Good morning, ladies and gentlemen. I'm Stuart Chambers, Chairman of Arm Holdings, and it's my pleasure to welcome you to Arm's prelims announcement of our results for 2014 full year, which I'll be handing over very shortly to Simon and Tim to run us through. Just before then, a couple of words of context from me. This is the end of my first year as Chairman of Arm. I had some pretty high expectations, I have to say, coming in, of how interesting, how fascinating indeed, and how enjoyable it was going to be, and haven't been at all disappointed. Had a pretty intense four months induction, courtesy of Patricia here and a few others. Spent that obviously traveling around, looking at the various Arm locations and also meeting a lot of people. I still, in between board meetings, I meet Simon's team and indeed their teams.
Every time I do, it reinforces the enthusiasm, the skill of the people at Arm who continue to sort of push forward this amazing Arm story. Very enjoyable first year. Just a quick summary of board changes last year. John Buchanan stepped down at the beginning of the year as the Chairman on medical grounds. We also had Philip Rowley and Eric Meurice stepping down as non-execs. Joining the board last year was myself, at the beginning of the year as the new Chairman, and also then John Liu joined from China as our newest Non-Executive Director, and he joined in the fourth quarter. The final, last but not least, board change was, as you may recall, Tim announced in May his desire to retire. Looks kind of young for that, but such are plans. His decision to retire this year.
Obviously we're very sad to be, or will be very sad to be saying farewell to Tim when the time comes. Nevertheless, we're absolutely thrilled and delighted to be able to welcome Tim's successor, which is Chris Kennedy. Chris, I'm going to embarrass by just asking to stand up and just wave. He's with us today. A very warm welcome, obviously, to Chris when he joins. In terms of the dates, the specifics of that, not set yet. We're still working that through. The handing of the baton will be defined, and as soon as we know those dates, we'll obviously announce and let you know. Please don't ask them about dates. Well, you can if you like, but they won't be able to tell you. I think that's probably enough from me. I'll hand over to Simon for our results.
Thanks, Stuart, and good morning, everyone. Thank you for joining us today for Q4 and full year 2014 results. Before we start, I'll just refer you to the usual cautionary statements, and we'll take these as read. It's been a great year for Arm. What we're going to do today, I'm going to talk, give an overview of the business, what's been going on, the progress that we've been making in our key markets. I'll hand over to Tim, who'll talk through the specifics of some of the numbers. We have time at the end for Q&A. Let me start with some of the highlights of what went on in 2014. It was a great year for execution. Our license revenues increased strongly, up 30% year-on-year. There were record units shipped by our customers.
12 billion chips containing Arm processors were shipped by our customers. That drove the royalty growth in the business. You'll see later on how those 12 billion chips have enabled us to grow market share in our key markets: mobile, embedded intelligence and enterprise infrastructure. As we look at this business going forwards, we see more opportunities to grow share, to make Arm technology relevant in more and more markets. We've been investing in the business to capitalize on that. We see further opportunities for continued investment as we go forwards. With the performance of the top line, with the investment in the business, still though, our EPS grew 17%, and we've recommended an increase in dividend of 23% for the full year. I think it's been a great year for execution.
What I'm going to do now is just look at some of the key markets and what's been going on. Let me start with mobile. Mobile is a market that's continuing to evolve. Through last year, there was a bit of a sentiment that there's no innovation in smartphones, and I think that's not the case at all. We've seen a lot of innovative devices come through, and we're going to see more coming up shortly. The way in which mobile has developed is you can think of it as kind of two main markets. You have developed economies where people have been using smartphones and mobile devices for a number of years. In those markets, the platforms are getting more and more sophisticated, higher and higher performance, bigger screens, more pixels, greater connectivity. The way in which people are using these devices is changing.
It's not just a device you make a call on and surf the web, look up the football results, which I wasn't enjoying last night. It's now a device that you can use to interact with many other things. You can use it to connect with your car to switch on the heating before you get in it. You can use it to connect with your home. You can use it to open your front door, set your heating. There are many ways in which these mobile devices that people are carrying around are being used in more and more ways, and we'll see that continue. On the other side is progress in developing countries where mobile phones are used very widely, but very few people have access to smartphones. That's changing as smartphones become less and less expensive.
We're seeing that advance in really quite phenomenal ways right now. You can buy an unsubsidized LTE phone for GBP 65. That is a price point that a few years ago, most people would think was just never going to happen. As people get access to these devices for the first time, it's going to open up a world of information. The kind of information people can access is very fundamental to quality of life. It's access to healthcare information. It's access to education in general. It's access to the market for products and services that they might create. It's really going to change the way in which people access the world. We're really excited about the developments in mobile, and long may that continue. As we look, though, at mobile computing, it's more than smartphones. It's about tablets, it's about clamshell-style computing.
We're seeing progress for Arm in those markets, particularly as we deliver more and more compute power in these very energy-efficient devices. Smartphones, that's a market where we've maintained our penetration. We place a lot of attention on the application processor, the main CPU that is driving the intelligence in these devices. The opportunity for Arm in mobile computing goes beyond that. It goes into the connectivity, the touch screen, these sensors that are interacting with other devices. That, in combination, drives the volume of Arm processors in mobile computing. If you leaf through the numbers, you'll see that about 47% of all the chips that were sold by our partners last year, that GBP 12 billion, was in mobile computing. Obviously, a lot more than the number of handsets. Great progress in mobile.
We'll go to embedded, where it has been another great year. We've seen many more chip designs by our licensees, very strong licensing of our Cortex-M series on which most of these products are based, and very high volumes of chips are shipped by our partners based on the existing technology. The number of microcontrollers shipped was 4.4 billion, which is a vast number of chips. That represents about a 24% market share. If you compare that to the same category definition last year, which was about 19%, you see a growth of about five percentage points there. It really is the number of companies in this space and the breadth of the product offering that is really, really impressive. There are over 3,500 different MCUs that you can buy right now from our customers.
Whether you want a lot of flash, a little bit of memory, peripherals, all sorts of combinations are out there provided by our partners, and that's creating a huge kind of portfolio that somebody looking to build an embedded product can now choose from. There's just vast choice. The right product for the thing that you want to build is out there. There's already a great starting point and scope for more performance, more integration as we go forward. When we look at the companies who are building microcontrollers, building these embedded intelligence devices, we see very strong penetration in Arm of everyone who's playing in this space. The top 10 companies are all shipping products based around Arm.
In total, we've licensed around 200 companies in total with Cortex-M. What many of those companies are doing is generating products for these new emerging markets, wearables, smart devices, IoT, which integrates different technologies in different ways to address new and growing markets. We think as we look to the future, there's new ways in which this technology will be used, and it's a really fascinating and dynamic landscape. Now let me talk about enterprise infrastructure, where again, I believe it's been a really exciting year for Arm. Our market share in networking has doubled year-over-year. This time last year, we were talking about a 5% market share. Here we are today, and our data suggests a 10% market share. Great growth. All of that is based around chips that have been designed over the last few years by our partners.
It's the culmination of work that's been going on for a long time. All those devices shipping today are based on version 7 of the Arm architecture. What's really exciting, therefore, is as we look forwards at new designs coming through based on version 8 of the architecture, which creates both greater volume, we believe, because with that technology, we can address a broader portion of the market, but also improve royalty rates, which will help our bottom line at the end of the day. A great example of that is what HiSilicon have been doing. They have a device based on 32 Cortex-A57s. It's a very high-performance, 64-bit compatible core built on the most advanced semiconductor process. They're shipping that.
It's a great example of the kind of work our partners are doing around Arm technology, creating energy efficiency in something that's typically not been a market that's cared too much about that before. Also in servers, which is the other side of enterprise infrastructure for us, it's been a great year of progress. We're seeing servers shipping right now based on Arm technology. We're seeing great design work by our licensees. New customers coming to the fore. Qualcomm recently announced some of the details about the work they're doing. We think there's going to be a great future here. In both these markets, though, one thing I do want to stress is we're at the very early stages. These are nascent markets that have required a lot of investment for us to get here, and we're very pleased with where we are.
It's an area where we think through greater investment and continued focus, we can really help penetrate these markets and grow our share more rapidly. This is a real area where we are continuing to place a lot of focus. What I've been talking about so far is mainly the progress around our CPU technology, but we have a lot of other technology as well. If I turn to our GPU, our graphics processors, it's been a great year for the adoption of our Mali GPUs. We believe we are now the number 1 graphics IP vendor in terms of volume, shipping 550 million chips, or our partners rather, shipping 550 million chips based on our Mali GPU. You can see there that the growth has been really quite phenomenal over the last few years.
We see no reason why that growth can't continue at similar sorts of rates through 2015. At the end of the day, people need to build chips, and for a long time, we've been focused on creating physical IP to allow people to take their designs, implement them on silicon with the highest performance and the lowest power. We've done some analysis looking at the volume of chips, which do actually contain our physical IP. This isn't data that we typically publish, but the analysis shows how the uptake of all the work we've been doing over the last few years is really paying off in terms of volumes. 8.9 billion chips, and this is based on wafer analysis, we believe last year, were containing our physical IP. When you look at the range of technologies, it's very broad.
We're seeing a lot of our IP being used to serve the microcontroller market. A lot of the volume here is based on microcontrollers, which are manufactured on more mature technologies, where the focus really is about small die size, low cost, and ultra-low power. It's an area where we've been innovating. We're also heavily focused on the overall performance and energy efficiency of the high-performance CPUs on the leading-edge processes. Because in those designs, people are trying to cram in as much technology as they possibly can within a power constraint. Whatever work we can do to facilitate the implementation of those designs, hitting gigahertz levels of performance, for being able to go into a device, a handheld device that runs on a small battery, that is really valuable to our customers. We have a lot of our engineers focused on that area.
We call this POP IP. We're seeing strong adoption of that technology, which will drive future royalty growth for the physical IP side of our business. That's kind of what's been going on over the year as a whole. If I look specifically about Q4 for a moment, the licensing performance, very strong. Through the full year last year, we signed 163 licenses. That is a record. That was helped by very strong licensing performance in Q4, 53 licenses signed, again, a quarterly record. Within those 53 licenses, we saw strong adoption of version 8 architecture products, and we saw more Mali graphics processors in there as well. If we look at the customers who are licensing this technology, it's a very broad range. There's a lot of business with the tier 1s.
We work very closely with them, and it gives us great confidence about how these licenses flow into royalty downstream. Now, our customer base, though, is evolving, and one of the fascinating things about my job is the range of different companies that I get to go and talk to about the range of different work that they're doing all based around our technology. When I think about the customers who are licensing our products, there's a very strong repeat business. We have a very loyal customer base because we're delivering the right technology. We always have new customers coming to Arm, people who maybe didn't need a microprocessor before in their products are adopting Arm for the first time. We're also seeing some diversification away from the traditional semiconductor companies.
We're seeing some OEM companies making end products wanting to get more involved in the design process, wanting to have more control over the technology that's in there. People are starting to access our technology and looking to see what they can do in a more vertically integrated way. This is a way of the world right now, the way that the industry is evolving, and we're evolving our business models and our focus on our customer base to move with that and to continue to prosper. We're also seeing through this licensing, the fruits of the investment in the technology roadmap over the last few years. Now, through last year, we kind of teased you a bit about a product called Maya. Just last week, we did the formal announcement of that. That is now Cortex-A72.
We rolled that out with, in fact, a complete suite of IP, graphics processors, optimized video and display processors, physical IP to facilitate that implementation, a complete platform. That is what our partners are asking us to deliver. We've organized the company around this platform of technology, this joined-up solution, because that's what our customers want, and that's what you saw in the announcement last week. Maya has become Cortex-A72 and is delivering phenomenal performance for the next generation. Cortex-M7, this is a very sophisticated processor designed for microcontrollers. I think probably the most sophisticated, the highest performance processor that you can find in an MCU. That again, is broadening the applicability of Arm technology. Then also, I want to tease you a little bit more about two new products that you'll be hearing about later this year, called Teal and Green.
You see we have a bird theme going through our code names here. These are more focused towards, again, the embedded space where we see a lot of opportunity. Now, the reason for pointing these out, these are products we've been working on, we are working on a roadmap. We're engaging, in some cases, with lead customers for the first time. All of these products were licensed in Q4 as part of the 53 licenses that we signed. Now, turning to the royalty. A year ago, we were forecasting a slower royalty growth in the first half of the year with an acceleration in the second half of the year. Feedback from you all was, "Okay." Let's see it happen. I think that has happened. We reported growth in Q3. We're reporting growth along the lines of what we anticipated in Q4 as well.
16% year-on-year growth for Q4 royalty. That was driven by very high volume. We have 3.5 billion chips containing Arm processors shipped by our customers in Q4, rounding out that 12 billion unit year. When we look at those key markets, we've seen growth, again, in adoption of our technology, growth in share. You can see across the three, mobile, embedded, and enterprise, a significant growth in share and volume from all of that activity by our customers. Again, within that, we've seen growth of version 8 architecture, product-based shipments. We now have seven customers shipping V8-based processors. That compares to five in Q3 last year. Another couple of partners shipping. The volume though is still small, about 50 million units of the 3.5 billion, so it's a small percentage.
A lot of those are shipping into mobile devices. We are seeing shipments into enterprise as well, and through 2015, we expect the volumes to grow, and we expect, again, the diversity of the use to increase as well. Mobile, though, still a very, very important part of our business. We're designing a lot of products specifically for mobile. It's really interesting to see how the products are developing right now. Mobile World Congress is just around the corner, and we're expecting to see many new products based on V8 announced by OEMs and shipping soon. Actually right now, here we have a collection of, in fact, a subset of all the V8-based phones which have been announced so far. These are announced products. You can see there are quite a lot of them. Also what really stands out is the price points.
Some of these devices cost as little as GBP 100, and they go all the way up to the premium point. We talked before at the Analyst Day last year about how version 8 should ripple through the entire spectrum of mobile devices very, very quickly. You're seeing here low-cost devices, premium devices adopting V8 already. That gives us confidence of the uptake of V8 in mobile. Great progress in 2014. I believe that points to future growth for the business as all of this work that we're doing comes through, and our partners take what we do, integrate it with what they do, and create compelling solutions for these various end markets. We've been doing work on how we see the markets growing ahead of us.
In the appendices of the slides that you've got there, you can see we've done some work updating the data, updating the anticipated volumes, updating our view on the market size out to 2020, which is two years further than the data that you've seen before. You see these key markets of mobile, embedded, and enterprise. We're now anticipating GBP 25 billion worth of silicon in each of those three markets. Quite significant growth compared to our previous forecast, which only went out to 2018.
Within those, again, we are more confident that we can achieve very high market shares than we were 12 months ago because of the progress of the business, because of the way our partners have taken what we've done and innovated around it, because of the way our ecosystem has come together to create all the other technology that you need to turn these microprocessors into an end product that is compelling for various different end-use cases. We have an opportunity to take a significant market share here, and that's what we're focused on delivering. In summary, we had great execution in 2014. I think it was a great year for Arm. That strong licensing helped deliver a 17% growth in earnings per share. That was against a headwind of foreign exchange rates against us and inventory issues, particularly in the mobile segment.
We're going to continue to invest. We're going to continue to innovate in our product roadmap, because that's what our customers want us to do. Everyone is looking to the future, the next generation, and the next great technology from Arm that they can use to build the compelling products that their customers want in turn. We see those opportunities to invest, and we're going to continue to do so. With that, let me hand over to Tim, and he's going to walk you through some of the numbers.
Thanks, Simon. Morning, everybody. I think that's a very comprehensive overview of both the fourth quarter and the full year. I will seek not to be too repetitive, and just provide a little bit of color on some of the numbers and probably more importantly, help us think about 2015 as we refine and tweak our models. As Simon said, in summary, fourth quarter, good quarter really across the board. 18% revenue growth in US dollars. That's in Q4. 19% GBP, so the FX is a little bit more favorable, but not on a full year basis. As Simon said, processor licensing up 30%, record number of licenses signed, big contribution from the licenses signed in the business to the revenue. You've heard me talk before about typically 40%-60% being contribution from backlog into revenue.
In this particular quarter, 60% of revenue came from the turns business. We were particularly encouraged that at the end of the year, the backlog is up sequentially about 5%. We said at the half and at Q3 that we would expect the backlog to be sequentially up at the end of the year and probably up from where it was at the half, and that is indeed the case. That's good news. I think Simon's given the narrative on royalty. We know it was a slow period in the first half. We pointed to acceleration when we were up in July, and we characterized that acceleration as a 10% growth in Q3 and a 15% growth in Q4. We delivered 11% in Q3 and 16% in Q4.
Very much in line with our expectation, and I'll touch on in a minute where we think that's going to go going forward. Normalized OpEx in Q4, higher than guided at the end of Q3 and higher than consensus, which emerged at about GBP 93 million. The difference is really twofold. One, strengthening dollar through Q4 has had an impact, obviously, on the translation of our US dollar cost into GBP. Also, with the strong revenue performance in Q4 and the very strong bookings performance, the incentivization provisions that we've been building up through the year trued up in Q4 because the Q4 outcome was stronger than the build up through the year. We'll talk about how that OpEx transitions into Q1 next year and full year 2015 in a moment.
That revenue growth gave rise to a 25% year-on-year improvement in PBT and a 36% improvement in earnings. Those of you who had a chance would have noted that the tax rate in Q4 2014 is lower than it was in Q4 2013. That, in summary, is as a result of the U.S. R&D Tax Credit getting legislated this side of the year-end, which doesn't always happen and you can't take that benefit into account in your forecasting through the year unless and until it happens. You may recall some in earlier years, it slipped over into the following year. That's why the tax rate is lower, because Q4 takes the benefit of the catch-up on that. That's the fourth quarter. In terms of the full year, as Simon said, the FX has actually been against us through the year on a full year basis.
Last year the effective rate was 156. This year it's 163 on a full year basis. 16% total group revenues in USD translated to 11% growth in GBP revenues in the full year. Strong licensing, I think that's been well covered. Strong market share gains in royalties, offset early on in the year by some of the inventory issues, giving us an 8% full year. Normalized PBT full year basis up 13%. Significant investments in the business at 461 net increase in our employee base, 16% increase in Arm's headcount over the year. A lot of investment in the future going on. That's driven a 17% increase in earnings. Back on the tax issue, as you know, we're on a sort of multi-year reducing tax rate trajectory.
Final rate this year, just under 17% versus 20% last year versus 25%-26% the year before. This is the phased introduction of the Patent Box, which is a five-year implementation program. Looking into next year, I would expect the rate to be lower again, probably around 16%, all other things being equal. Maybe marginally under. Strong performance again in cash, ending the year just over GBP 860 million. As Simon said, the full year dividend increased by 23%, obviously subject to the shareholder approval at the AGM. You may recall that we increased the interim dividend by 20%, we've increased the final dividend by 25%. A manifestation, I think, of the board's confidence in the prospects as we go into 2015. As signaled at the beginning of last year, we reintroduced the share buyback program, limited share buyback to maintain a flat share count.
During 2014, we bought back just under 8 million shares. Total outlay, about GBP 67 million. Just a quick reminder on the sort of correlation and the relationship between our investments in R&D and license revenue growth trajectory. You can see, looking back on a sort of seven-year view, that we've had group license revenue 20% CAGR. It's a 30% CAGR post the downturn. Through that period, our investment in R&D has been 17%. You can see from the top chart that there have been periods where we've been investing in R&D through the cycle, even though license revenue has been down year-on-year, as it was in 2009, and as indeed it was in 2008. As we look forward, this correlation between license revenue growth and the increase in R&D will continue. It's not a direct time match.
As we think about the license revenue growth normalizing to the level that we've been signposting now or through these four or five years, really, of mid to high single digits, 10%, you would expect over time that the investment in R&D will normalize along with that. That won't be necessarily this year, next year. As Simon said, we have significant opportunities to invest to accelerate our penetration, particularly in some of these new markets. With that, let's look forward to full year 2015. Just sort of talk you through the outlook and how we're seeing it at this stage. It's early days. It's February the 11th. We've got a long year ahead. We are comfortable with the consensus that's out in the market. It's 1.7, 1.4. We tend to shy away from guiding specific revenue streams.
We do believe that the growth that's assumed, both in terms of licensing and royalty currently in those expectations, is broadly sensible. License revenue growth is in there at about 9%. Royalty revenue growth is in there at about 21%. I say it is early days. I think if we were doing a sort of probability analysis on that right now, we'd probably point to upside risk being on the royalty side against that consensus, with maybe some offsetting downside risk in licensing. Really early days to call that. Overall, USD 1.47 billion seems like a good place to be going into 2015. Then looking at that on a Q1 basis, we see that the acceleration in royalty revenue growth continuing. On an overall revenue basis, we're pointing to about 10% year-over-year.
As you punch those into your models, you'll see that that really points to license revenue being at a similar level to Q1 last year. Again, it's early days in the quarter. Some licenses will sign in Q1 that we thought might sign later and vice versa. I think, as we look at this full-year license revenue growth of about 9% or 10%, and as we look at the quarterly phasing of how backlog turns into revenue and how we see our pipeline turning from deals in view to signed deals, we think that at this stage, positioning Q1 as broadly flat with Q1 last year, I think is a sensible place for us to start the year. That's kind of how we're seeing the full year and the sort of Q1 relationship.
On costs, we're guiding to sort of at broadly current rates, i.e., in the early 1.50s. We're about 1.53 today. We would see Q1 costs in that sort of GBP 98 million to GBP 100 million range. Relative to Q4, the truing up of the incentive stuff will in a sense disappear and normalize. Obviously, most of our people there was some wage inflation going on with effect from January 1. That gets built into the Q1 numbers as we think of the transition from Q4 to Q1. We're pointing to round about GBP 100 million, just under, for Q1 OpEx. With that, I think we'll move to the Q&A. Thank you.
Thank you.
Just ask if we go around the Q&A, maybe if we all just ask one question first, Francois? We'll have room to get around the whole room.
Three questions to start with. Yeah. How do you feel about the phasing of royalties and licensing during the year? It looks to me that the comps are relatively easy for royalties in Q1 and Q2, because last year it was like single digit, which was very low by your very high standards. Is it more like basically 25% H1 and then going to 15 and maybe licensing a bit different, so quite low start and then accelerating at the end of the year?
Yeah, that sort of shape wouldn't surprise me, as you say, the comps on royalty in the first half of the year are a bit easier. Percentage-wise, that will look definitely flattering. The strength is there in terms of the licensing that's been happening. I think in terms of the overall trend, if you look at absolute, we're on a growth trend here for the future. Licensing is always a lumpy business. I know we've had, you're sort of bored of us telling you that, but it keeps growing 30%. It is a lumpy business. We have licensed a lot of technology over the last few years. As we exit what you characterize as easier comps in the first half, into the second half is more likely to benefit from the version eight.
The growth may not be as lopsided as you may think, just by looking at the eight compares. The exit rate of the eight is clearly going to be much higher than the entry rate, and it's going to be on a, we don't know the trajectory because it depends what consumer enthusiasm is for these products. It's certainly going to be considerably higher at the exit than at the entry, and that is obviously going to help the royalty through the year.
Okay. Back of the room now.
Sorry. Hi, it's Gareth Jenkins, UBS. Couple of quick ones, if I could. Just on the 64-bit while we're on that subject. I think on a recent call you talked about, if it were maybe one of your managers talked about 50% of unit volumes potentially coming from 64-bit through the average of this year. I just wonder whether that's a number that you feel is appropriate or where you see the exit rate in terms of penetration on mobile specifically. Just secondly, I think in the past you've talked about continuing Sorry. No. It is a very quick answer. It's a long question, but a quick answer. Market share gains. Historically, you've talked about historic 3% you've delivered. Is that still the thought process going forward?
On the 64-bit rate, I think that the 50% more points to the exit rate. As we think about that blending through the year, is probably more about 30% on the 64-bit rate. I think that the 50% more points to the exit rate. As we think about that blending through the year, is probably more about 30% of mobile devices being based on V8 on average through the year. In terms of the year-on-year increase in share, I think a couple of percent per year shouldn't be surprising over the next few years. Hello. Keep going to the back. Gentleman there.
Thank you. It's Kai Korschelt from Merrill Lynch. I just want to connect on the royalty growth commentary. I understand, obviously, the comps are different. If I look at smartphones, though, even in your Q1, which is basically the industry's Q4, I believe Apple's the only large volume player that really is shipping 64-bit volume. I think Apple's about 20% of the units. As we go through this year and the adoption in those other 80% of 64-bit materializes, couldn't your royalty growth accelerate meaningfully north of 20%, if you're assuming you're doing 20% in Q1, just because we're still at a very low point in the adoption cycle? Thank you.
The point Tim was making, we're expecting to see more products launched based on version 8 of the Arm architecture. Mobile World Congress is the first week of March. Typically, products start shipping thereafter. Unit volumes tend to be more back-end loaded. You've got various seasonality that goes on. Hence, that's why we're anticipating a greater run rate of V8-based devices handsets towards the end of the year. Exit rate at about half the handsets being based on V8, half based on V7 and earlier, but growing through the year, 30% on average. That's what we mean. David.
Morning. It's Nick James from Numis. You talked about the customer base kind of broadening and engaging with OEMs and having vertically integrated OEMs, you evolving the business model to respond to that. Can you just expand in terms of what that means in terms of the deals that you're doing, the activities that you're doing with those types of customers?
Yeah. As you go talk to people at different points of the supply chain, exactly what they want to do, how they want to go about design, how they want to engage with other people in the supply chain is slightly different. We have to think about what deliverables we provide to a customer who might not want to go all the way to manufacturing, might want to do more of the upfront design. We kind of think about the package of components that we deliver to them, the rights they have to use those, the rights they have for other people then to supply silicon. It gets boringly complex in our license agreements. We're trying to get that right, recognizing that the way in which the industry is evolving is something that we have to respond to.
I guess the question is your value-add higher in that relationship than it is in a traditional chipmaker relationship? Can you take more value out of that relationship?
I don't think that the value per device changes necessarily as we move around the supply chain there. We have a view of what an Arm processor is worth in a chip that goes into a device, and who is doing the design on that doesn't make much difference to that. Yep.
Thank you. Good morning. It's Andrew Dunn, RBC. Actually networking, because you've doubled your share this year from 5% to 10%, and it's all pretty much been V7. First of all, do you have a target for this year for networking in terms of market share? Secondly, when do you think we start to see V8 shipping in volume in that market? Thanks.
We haven't put a specific number out there for market share. Excuse me. Jesus. I will look away next time. I could just hear a market share prediction coming on.
Yeah.
Over you.
That's all right. Yeah.
We haven't put a concrete number out for growth this year, but what I was talking about in the presentation was about the adoption of the technology and how you get significant market share and what we need to do with the ecosystem to really make that work. We've seen good growth. To maintain high growth levels and get to the kind of share that ultimately we want to get to is going to require more of all the other technologies to come together, especially with this move to 64-bit.
Just a very quick follow-up, if I could. You said, I think you can target a sort of 30-ish% by 2018 market share. Do you still stand by that target?
Yeah, we think that's achievable.
Thanks.
Andrew.
Morning. Andrew Gardner from Barclays. A question on the licensing side. At the last set of results, you gave us a bit more detail or insight into some of the drivers beyond just turns and backlog, talking about sort of moving parts behind subscription and renewals and those kind of factors. Given the strength that you've seen in fourth quarter and then perhaps sort of slightly seasonal weakness into first quarter and expecting it to accelerate through the year. Can you give us any sort of further insight into your visibility there around some of these factors, be it renewals or subscriptions and how that's flowing through?
I don't think anything's fundamentally changed. I mean, the longer-term agreements, as we sign those, we know when they come up for renewal. We know the kind of technologies that we're bringing to market. I don't think there's been any change in the sorts of agreements that customers want to engage with us on. We've seen a big uptake in subscription licensing. We saw more architecture licensing. There's been a sort of surge of that. I think things are more kind of stabilized back to normal kind of mix right now.
I wouldn't really use the word seasonal, although we are obviously guiding Q1 license [inaudible] I don't think it's really a seasonal issue per se. I mean, it's more, you look at product deliveries and therefore, how does revenue recognition get transferred from backlog into revenue and what is the timing of your opportunity pipeline of licenses that you've got in view and you're negotiating, when do they crystallize in? I think we've talked, as Simon said earlier, a lot about kind of lumpiness of licensing. I think as we embark on a period of 7%, 8%, 9%, 10% license growth. I think the market's got an 8% compound in there for the next four years. I think we have to expect a little bit more lumpiness relative to this four- or five-year period of 13% license revenue growth.
Hopefully as we go forward, there won't be too much angst and over-analysis of the quarterly licensing move. I think a few years ago, we introduced those charts that sort of showed total licenses signed per annum. I mean, there's five or six years of steady increase. You compare that with the amount of quarterly angst that we had around, "Oh my God, Arm's gone from 20 licenses to 17." I don't want to trivialize it kind of really doesn't matter. It's really more about the timing of when deals turn into revenue and when backlog turns into revenue. The deals themselves, I don't think there's any particular change in the mix of them.
Things like renewals or subscriptions, they have much more inbuilt on backlog than they do on revenue because as we know, subscription five-year deal, full value goes into backlog on day one, gets transferred into revenue linearly over five years. Assuming it's renewed, which it usually is, backlog goes back up here and comes down gently over five years. You've got all those things going on within it.
Sandeep Deshpande, J.P. Morgan. Just a quick question on the server market and other new markets that you are targeting, and what progress you are making in those markets, and what you see particularly in terms of customer engagement in those markets.
In terms of server and progress there, we've seen the Moonshot product launched. That's great. There are two flavors of that. One is based around the Applied Micro chip, and there's another based around the chip from Texas Instruments. People are experimenting and putting those boxes in their data centers and using them to analyze data in really quite interesting ways. The TI-based flavor is being used to do analytics in a way that you just can't do on a conventional server system. It's really demonstrating the benefits of taking an SoC approach to building a chip for a server that's sitting in a data center. That's obviously focusing around the main kind of computing element of the data center.
We have other customers looking at all the other places where intelligence sits in a data center, how you manage the network, how you manage storage, how you just manage data flow around these. Lots of experimentation going on there. The breadth of that product offering on the microcontroller front is really lowering the barriers to innovation because there is a chip that's fit for purpose for so many different end applications right now. If you went to CES, you can see how people are very rapidly innovating and building products and can get them to market quickly using low-cost devices. All the infrastructure that exists in the cloud for cloud-based services. You can use your smartphone to talk to it. You can manufacture it out with a 3D printer on your desk. It is driving a rate of innovation that's really quite fascinating.
Some of the products won't work, some of the products will work, and then they'll move into bigger scale production. I think it's fascinating right now to look at how much innovation is going on and how the advent of all this technology is lowering the cost of doing that, and it's allowing more people to bring their ideas to life.
Sorry, just in that server comment that you made, are you saying that you are more focused now on the networking side? You talk more about fabric and monitoring and storage rather than the core of the server market. Has Arm changed its focus, or is it that you're first trying to take the networking side?
No, these things are going on in parallel. I mean, the beauty of what we do is that our products can be used in combination with our licensees' technology to create a different solution. A Cortex-A57 could be used in a networking application. It could be used in a core processing application, and it's down to the configurations that our customers build and the integration of their own IP and the market that they want to go focus on. The great thing is we can build a product that's suitable for those different areas. For us, what we then work on is the rest of the ecosystem and what do we need to do to support growth in these markets?
What do we need third parties to work on? How do we go and make that happen? It's not like we were focused there and now we're focused there. We're trying to be as broad as we can, and I think our model promotes that. Just behind you.
Thanks. It's Achal Sultania from Credit Suisse. Tim, on one of the slides on OpEx, you talked about OpEx growing more or less in line with the licensing revenue growth over the last few years. Obviously, licensing growth is likely to slow down to more normalized levels. Royalties is probably going to grow much faster than that. Should we expect that the gap between your overall revenues and OpEx should start to widen from here going forward over a long period of time?
I think over a long period of time, that's probably fair. Before the downturn of 2008 and 2009, the model was licensing grows at 10%, royalties grows at 20%, and there's margin leverage. In the last five years, we've, in many of those years, seen very strong license revenue growth and good royalty growth. I think the point is that, over time, there clearly is a correlation between the amount of products you're developing for the market, and therefore the license revenue you're generating, and the amount of investment you make in R&D. I think it needs to be clear that it's not time aligned specifically. We've been painting a picture here of a lot of opportunity to increase and hopefully accelerate our share into some of these nascent markets. Clearly to do that, we need to invest.
I think the general way to think about the Arm model holds good, that investment in R&D relates to license revenue growth rate. If you want to return to the royalty rates that we've typically seen before the first half of last year, cycles aside, there are always cycles, and we always go through short periods of lower royalty growth rate. I think as we return to the royalty growth rates that we're painting here, I think the margin leverage and increase in profitability continues. Over the period we've just been discussing, the operating margin's gone from the early 30s to now getting its head over 50.
Thanks. In front.
Hi there. It's Eoin from Liberum. A question on your royalty expectations for 2015. In your 21% year-over-year growth in royalties, how much implicit share do you expect Intel to get in smartphones in the second half of this year, Intel, Spreadtrum, Rockchip? I'm just trying to bake in what's in your view of 21% share. Do you believe they can get similar share as they got in tablets in the second half of this year? If they don't get there, is there upside to your 21% royalty growth this year?
We're not assuming any significant market share loss in smartphone in any of the numbers and outlook that we've given today. I think we've got an excellent portfolio of technology that's going to enable Arm and our licensees to continue to succeed in that space. We're not baking in some big share loss, and therefore there's not the opportunity if that doesn't materialize. I think over the foreseeable future here, over the lifetime of the products that we have, our shares there are looking good. It's all about the long term, and that comes down to the amount we invest in R&D to make sure that we've got the most competitive technology out there.
Just a very brief follow-up. You gave the backlog went up quarter-over-quarter, year-over-year. How did the backlog change between 2014 and 2013?
It was lower at the end of 2014 than 2013. Like license revenue, backlog has gone up relentlessly for three or four years through the V8 upgrade and licensing cycle. Now, the way the business works, if that backlog didn't go down, then we'd have some very disappointed customers because we wouldn't be delivering any technology to them. The reason you get fluctuations in backlog is because you license new technology that you can't deliver it in full, so it goes partly into backlog, partly into revenue. You deliver it, the backlog goes down, and then you introduce new technology. As Simon said, we've got a rich portfolio and rich plans for introductions of new technology over the next few years, which are going to drive the backlog up. It's going to be a wave effect.
I think in an environment of 8%-10% license revenue growth, you're probably going to see more lumpiness in backlog on a quarterly basis than you saw through the relentless period of accelerated licensing and backlog growth.
The lower end and the impact it has on your royalties growth. Clearly, as you've shown today, for example, there's a GBP 100 phone as well out there that boasts of V8 architecture. Could you maybe just comment on the adoption trends you're seeing in low-end smartphones in terms of Arm-based content and how that is potentially helping you mitigate the so-called pressure or the pressure that we've seen with smartphone ASPs overall? Thank you.
Some time ago, we kind of spelled out how we thought we were going to see the ASPs in the various sectors of smartphone kind of evolve, and how we expected the market to mature into premium devices, mid-range, and entry-level. That is fairly accurately playing out to what we expected. We haven't seen a big shift in ASPs in those three categories, and we have seen an uptake in the use of our technology in those three categories as we were anticipating. You've seen from the devices I showed earlier, low-cost handsets, premium handsets based on V8-based chips.
Over the last year or so, we've seen our licensees who ship very high volumes into this space come out with a portfolio of products to address those different tiers with different amounts of content, multi-core, big.LITTLE, varying amounts of GPU capability and video. For us, it's about keeping our roadmaps up to date, bringing out the next generation of technology so that we can continue to be of assistance to our customers as they're looking to integrate new technology to maintain the ASP of their devices. The innovation that we do helps keep those ASPs high. If we didn't, you would just see a big reduction in the price.
I think this market is playing out, certainly over the last 18 months or so, as we expected it to. There are obviously lower cost ASP devices going into the lower-end handsets. The variation that you see in handset price, I don't think completely ripples through into bill of materials, of course. Just behind.
Thank you, sir. Andrew Dunn at RBC. Cheeky second question. Just on the balance sheet, because you've sort of tantalized us at Q3 by saying that GBP 800 million was roughly the right level for Arm, and you wouldn't expect to grow that. You're at GBP 860 million now. You've upped the dividend, and you talked about a limited share buyback. Those two together won't be enough to keep your cash level down. You probably generate around GBP 400 million free cash this year. What extra steps might you be considering?
I think you're right. We are indicating that we don't plan to build a cash pile for a cash pile's sake, right? You can debate in the context of Arm's growth opportunity and Arm's business model, what is a cash pile. I think it's fair to say that we don't really see any need to meaningfully increase the cash from here. Our plan of record is to increase the payout ratio of the dividend, which you've seen in today's numbers. There's some movement on that, 23% increase in dividend relative to 17% increase in earnings. You've seen us do more buyback. In due course, when we think the time is right, we will introduce, if you like, the next level of our capital structure messaging. It'll be consistent with what we've said to date. It will clearly evolve. It's regularly reviewed by the board.
It will clearly be closely related to our investment opportunity in the business as we go forward. Nothing new and incremental to add today, I think it's fair to say that if you look out over the next few years, cash returns, in addition to the investment in growth, are going to be higher than they've been in the past. The timing and the mechanic of that, TBD. Just behind you.
Thanks. Johannes Schaller from Deutsche Bank. I was just wondering if you could give us a bit more of an update on Mali going into 2015. You had about 40% unit growth, I think, in 2014. Certainly, there was a nice royalty tailwind. We look into 2015, the trends for you at Samsung arguably look quite good. Maybe at some of the other licensees, maybe MediaTek, they have a bit of a tougher start to the year. Could you just give us an idea in terms of unit growth, what we should expect from Mali, and also in terms of deep GBP growth contribution to royalties? Thanks.
In terms of units, when I look at the growth year on year in terms of units from 2013 into 2014 and think forwards based on the licensing activity that's gone on, I think that those growth rates could be quite similar for 2015. We've just announced more technology, more GPU technology. Part of that wasn't just about the graphics processor, it's about having a complementary video engine, a display processor, which allows you to bring all these visual streams together and put them on a screen. That's making our product offering more compelling for our customers. We feel good about the product roadmap. In the medium term anyway, we feel good about the volume that's going to flow through from all of that licensing activity and those units.
We've never broken out the royalty GBP contribution between the GPUs and the CPUs, except in terms of round numbers, percentage gains of extra royalty that we might get.
We shouldn't basically think about a massive acceleration or deceleration from the trends you're currently seeing. They're roughly the same growth rate you've seen in 2014.
Yeah. The growth rate's not bad year-over-year. I think that's kind of steady as she goes.
Got it. Thank you.
A little more.
David Mulholland from UBS. Just following on a bit from the last question and one of your comments earlier on the use multiple bits of the Arm IP.
In the product launch that we did around Cortex-A72 last week, we talked about our interconnect products. At one level, that's wire that hooks all these things together. It's actually much more sophisticated than that. We have intelligence in that interconnect that's helping manage the on-chip data flow, which can result in quite high power consumption. We've put a lot of thought into how the building blocks, which on a PowerPoint slide look really simple and wire them up, how hard can it be? Actually, there's a lot of thought that goes into doing that intelligently that minimize power consumption. Bit hard to put into a presentation.
When we're in front of our customers, they get this, they understand the details of it, and those interconnect products are of growing complexity to manage the amount of data that's flowing around the chips and the size and pixel depth of the displays that we're trying to drive. Seems like it. Any more? No? Well, if there are no more questions, thanks very much for coming today, and look forward to seeing you. Oh yes, one last thing. Thank you. Any more? No? Well, if there are no more questions, thank you very much for coming today, and look forward to seeing you. Oh yes, one last thing. Thank you. Our analyst day is normally in May.
Obviously, with the handover between Tim and Chris, we thought we better give Chris a bit of time to get his feet under the desk and eject all that junk out of Tim's office. We're going to delay that a little bit and do that on the 15th of September. We'll let you all know where the venue's going to be.