Good morning, everyone. Welcome to Arm's full year 2013 and Q4 2013 results. We're going to do this morning, I'm going to talk about the business, the progress we've made towards our strategic objectives. I'm going to hand over to Tim, who's going to talk through some of the numbers, then we'll go to Q&A. I'll take as read the usual cautionary statement, we'll start talking about the business. I think it was a very exciting year for Arm in 2013. We achieved some great milestones in our business with our technology. I think one of the most significant being that our licensees sold over 10 billion chips containing Arm processors. That was a key milestone for us that takes the cumulative total now to over 50 billion chips containing Arm processors since the company was founded just over 23 years ago.
That's a huge number. It represents a massive diversity in the end markets in which Arm is designed. You see Arm technology being used in big, high compute performance applications such as servers, in networking equipment, in phones and smartphones and tablets, obviously, also a growing proportion of Arm processors in very small, very low power, deeply embedded microcontrollers that are in a vast range of new and exciting markets. Mobility, smartphones, tablets remains a very important market for Arm. A year ago, we were forecasting there would be about 1 billion smartphones sold in 2013. That indeed seems to have been the case. Our estimates show about 1.1 billion smartphones were sold, taking the total now in use around the world to about 2 billion. The growth of smartphones, the growth of tablets has been very strong.
We've seen tablets out shipping PC notebooks in 2013 as well. These devices are the way in which people want to interact with the internet, want to interact with each other. I think there's going to be a continued growth of these markets that we'll come back to. In 2013, we also saw some key milestones in terms of our technology and the growth of new markets. As far as our technology goes, we saw the first Arm ARMv8 architecture product shipped. These are our processors that add 64-bit processing capability, we received our first royalty from those. We saw our first royalties from 20 nanometer physical IP as our partners moved to advanced processes. We've also seen growth in new markets for Arm. There's a lot of buzz and hype at the moment around the category of wearable devices.
Recently at CES, you just saw a plethora of new devices all based around Arm processors, which are exploring this product category. I think we're going to see a lot of growth there, in years to come. That is a subcategory of the Internet of Things, which again, is in its infancy, but I think set to grow very strongly. Right now what people are using are Arm-based processors. They're using the chips that have been designed by our licensees. They're embedding them in devices, they're very quickly getting to market with new and exciting products. As these markets mature, we're expecting that as the Internet of Things runs on Arm today, a vast majority of that is going to run on Arm, in the future. At the other end of the compute spectrum, we've seen progress in enterprise networking and servers.
All of this data that smartphones and tablets and IoT devices are producing needs moving around. It needs storing and processing. We're seeing traction for Arm technology into these markets as well. With the success of this business, that's led to record revenues, record profits, and we've been continuing to invest in our technology. We added over 400 people into the business in 2013 to help us capitalize on the new opportunities in front of us. It was a very exciting year, very busy year. Q4 rounded off that year with great progress. A very strong quarter, again, in terms of licensing. We saw 26 processor licenses, and that helped create 121 licenses for the full year. Within that, there were four Mali licenses, and in the quarter, our licensees shipped 2.9 billion chips. That's the most ever shipped in one quarter.
There's a lot made at the moment of the slowing of growth in the high end of smartphones, and indeed that appears to have been the case. Despite that, still, we have record unit shipments because of the vast number of end markets in which Arm technology is used. The performance of the licensing and the royalty growth led to 15% year-on-year revenue growth. Through that performance of the business, we're able to keep investing in the business while we increase returns to our shareholders. We've just announced we increased our dividends by 27%. If we look into what's behind royalty at the moment. Typically, you're used to hearing us talk about how Arm has outperformed the industry, and indeed through 2013, that was the case. The semiconductor industry only grew by about 1% in 2013, yet Arm's royalty revenues grew 19%.
We've continued to strongly outperform the industry. In Q4, that outperformance was lower. Arm's royalties grew about 7% year-on-year. The industry grew about 2%-3% in that time. Overall for the full year, another very strong performance for royalty growth. Again, you're used to hearing us talk about progress within mobile and progress outside of mobile. Now over half the chips that are shipped by our customers are used in applications that aren't mobile devices. They aren't smartphones or basic cell phones or tablets. They're in things such as consumer electronics. It was about 2 billion chips shipped into products like digital cameras, digital TVs, DVD players, vast range of consumer electronics, and about three and a half billion shipped into enterprise networking applications and embedded computing, these very small microcontrollers I was talking about.
Smartphones, though, remain a very important and a very valuable market to Arm, and one we're expecting to continue to grow. Growth in 2013 was strong. We're expecting 15%, maybe as much as 20% growth, in 2014. Long term, a compound annual growth rate of about 10%. This market is going to continue to grow and continue to be profitable for Arm. Despite the slowing of growth in the high-end, what we're expecting over time is a rapid growth of entry-level and mid-range phones priced such that literally billions more people can get access to these devices. That creates a very valuable opportunity for Arm. We look into that. You can see the graph on the right there showing our expectation of growth rates of these different market segments of smartphones.
At the top, the premium end, we're predicting about a 4% CAGR out to 2018, with much higher growth rates at the entry-level and at the mid-range. Right now, if you go out to buy a smartphone, about 95% of them have an Arm processor in the application processor. Almost all of those now are Cortex-A processors. You'd find it really difficult to buy any sort of phone that doesn't have an Arm processor in its modem. In combination, you're going to struggle to buy a phone that does not have at least one Arm processor in it, and usually many more. As these markets grow, we're seeing a greater opportunity. About 70% of the chips inside smartphones integrate both the modem and the apps processor, and that helps reach a lower price point that is helping enable growth at the entry-level and the mid-range.
The common complaint we get is that while those devices are cheaper, you're going to take less money than you would at the high end. While it'd be a glorious thing if everybody on the planet spent $700 on smartphones, it just isn't going to happen. The growth of the entry-level, the growth of the mid-range is a great thing because they represent a big opportunity for additional Arm content in these devices. What they're replacing are very simple voice-only and feature phones, which have a single Arm processor. They're going to be replaced over time by smartphones. Some of them may be more basic than the high end that we know today. There's more Arm content in there. Cortex-A processors, graphics cores, the chips that build on advanced processes, which often use our physical IP.
As we go forwards, we'll see big.LITTLE configurations of processors, again, increasing Arm's royalty, the adoption of 64-bit processors across the entire product range at some point in the future. This growth represents a very valuable revenue potential for Arm, and we believe mathematically supports our estimated growth of revenues from smartphones out to 2018. The other side of mobile computing right now is tablets. Tablets had a very strong year in 2013. As we can see there, Tablets and ultra-mobiles between them out shipping laptops. Again, the growth rates expected in this category are very strong. We're expecting a solid growth in 2014, and out to 2018, again, this five-year window that we look at, a 20% compound annual growth rate of tablets. You can see how the growth just starts to swamp conventional legacy laptops over time.
When you look at desktop PCs, again, you see the growth there very slow, minus 5% CAGR. Desktop PCs aren't going to disappear. There's always going to be a need for those. The way in which most people are using computers most of the time is via smartphones, via tablets, and that's why the growth rate is so strong. Again, virtually all of those are using Arm processors today, and we expect that to continue. Again, the market splits into different categories. There will be premium tablets. There will be low-cost tablets. The low-cost tablets today are all using Arm, and that is enabling billions of more people to get access to this technology. One area I've been particularly pleased with progress on in 2013 is the adoption of Arm technology into enterprise networking.
This is a big market today and will grow to about a GBP 20 billion silicon TAM in 2018. That's about the same size as smartphone application processors. This is a big semiconductor market. It's for products that form the infrastructure of the cloud and the high-performance switches that are required to move all the data around the internet. This is a market that's served by a relatively small number of semiconductor companies, and all of them, or almost all of them, have already announced products based on Arm technology. In fact, four of them are shipping and paying royalty. What that led to in 2013 is about a 5% market share for Arm in enterprise networking, and that may not seem like much, but that's 70 million chips. That's about twice what it was in 2012. The growth rate is strong.
We've been working with our partners. Our partners have been working very hard on designing the right technology, winning the sockets in the enterprise networking equipment companies, and we're starting to see growth. We expect to see more products brought to market based on Cortex-A15, based on ARMv8 architecture in 2014. This is a market we've had good success in 2013, and one we expect to continue to grow. Similarly, in servers, I think 2013 and very recently has been a good time for progress for Arm technology in servers. This was a strategy that was always going to take a long time to play out. There is a discontinuity that comes from the changing workloads of servers, and that creates an opportunity to change the way in which servers are built.
To move to much more customized chips that are specific to the task that the server is trying to process. That is what the Arm partnership does very well. Taking high performance processing technology from Arm, integrating it with IP and knowhow in our customers to produce a tailored solution that's optimized for the workload. If we look back over the last couple of years, prior to 2013, we were in a very early pioneering phase where some of our partners were looking at this discontinuity and looking at how to take advantage of the opportunity that came from that discontinuity. Through last year, there were about 15 different companies building chips and looking at designs based on Arm technology. What we're expecting this year is that growth to accelerate, units start to ship, and based on all the software work that we're doing, real servers actually deployed.
I think what was one of the most exciting developments very recently was last week we announced a partnership with about 12, 14 other companies, that really demonstrates the Arm ecosystem in action. What we've been looking at for a long time now, this activity has been going for about a year, is how to ensure some degree of standardization to make it much easier to port software to an Arm-based server. In conjunction with some big companies that play in this space, people who build servers, people who buy servers, people who write software for servers, a collaboration has formed to create a standardization effort called the Server Base System Architecture, that many Arm-based server designs will conform to, and therefore make it easier to port and run software.
Of a couple of announcements recently, AMD just announced their Opteron A1100 series that's compatible with this new standard. Applied Micro, similarly, their designs are compatible with this standard. That's going to make it easier for the industry to adopt Arm-based servers and get software running more quickly. This is a great demonstration of the kind of thing Arm can uniquely do. Help work with people who compete with each other to solve common problems and make it more efficient to deploy new technology. The other end of the computing spectrum are these very small microcontrollers. Through last year in this embedded space, there were over 3 billion chips sold containing Arm embedded processors. A lot of those are based around the Cortex-M series.
This is a series of products we've designed specifically with this market in mind. There's now about 160 companies with licenses to Cortex-M technology. What many of them are doing are building devices that integrate a sensor, which is largely an analog device that's sensing something from the environment, be it temperature or humidity or the road shaking. Integrating those sensors with a processor that can make sense of the data that's coming from the sensor, with a radio that's performing some form of connectivity up to the cloud. Arm's share in MCUs is about 20%, over 20% now. What we've seen over the last few years is a migration from very, very simple, what's called eight-bit processors, to much more sophisticated 32-bit processors, i.e. Arm. That gives more compute power. It makes it easier to write software.
It makes it easier to maintain software. Because we're able to deliver 32-bit performance and all those benefits at very, very, very low cost, the fastest-growing segment of MCUs is 32-bit. We're very well placed to take advantage of that. Similarly, with the radios from our heritage in mobile, many of Arm's partners are using Arm processors in their radio devices and processing the protocol stack, processing the data that comes in, and forming the connectivity of this sensor MCU combo up into the cloud. There are many examples of how these devices are being used on display at CES a couple of weeks ago. Everything from basketballs to watches and wearable devices, just a whole plethora of end devices that are being made. This category, I think is going to grow very, very strongly. We're seeing lots of experimentation in wearables.
Which ones take off, which ones fail is hard to say. I think what's really interesting right now is just how much experimentation is happening. It's enabled because the access to the technology is very low cost. The access to cloud-based storage and processing is very low cost. All of this is coming together to create an environment where many product experiments can be run simultaneously, and we'll find out which ones are the most successful more quickly. This is a really interesting space. Wearables is going to grow to a very large market. More than 200 million units is our expectation in 2018. We expect that Arm will have a very large market share of that.
As we look generally at the expanding opportunity for Arm, we see within smartphones, application processors, that's a large silicon market today, about GBP 13 billion, growing to about GBP 20 billion in 2018. There are multiple opportunities for Arm royalties within that segment. We have a very large market share today. As the market grows, there are more opportunities for Arm royalty-bearing units as these devices move to 64-bit and as they become more sophisticated. Enterprise networking, as I mentioned earlier, grows to a market that's about GBP 20 billion in 2018 as well, from about GBP 13 billion today. We have a very low market share there right now, as I explained earlier, that's grown from about 2.5% the year before. As those devices are built out, again, there's more opportunities for Arm royalty. Cortex-A15 moving to version 8 architecture, multiple processors in these very sophisticated devices.
Within embedded, this is a market where we have a pretty respectable market share today, about 22%. The size of that market today is about GBP 14 billion of silicon. Now, these devices are very low cost, so that's billions, literally tens of billions of different devices. The semiconductor TAM grows again to about GBP 20 billion out in 2018, and we expect to see Arm's market share grow, again, generating very profitable royalty streams from that market as it grows. These are the three ways that we look at the expanding opportunity for Arm. The Arm content in all of these devices is different, but represents a big opportunity for Arm's royalty stream to continue to grow. In summary, I think 2013 was a very exciting year for Arm.
We made great progress on a number of fronts in delivering our technology, in working into these new and exciting growing markets, and that has led to a very strong performance of our business. Right now, the smartphone market may be slowing at the high end, but it's still a very large and very valuable market, and I've explained how, as the entry level, as the mid-range grows, that represents a valuable royalty stream to us despite that growth and slowdown of the premium end in the second half of the year. The design wins that we've been working towards in enterprise networking particularly, and servers starting now, are a great opportunity for future royalty growth, and that embedded market is growing very fast. The uptake of Cortex-M at that end of the market has been really strong, and we're seeing just lots of devices start to ship.
All of this blends together, I think to a great opportunity. Arm's fundamental business model is intact. We're able to respond quickly to new opportunities as they emerge, and I think that creates a great potential for future very strong business as we had in the past. With that, I'll hand over to Tim.
Thank you, Simon. Good morning, everybody. If I look a bit miserable, I've developed a little bit of an allergy over the last 12 or 18 hours. If you see me sprinting from that platform out that door, it's not because I don't like you or I'm not enjoying myself. I think we're okay for the next hour or so. As usual, I will be fairly brief. There's a lot of financial information in the release. There's a lot of financial information in the slide deck to help you with your models. I'm just going to focus on the brief highlights of Q4, talk a little bit about the exceptional item that most of you might have noticed going through the Q4 results, look at the full year, and then focus in on the outlook. For Q4, you've read the headlines. Revenue growth at 15%.
Again, very strong licensing. As I said, well ahead of expectation. Royalty, slower growth, little bit behind expectation. Combination, about $10 ahead in dollar terms on the market. Despite the normalized OpEx in Q4 being a little bit higher than we and you were expecting, that's driving PBT growth of 19%. In that GBP 88 million, you've got some mark-to-market impact of the normal revaluation of the monetary items and the forward contracts, et cetera, and a little bit of bad debt provision, some truing up of bonuses and commission payments. It's a little bit higher.
Most of that is not run rate. When we look into Q1, despite the fact that obviously there's some wage inflation coming in from 1 Jan, which is when most of our folks get their pay increases, the guidance for next quarter's OpEx is obviously somewhat lower than the Q4 outturn in the sort of GBP 84 million-GBP 86 million range. The combination of those two factors has driven earnings growth of 30%, benefiting to some extent. The difference between the 19% and 30% is to a large extent made up by tax, because we're now getting, if you like, the first 60% of the benefit of the patent box regime that was introduced last year. Our tax rate is down from late 20s to around 20%. We'll talk about the full year a bit later on that.
You may have seen in the release in the IFRS numbers, there was an impairment charge, exceptional item, non-cash. I just wanted to give you the context. Briefly cast your mind back to this time a year ago. Many of you will recall that MIPS Technologies had an activist shareholder on their register, and it was very clear that business was going to change its shape. Really the headline value in MIPS was the patent portfolio. That is a patent portfolio that had been developed over the last 20 years, very close to our heart, 32-bit processors. We, along with other companies, were not very keen about the notion of that portfolio getting into the hands of, for example, a patent troll, because over the long term, that could have caused quite a lot of disruption to the Arm ecosystem.
We felt it was very important to neutralize the potential impact of that, and we clubbed up with a couple of handfuls of other major technology companies, and acquired rights to that portfolio for GBP 350 million. Our share, you may recall, was GBP 167.5. We viewed that, and how we told you about it last year, was effectively that was a lifetime insurance policy as far as we were concerned, against costs that would have been incurred down the road had that patent portfolio got into the wrong hands.
The way that was accounted for, is in the balance sheet split between a GBP 67 million intangible asset, as it says there, which is being written off over eight years, which is the average life of the patents, and an available-for-sale financial asset, which represented the fact that typically, these trusts that hold these patents often embark on licensing programs that yield cash, and the cash is returned to the participants. Obviously, given Arm's contribution, we were well-placed to receive that cash. Rather than it being written off at that point or a larger intangible asset, it was an available for sale.
During this year, for a number of reasons around the participants in the consortium and around discussions at what the parameters of a licensing program might look like, and the type and size of companies who might take licenses, it was decided by the trust to not license the program, to not do a license program, but to put it up for sale. For us, we see incremental long-term strategic benefit in owning those patents outright than having the rights to license them. For a fairly nominal incremental cash outflow of GBP 4 million, we have now brought that patent portfolio inside of Arm, and it's adding broadly 500 patents to an existing portfolio that was about 3,000. It has the effect of, in accounting terms, of impairing that financial asset.
In cash terms, what's actually happened is the lifetime insurance policy we bought for GBP 167 million for an extra GBP 4, we've now improved. Because there are lots of things you can do when you own patents in terms of defending your position out into the future, that you can't do if you only have rights to them. That's the detail on the exceptional item, non-cash and non-recurring. For the full year, briefly, 22% revenue growth overall, 32% licensing. I mean, licensing, as we all know in here, is a revenue stream that, in the history of Arm, has grown closer to 10% than 30%. For the last four years, it's grown to 30%, which I think represents the fairly dramatic increase in Arm's addressable market.
The fact that a very wide range of semiconductor companies now feel as though they have the opportunity to deploy our technology in multiple markets. That has been driving very strong licensing growth. We have been fairly serially beating licensing expectations. But it remains a lumpy concept. We only do 20 or 30 licenses a quarter. But it looks strong, and we will cover that when we look at the outlook. Overall, for the full year, as Simon says, another strong year for royalties overall. Clearly, as Simon said, we have seen some slowdown in the back end. Again, we will look at that when we come to outlook. 22% revenue growth, 32% growth in normalized PBT, in spite of the fact that this is probably the singular biggest year of investment that we have made in the business.
We've increased our overall headcount by over 400 people this year, and we've invested in a sort of business infrastructure that is absolutely crucial for supporting this rapid growth that we're enjoying. So it has been a big year of investment, but the operating margin has gone up from about 46% to 49%, notwithstanding that investment. Again, because of the tax, the 32% increase in PBT becomes a 40% increase in earnings. Effective tax rate on a normalized basis in 2013, about 20%. Expectations for 2014, about 18%, because we get-- Well, the patent box regime is being implemented over five years. You got 60% of the benefit in year one and another 10% in each of the next few years. You can expect our tax rate, as I've said before, to gradually edge down over time. Strong year for cash. Net cash generation over GBP 344 million.
We end the year with GBP 700 million net. No debt. Simon said we've increased the dividend by 27%. We've also said in the statement that we intend to-- We're confirming that we intend to maintain a flat share count over time. In reality, if you look back on a nine-year view, which is-- I know there are some faces in here who've been throughout that journey, but on a nine-year view, we actually have a flat share count. It just came in one particular period in 2005 to 2008, when we bought back 60% of the stock. The issued share capital is now just coming back to the level of where we started that program. So the share count's been flat over nine years. We intend to keep it flat. In reality, what that will mean is a limited share buyback program to achieve that.
Which will be sort of an ongoing item. Quickly looking forward then before we get into Q&A. The release said the order backlog, which as we know has been growing very strongly, was marginally down Q4 versus Q3, but up about 17% end of 2013 compared to the end of 2012. A combination of that high order backlog and looking forward into our opportunity pipeline of licensing looks encouraging, and we see another positive year for licensing in 2014. Of course, what this means is, licensing is the key leading indicator for Arm's growth and for Arm's value. There are no royalties without licensing, and so licensing is completely key. The fact that we've grown licensing broadly 30% per annum over the last four years, I think bodes very well for royalties in the medium and longer term.
What we're saying in this particular year, despite some of the slowness at the high end that we've been seeing, because of some of the penetration we're making to the new markets, because we still see mid and low-end phones growing nicely next year, we see the full year growing at a similar rate that we've seen over the last three years. Which for those who have added up the numbers and divided by three, as one way of assessing that, it's about 19.3% or 19.5%. What we're actually saying is similar rate to our normal growth in royalties is what we expect. Therefore, overall, we expect the full-year dollar revenues to be in line with market expectations, which are currently in the sort of $12.80s, high $12.80s, $1.28. That said, assuming semiconductor industry improves as generally anticipated.
I think most of us around here are expecting a stronger second half than the first half. We see no reason why that isn't going to materialize for us. Yeah, we'll obviously have to see because early on in the year. With that, I will throw it open to the floor.
If I can just say before we get into Q&A. Out of respect to everybody in the room, if you'd ask a question as opposed to 17, I'm not just looking at you, Francois. We could move around and come back to another go. Go ahead, Francois.
Well, I will have 16 questions today. Right. I would like to understand what's going on with the ASP going from cents 4.8 or cents 4.9 to cents 4.5, so it's kind of 10% down. I understand the mix effect, but at the same time, I think Cortex-A shipments have doubled and the embedded or microprocessor microcontrollers have gone up by only 35%. If you could help me with some size, those numbers that would be really helpful.
Yeah. As you say, it is about mix. That Cortex-M growth is in chips which are much less expensive than typically what Cortex-A goes into. As we've seen, there's slowing of growth in the premium end of smartphones. One would expect that the ASPs will come down a bit.
As well in Cortex-A.
Well, as we all know, some of the devices are being built for the low-cost handsets sell at lower cost than some of the chips for the higher end. They are physically smaller, they integrate less technology, they sell for a lower ASP. That's good to be expected.
Thank you.
Yes.
Sorry. Thanks. It's Sumant from Redburn. I guess if you don't mind, could you give us how many number of your licenses, both in number as well as in value, are coming out of China and Taiwan? The related question I actually have is, if I look at your licenses growth for the past three years or the three years before that and compare it to the royalty growth, and I look at the licensing surprise this time around, I'm just wondering whether the royalties you make with the incremental licenses you're getting in the last two, three years is lower than the average royalties you got from licenses previously.
That's an interesting question. I think that's not necessarily the case. A lot of the licensing that we've done over the last couple of years has been version A of the architecture, both the architecture itself and the processes that were built around that. They are typically going into higher ASP devices. I think one would expect the royalties from that to be at the higher end. In terms of the split of licenses into China, I'm afraid I don't have that off the top of my head, but roughly half our total business comes out of Asia. China is a growing area for that. We're seeing increased design activity in China. We're seeing the sophistication of the chips designed in China going up and up over time. I think that's partly a overall trend of the technology industry.
Could I assume more than a quarter?
I'm sorry.
Could I assume more than a quarter then from China and Taiwan? In terms of numbers.
Yeah. Maybe. I think if the general line of inquiry is, has your licensing growth really been fueled by China, is that likely to lead to a lower conversion into royalty over time? I think the answer is not, we don't see that. We've been licensing in China for really many years. In the last two or three years, material royalties are now being earned from Chinese licensees. These were licensed four, five, six years ago. This is not a new phenomenon. I think if you look into this, what underpins this 30% licensing growth, it is very broad in territory, it's very broad in
The product that the market is being aimed at. I don't think there's a particular China focus to it, but it's helpful.
Yeah, thanks. Didier Scemama from Merrill Lynch. I'll ask two questions. First one, can you maybe just talk about your Processor Division royalty guidance, 19%, 20%? Maybe just give us the puts and takes in that. How much of that is driven by share gains in non-mobile? How much of that is driven by royalty rate expansion? How much of that is driven by specific share gains in maybe graphics or other elements? I think, given your recent performance in revenue growth, probably some people will look at that number and think it's a bit of a stretch. The second part on licensing, I was just curious to have a bit of color from you on the comments you made on the press release regarding licensing that has been done with telcos and software companies. How meaningful is that going forward? Thanks.
Sorry, I've actually forgot the first part of the question. Oh, sorry. As you know, Arm growth is about layers. You heard Simon say that it is our expectation that smartphones generally continue to drive very meaningful growth in Arm's royalties, which is broadly consistent with the overall royalty growth rates we see. In a five-year view, you've heard us talk about driving a 15%-25% overall royalty growth from smartphones. We do see strong growth in 2014 in the medium and low end. I think you've got smartphones are still contributing well. I think Simon touched on a lot of things like enterprise networking are now starting to be meaningful in 2014. We've signed 200 Cortex-M licenses in the last few years. A relatively small proportion of those are shipping to date. There's a lot of them just about to come to market.
I know it's low end. It's high volume, low value, but it all adds up nicely. I think towards the back end, we're going to start seeing some contributions from the V8 in higher priced, higher chip price. It's a combination of those factors that support our confidence.
In terms of the growth of type of licensee, that's something that's been going on for some time. Arm's history is obviously about licensing semiconductor companies. Over time, our commercial relationships with other people in the supply chain have grown. We've seen really just an extension of that, us developing relationships with other people who are either buying silicon or contemplating doing designs themselves or wanting to engage differently with the overall semiconductor supply chain as they look to optimize their products, maximize their differentiation. We're a partner in that.
Yes.
One question to each of you, if I could. Simon, it's Gareth from UBS. Simon, I just wondered if you could talk about Calxeda's exit from the market and I guess whether you see more of an opportunity in microservers and sort of the data center opportunity rather than necessarily enterprise class servers in the kind of medium term. One for Tim, just on the OpEx progression through the year, should we expect a kind of steady progress in OpEx through 2014 upwards and related to that headcount forecast for the year?
In terms of the opportunity around servers, as I talked about in the presentation, I do see a big opportunity for us there. I think when I look at the amount of activity that's going on in that space, I'm confident the products are going to come out, and they're going to start to ship, and that they're going to grow. Calxeda was a pioneer in that field. They entered this market very quickly, and like many startup companies they didn't get there. I don't think that means that's the end of the foray into servers with Arm licensees. Quite the opposite. You've seen from the press release about the System Architecture that's just come out. Big players are looking at this space. We talk to lots of people about the adoption of Arm in the data center because people want a different approach.
They want to take advantage of this discontinuity about workload, and they want a different approach to solving the problem. I think the opportunity is very real, and I think the Arm partnership in totality has made great progress on that in 2014.
Generally on OpEx, we are still investing in the business. Therefore from the Q1 base, you would expect OpEx to gradually go up quarterly with other things being equal. To what extent we invest in the business in this 12 months will depend to some extent on our view of the market and the speed of the appetite and how the overall market's playing out. We obviously have got a lot of control or almost total control, really, over the timing of our investment in people. Certainly, there are a lot of opportunities here that we need to seize. We would expect it to be an ongoing investment period in 2014.
Just coming to the gentleman in front here.
Morning. Nick James from Numis. Just on the smartphone thing, I guess we're having two changes in the market. One is the growth more from the low and the mid end. Also, there's a change in the supplier base to the Asian chip makers from the Western chip makers. Asians tend to accept lower gross margins, lower ASPs. Just wanted to understand how you're thinking about this impacting on chip prices and royalty rates.
Yeah. Primarily, the way we think about it is an overall expansion of the end market. I think it's a great thing that you can today buy a smartphone out the door, zero subsidies for $50 in China. That is going to put smartphones in the hands of many more people and drive an overall long-term upgrade in both the handset devices themselves, in the network infrastructure that's going to process and move around this increased amount of data. We look at the long-term trend of an increased number of units out in the field and see that as the good thing. There will be pricing pressure. It's a maturing market overall, so you would expect that. With these low-end devices, there's just more opportunity for Arm content. The volume is good.
The sophistication of the product is going up consistently over time, is a good thing for us. Yes.
Thanks. It's Arjun Sukania from Credit Suisse. Just a clarification on the smartphone question earlier. Are you basically saying talking about 10% CAGR growth in smartphone volumes, and you expect your smartphone royalties to grow at about 15%-25% CAGR over the next four or five years. Given that what we know about the volume growth, and obviously bulk of the growth is going to actually come from the lower end of the smartphone market, would it be fair to say that all this growth in smartphone royalties is actually going to come from increase in royalty rates as opposed to increase in the application processor market, size of the application processor market? Would it be fair statement to say, Segars?
Well, I think it's mainly a growth in the Arm content. As I said in the presentation, right now we're looking at very low-cost smartphones replacing basic voice-only phones, which have a single Arm processor in them doing not very much. As the smartphones get more sophisticated, there's opportunity for a new Cortex-A processor replacing probably an ARM7TDMI that was designed 20 years ago. There's opportunity for graphics processors. Over time, I think the whole smartphone market moves to 64-bit. It's a blend of more opportunities for Arm royalty-bearing processors within a smartphone that today doesn't even exist, then over time, the royalty rate going up as those 64-bit comes in. Yes.
It's Andrew Dunn from RBC. If I can just ask the smartphone question perhaps a different way. Your outlook for smartphone royalties hasn't changed in the last two, three quarters for us, I can tell. You did single out high-end slowing in the second half of last year. Were there any perhaps customer or industry specific issues in the second half of last year that you perhaps wouldn't expect to repeat going forward? Thanks.
Well, I think what we've seen is, for now anyway, a maturing of that high end. Now, I wouldn't write off the possibility of innovation at the high end at all. In fact, I'd be surprised if there isn't continued innovation at the high end, which creates demand for those products. For now, we're in a period where constant replacement seems to have slowed down. Therefore, people are living with their devices for longer. Isn't necessarily going to be the case that will be like that forever. That's the period that we're in at the moment. That's been anticipated through last year, and we saw it really happening in the second half. Yes.
Thank you. Andrew Gardiner from Barclays. Just a question around licensing. Clearly a very strong year in 2013. You're guiding for another one in 2014. Can you just give us a bit of color about some levels coming out of backlog? We're off the record level in the fourth quarter. As we look into 2014, can we continue to expect on the order of 20 to 30 per quarter and similar kind of booking levels to be coming through from that? Can you put it in the context of obviously a very good couple of years for V8, big.LITTLE, Mali coming on stream that kind of thing. Is there enough activity going forward to continue to support that backlog? Thank you.
Typically, when we look forward 12 months, we would expect about 50% of target license revenue to be in opening backlog. Some quarters it's 40%, some quarters it's 60. I think 50 is good. Although the backlog is a little bit off marginally Q4 versus Q3, if you actually look at how the backlog has developed over the last three or four years, it's actually grown faster than the license revenue. Obviously, in a sense, that gap needs to be changed. I think combination of that and combination of how you would expect us to look at this is, which is by product, by customer, by value. What does our pipeline look like in 2014? That's what gives us the confidence. I think V8 is relatively early in its licensing cycle. We've done many more V7 licenses than we have V8.
Although it's been grabbing a few headlines, it's still very early in its licensing cycle. That will be a continued grower. Cortex-M is by no means, for example. There will be new processors coming out all the time, as you know. Having said that, I don't think we've ever tried to position license revenue as a 30% grower forever. It will revert closer to historic growth rate before 2009, which you've often heard me talk about as mid-high single digits. I think there's going to be a journey from what we've seen to there over the next two, three years.
Simon Schafer. Thanks. Actually, a follow-up question, Simon. I think you said your backlog can grow significantly faster than even licensing. Of course, there's been a strong amount of momentum as a lot of non-mobile people have been signing up your architecture. What is the risk that backlog actually starts to descend rates a lot in an environment that you alluded to that licensing only grows mid-single digits and backlog actually starts to go down? I'm just wondering just because when you look at a 15-year trend line of licensing, the run rate of GBP 100 million is standard deviation above what we're used to. Again, just any concerns that backlog actually may start to fall?
What we'll probably move into is a world where backlog is a bit more lumpy and goes up some quarters and goes down. We've seen a fairly relentless period of backlog growing really very fast. Some of which has come into license revenue, obviously a lot hasn't. I think part of the move from growing license revenue at 30% to growing license revenue at, let's say, 10 in the out years, part of that will be a kind of flattening out of the backlog. In the end, ultimately, of course, backlog has to grow in sync with license revenue, albeit removed in time.
From the gentleman in the front.
Thanks. It's Vijay Anand from Espirito Santo. A question on the networking market. You have a 5% share today. The analysts say you talked about a target of 25%-30% share by 2017. I guess barring Intel, pretty much all the major networking semis have licensed Arm. The question is, are there any major roadblocks or major uncertainties which can, I guess, prohibit you from hitting that target? Or is it just about steady share gains from here? Related to that, the vast majority of the networking market is based on 64-bit. Would it be fair to say that the average royalty rate in the networking market is going to be higher than what we've seen in the smartphone market?
Yeah. To your question, right now our market share is small, though the chips that we expect Arm-based chips to replace are using a variety of other architectures, PowerPC, MIPS, some Intel. As you saw on the slide there, most of the people who are playing in this space are using Arm technology, and we would expect share gains over time. As you also point out, a lot of those applications run 64-bit code. It's only very recently that Arm 64-bit products are out there. Again, that's going to take some time to come through into silicon and to broaden out the number of sockets that can be addressed. As that happens, obviously with a higher royalty rate for 64-bit technology, that benefits us. In the meantime, there are a number of designs out there with multiple Cortex-A15s in them.
That generates a good royalty per chip, and those chips are physically large and obviously have a higher ASP that helps push the average up for these, and that's important. Those chips, whilst they're much lower in volume than, say, the embedded space, have quite a lot of Arm content in them, and I think represent a good profit stream for us in terms of royalties.
Hi, it's Adam Bowler from Deutsche Bank. You previously talked about 15%-20% PD royalty rate outperformance versus the wider semi industry. You obviously came in at the higher end of that this year. When we look ahead to 2014 based on the outlook that you've given us today, is that still the case or should we think 15% and slightly lower, if anything, now?
Well, I think if you take the 19% that I talked about and add your forecast for the semiconductor industry, you'll get the answer. Probably that implies around 15%, right? I think these are sort of medium-term measures and trends we're seeing. Arm does not grow its royalties at that rate every quarter. I think I remember the first half of 2012, it was growing at 5% or something. Certainly we don't see any reason why our long-term relationship to the industry growth would change. In fact, I could probably build a case for why it would improve given our penetration across these multiple markets, notwithstanding the fact that high-end smartphones and smartphones in total slowed down over the last two or three years.
Don't forget, in a period where smartphone growth overall has already been slowing from three or four years ago, Arm's royalty revenues have been growing at a similar rate. We've already been more than making up for the slowdown in smartphone growth in penetration in other markets.
Ben from .
Thanks. Ben Gardner, Arete. Given how fast Cortex A has ramped over the last three years, 100% penetration of smartphones, 18% units and 500 million units per quarter, how quick do you think V8 will ramp based on your understanding of when Android will adopt it? If you can frame that for 2014 and 2015, that'd be very helpful. Thanks.
I think the growth of V8 and the growth of Cortex A aren't necessarily. The growth rates aren't necessarily going to look the same. Cortex A is used in a very wide range of end products where most of them may move to 64-bit over time, but there's no burning need to do that. I think where we'll see V8 adopted is strongly in networking, in the server space, and over time, across pretty much, I think, the whole market of smartphones. Once software starts to move over to 64-bit, it's just easy to be compatible or easier if every device is compatible and running the same version of the software. I think that's a long-term trend. I think given the general very broad applicability of Cortex and our 32-bit architecture, I still expect very strong performance there, and with a more gradual shift to 64-bit.
Good.
I'd like to go back to royalties for Q4. It'd be a feel for Q1 if you've got any. Just looking at the thesis for Arm over the last, I think the Analyst Day 2011, was talking about the expansion of royalty revenue, more recently V8. Your very recent royalty revenue growth has not been totally de-correlated really from unit growth and the deceleration we saw in end markets. I'm just wondering, have you seen any meaningful impact from 64-bit in Q4, or is that really something that's going to kick in Q1 going forward? Is there anything slightly different underlying that we should be aware of that would justify the deceleration in your royalty revenue growth?
Well, I think the total number of units containing V8 in Q4 was small. Compared to 2.9 billion, it was a small amount. To start seeing that come through appreciably is going to take some time. I expect the V8 shipments will increase through 2014, and it's probably beyond then that we'll really start to see meaningful volumes and be able to spot that. Given the vast number of Arm-based chips we're selling to other markets, 2.9 billion is a lot to make an impact on. Yeah, I think we're going to see it over time. Exactly when is a bit hard to call because it's based on product shipments that are being made by companies a long way further up the supply chain than we are. When it happens, obviously, we stand to benefit from it.
I think we've got time for one more. Gareth, are you keen? Let's do Tim from there, and then Gareth. We'll round out with Gareth.
Two more questions. There we go.
Sorry. Repeat question. These are repeat questions now. Those are repeat questions.
You've given a very useful, I think, baskets of three end markets, essentially, where you see your royalty revenue potential being the application processor, mobile computing, enterprise networking, and then the IoT and MCUs. In fact, that kind of highlights how the traditional non-mobile is almost double in revenue potential versus the old mobile going forward. If I look at your royalty revenue guidance in the medium term, you talk about 15%-20% performance versus 10. If you take a 3% semiconductor growth, that means roughly about 18%-20% royalty revenue growth. If I look at these particular baskets and I look at the mobile growth potential, that's about 10% CAGR unless you go get a lot more content growth over there.
I would bet that a lot of your growth essentially is going to be coming in the future from the non-mobile potential, essentially. I was wondering, when you look at that guidance which you've given on enterprise networking or MCUs, would it be fair to assume that you're expecting sort of a quadruple in market share or something of that sort in this particular space? What would be the underlying market share growth potential there, or what are you seeing from your licensing which gives you this confidence of growth?
Well, if you take the two non-mobile buckets, we showed a 5% market share for enterprise networking and servers. As one of the other gentlemen pointed out, the numbers we put in the Analyst Day last May were more than kind of 17-ish% in 2018. That's quite a significant growth there. In embedded, we're at 22% growth for 2013. That was up from, I think, about 18% in 2012. We're seeing steady gains there. We've done a lot of Cortex-M licensing over the last couple of years. Designs, as we always say, take three, maybe five years to come into mass production. We've been on a licensing program with Cortex-M for longer than that, so you would expect to see increased share gains there in the embedded segment. I think hopefully that answers your question in terms of where we see that going.
We do expect an expansion of our market share in those two areas, I think we're well primed for it.
I guess given the Internet of Things and since you don't have much of a competition, so to speak, versus just the internal customers within, would it be fair to say that you could double your market share within the next three years in that?
I'd be very happy if we did. It's not to say that there isn't competition there. It is a large market. It's a new market. There are internal architectures competing for it. There are other third-party IP vendors out there trying to take a share as well. I think we've got a strong story that sets us up well to take a significant share.
Gareth.
Thanks. It's Gareth Jenkins again from UBS. Last year, you gave some very useful metrics around graphics and your units in terms of graphics development. I just wondered whether you could give us a sense of your graphics market share this year, the growth that's expected in that market. I guess you've got very high attach rates with Chinese mobile players, so you'll benefit from the low end. You've got high attach rates in DTV. I just wondered if you could give us a sense of what you think your graphics market share will be through the course of this year. Thank you.
The growth has been strong. I'm not sure we put a number in the
150.
Right. There you go. It was in there after all.
You want this one.
150 million units in 2013 going to about 400. Sorry, 2012 to 2013. I'm sorry, you're going to have to point me the page in.
It's in the growth section. 2012.
Right. Total through the year was about 400 million in 2013, up from about 150 million in 2012. 2013. Strong trajectory, Gareth. If you think of where it's come from, if you look at the Mali licensing in the last few years, and you look at where the unit shipments have gone back to 12 and 11 and 10 and virtually nothing, it is a pretty strong trajectory. I'll just say that again. Total for 2013, about 400 million, up from 150 in 2012.
The rest?
We haven't guided a number, but I would expect solid growth.
All right. Well, thank you, everyone.