Thank you for standing by, and welcome to the ARM Holdings plc Q3 analyst results conference call. At this time, all participants are in listen only mode. There will be a presentation followed by a question and answer session, at which time, if you wish to ask a question, you will need to press star one on your telephone. I must advise you that this conference is being recorded today, Tuesday the 22nd of October, 2013. I would now like to hand the conference over to your speaker today, Ian Thornton. Please go ahead.
Thank you, Will. Good morning, everyone. This is Ian Thornton, VP of Investor Relations at Arm. On today's Q3 results conference call, we have Simon Segars, Chief Executive Officer, and Tim Score, Chief Financial Officer. On today's call, Simon and Tim will take us through the highlights and comments from the quarter's results. Then we will open up the call to a Q&A session. As a reminder, the presentation and release can be found on the Arm Investor Relations website at www.arm.com/ir. Before I hand over to them, I have to just read out a few words with respect to this conference call and what we are about to discuss. The contents of this conference call are being directed only to those of you who have professional experience in matters relating to investments. The information communicated on this call is being made available only to investment professionals.
Any persons present on this call who does not have professional experience in matters relating to investments should not act or rely on the contents of this call. The following conference call will contain forward-looking statements, which are other than statements of historical fact. The company's actual results for future periods may differ materially from these statements, as they are based on current expectations and are subject to a number of risks and uncertainties. On this note, I will hand over to Simon.
Thanks, Ian. Good morning, everyone, and thank you for joining our Q3 2013 results conference call. What I will do this morning is run through the business highlights. Then hand over to Tim to provide some more detail on the numbers. Following Tim, there will be some time for Q&A. Let us start with the business overview. Three months ago, we highlighted how leading companies in multiple markets were licensing Arm technology and introducing new products into a wide range of end markets. We have seen this activity continue in Q3. In the quarter, we have seen very strong demand for our processor technology, which has resulted in record licensing revenue. A record 48 processor licenses were signed in the quarter with 24 companies.
This included 11 companies taking their first ever Arm license, many of whom are planning to take our technology into new and emerging markets, such as fingerprint recognition and wearable devices. Established thought leading technology companies also continue to make long-term investments in Arm, and this quarter, Arm signed four large multi IP licensing deals, including a significant licensing deal with MediaTek. This quarter has also seen several important design wins. MediaTek and Samsung both announced chips that support Arm big.LITTLE technology. AMD, Broadcom, and Huawei announced Arm-based chips that support 64-bit for mobile infrastructure and enterprise networking applications. In January 2013, we increased our medium-term guidance on how Arm's royalty revenues would outperform the wider semiconductor industry from a 10%-15% to 15%-20%. This was to take account of the increasing opportunity for Arm's Cortex-A and Mali graphics technology, which yield a higher royalty.
Over the last few quarters, the rate of Arm's outperformance significantly exceeded our raised guidance as chips containing Cortex-A and Mali graphics processors grew strongly. This quarter, Arm's royalties outperformed the semiconductor industry by 16 percentage points. This is in line with our raised guidance. The revenue momentum in both licensing and royalty has enabled Arm to grow earnings by 38% and deliver record quarterly cash generation in Q3. This has been another quarter that underpins the long-term growth opportunity of the business and enables us to continue to invest in our R&D capability, enhancing our ability to innovate and develop new products. Looking forward to Q4, the combination of record order backlog and a robust opportunity pipeline points to another strong quarter for licensing revenue. We also expect to see a sequential increase in Arm's royalty revenue of similar dollar value to previous years.
Assuming that the macroeconomic situation does not deteriorate significantly in the remainder of the quarter, we expect group dollar revenues for the fourth quarter to be in line with current market expectations of approximately $290 million. I'll discuss the revenue drivers in the different parts of the business in more detail, starting with technology licensing. As I mentioned, we signed 48 processor licenses this quarter. These licenses were signed with a view to Arm technology being deployed in a broad range of end markets. These included 14 licenses for mobile and computing applications, including entry-level smartphones, tablets, and two-in-one laptops. 18 licenses for microcontrollers and sensors for the Internet of Things applications. Seven licenses for wired and wireless enterprise networking. Nine licenses for other consumer electronic applications such as digital TVs.
Included in these licenses were a further three for Cortex-A50 series processors, which include support for 64-bit computing. To date, Arm has now signed 24 Cortex-A50 series licenses, and we are in discussions with many of our partners who are considering licensing the latest technology. Cortex-A50 series processors can be used in a multitude of high performance applications, from mobile computing to enterprise networking and servers. This quarter, Arm also signed another 18 Cortex-M series licenses. With close to 200 Cortex-M licenses now signed, Arm's partners are well-positioned to continue to gain share in the general purpose microcontroller market, as well as to take advantage of the emerging Internet of Things opportunity. Arm signed five more Mali graphics licenses and four more POPs during Q3. POP IP is physical IP that has been optimized to enhance the performance of Cortex-A and Mali processors. Switching to royalty.
Arm's royalty revenues are reported one quarter in arrears, so our royalty for Q3 was generated from chips sold by our licensees in Q2. Arm continues to outperform the semiconductor industry, despite the industry being down 2% year-on-year in the relevant period, processor royalty revenue was up by 14%. Just over 2.5 billion Arm processor-based chips were reported in the quarter. This 13% year-on-year increase was driven by growth across all target end markets. Shipments of embedded Arm-based chips were up particularly strongly, non-mobile chip shipments now account for more than half of Arm's total royalty shipments. Arm is also beginning to see traction of shipments in enterprise networking applications as the first of the early adopters start to ramp into high volume. In fact, more than 15 million Cortex-A enterprise networking chips were reported in Q3.
The growth in functionality of consumer devices such as smartphones, tablets, and digital TVs also continues to benefit Arm. This quarter, Arm saw a doubling of both its Cortex-A and Mali processor shipments year-on-year. Typically, Arm receives a higher royalty percentage for its Cortex-A class range of products and an additional royalty percentage for chips that contain Mali. Notwithstanding the strong growth in both Cortex-A and Mali chip shipments, Arm's average royalty revenue per chip was flat year-on-year as the strong growth in the higher value, lower volume application processors was balanced by the strong growth in shipments of higher volume, lower cost chips such as microcontrollers, smart cards, and touchscreen controllers. Turning to the operational side of the business, we've continued to invest in our R&D and commercial teams and in our business infrastructure. So far this year, Arm has added an extra 364 people, including 142 graduates.
We expect to continue to invest across our business to ensure that we are well-placed to seize the growth opportunities that lie ahead of us. It was a busy quarter for marketing, with record numbers attending our annual partner meeting in Cambridge, where we shared our roadmap plans with over 500 representatives from our customers and key ecosystem partners. Next week, we will be holding our annual technology conference in Silicon Valley in California, where we expect to host around 5,000 developers over three days, where we will present new technologies and have exhibits and demonstrations from our engineering teams and across the ecosystem. There is an investor event on the Wednesday afternoon, if you can join us, more information on that is on the IR website. I'm going to hand over to Tim, who'll provide some further detail on the numbers.
Thanks, Simon. Good morning, everyone. Hopefully, many of you will have had the chance to have a quick look at the Q3 earnings release. Just to remind you that the quarterly slide set is available on our website as usual. Simon's touched on a lot of the financial headlines, I won't repeat all of those, I will add a little bit of color as we think about the sell side models going forward. Overall, Q3 dollar revenues at $286.7 million, up 26% year-on-year. You will obviously have noted that particularly strong growth coming from processor licensing, up 52% year-on-year to $106 million. Clearly well ahead of the $80 million that I pointed to at our Q2 results presentation.
As Simon noted, we signed 48 licenses in the quarter, much higher than the normal quarterly run rate, where over multiple years, the average is more in the sort of 25 to 30. Further, these 48 licenses included some licenses which yielded significant revenue in the quarter. Approximately 50% of PD license revenue in Q3 was generated from backlog. In the normal range of 40 to 60, but actually a lower contribution from backlog than we've seen in the last couple of quarters. The third quarter was also a strong bookings quarter. Notwithstanding the strong license revenue result, we exited the quarter with backlog up about 3% sequentially. Given this healthy backlog position and the quality of the licensing opportunity pipeline, $90 million ± looks to be a realistic base for Q4.
The usual analysis of backlog maturity and composition is included in the Q3 slide set on our website. That shows that approximately 25% of total backlog is expected to be recognizable as revenue over the next two quarters. Simon's gone into some detail on royalties. Again, the headlines up 14% year-on-year in the third quarter, about 16% ahead of the industry, which was down 2% over the same period, consistent with our long-term guidance of outperformance of 15%-20%, albeit at the lower end and lower than in recent quarters. Normalized OpEx in the third quarter headline was GBP 85.6 million. This included a mark-to-market charge of GBP 5.5 million in the quarter. Which actually gives us a mark-to-market charge of nil in the nine months to date, but GBP 5.5 in the third quarter, reflecting the weaker dollar at the end of Q3, compared to the end of Q2.
Underlying costs, therefore, were around GBP 80 million, broadly in line with consensus. Normalized OpEx in the fourth quarter, assuming effective exchange rates are similar to current levels, are expected to be in the range of GBP 82 million-GBP 84 million. Again, consistent with current market expectations. The group's normalized tax rate in the third quarter was 22%. Our guidance for the full-year normalized effective tax rate remains unchanged at just under 20%. As the Patent Box regime in the U.K. transitional implementation continues over the next few years, that tax rate is expected to gradually reduce, all other things being equal, from the 2013 base of just under 20%. Reiterating Simon's comments on the outlook. We enter the final quarter of the year with a record order backlog, healthy opportunity pipeline for licensing.
As I say, that points to a strong quarter of $90 million or so for licensing. On the royalty side, the relevant industry data that we've seen and some of the customer guidance data for the third quarter, which is our shipment period for our Q4 royalties, suggests a sequential increase, similar dollar value to prior years. Last couple of years, it's been up $15 million or so. Looking back over three to five-year average, it's somewhere in the region of GBP 10 million-GBP 15 million. In this context, taking licensing and royalty together, we expect group revenues for the fourth quarter to be in line with current expectations of approximately $290 million. With that, we'll throw it open to questions.
We have some time for questions now. If I could request, please, that when you come on the line to ask a question, you ask just one question, that we do have a chance to get through everyone on the call. With that, operator, could we have the first question, please?
Thank you. Your first question comes from the line of Gareth Jenkins. Please ask your question.
Thanks, gentlemen. We'll probably hear more on this next week, I just wondered if you could give us an update on the server market and the developments there. Obviously, we've seen the recent developments at Broadcom and also Calxeda taking a Facebook exec onto the board. I just wondered whether you could give us an update on how things like Project Moonshot are coming and really the expectations around 64-bit in the market. Thank you.
I think generally, progress has been very good in the server market over the last little while. We're seeing hardware platforms come to maturity. We're seeing some of the software components that are going to be required to run on the server hardware become developed across the ecosystem. I think over the next 12 months, we should start to see some shipments of real server products to real customers being deployed in real applications. I think progress has been good, and as you say, I think you'll see some more on that at next week's tech con.
Your next question comes from the line of Simon Schafer. Please ask your question.
Yes, thanks so much. Just want to get your sense on market share and non-mobile going forward. Obviously, a very strong beat once again on the core licensing side. If anything, I guess that should be a precursor of future market share. I think last year you had roughly 20% market share non-mobile. Any sense as to how that might look in more like 2014, 2015 type timeframe based on the licensing that you've exceeded this year?
Well, I'd like to think that overall there's a healthy adoption of Arm technology in many markets. One of the particular standout features of our licensing in Q3 was the number of Cortex-M devices. In the microcontroller space, our market share is today relatively small. I'd like to think that's going to grow quite strongly over the next couple of years off the back of this licensing and the opportunity that IoT represents. The other side, I would say, where we are poised for some good growth is in enterprise networking. We've already had early adopters such as LSI start to ship products. You saw in the quarter that we signed licenses. Well, sorry, the products were announced by Huawei, Broadcom, Freescale talking about new enterprise networking applications that they're using Arm technology to develop with.
I think we should see some good market share gains there.
Great. Thanks, Simon.
Your next question comes from the line of Pierre Ferragu. Please ask your question.
Hi. Good morning. Thank you for taking my question. I'd like to come back to the royalty growth this quarter. Your 15% was clearly ahead of the overall semiconductor market. If you think that about half your revenues come from smartphones, and smartphones have been growing like in the high 30s this quarter, it's kind of puzzling because it doesn't seem that your actual royalty revenues have been growing faster than your addressed market, even if it's growing much faster than your addressable market. The reason why I'm pointing to that is that next year we'll probably have a slowdown in smartphone growth For you to stay 15%-20% above the overall semiconductor market in terms of growth, I need to see something coming into play and relaying the growth that is coming from smartphones today.
I'm not sure where we should see that coming. I would assume that increased penetration in non-smartphone segment is already quite strong this year. Do you expect a steep acceleration? If that's the case, what would be the segments where most of this acceleration would come from?
I think in smartphones generally, we're expecting quite strong growth next year. We're expecting about 20%-25% growth in smartphones. There's going to be a range of end devices there. It's going to be growth in the high end. Most growth will be in entry-level and mid-range phones, where there's an opportunity for a lot of Arm technology. Opportunity for Cortex processors, opportunity for Mali, opportunity for our physical IP. The growth, and whilst they may be lower cost devices, there's an opportunity for a lot of silicon content there for Arm. We're expecting mobile to represent strong growth in our royalties through next year. In non-mobile, we've seen in this last quarter, the overall split of unit shipments. Now it's 52% in non-mobile devices.
A lot of growth there in embedded, where we're seeing strong growth of our Cortex-M based parts from many of our licensees. We would expect that to continue as well through next year.
I suppose as a sort of a more general point, looking sort of longer term. License revenue has obviously been well above trend growth now for three and a half years, and continues into 2013. Really, given the normal lag of three to four years between licensing and royalty, looking further out into royalty, we haven't yet really seen the benefit of much higher than trend licensing growth. I think the sort of long-term picture for royalty that is being painted by license revenue growth in the last three and a half years is very encouraging.
The other end market I'd point to would be enterprise networking. We just touched on that, but with the licensing that's gone on, and with the first products that are now being shipped, that is an area where we do expect to see some good market share gains. Of course, that translates to volume. Given that those devices tend to be at the more expensive end of the spectrum, then that should be a good GBP contributor to our overall royalties.
Thank you.
Your next question comes from the line of Didier Scemama. Please ask your question.
Hey, good morning, gentlemen. Thanks for taking my question. Very simple question. You mentioned in your press release that you've got a major networking OEM signing a license. My question would be, what sort of applications are we looking at? Number two, do you think that the trend of disintermediating the semiconductor vendors and going straight to OEMs is going to accelerate, whether in network or in other applications, going forward? Thank you.
I think that the licensing we've done is with the silicon partners who are going to provide devices into OEMs. To your question about OEMs taking licensees. There are certainly some who want to be able to guide what their system products look like. I think generally, whilst that was a trend that people were anticipating a few years ago, I don't actually think we've seen much of that. I think the semiconductor partners are working closely with the OEMs on defining what the products need to look like, and then going off and doing what semiconductor companies do well. That is designing semiconductors and providing them to their end customers.
Your next question comes from Amit Harchandani. Please ask your question.
Good morning, gentlemen. Amit Harchandani from Citigroup, and thanks for taking my question. My question is around looking at licensing going forward. As Tim pointed out, we have been above the usual trend for close to three and a half years now. Particularly when I look at 64-bit, and a scenario wherein given that one silicon maker out there has already got a 64-bit product in mobile, and others looking to accelerate their attempt in 64-bit. Does that potentially lead to an accelerated revenue recognition coming out of backlog as well as milestones that are achieved? Potentially, how should we then think about the licensing revenue that you declare going forward over the next 12-18 months? Thank you.
Amit, Tim. I think generally when we discuss the trend that we've seen in licensing with investors. Our message remains the same. If you look out medium and long term, we do expect license revenue to grow in the sort of mid to high single digit. I think the thing that's interesting is how we transition from the growth we've been seeing in the last three and a half years to that steady state. I think that is quite a difficult transition to call. Obviously in recent quarters, we've been performing ahead of our guideline and guidance. I think the 64-bit and licensing of version 8 is all a very strong driver for Arm's licensing going forward. I don't think it really materially changes the overall profile of how backlog gets converted into revenue
New technology, when it's adopted by lead licensees, takes a good few quarters to get into revenue. Then when it becomes available for general licensing, it's a much shorter period into revenue. I don't think that overall dynamic changes. As I said earlier, there is the backlog pie chart in the slides shows the recognition profile next couple of quarters, three and four quarters out, and more than 12 months. I don't think the shape of that is particularly different from what it's been in the past.
Your next question comes from Andrew Gardiner. Please ask your question.
Good morning. Thank you. Tim, I was just looking at some of the comments that have come across the wires, where you're acknowledging inventory correction in the mobile space. I'm just wondering if you can give us any more color on what you're seeing, sort of the early stages of the three Q royalty reports coming through and, just signs of your confidence in that business coming back in the fourth quarter. Any further detail there would be helpful.
Well, I think, Andrew, the guidance we've given is that we expect a broadly similar seasonal picture than we've seen in recent years of an uptick in the mid-teens area. Actually, 22nd of October, it's even earlier than usual for us to be reporting. Therefore, we've actually seen less royalty reports than we may normally have done by the time we report. Obviously what we're looking at, as you do, is guidance from some of our big shipping partners. Inevitably, it's a somewhat mixed picture. Some strongly up, some less so. Taking that all in the round, we see a picture probably emerging that's similar shape to previous Q3 to Q4 transitions.
Thank you.
Your next question comes from Sandeep Deshpande. Please ask your question.
Yeah. Hi. Thanks. Can I ask a question about PIPD royalties? PIPD royalties were showing quite nice signs of growth in the first half of the year. That growth doesn't seem to be shown in this report as such. What is exactly happening? You did seem to be showing some signs on traction associated with POPs. At this point, it does seem very lumpy because normally, historically, licensing has been lumpy in ways, but once the royalties started coming in, they keep coming in. Is there anything happening on the PIPD royalty side?
I think POP licensing continues to perform well. There were another four deals that we did in Q3, and plenty of opportunities we see ahead of us. The royalties that those license deals will generate follow a similar trend to processor licensing as well. With all the POP deals that we've done, over 50 now, I believe, they will take time to come through on the royalty line. The Q3 royalties in PIPD last year were very strong. The comparison year-on-year is unhelpful in terms of percentage-wise. Year to date, PIPD royalties are up 16%, which I think reflects the overall success we've had in design wins and delivering leading edge technology over the last five years or so.
You're saying that there is no reason why, because you were almost at 51% year-on-year growth in Q1, and now it's 5.6% year-on-year growth. Is it just the base effect, or is it nothing else happening there as such, really?
No, there's nothing else affecting that. You're purely looking at a mathematical artifact of the comparison.
Thank you.
Your next question comes from François Munier. Please ask your question.
Yes, thank you. Actually, I'd like to ask a question about the monster license number you printed this quarter. If I understand well, there's been a few big deals in this number, maybe like three or four versus the normal one or two. I was wondering, is it basically big deals which have been pulled in into Q3? Is there a limited number of big deals you can sign every year? Did you give the normal rate or the normal level of discount for big deals this quarter, or maybe more than usual? I'm trying to understand basically, because I think everyone here has been surprised with licensing numbers for the past three years. I think Tim has been quite good at managing expectations. Every quarter, basically, the number is way ahead of what we've been thinking.
Yeah. François, you're right. I think there are a number of things to note. The most significant ones are there were 48 licenses, and we haven't really been north of the mid-30s before. A lot of licensing across a very broad range of customers and end markets. There were some significant revenue contributors. There are in every quarter. There were two or three here. I can categorically state that none of these were brought into Q3 from future periods in terms of commercial arrangements to pull that in. If underlying is a sense of, were you trying very hard to offset the royalty shortfall with licensing? If we'd been trying to do that, we probably would have stopped well before we went 20 over. That's just not a factor at all. The reality is there's a very big demand for Arm technology at the moment.
Whilst I think 106 is clearly a high base, we are pointing now to a base of 90 rather than 80 coming into this quarter. This is not a one quarter wonder. The licensing continues to be strong and the backlogs are up and the pipeline's still healthy.
Are we going to be digging into a period where, because as you've said many times before, you say maybe it's going to be 90, 85 next quarter and then it'll be 105 again. It's just like, what's your feeling now?
Well, I think the reality is when you've got a period where there is a really strong demand for Arm technology, it is actually quite tough to call precisely which deals are going to close in which period. Deals can close very quickly. They can take longer to negotiate. As you I think said earlier, I view our role to err on the cautious side in guiding. A quarter is a short period for licensing in this type of business model. I think we need to err on the side of caution when we're trying to predict precisely which deals will close.
Okay, thank you, Tim.
Your next question comes from Matt Ramsay. Please ask your question.
Yes, thank you very much. Just wanted to dig in a little bit further on the licensing. You guys mentioned that about half of the licenses signed in the period were from new licensees. Maybe you could give a bit more color of where these new licensees are coming from. Are these processor makers that formerly used a different instruction set architecture, maybe PowerPC or MIPS in the embedded space? Are these sort of new upstart processor companies from Asian markets or whatever emerging market might be? A little color on that would be great. Thank you.
There's a very broad range of customers that we've licensed to. Of these 11 customers who've never taken a licensed Arm technology before. Broad range of end markets that people are looking at. Although many are looking at microcontroller and IoT applications, there are companies in the U.K., companies across Europe, companies in China, companies in other parts of Asia, companies in the U.S. A very wide range of end customers that we've licensed to addressing a broad range of markets.
Matt, many of these companies are established semiconductor companies who just because they've been developing, say, analog sensors or something that didn't necessarily require a smart processor before. Quite often I dig into these companies and find that they've been around for years, but they just never needed to have any smart technology in their chips before. They're coming to Arm and just getting their first ever processor.
Your next question comes from Johannes Schaller. Please ask your question.
Thank you. Thanks for taking my question. Just really one on 64-bit driving royalties. I mean, we obviously have one major smart device OEM that is shipping volumes here now, and probably see a few more 64-bit products on the networking server space in 2014. Just if you could give us a bit of a feeling on how your conversations with your licensees are currently going on 64-bit, both smart devices and other areas, and how we should think about volume shipments here. I guess that smart devices next year will still predominantly be Apple. How should we think about the other markets and also into 2015 on the volume side? Thank you.
We're seeing demand for Arm 64-bit compatible technology across a wide range of markets. In mobile, in smartphones and tablets. We've said for quite some time that we see it as an inevitable shift to 64-bit at some point in the future. In this last quarter, we've seen Apple come out with a first 64-bit device. I think we're going to see the start of a transition there. How quickly that migrates in the high end, the mid-range, the entry level, time will tell. On the other side, we see strong interest for 64-bit Arm processors in the server space and in enterprise networking. The technology we've designed, we've designed in a way that it can target a very wide range of end applications. We have different products for different markets. We're seeing strong uptake of that in the licensing.
In terms of volumes next year, I think realistically that's going to be quite modest for 64-bit. These products do take time to design, to deploy into the field, and for volumes to grow. I think 64-bit volumes will be quite modest in 2014.
We should likely get more clarity on the 2015 outlook then, and maybe a few of your licensees giving us an update on new products over the course of next year. Is that fair to assume?
Yeah. I would expect that you would be seeing product announcements from our licensees talking about new devices featuring 64-bit Arm technology through next year.
That should be both in mobile and outside.
I would expect that across a wide range of markets, yes.
Understood. Thank you very much.
Your next question comes from Sumanth Vahee. Please ask your question.
Good morning, guys. Thanks for taking my question. I had a bit of a broader, longer term strategy question around Intel and the threat of that. In September, Intel announced that it's entering the Internet of Things with its Quark family. Yet another market, I suppose, where you and Intel will compete head to head. What I was quite interested in that was their comment about the fact that they will offer Quark family or Quark solution as a synthesizable product. If I compare that with the smartphone market, you've clearly won because of probably three reasons, the right technology, the right business model, and of course, that your cores were standard by the time Intel really decided to come in.
In the Internet of Things, when I look at it from where we are today, the standards are not set, and they suggest that they're probably emulating your model to a certain extent by selling IP, essentially. How do you think you'll be able to defend yourself in that? I know that your technology is much further ahead, but is that the only focus which should help you stay ahead of them? Or is there anything else I should be looking at? While you're at it, could you also tell us what you think is your latest view on how big in dollar terms is the Internet of Things today?
I think you have to be careful not to confuse a processor being synthesizable with a processor being licensable.
Okay.
I may be wrong about this, but I don't think I've heard Intel say that they're going to be licensing their processors to other people to design chips with. I think our business models remain very different.
I think the Internet of Things space is very broad. There will be a wide range of processors required to address it in its entirety. Internet of Things can mean processors connected to large pieces of equipment sitting in factories, which are being monitored to see when they need repairing. It can mean wearable devices which can require high-performance processors or very small processors. It can mean sensors embedded in light bulbs to work out when they're going to need replacing and what color temperature the light is going to be coming out of it. There's a very wide range of end markets which comprise the Internet of Things. All of them pretty much have characteristics of there being a processor, a sensor of some sort, and some wireless technology, all of which Arm has a very deep penetration of into already.
If you look at the microcontroller shipments by Arm's licensees, there's cumulatively over 5 billion of them that have already been shipped. While the Internet of Things is a new space and there's scope for many people to go and play in it, and there is scope for new standards to be set, I think Arm has already been successful with our range of microcontroller processors and achieved very large penetration into companies that produce chips for the Internet of Things or will produce chips for the Internet of Things.
Okay. Any view over the next 5 years how big this market could be in dollar terms? Any new view, essentially, given the fact how quickly it's growing?
Well, I don't think we've changed our view on that. Exactly how big it's going to be, again, remains to be seen. How quickly standards for security get deployed, for example, how quickly services can be built up around these interconnected devices remains to be seen. This is a tens of billions of unit opportunity. The really small sensors in the light bulbs will be $0.50 and below. Products, larger processors going into smartwatches are probably going to be in the $5 range. Industrial applications might be that and maybe more. Exactly the blend of that, hard to call right now.
Thank you very much.
Your next question comes from Achal Sultania. Please ask your question.
Hi, guys. Yes. On mobile royalties, you mentioned in the press release that mobile royalties were up 20% year-on-year. Volumes were up 8%. That would imply that your royalty rate per unit within mobiles was up about 12%. Can you give us some sense of, obviously it's being held by rising royalty rates, but that's been offset also by declining chipset ASPs. Can you give some sense of how much of that is actually being driven by the increase in royalty rates? Is that trend that we should actually expect going forward? I have a link to that. How do you actually see the adoption of both 64-bit and also big.LITTLE technology?
We've seen some initial traction in the high end, do you actually believe that these technologies will be needed in the mid-end of the smartphone market three or four years down the line?
Okay. Well, I'll take the first part of that one. Clearly, as we say in the release, look at Cortex-A processors and Mali graphics processors, both of those up about twofold year-on-year. Cortex-A class processors, when compared to sort of the ARM11s and ARM9s have a slightly higher royalty rate. Mali graphics adds another sort of about 1% as well. The combination of those together does mean then that we are going to be seeing slightly higher royalty percentages from mobile from application processors going into mobile phones. That's obviously going to have to be put into the overall mix with what's going into microcontrollers and Internet of Things where we typically will have a say Cortex-M class processor. In the end, what happens to the average going forward will very much be dependent upon the mix.
I certainly think you can sort of separate your Cortex-A your application processors going into smartphones, into tablets, into computers, into TVs, and also into things like enterprise networking and servers and think that those will probably have a growing royalty percentage over time. The other markets maybe We hope we get a slightly higher royalty percentage, but probably a slower growth trend going forwards. Your follow-on question was about big.LITTLE. We've seen a couple of announcements in the last quarter, one from MediaTek, one from Samsung, about new devices using big.LITTLE. As you said initially, that's been successful in the high end. I think over time we should expect to see that migrate into the mid-range. big.LITTLE is most suitable where you have a broad range of processing performance requirements.
In an application where sometimes you need a lot of performance, sometimes you don't need much performance, big.LITTLE is a great technology for helping lower power and smartphones exhibit those characteristics. No reason why it's not a suitable technology across the entire spectrum. In a low-cost entry-level device, people are probably initially at first anyway, likely to go for the lowest cost implementation point, which is probably a single or dual core processor. Over time, as devices continue to shrink, as silicon continues to get less expensive to manufacture, there's no reason why big.LITTLE can't span the entire range of smart computers.
Great. Thanks a lot.
Your next question comes from Jerome Ramel. Please ask your question.
Good morning. Just coming back to the royalty per chip flat year-on-year. If I look at the mix between Cortex and the rest, Cortex was a little bit ahead of 50%, 53%, significantly higher than last year. And if I look in the Cortex-A was up 127%, Cortex-M only 74%. I'm just wondering why. I understand the mix effect, but if I look at the Cortex-A grew significantly higher and quicker than the Cortex-M. I'm just wondering why we had only flat royalty per chip. I'd just like to understand what was the underlying price decline you saw per chip?
I don't know about price decline, but there's potentially a large price difference in the cost of an application processor with multi-core Cortex-A in it and the price of a microcontroller with a single Cortex-M in it. One could be five, 10, 15, maybe GBP 20. The other could be GBP 0.50 and below. In terms of the volumes we've seen large volume growth in Cortex-M. That has pretty rapid effect on pulling down the average when you just divide the royalty GBP by the royalty volume.
I understand. I know there are many inputs, but what could be your best guess for the royalty per chip going forward?
It depends on the mix.
Jerome, this has obviously been a long-standing discussion. Our guidance, as you heard before, tends to be we sort of think broadly flat. What we've seen over the last two or three years, well, in the many years before that, we saw a sort of a gradual decline as the weighted average of chip prices Arm got designed into went down. Then we've seen in the last two or three years sort of flat to slightly up. When we look out forward, we obviously see some trends that are taking in different directions. Very strong growth, Cortex-A, version 8, higher %, higher chip prices. Simon's just sort of painted a picture of what the Internet of Things opportunity looks like. Massively high volumes but obviously characterized by lower chip prices on average.
It really does depend on one's assumptions about the growth rates of these markets and Arm's penetration of these markets as to where you get to on the average. The good news is it's all 100% margin revenue.
Thank you very much.
Your next question comes from Janardan Menon. Please ask your question.
Just to go on your guidance on royalties into the fourth quarter. You said it's a similar level in dollar value. Obviously that would imply a declining quarter-over-quarter growth rate, which taking sort of the midpoint of your guidance about $13 million-$14 million would imply about 10% versus 12%-14% that you've seen in previous years. I was just wondering what one should read into that. Would you say that that's a sort of a temporary kind of a deceleration that we are seeing because of inventory, et cetera? Or does that signal a more structural slowdown in key markets like smartphones and tablets and until some of the other applications start becoming bigger we could be in a slower growth phase for some time?
I think when you look at the end markets, consumer products go through cycles. What we've seen in Q3 is many devices kind of coming to the end of their life and consumers waiting for the next version to come out before upgrading. We've seen that in some smartphones. We've seen that in tablets. We've seen that in game platforms. Kind of a lot of that has happened at once. We're expecting something of an uptick in Q4. We're making a judgment on that based on the information that we have in front of us, which is obviously imperfect, and we'll see how the quarter plays out.
In terms of long-term structurally, though, I think when you look at the number of licenses that we've just done, when you look at the strength of licensing over the last couple of years, that points very strongly to an uptick in design wins for Arm technology, which drives royalty, which drives royalty dollars. I think for the long-term, and we are in this for the long-term, we should continue to see gains in market share, royalty volumes and royalty dollars.
Got it. Thanks for that. Thanks, Simon.
Your next question comes from Vijay Anand. Please ask your question.
Thanks, guys. I had a question on the server market. Simon, could you maybe talk about the 64-bit ecosystem, in terms of how far or how close we are for the ecosystem to be mature enough to support commercial deployments? In that context, maybe could you also talk about the licensing traction you're seeing in the server market, maybe in the third quarter? Both some quantitative as well as qualitative color will be helpful. Thanks.
In terms of the ecosystem, a lot of work's been going on for a long time to be ready for 64-bit commercial deployments of servers. We're starting to get to the point where that can happen. We are expecting this quarter for HP's Moonshot products to ship based on 64-bit Arm silicon from partners like AMCC. At that time, the required software and hardware will come together. We've seen much of that being developed by numerous partners in the Arm ecosystem. It's been a work in progress, and I'd say at the moment it is coming to fruition. We continue to see very strong interest in putting Arm in the data center. Many people who build out data centers for specific applications are very interested in doing trial deployments right now.
I would expect based on the number of trials that are going on, to start seeing some meaningful commercial deployments probably starting in about, well, sometime through next year. I think we're in pretty good shape, and it's one of the things that's helping drive licensing.
Your next question comes from Lee Simpson. Please ask your question.
Great. Good morning. Thanks for letting me on. I just want to ask a quick question about networking, if I could. As far as looking at trends in that space, it looks as though there's an ongoing shift towards low-cost wireless base stations. I just want to try and understand beyond help that it has for FPGAs near term, and some of that obviously next phase rollout infrastructure in China. Just trying to understand where Arm licensees are coming into the mix here. Maybe associated with that, are you encouraged by the moves that some have towards a unified stack across macro and lower cost base stations as a positive for the Arm architecture?
I think in order to help lower costs, some form of standardization is required. There are many Arm licensees now taking Arm's technology and targeting this market. I think, as you say, that there is a shift towards the combination of macro base stations and small cell base stations in order to achieve a number of things. Delivering the bandwidth that 4G technologies offer the capability of is going to require an upgrade in the infrastructure. A greater shift to the combination of large and small cells. Enabling people to do more and more communication from mobile devices indoors is something else that's driving that. When original mobile networks were deployed, they were deployed to enable people to make phone calls whilst they were out of the office and away from home.
Now increasingly, most calls are initiated inside buildings, which the conventional infrastructure doesn't well suit. That is one of the things that's driving the adoption there. To make all of that a reality, these devices need to be low cost if we're going to have high volume. The Arm business model, the technology delivered through Arm's licensees, helps enable lower cost end devices and helps drive volume, which is good for everybody who plays in the space, and it's certainly good for Arm's volumes and Arm's royalties.
Perfect. Thank you very much.
Your next question comes from Andrew Dunn. Please ask your question.
Good morning. I had one actually on headcount. You've increased your headcount sequentially this quarter by around 8%. I think that's the highest you've done for a number of years, particularly in Asia Pacific, I noticed. Are there any specific areas that you're investing in? I know you've spent a lot on Mali and getting that up to a competitive product. Are there any particular areas that you are investing in now going forward? Thanks.
Well, there's a number of areas. A lot of the engineers that we're hiring are going on to developing next generation of processor technology. We're also increasing the number of people that we have in the field, working closely with our customers to help with the integration of what we license to them, and help accelerate time to market for our customers and help support them. We're also investing in our internal infrastructure, so that the company can scale and grow to deliver on the opportunity that we have.
Thanks.
Your next question comes from Brett Simpson. Please ask your question.
Just go back to the second quarter, looking at the smartphone market, some of the trends. Quite a lot of changes with your Qualcomm and MediaTek spread had very strong sell in in the June quarter. Apple clearly working down inventory on the iPhone side ahead of the new product launches, I think also had weak iPad numbers. Quite a big mix shift in the second quarter to the low end. I just wanted to ask, can you just talk a little bit about Cortex-A royalty ASPs? With this mix shift we saw in the June quarter, have you seen any major changes in your Cortex-A royalty ASPs?
Hi, Brett. Stephen here. When we look back at the ASPs of the chips, then they've been flattish, as they often move around a bit, but they're flattish for the last few quarters. Obviously, the royalty percentages that we get from those as we're starting to see some of the Cortex-A15, some of the big.LITTLE ones starting to come through. The very first of those, they obviously add a little bit of an increment, a slightly higher increment. Obviously with Mali graphics doubling units year-over-year, that also adds an extra %. There's a number of upward drivers within that hopefully getting us a higher royalty percentage over time.
That's helpful. Just maybe a follow-up here. Looking at the backlog on PD licensing, what portion will be 64-bit V8s? Can you give us a sense what typical license ASPs, the bump in ASP you're getting for V8 versus V7 on a like for like basis?
Well, we haven't really provided guidance on ASPs of our licenses. That's not something that's particularly going to help us, I don't think, in liaising with our customers. We haven't really broken that out.
The more sophisticated the processor and the more complex it is to develop, the higher the license fee, Brett. That normal trend continues. We don't go down to the next level of detail around the backlog. As you know, at any one time, the latest generation of technology that we are licensing forms a portion of the backlog, and V8 is no different. You've seen how many licenses we've done, and the revenue recognition on those is ahead of us, not just behind us.
Maybe just another way of asking that question, Tim. When you look at V8 and the deals you've signed so far and compare to the same time with Cortex-A V7, is the take rates from the industry for your latest architecture much faster than we saw in V7 Cortex-A?
I think there's distinct enthusiasm for the version 8 of the architecture. There are other products that have licensed very rapidly, like Cortex-M. You can tell generally, if you look at what's happened in licensing in the last year or so since the V8's been introduced. It is benefiting from the same level of enthusiasm of our technology as some of our existing portfolio.
I would say probably numerically, it's the case that the uptake has been faster, but compared to when we introduced V7, the business is bigger, the number of customers we have is bigger as well commensurately. Yeah.
Thanks very much.
We're going to make this the last question.
Your last question comes from Kai Korschelt. Please ask your question.
Hello, Kai.
Yeah. Hello, can you hear me?
Yeah.
Sorry. Thanks for squeezing me in. I just had a couple on some of the incremental markets. The first one was on networking. I know there are a couple of questions on this, any color you can give on maybe the value and essentially unit opportunity, short and midterm here. The second was really on the iPhone 5s. We saw further Arm content growth. We have the M7, that's obviously a low ASP chip. Should we expect this sort of processor growth or content growth to continue in other high-end devices going forward? Thank you.
I'll take the last bit first. In terms of content in high-end smartphones, I think what you are seeing is an increase in the number of sensors that are embedded in phones as well as just the raw application processor performance. You've seen that in a number of the high-end devices that have shipped over the last year. I think that's going to be a growing trend. That does represent a further opportunity for more Arm technology. Could you repeat the first part of your question?
Sorry, the first question was on networking. Just if you can maybe give any color on maybe the value or unit opportunity maybe next year and then on a three-year view potentially. Thank you.
Well, Kai, as you probably have seen before in our roadshow slides, we do tend to lay out quite a bit of detail. If you go to the roadshow slides from our website on slide 22, we do have a slide just on enterprise networking, including value and size. There's a table there which gives you the detail. The headline basically is for 2017, Charlene Marini, who runs that team there, is estimating that 700 million chips is the size of the TAM, and the value of that is about GBP 17 billion. Which when I get my calculator out, that's very similar to sort of the application processors going into smartphones.
Billion.
For GBP billion. Sorry. The team beside me is mouthing billion, not million.
We're talking about between friends.
Yeah. I get my calculator out, that is about the same sort of size as application processors going into smartphones, yet it is a market where we have practically today, well, I would not say zero, we now actually do have one company that is shipping millions of units. Even so, it is a very, very small market share. Hopefully, as we have been discussing already about the licensing, this is a market where we hope to gain share rapidly with some of our most advanced processors, which have the higher royalty percentages typically.
Based on the licenses you have been signing recently, what would be your sort of market share target maybe by 2017?
Well, that is also on page 22 of the roadshow slides. I guess you are looking at sort of 20%-30% by the time you get to 2017 on average.
Okay. Thank you very much.
Okay.
Okay. Thank you all very much for your questions and for your time this morning. We'll see you at the end of January and early February.