Good morning, everybody. Olympic week, it's a fairly select bunch. Thank you all very much for coming along, braving the London transport conditions. They were fine at 6:00 A.M. this morning, by the way. As usual, I will step through a business update, and Tim will follow up with some detail on the numbers. It's been a really good quarter. Not one that I would say, "Here's the standout one thing that we talk about." Actually, if you look right across our business, whether you cut it up by application sector, whether you cut it up by product sector, then it's been an excellent quarter with some good news. If we're to highlight some things at all, then on our highlight slides, it has been a continuation of big semiconductor companies continuing to invest in our technology. Not just sort of tier 2-type semiconductors.
We've had some very serious semiconductor companies, who you'd call market leaders, who have incumbent architectures that they've been with for many years, adopting Arm in their next generation of their products, which is very good. We've signed another ARMv8 architecture license, that particular one will be announced in due course, as and when they want to announce it. What we can say is that it's in networking applications. If there is a little bit of a theme, particularly behind my excitement in these results, then it is the fact that we're starting to see quite a lot more Arm activity in the networking space. More on that as we go.
If you look at it from a volume shipments and a royalty point of view, again, we'll come on some detail on that, across all our target markets, we're continuing to outperform the industry at large, and I'll come back to the numbers on that. Just earlier this week, we had an announcement with TSMC about pushing forward on semiconductor process technology and involvement there in terms of optimizing our Physical IP, optimizing their semiconductor process, and working with our leading edge new 64-bit product. A bit more on that later as well. Net net, very good quarter, which helped us generate significant increases in levels of profitability. We're maintaining our dividend going forward. That's the sort of summary picture. Now, if we sort of cut up and start looking at the business.
We'll look at the business to start with from the outside in, so look at it by market sector to start with. Obviously, mobile and computing remains very important for Arm, and it's nice to see a flagship mobile product like the Galaxy S III, which was launched during the quarter, containing so much Arm technology. In the apps processor, we've got Arm microprocessors. We've got Arm graphics. We've got Arm Physical IP, I played with one of these products for the first time only yesterday, actually. It really does perform. It's an outstanding performer. I can see why everybody is very excited about the Galaxy S III. For us, of course, having all that Arm technology in it is good. By the way, there's also Arm in the touch screen. That's encouraging to see our processors getting into those more sophisticated touch screens.
One thing we are seeing is the value coming through in mobile generally. The increasing number of smartphones, and within the smartphones themselves, an increasing number of Cortex-A products. You can see from the little histogram halfway down the slide, the top bar there is the ARM11. ARM11 is still accounting for 40%, roughly, of the apps processors. The Cortex-A is accounting for roughly 60% of the apps processors. Within that Cortex-A, you can see dual-core Cortex-A increasing significantly if you compare the situation with a year ago. That's good news from a value point of view for Arm's royalty, because typically, these chips are more expensive. Single-core moving to dual-core and quad-core is a good trend for us. Note also the underlying growth in sheer volume of apps processors in smartphones.
Don't forget, with all this gloom and doom around, smartphones continues to be an area of significant growth for the business, and we're looking forward to 30% there or thereabouts growth in smartphones year-over-year for the year as a whole. A computer is really like a smartphone in a different form factor. Obviously, the big news of 2012 in the computing front for Arm will be the arrival of the Windows operating system on Arm. This quarter saw the Computex show. We had an excellent Computex show. The Arm partners who were engaged in the Windows launch were showing off their products. They went down very well. We're looking forward to that launch which is going to happen in the fourth quarter of this year. That's yet to come, but we've seen the products and they're looking good.
Just while we're on PCs, of course, it was a quarter in which AMD made an interesting announcement about adopting Arm technology for use in their computer chips. This is an example of how Arm is actually sitting alongside an x86. It isn't always a question of either/or. This is Arm adding significant value to AMD's x86 product line as security becomes more of an issue in these products. Moving on to sort of down the chain now. These smartphones, computers, and everything, they communicate, and that communication means that they're getting data from somewhere or they're sending data somewhere, and they're sending it over some data handling infrastructure. The explosion in smartphones and more mobile computing and prevalence of the internet is generating much more data.
Some studies suggest as much as 20 times as much data over the sort of 10-year period from 2010 to 2020. Clearly, if that data is handled with the existing architecture, it's going to consume 20 times as much power, which is not a very sustainable situation. If you look at all the electricity generated in the world, then IT equipment accounts for about 10% of it. If that is going to increase by a factor of 20, then we're going to have to build a lot more power stations. That isn't going to happen. People are going to look for more power efficient ways of designing this stuff. Here is the opportunity for Arm in networking. You see, as I mentioned a moment ago, a new ARMv8 architecture licensee engaged in Arm networking.
Freescale, I was at their Freescale Technology Forum a few weeks ago. Freescale busy announcing their extensive networking product range, switching to adopt the Arm architecture. We've seen similar indications from HiSilicon, LSI, TI, Xilinx, and so on. Everybody is realizing that in order to get more power efficient products here, Arm is a great solution. It's the same power efficiency story which is behind Arm's activity in servers. We've talked about servers before. No great stunning news this quarter from Arm in servers, just continuation of the six-year journey that we're on, that we're now getting towards the end of this six-year journey from 2008 to when you can actually go out and buy commercially sensible Arm-based servers. It's actually been a good quarter in terms of data points on the way.
Calxeda, we had a demonstration of some of the Calxeda equipment at our analyst day here earlier in the quarter. You can see an example there in the histogram on the right of the slide showing the energy saving for a similar performance produced by an Arm-based Calxeda product. It's not Calxeda alone. You saw Dell announcing products based on Marvell. The hardware alone is, of course, not sufficient. One of the reasons why it's an eight-year journey is that the software ecosystem existing in the server world needs to become Arm flavored. We've also had some announcements this quarter about that software ecosystem inching its way towards the Arm world as well. With Canonical's announcement, that was very encouraging. Continued proof points then of Arm in servers.
Next sort of high volume wave that comes along after the wave of networking is going to be the Internet of Things. We've talked about the Internet of Things. Obviously, it's not our catchphrase. Certainly Arm technology is sitting behind a lot of the products that comprise the Internet of Things. We've seen continuation of our Cortex-M series licensing. We've now over 100 companies using that product. Collectively, if you look at the line cards from the Arm partners, there are over 1,400 different Arm microcontroller products that you can go out and buy from Arm partners today. That's going to be a much bigger number by the time all of that licensing that we've been doing gets into silicon production. Earlier this year, we launched the Cortex-M0 product.
I think probably I stood here in February and said we've got a new product coming down. One of them is going to be a small one, we're going to launch it later in the quarter. We did. That became the Cortex-M0, which is the most energy efficient processor for microcontrollers that is out there. Again, at the Freescale Technology Forum, we saw an excellent demonstration of that power efficiency, where they literally had an Arm-powered charger, crank it up with a crank handle, charge a few capacitors up and a range of different microcontrollers. Of course, the Cortex-M0+ went on and on and on. That's a great product. As far as the range of opportunities is concerned, it's huge, we're starting to get design-ins.
As we start to get design-ins, so more and more semiconductor companies are jumping onto the Arm-based microcontroller party, and they're making these decisions in order to position themselves for the Internet of Things wave. In terms of volume shipments at the moment, we saw another great quarter where if we look year-over-year, Arm microcontroller shipments up about 20%, compared with industry shipments up about 8%. Looking ahead to more leading-edge technologies.
As I said, we had an announcement earlier this week with TSMC. This is Arm and the biggest independent semiconductor foundry company in the world getting together to actually continue work that's been ongoing together for quite a long time in terms of optimizing their process technology, working with our Physical IP division to optimize our Physical IP on their new FinFET process, and using our new 64-bit processor as a vehicle for that development. It's world-leading companies getting together to work from transistors right through to microprocessors, to enable our joint partners to produce world-leading products. I'll look at the business and sort of cut it the other way. Rather than by end products outside in, let's look inside out, and let's look at processor licensing to start with. Solid quarter, up 15%.
Importantly, coming out of the quarter, our backlog up again, and now back at record levels. I think we've got some data on that a bit later on. We now have nearly 900 licenses. That continues to grow the pool of licenses that are out there to generate royalties for the future. If I look at just the quarter on its own, 23 licenses in total. Collection of Cortex-A licenses, including our 12th big.LITTLE licensee. So we've now got 12 partners signed up for big.LITTLE. At the other end of the scale, the microcontroller end, I was just talking about the Internet of Things. Yes, more licensing of our Cortex-M products. New architecture, the V8 architecture, the 64-bit stuff. We've now got nine V8 licensees, including the latest architecture licensee. We've got this rather ill-defined horizontal axis of time going along the slide here.
We are at the stage where we've done a lot of lead licensing now. We are approaching the first silicon, the product launch type phase. The 64-bit program is on track. The interesting thing about our 64-bit architecture, it is not just about high-end computing and servers. It's actually people talking about using it in mobile as well, talking about using it in infrastructure applications, some of the networking applications that I talked about a moment or two ago. Licensing is, of course, essential to drive royalties going forward, and here's an update of the chart that we've shown several times. You can see the licensing on the bottom left of the slide. A continuation of growing our base of licenses. Nearly 900 licenses out there.
Interesting thing to note is that if we look at the chart on the right-hand side and the bullet halfway down the slide, only 5% of Arm's royalty in the last quarter came from the 350 licenses sold in the last few years. Now 95% of Arm's current royalty is coming from legacy licensing. All that new licensing is royalty yet to be delivered. That's a good picture for the future. Now look at royalties actually and how they sort of panned out during the year. The story here is consistent gain in market share from a volume point of view. Up nearly 10% against an industry which is down, actually. That's a continuation of the gaining market share story. If I translate that into value, however, value up 14% versus industry value down 7%. It's a wider gap.
One of the things that's driving that increase in value is the increasing activity in the Cortex-A land, where increased adoption of Cortex-A in smartphones is helping us. If you look at the waterfall chart, then you can see, yes, penetration in microcontrollers. We've always said these are lower-priced devices. Significant growth in microcontrollers, you can see, is putting some downward pressure on the average amount that we earn per chip. Similarly, in the mobile world, some of the Bluetooth, some of the Wi-Fi baseband modems, these sorts of things, having a similar sort of effect. They're similar low-priced chips. If we look at the higher priced chips, the networking, it's a small contributor at the moment. Hopefully, with all that licensing, it's going to be a more significant contributor going forward.
If we look at the Cortex-As in consumer products, if we look at the Cortex-A in mobile apps, you can see significant upward pressure there on the royalty that Arm is earning per chip. Put all that together, that's why you get a 14% increase in value versus a 7% decline for the market when we look at Arm's royalty revenues. I should just highlight, we've put on the slide, of course, millions now of Mali devices as well, going into those Cortex-A based chips. As far as Mali is concerned, then we're very much on track for the 100 million plus units that we expect to deliver this year. If I look at our Physical IP, the story here is our Physical IP is being used right across the different sectors that Arm's processors are used in.
We're continuing with the processor optimization package activity. It was a record quarter for POPs, the best quarter we've had, total of over 32 POPs sold now. Still about a 50% attach rate with Cortex-A licensees. That's good in terms of generating royalty for the future. Also good in terms of generating royalty for the future is that this quarter we had four new fabless semiconductor companies adopting Arm Physical IP for their 28 nanometer designs and beyond. That is good for royalty growth going forwards. If we look at, to sort of summarize really, Arm is getting designed into a range of new products. The exciting thing about this is that when we look at the pictures, they're not all phones and tablets. We are seeing design-ins in the automotive space, in the networking space, the infrastructures, the base stations, and so on.
We are seeing new designs in the embedded world, driving the Internet of Things. Of course, on top of that, if you look at the top of the slide, we're seeing Arm getting designed into a great range of new, exciting products in the mobile and computing space, in that world of internet-connected screens. Right across the different market sectors, we're getting more design-ins, growing market share. We have strong licensing activity sat behind that, which is building a continuation of this picture for the future. That's all generating enough revenue and profit for us to continue to invest in our roadmap back in Arm.
We've grown our headcount by 6% since the start of the year, and we are going to continue to grow our headcount into the back part of the year, so that we can actually put the resource behind the Arm activity, where we're forming ever richer partnerships with leading-edge technology players to develop more leading-edge technology for the future. With that, I will hand over to Tim, who will put some color on the numbers. Thank you.
Thanks, Warren. Morning, everyone. Good news is we've only got four slides. Warren, I think, has given a pretty good overview coverage of the numbers in high level, and you've seen them in the detail. What I'll do is focus, as he says, a little bit of color and a couple of comments about the sort of forward looking as we think about our models in the second half of the year for the full year. Licensing, $67 million. You will recall, when we stood up Q3 2011 last year, and we'd just done a sort of a 58 and a 59, and we were sort of well up on the run rate before that, I gave guidance that said, looking out now, it looks like a sort of $60 million plus or minus, i.e., higher level, but still potentially lumpy around a base.
Since then, we've done a 67, a 65, and a 67. I think if I stand up here and say 60 plus or minus, you're probably going to yawn ever so slightly. What I'm going to say today is that I think looking at the level of the backlog and looking at the pipeline, I think to move that base up a little bit to think sort of $65 million plus or minus as we look forward, I think is the right way to think about it. The reason I'm saying that is not just because the opportunity pipeline for the second half specifically is, as we say, they're healthy, and the order backlog at the end of June is, as Warren says, at a record level, having gone up sort of small single digits in Q2 after a very slight fall in Q1.
It's also about the picture on the bottom right, which shows the relationship between how license revenue has developed since the first half of 2007 and how the backlog has developed since the first half of 2007. That's PD licensing and PIPD licensing combined. In a nutshell, what that's telling you is, in H1 2012, licensing is just under 1.5 times higher indexed than it was H1'07, whereas the backlog is broadly three and a half times higher. Obviously what happens over time, an Arm backlog, remember, is a contractual value of licenses signed. It's not some discretionary list of things that people may or may not choose to do in time. This is contract value, so this gets into license revenue in due course. What's happening, that gap there is obviously going to feed its way into license revenue over time.
Looking out now quite a long way, the underpin of future license revenue is very encouraging. Specifically, we do say in the release that prospects for backlog in the second half are promising, which is language we've used before suggesting that we think probably we exit the year with the backlog higher than it currently is. Clearly, that's not guaranteed because licensing is lumpy, but that would be our expectation. Bottom line is outlook for licensing is pretty healthy from here. Looking at the royalty, Warren sort of touched on most of the numbers. Clearly, it's another period of outperformance, +14 versus -7 in value terms. +9 versus -4 in unit terms against the comparable periods. Gains in all markets and increasing average royalty per chip.
Quite encouraging really that the 4.8 average royalty per chip this quarter is at the same level as last quarter. Because last quarter, microcontroller growth was relatively muted compared to previous quarters. Whereas this quarter, microcontroller growth is back strong again, and yet the overall average has stayed up notwithstanding that suppressing influence. I think that's pretty encouraging. As Warren says, we continue to invest. We're hiring people both in the R&D capability, notably in our processor division and in our graphics division. We're also obviously scaling the commercial and infrastructure part of the organization to support the growth. There's investment going on there. OpEx overall, you've seen this morning, GBP 66 million against consensus GBP 68. There's a reference in there to the mark-to-market FX impact, which every quarter or in most quarters, unless you have significant volatility, is in the sort of ±GBP 2 million.
This quarter because the dollar strengthened when we were marking to market contracts in the balance sheet, saw a small credit. Basically underlying in line with consensus and our guidance for next quarter, broadly similar exchange rates, GBP 68-GBP 70, again, in line with current consensus of GBP 69 for Q3. Full year consensus for OpEx around the GBP 272, GBP 273 mark, that would still be a good best estimate, I think. Notwithstanding this investment in our R&D capability and the growth of the company margins about four percentage points ahead of consensus and a couple of percentage points ahead of where we were this time last year. Earnings up 20% in the quarter, 22% in the half, and interim dividend increased by 20%, consistent with where it's been for the last couple of years since we came out of the downturn.
When we were reminded, we increased it by 10% through the downturn when the world was falling off a cliff. Outlook summary, putting our sort of bit of color behind our guidance. We're in good shape. We're entering the second half, as we say, with a record backlog, healthy pipeline, licensing looks robust. Underpinning this, of course, is the dynamics that Warren was talking about in terms of more companies using more Arm technology in more markets as a long-term trend. From the macro standpoint, we're all living in the real world. We're all reading about the uncertainties that surround us. The industry in Q2 was up mid-single digit, 5% or 6%. That's obviously the context of our Q3 royalty. You would expect to see a sequential increase in Arm's Q3 royalty.
If you look back over the last five years at Arm's Q4 royalty, you will have seen numbers between a GBP 10 million and a GBP 16 million uptick on a quarterly basis, reflecting normal seasonality and Arm's growing faster than the market. It looks from here as though we're going to see something overall less than normal seasonality. A number of the semiconductor companies who have guided their Q3 to date are guiding sub-seasonally. Qualcomms, TIs, STs, these sorts of people. Others are in a slightly different place. I think as we look out over the second half and we look at the GBP 875 million full-year revenue consensus coming into these results, with that uncertainty about the seasonality effect of Q4, we think that's a reasonable place to be thinking about Arm's revenue in the full year.
If there is a risk in Q4 royalty, we think it would be offset by licensing. GBP 875 seems about the right full-year number which is why our conclusion is we expect to be in line with market expectations. Summary. Strong licensing, lots of design activity. Arm in an early stage of penetrating a lot of new markets, which is very encouraging. Expect to continue to outperform the industry. We're still investing, but we're increasing profitability and generating strong cash as we do that. It's a good picture. A quick advert at the bottom. Many of you would have gone to our TechCon in the Valley last year. We're running a similar program this year. It's an industry event, but there's an investor activity in there on the 31st of October.
Our partners are involved, so it's quite interesting to get their perspective of how Arm is helping them and how they work with Arm. Focus on advanced manufacturing, mobile computing, networking, which Warren was talking about, and security. Anyone interested in going down to that, contact Ian or Jonathan. With that, we'll throw it open to questions. Thank you. Didier, first hand, first up. That's a long way from the mic.
Thanks. It's Didier Scemama from Merrill Lynch. First question, just wanted to have your views on the ASML-Intel transaction. What are your thoughts as to what Arm needs to do, can do in response to that? In particular, I'm just wondering whether the Arm partners have talked to you about it, and whether they would consider a similar transaction with Arm. I've got a follow-up.
Yeah. Well, I'll sort of kick off. It's not something we've spent a huge amount of time discussing and cogitating on in Arm, to be honest. We know that the world of equipment supply is small, the number of customers is small, and in order to move to 450 millimeter wafers, then significant investment is required. ASML have a business model where they get money from their customers for the products that they sell, and running that business model, they reckoned that they were unprepared to invest in 450 nanometer development. Intel wanted them to invest in 450 nanometer development, and ASML presumably said, "Well, if you want us to do it, you have to fund it." That's how I would imagine it. Well, we have a similar business model where we get money from our partners for selling Arm licenses.
They'd love us to do many more products, but we have to balance the books, and so we say, "No. If you want us to invest some more, then you have to pay some more." We generally manage to get them to pay for the extra value of higher value products, which involve more investment from Arm. All the 64-bit products, for instance, we are selling these licenses at a higher level. We've not had a reason to turn to our partners and say, "You need a different kind of commercial arrangement," i.e, an investment in Arm to fund that sort of development. They invest in Arm through writing us purchase orders, and they agree to pay us higher levels of money for a higher value license.
I think what you're saying is that licensing could be sort of brought forward as a result of this transaction as a retaliation from the Arm partners someday?
Well, I don't anticipate a big sort of retaliation. I don't think it is about retaliation. It's about people wanting to build 450 millimeter wafers instead of 300 millimeter wafers. Whether Intel and ASML decide they need some different kind of commercial structure to make that happen is up to Intel and ASML. As I say, you asked, did we have discussions with our partners about something similar? No, we do it through the standard business model. The standard business model, in our case, seems to work because people are prepared to pay more money for products that cost us more to develop because we're able to show that those products deliver more value.
Also, ASML are in the business of supplying to the whole industry, and that's not going to change. The fact that their customers, Intel and/or others in the future, might be funding certain pieces of it doesn't mean that technology is not going to be available at all, which in a way is the key thing from Arm's perspective and from the Arm partnership perspective.
The counterargument to aggressive CapEx and process technology is CPU innovation, as you mentioned in the past. I was just wondering what you can say about the progress of big.LITTLE. You've announced a new customer also for the ARMv8. What sort of feedback do you get from not just the semiconductor companies, but also the OEMs and the broader ecosystem about big.LITTLE, whether they think that's enough to keep Intel at bay when it comes to power consumption?
Well, a couple of things. I don't think it's all about keeping Intel at bay when it comes to power consumption. There is an underlying demand for making microprocessors more efficient. If we look over the next decade, there's about a 30-fold increase in the amount of processing power that people expect to find in things like phones and tablets and those sorts of things. There's about a two times improvement in battery capacity. There's a massive gap to close by intelligent design of the products. You can't just rely on the batteries to get better. That applies whether it's Arm, whether it's Intel, whatever. That I think is what's motivating us to create microprocessors which are ever more efficient.
Big.LITTLE achieves We have yet to see the silicon, but we anticipate that we can achieve multiples in terms of efficiency rather than 10% improvement here, 20% improvement there. Big.LITTLE, we expect to see about a fivefold improvement in efficiency when you look at the whole system. Our partners are buying into that. Their customers are buying into that because they want to close this massive gap that needs to be closed in terms of performance of the thing and amount of energy you can get out of a battery.
Final one. Just maybe a quick one, housekeeping question for Tim. If I remember correctly, reading the press release, you was talking about, I think, GBP 18 million for a data center in Cambridge. Why do we need to build a data center? What's behind it?
If you think about what the Arm R&D people actually do is they're designing products using machines. They need compute capacity. The more engineers we have and the more complex the things they're designing, the more capacity you need. What this is, every so often you're going to get a step change in your production requirements. That's what we're doing. In due course, over the next three or four years, there may be another data center built somewhere else in the world. It looks like it's a fairly significant step-up, but it's kind of an investment for the next 10 years. This is not something you're going to see keep repeating.
Thank you very much.
Thank you. Let's move along. Front row here, and then we'll move over to this side.
Yeah, we try to be shorter. On the FinFETs with TSMC, can you give us maybe a bit more comments about this? How do you think it compares with Intel 3D, or whatever they call it? How involved your PIPD team is involved to extract the transistors, characteristics of those transistors? Also, I think the timing has been brought forward by one year, I think. That's the first question. The second question is, you've been talking about 64-bit. The V8 architecture taping out relatively soon. Maybe if you could give us a bit more details on what type of products will come on the market in the next 12 months for this 64-bit, if it's on the servers and other things. Thank you.
Okay. Dealing with the FinFETs first. A year or so ago when Intel talked technology, we said, "Yes. All right. This is something which has been around in the semiconductor industry for the last decade or more." It's one of the ways of making transistors more efficient, but it comes with a load of associated challenges about actually making the stuff and making them yield. That sort of holds back the semiconductor industry from taking that step. Intel took the step and announced that they'd taken the step. They were the first ones out of the gate announcing that they were doing this. Of course, everybody else has been the same, researching it and playing with it for the best part of the last decade. TSMC had their plans in place. They just were not choosing to go public on FinFET until they were choosing to go public.
We've been working with TSMC on next generation processes for some time. We've stood here and done presentations and talked about tape-outs on 20 nanometers. The first Arm tape-out on 20 nanometers was well over 1 year ago. We've taped out first 14-nanometer designs already with some of these players. It's R&D activity. As and when the foundry wants to make some of these things public, then they will, and that's what TSMC have chosen to do this week. They chose to, I guess, communicate particularly with their customers who are Arm partners by saying, "Not only are we doing some process development in the back room, but we're also thinking about how you're going to take this technology to market, the sort of products you're going to build with it.
You're probably going to build Arm-based products with it." We've been working with Arm and Arm's Physical IP division to make sure that their Physical IP, their microprocessors, and our semiconductor process technology works well together. That's all there is to it. On the second question about 64-bits, as I said in the presentation, it's being used across a range of different applications, including mobile and computing. Servers is a very visible application area where, as we've said before, our penetration in the server market is limited until such time as we deploy 64-bit solutions. I think it's well known that one of our early 64-bit architecture licensees is targeting server applications. Probably you'll see that silicon fairly early on.
Thank you.
I think if we move along and move back.
Thanks. I guess firstly, just a related question. I think Calxeda provided some interesting milestones this quarter in terms of the server progress, I just wonder whether you can talk to how you feel the progress is going there in terms of actual sort of processing. Secondly, I just wondered whether you provided the interesting slide just on the multi-core effect in quarter. I just wondered whether you have a sense of how much of your unit shipped in mobile today is actually on quad-core based devices versus dual core. So the impact of quad-core presumably still to come. Just secondly, some housekeeping, maybe a couple for Tim. Sorry for Tim, yeah. Just in terms of the cash is creeping up.
It's been creeping up for several years. I know the historic argument that as a percentage of your market cap, maybe it's not quite as high as some of the technology peers, your business model is quite different. What's the intention for the cash pile going forward?
Okay. On Calxeda and the server activity, I haven't really got anything else to say. We're very pleased with the progress. The data that's coming out suggests that all the experiments that we did before and all the simulation that we did before is being proven in silicon. Bear in mind, this first Calxeda silicon is actually Cortex-A9 based. I think I've said Cortex-A9 was a core we developed very much with mobile in mind. Calxeda have added System on a Chip infrastructure to turn it into a server chip, it's still a microprocessor core that was designed for mobile. When you put that server infrastructure around a microprocessor core that's been a bit more designed with server applications in mind, like for instance, Cortex-A15 or moving on to V8, then you're going to see even better performance at these levels of power consumption.
We're very pleased with the data that's come out so far. We're also pleased to see other Arm silicon partners starting to get a bit more public with their activity on servers. The dual-core, quad-core, I don't know that I can talk specifically about numbers, I'd just point you to shows like Mobile World Congress and CES, where what tends to happen is that you sort of have an announcement about products one year, they turn into reality the next year. We saw, in the 2011 season, a load of dual-core devices being announced, they've now sort of materialized into phones. It was about a year later at these shows that we saw the quad-core products announced, we'd expect that sort of trajectory to continue.
Over and above that, some people have gone a little bit further ahead with the quad-core, they're using it as a sort of marketing tool and saying, "Core is better than dual." It's a bit of a marketing thing. It's up to our semiconductor partners to see what performance for a given level of power consumption they can actually achieve. We put it up on the slide as multi-core, put the two together because that's really how we view it.
On cash, nothing particularly new to say about that today. You're right. We managed cash down before the downturn to around GBP 50 million by the end of 2007. We bought back 16% of the stock in the period leading up to that and a little bit into 2008 at about GBP 1.20. We have let the cash ride up through the downturn. Since we've come out of the downturn, we're now just under GBP 500 million, which as you say is sort of 7% or 8% of our market cap, although that is in a sense quite an academic point. Certainly if you benchmark the sector, it doesn't look an unsightly amount. The bigger point is, what are you doing with your cash and where do you need it and where is it going?
The plan of record remains, as the base case, a progressive dividend with a gradually growing payout ratio. At the moment, you look at half one, our interim dividend is up 20%, broadly the same as our earnings. I think in a period of fairly considerable macro uncertainty, I think growing a dividend 20% in line with your earnings is not an unreasonable place to be, especially at the interim. The trajectory of how this payout ratio increases is sort of constantly reviewed by the board. At the moment, we're comfortable with the level. It may well be when we look out that an increasing payout ratio, depending on the trajectory, is insufficient to stop the cash pile from growing beyond the level that we feel comfortable with or need, and in which case, we could do other things. We could step up dividend.
We could return via special. There are other things that we consider. The plan of record at the moment remains a gradually increasing payout ratio from here, i.e. dividend growing faster than earnings over time.
I think if you just sort of move it back and along, we'll go zig zag.
Thank you. It's Kai Korschelt, Deutsche Bank. A couple of questions I had. The first one was just on the attach rates for GPU or for Mali for that matter in set-top box and digital TV market. Where do you think, say at around this time, where we are in terms of %, in terms of overall volumes, and where do you think that goes maybe on a five-year view? That would be my first question. The second one is, looks like particularly in the lower end of the smartphone market, there are sort of increasingly platform solutions which rather than using an apps processor and a GPU, use a beefed up GPU as basically to drive the only processing function. I think Broadcom has one of these platforms out there.
I'm just wondering how or whether you see a risk that maybe particularly in the low end of the market, call them beefed up GPUs take over more of the historical or traditional application processor functions. The third one was just on a like for like perspective, if you could remind us maybe of the potential royalty premium for 64-bit versus 32-bit, please. Thank you.
Okay. On the Mali attach in digital TVs, we expect to be in 70%-ish of the digital TVs that we have processors in to have Mali. We expect to retain our number one slot in TVs this year. Exactly where we are at the moment, I'm afraid I can't tell you. You looked at Ian. Do we have a sort of mid-year view?
I'd say the 70% capacity we expect is 2012.
We don't really have an update on the full year view at the mid-year on that, I'm afraid. As I said in the presentation, we're very happy with the Mali shipments that our partners are reporting in the latest sort of quarter just gone. We think we're on track for the 100 million units plus, and obviously, digital TVs are part of that. On the low-end smartphones and the GPUs, that's an interesting question. I think you will see people experiment with this concept of GPUs being used for computing functions. I personally am not seeing it as a serious threat to our applications business at the moment. Our applications processor business in mobile phones. Probably even less so in the low-end phones where you want to produce something really cost effectively.
We have Cortex-A products like Cortex-A5, like Cortex-A7, which are really designed. I mean, Cortex-A7 is the little half of a big.LITTLE combo, but it can also be used on its own, and on its own, it's our most efficient apps processor yet. In terms of delivering into that low-end smartphone space, I'd say it's a lot less effort to use a Cortex-A7 efficient processor on its own and coming with all the software ecosystem that already exists around the Cortex-A series products. That said, if the world does at some stage turn to GPGPU, then if we look at our Mali graphics family, we do have two parts to our family. We have the pure graphics, and we have the GPGPU series as well. To an extent, we're sort of hedging our bets, but I don't see it as a serious threat.
On 64-bit premium for, or sort of royalty premium for 64-bit, I mean, this is a continuation of the trend we've been on for a while, where basically if there's more value in the microprocessor, the royalty comes through at a higher rate. We've talked about Cortex-A being sort of typically in the 1.5%-2% range, compared with pre-Cortex-A being more in the sort of 1%-1.5% range. That trend will continue with our V8 architecture. It's going to be at the higher end of that range. I think we got one here, but yes. Got one here first, and then we're coming over next one, Simon.
Just a very quick one, if I may. Andrew Dunn from RBC. Just following up on royalties. Your slide number five on smartphone application processor types from single-core, multi-core, et cetera. Could you just remind us on either the absolute or relative royalty rates you might get for each of those segments?
Sorry.
Multi-core, single core, and ARM11, just the relative royalty rates for each.
Well, typically, ARM11 is at the in the sort of 1%-1.5%, and typically Cortex-A, as I say, is in the sort of 1.5%-2%. Cortex-A9 processors are less than Cortex-A15, so they'll be more at the sort of 1.5% of that top range. That's the percentage. You then multiply that by the chip price, if it's a dual core chip or a quad core chip, typically it's more expensive than a single core chip. It really is a matter for the semiconductor partners and where they are sort of positioning their chips. If you've got a single core A9 with some very fancy video accelerators and graphics accelerators and some integrated Wi-Fi or something like that, it could be just as expensive as a dual core chip with fewer fancy bits on the side.
I don't think we can clinically say single core is this price, dual core is this price, and quad core is a higher price. Clearly, there must be a sort of trend there that otherwise people wouldn't do it. I can't tell you the data points. Simon, I think we're over here.
Hi. Thanks. Simon Schafer, Goldman Sachs. I want to actually stick on this royalty debate. I guess in just in basic terms, if I look at slide 10, I think you talk about 95% of your royalties still being from licenses signed four years ago. In rough terms, I mean, is it fair to say that 95% are collecting a 1% rate and then the 5% are collecting a 2% rate, or how should we think about that?
Yes and no. What I would do to refine that model slightly, I would make one refinement, and that is in your sort of 5%, I would split it into the microcontrollers and the non-microcontrollers. It might be fair to look at, or you could split it as Cortex-A and others. The others, you'd still count at the sort of 1% end, the Cortex-A, you could allow to go a bit higher.
When do you think that 5% will be 25%? How long will it take?
That's why we published the little bit, the chart on the bottom. We would expect a similar sort of trajectory to what we've seen in the past.
Okay, great. Then just on the licensing side, Tim, for you sort of alluded to it, but the $65 million number, yet another step up. In essence, is this just because people are paying a higher value per license? There's not necessarily more people. I appreciate there's a few more different verticals that are being added to the licensee base, but what's really the delta? Is it just a higher price per license? How should we think about it?
Well, it's actually multiple factors. Clearly, there are more companies buying Arm licenses today than there were before. If you look back over the last couple of years, about a quarter of the licenses we've signed are with new companies. That's not insignificant. Albeit quite a lot of them at the Cortex-M end. I think in a way, the most interesting thing is that backlog line that we were looking at, because a lot of the bigger deals that we've signed in the last couple of years have really found their way mostly into backlog, not yet into license revenue. It's quite interesting looking at the If you look at the license growth in that bar chart, you can see that the growth is coming from backlog, but the turns business is staying reasonably consistent.
You could argue with lots of big companies entering into things like more subscriptions. It's actually very encouraging that the turns business is staying at the same level on top of an increasing backlog. I think, it's multiple dynamics, but underpinning it is Arm's addressable market is much broader across the computing spectrum than it was viewed to be two, three, four years ago. Now companies are licensing Arm. The traditional companies are using Arm more and more in more and more end product divisions because of the, in a sense, the typical outsourcing dynamics and the technology pressures. There are also new companies coming into the frame for the first time, both at the top end towards the servers and at the lower end towards the Internet of Things. All those factors are driving it.
Great. I think we'll continue to move back to the AV if we get time for a second one. I haven't forgotten, we've got one at the front here after this one.
Hi, it's Janardan from Liberum Capital. Two questions. One is on the FinFET agreement with TSMC. It's on 64-bit. I was just wondering what plans you have on moving the 32-bit Cortex-A15 kind of products to FinFET. Do you have another agreement with them which we don't know about? Will the timing of introduction of that be roughly the same as the 64-bit side? Second question is on Windows on Arm. There have been not that many companies and devices being announced so far for Q4. Presumably some of that is because of software compatibility issues and things like that. What is your estimate of how much time it'll take for that software ecosystem to come up to comparable levels with x86 architecture? Are we talking about a couple of quarters, or is it going to be a year, or is it more than that?
What's your view on that?
Well, let's answer the first one. The FinFET. Yes, the announcement is with our 64-bit processor because just as we want to work with TSMC's most advanced process technology, they want to work with our most advanced microprocessor. Making a 20-nanometer FinFET and later 16-nanometer FinFET implementations of our 32-bit processors will fall naturally out of that development activity. We're optimizing our Physical IP to build microprocessors. We just happen to be using our new 64-bit processor as the vehicle for it. The same Physical IP will be very easily used to implement our 32-bit processors.
What's your estimate of the timescale of introduction? Is that a 2014 introduction or is it 2015?
Well, we have to stick with the announcements for now. I think as and when TSMC want to make more comments on when these things are available, then they'll make more comments. As I say, from a development point of view, we're taping out stuff all the time. Windows on Arm, not that many announced. Our estimations for when things become widespread. As I said, we're very excited to see the Windows on Arm products that are going to be launched in the fourth quarter of this year. My expectation for volumes of those things in the fourth quarter of this year, incredibly modest. Frankly, I'd rather see fewer good quality products than many products that don't go down very well. The fact that Microsoft are controlling the launch very tightly and working with a limited number of partners, we see as a good thing.
We're not changing our estimates for an expectation that by about 2015, we expect that we'll have a share of the laptop market of maybe 15%-20%. That hasn't changed. I think that's a sort of measure of how we see the rollout of that going, because 2015 is still a few years away. I haven't forgotten, let's keep going back because you've got the microphone there.
Thanks. It's Julian Yates from Investec. Just a quick question on the licenses and backlog, Tim. If we look at the backlog, it seems as though you may be taking more from the backlog into the revenue line over a period of time. I'm trying to understand, should we assume the backlog at some stage levels off and starts declining as more of that backlog gets recognized into P&L? Do you think over the medium term, that backlog should continue to increase?
Well, backlog contribution to license revenue is typically in the 40%-60% range per quarter. As it happens, this quarter is up at the 60% range. Therefore, in a sense, even more encouraging that the backlog is up quarter on quarter, even with a reasonable drawdown into license revenue. Therefore, in a sense, whatever the recent dynamic has been, the fact is the backlog is three and a half times higher than it was at the first half of 2007. Frankly, that gap that I showed between license revenue growth and backlog growth takes a very long time to unwind because of license revenue recognition through periods. No. Obviously, the backlog is inevitably lumpy quarter on quarter, depending on which profile of deals are signed in a given quarter. We're not just about to stare at a reduction in that backlog.
We shouldn't worry about the increase in license revenues of $ 65 million being a reflection of yourselves taking more of a bigger backlog?
No. We've moved from a world of 30 million a quarter to 45 million a quarter to 60 million plus, and the backlog's still going up.
Thank you.
Sorry, where did my one go? Let's come over here to the front then.
Thanks. It's Sumant from Redburn Partners. Just three quick questions, if I may. First is on your sort of market share. Given that about 1/3 of your licenses are still sort of not generating much royalties yet, and majority of them are in the non-mobile sector, what do you see in the next five years assuming if you don't attain any more licensing, license in this space, what do you see your market share growing to just from these licenses which you've already sold out in terms of royalty market share? The second question is to do with the FinFET again. Most of the foundries are sort of offering different node transitions, and in between, I assume FinFET would be an option in between 20 nanometer and probably 16 nanometer.
My question really was that would you be licensing FinFET technology separately as well, or is this an exclusive collaboration with TSMC? Is there a royalty increase coming from products based on FinFET PIPD, so to speak? My final thing is on network opportunity, in networking equipment opportunity over there. Just trying to understand in the networking equipment, how big as a percentage of your royalty revenue in about five years' time do you see networking, so to speak? I'm talking separate from servers. You're talking about there being a networking opportunity in terms of power saving. I'm just trying to understand what kind of chips are we talking about and what kind of cores are we talking about within them?
Okay. Yeah.
I think if I understand your question correctly, Sumant, talking about the overall market share, which currently in 2011, our estimation is we're about 30% overall market share in the embedded processor market. You can see from our market segment slides, which I think in this pack are in 21 and 22. If you look back over five years, that's grown from 17% to 30%. There are some years it's growing 3%, sometimes four, last year, five. We don't see any reason to suggest that the growth rate of our market share doesn't grow at least that rate as we look out over the next kind of five years. A good number to have in your head might be about 50% in a five-year view, which would obviously be around 4% per annum.
There are probably more reasons to think it's going to be at the high end of that than the low end, given microcontrollers and the volumes associated with it. In a sense, if we were talking about this nine months ago, a year ago, we wouldn't have been talking about networking in the same way. There's a little bit of flicking of a switch going on. Actually, that's a bit of a pun, isn't it? There's a bit of a flicking of a switch going on, I think, in networking, which might drive it. I say on balance, Ian and Jonathan update this every year. Most years we update it, the actual market size grows a bit because the addressable market for Arm grows. Bottom line, around about 50% seems to us pretty reasonable in a five-year view.
Okay. The next question was about FinFET and whether it's essentially a different Physical IP product from Arm. The answer is, well, it's a different flavor. We have different flavors of our Physical IP for each semiconductor process. A low power version of a given node is a different Physical IP bundle than a high performance version. A FinFET is another flavor again. It would be an incremental licensing opportunity. The fact that our Physical IP is used would generate the royalty opportunity. It's not an incremental royalty opportunity. The fact that it's FinFET, it's just another flavor. If we're going to have a 20 nanometer low-power planar flavor and a FinFET flavor, the chip's only going to be made out of one process technology. The royalty opportunity is the same.
Networking equipment and what sort of share can we expect in that? I think to Tim's point, we update the numbers every year. We will be updating the chart in the end of this year. Probably when we look at the networking line, and you'll see on the slide, which was produced at the beginning of this year, we do have one of the green arrows, as in key growth area for Arm, beside networking on the slide. We are excited and we have been about this area. Maybe the absolute numbers there might be one of the areas where, as he says, when we update, we quite often revise the size of the market as we see bigger opportunity.
I'm very encouraged by the reception that our customers' customers are providing to Arm technology in this space, and I don't see why we can't have a very large share of this space as the current incumbents have a large share today. I think the power efficiency benefits of Arm and the ecosystem benefits of Arm and the business model benefits of Arm apply just as well to this networking space as they do to mobile computing. I think we're on one more. Was there anybody at the back who didn't get a go yet who wanted? No. Didier, bonus ball.
Final keynote question.
Yeah, thanks. I just wanted to understand, we all talk about application processors and so on. I was just trying to have your views as to whether we need higher performance CPUs in modem. e.g., does LTE require a higher performance modem? Does Wi-Fi AC require a higher performance modem? If the answer is yes, what sort of upside in royalty rate, broadly, are we talking about? The second bit is regarding basically the ARMv8. The bears, or let's say, those who are skeptical about Arm penetrating the server market and the notebook market, talk about lack of ecosystem support. They talk about the lack of ISV software, these sort of things. What's your response? More importantly, what's the feedback you get from the likes of HP and Dell that are dabbling with the technology? Thank you.
Okay. Perceptive question on the modems. In fact, what you're trying to do with these modems is squeeze an awful lot more data down the same size pipe. That means a much more complex coding system is required. Just to put it in perspective, some of the LTE plus type modems, if we look at the complexity of that and compare it with, say, a 2G modem, you're talking about a factor of 500 or so in the complexity. Actually, they do need higher performance microprocessors. If we look at the modems which are shipping today, so the 3G type modems, then it's not a linear journey from 2G to LTE plus. Actually, the 3G ones are just a little bit more complex than the 2G ones.
Today, most of those sorts of products would be ARM11 type products versus ARM7 or ARM9 type products. A little bit more performance. From a royalty point of view, no real significant incremental royalty opportunity. We've had a lot of design-ins for a Cortex-R4 type products and even some Cortex-A products in modems, but these aren't really shipping in volume today, and these are for initial LTE type modems. As we look to the more sophisticated LTE type modems, then yes, they will certainly use more sophisticated processors because, as I say, they are much, much more complex. There will be a corresponding royalty upside opportunity there. It's just not really in the numbers yet because, I haven't got the numbers off the top of my head, but the vast majority are ARM11 based today.
The last question was about?
What's the feedback on ARMv8 from the OEMs? Hewlett-Packard, Dell?
Pretty positive so far, but they have to see the silicon. The other thing, as you pointed out, 64-bit, it's a new instruction set. The software has to be written and so on. It's early days. They wouldn't be engaged in these programs with us if they weren't actually quite positive about it. Thank you. That is the last question. We've actually finished roughly on time. Thank you all very much.