but actually possibly have lost some share within the mobile part of the graphics market. What do you see the dynamics there in the mobile part of the graphics market?Finally, actually, I'll leave it at that.

Warren East (CEO)

Okay. Quick answers then. Windows, I referred to Steve Ballmer and his keynote at CES. When asked, you know, what's the future, he said, "Metro, Metro, Metro." I think Microsoft are launching their product, and you'll have to talk to them about which applications are going to appear and what the way to get those applications is going to be. They were very, very positive about Metro when they were doing their keynotes. I don't think we've lost share in mobile in graphics because, as I pointed out, we're coming from behind and we don't have share to lose in mobile. What are the dynamics there? It's going to be the same as the dynamics across the market. Arm technology is not more or less applicable in mobile than in digital TVs, but we are number two in that market space.

Hi, it's Gunnar from Citi. Just on the average royalty rates, I know that you're not guiding for this, but last couple of quarters were stabilizing. We had a sequential pickup this quarter. How should we think conceptually about royalty rates going into 2012? Secondly, on development systems, you're rebuilding your system, you are guiding to flat revenues. How much visibility do you have along this process? Thank you.

Tim Score (CFO)

I mean, you're right. We don't guide on average royalty rates because clearly it is a mathematical output of blended growth rates across a variety of sectors. I think it's worth noting that if you look back over the last two years, eight quarters, it's been between sort of 4.4 and 4.8. In some quarters it's notched up and in some quarters it's notched down. The current level is not a bad way to think about it, but ultimately, it will depend on the relative growth rates of the top-end processors and the microcontrollers. As we've said many times before, the best outcome for our total royalty revenue and our margins and our earnings may well be a lower average royalty rate because faster penetration into microcontrollers. Certainly, looking into 2012, four point something looks sensible.

Warren East (CEO)

With regard to our development systems business, we are doing a bit of restructuring there. You are seeing the world changing in that space for non-microcontroller applications, increasing use of open-source tools. Arm is not open-source tools. Our tools are higher-value tools. Undoubtedly, some of the market where we would normally have expected to sell tools, people are using free open-source tools. We need to adjust our product line to cope with that. At the other end of the spectrum, we have growth in microcontrollers. Microcontroller tools are much lower priced, but obviously, we're selling into a much higher volume market. We are seeing growth in terms of number of products sold, but they're at a lower value. That is why you see this flattening of the revenue in that business.

I pointed at the 130-odd Cortex M licenses and the vast growth that we are seeing in microcontrollers and that whole space around Arm growing significantly. We are playing into a much larger market.

Tim Score (CFO)

I think we should have two more questions, unfortunately.

Gareth Jenkins (Head of Research)

Thanks. Gareth Jenkins from UBS. I'll take up the last two then. No, I'm just checking. Two very quick ones if I could. One for Tim. Just on the gross margins, given what you've just said on development systems, is 96% the kind of new baseline that we should assume going forward through this year? Then just secondly, Warren on Big.LITTLE, I just wondered whether you can give us some initial sort of industry feedback and the kind of timetable for implementation. Thank you.

Tim Score (CFO)

I'm not too good on baselines because they sort of hostage us to it. If you look at Q4 and you look at what we've told you is some very, very strong licensing, very strong royalty yielding a 96% gross margin. I think if, given the guidance I'm saying of 60± on licensing, if the licensing is around the 60 area and the royalties are lower because of what the market's done in Q4, then all other things being equal, the gross margin will be a bit lower. The main point is the gradual increase over multiple periods that we've seen in gross margin, which is essentially a reflection of royalty revenue increasing as a proportion of total, is going to continue over time. Are there going to be quarters when it's lower than 96%? Yes. It's always going to be in the mid-90s on a medium-term rising trajectory.

Warren East (CEO)

With regard to Big.LITTLE, it is a little early to see anything coming out there. We have sold licenses for Cortex, what's become Cortex-A7, which is the product that we announced, the little core that we announced at the time. I think you should probably ask that question again in July when we're back at the half year, because at that stage we'll be a little bit, or we'll have better visibility of our partners' plans to bring Big.LITTLE products to market. However, it's been very enthusiastically sort of received by semiconductor partners. Yeah.

Tim Score (CFO)

Yes, I think we got more.

Andrew Gardiner (Analyst)

Thanks very much. Andrew Gardiner from Barclays Capital. Just on the licensing side, Warren, you said that the number of partners that you're now working with are something like 290, I think up from 220 something the year ago. A fairly significant, almost 30% increase year-on-year.

Warren East (CEO)

256, actually.

Andrew Gardiner (Analyst)

In terms of the increase then year-on-year, how much of that is from either, say, sort of new startup semiconductor companies versus more established ones who are perhaps changing their architecture choices?

Warren East (CEO)

It is almost exclusively new ones, and it's almost exclusively in the microcontroller area. What we're seeing is Arm gaining share in microcontrollers where hitherto people would use an 8-bit architecture, and they're now adopting a 32-bit architecture, and Arm is the architecture of choice.

Andrew Gardiner (Analyst)

Just given the structure of that market, do you see similar opportunities in 2012, 2013 for further?

Warren East (CEO)

Yeah, that is going to continue. There are many companies there, and that's why we're continuing with our microcontroller product lines, introducing things like the Flycatcher product, because we can see scope for more product licensing into that marketplace. I don't know if anybody has any other questions. We're probably around for just a few minutes, but at this point, we ought to thank you very much for your questions and say thank you for coming, and we'll see you in July.