Ladies and gentlemen, thank you for standing by, welcome to the Arm Holdings Q1 Results Analyst Conference Call. At this time, all participant lines are in a 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 can do so by pressing star one on your phone keypad. I must advise the conference is being recorded today, Wednesday the 20th of April, 2016. I would now like to hand the conference over to your speaker today, Mr. Ian Thornton. Please go ahead, sir.
Thanks, Callum. Good morning, everybody. This is Ian Thornton, Head of Investor Relations at Arm. On today's Q1 results conference call, we have Simon Segars, Chief Executive Officer, and Chris Kennedy, Chief Financial Officer. On today's call, Simon and Chris will take us through the highlights and comments from the quarter's results, then we'll 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 just have to read out a few words with respect to this conference call and what we're 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 investment, 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 investment should not act or rely on the content 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 that note, I'll hand over to Simon.
Thanks, Ian. Good morning, everyone. Thank you for joining our Q1 2016 results conference call. I will run us through the business highlights, then hand over to Chris to provide some more detail on the numbers. Following Chris, there'll be some time for your questions. Let me start with the business overview. Following last year's strong performance, we've had an encouraging start to 2016. We've seen healthy demand for our technology from semiconductor companies, our bookings were the highest we've ever seen in a first quarter. Total technology licensing revenues were up 11% year-on-year, technology royalty revenues were up 17% year-on-year. During the quarter, I saw many of you at the Consumer Electronics Show in Las Vegas and at the Mobile World Congress in Barcelona.
One thing that struck me at these events was the sheer ambition of our partners' roadmaps and how they intend to use our current and future technology. For example, we've seen increased activity around next-generation technologies such as 5G communications and Network Functions Virtualization, autonomous driving and computer vision, virtual and augmented reality, artificial intelligence and machine learning. Arm has been investing for many years to create the technologies that will enable many of these future products. Those that follow Arm closely will know that at our capital markets day in September last year, we announced a step up in investment to accelerate share gains in networking and servers, and also to create new revenues in nascent markets such as the Internet of Things. Having made the step up in costs, we've now returned to our longer term cost trajectory. Where is this investment going?
Over the past year, we have increased our headcount by 20%. We have more engineers working on our core roadmaps, working closely with our ecosystem partners and helping to create and optimize open source software for Arm-based systems. We're seeing the results of this. In networking, Arm software teams have been working as part of the OPNFV project to optimize their recently released Brahmaputra software. We've been collaborating with Enea on the opening of the Pharos lab in Sweden to enable the testing of networking applications. With the availability of Arm-based chips and Arm-optimized software, our engineers and marketing teams are working with OEMs, operators, and end users to accelerate the adoption and deployment of Arm technology into our target markets. Our partners are also investing in the enterprise ecosystem.
By way of two examples, in February, Qualcomm announced a $280 million joint venture which will develop server technology for the Chinese market. In March, Red Hat previewed its Enterprise Linux server software running on Qualcomm's Arm-based server chips. These are two of many examples across the ecosystem. In the Internet of Things, we have seen Arm-based chips being used to improve productivity and efficiency in traditional industries such as agriculture and construction. Other examples, Verizon's ThingSpace platform has signed up 4,000 developers and is boosting productivity in precision agriculture from vineyards to oyster farms. This month, a leading building contractor in the U.K. deployed smart sensors to reduce construction time on major projects.
A whole host of telecoms companies such as SK Telecom in Korea, Orange and Bouygues Telecom in France, Swisscom in Switzerland, Tata Communications in India, all announced plans to deploy nationwide, low power, wide area networks for IoT devices. All these exciting industry developments place exacting demands on processor performance and power consumption. We will continue to invest in our roadmap, and in the next few years, we will deliver new technologies that will enable our partners to meet their ambitious goals for innovation. Now I'll discuss the revenue drivers in different parts of the business in more detail, starting with technology licensing. We signed 39 licenses with 27 customers in the quarter. Over half of those customers were new to Arm, and two of those new customers were OEMs that are just starting to build up their expertise in chip development.
Our Cortex-A processors run rich operating systems and applications in consumer and enterprise devices. We signed eight licenses for Cortex-A processors in Q1. Four of these were Armv8-A based processors that will be designed into chips for networking infrastructure applications and future consumer devices. Our Cortex-R processors run real-time and deterministic control software for applications such as communication protocol stacks and safety-critical systems. We signed five licenses for Cortex-R processors in Q1, and our customers will use these to develop chips for 5G modems, network storage, and automotive controllers. Our Cortex-M processors are designed for low-cost, low-power chips, such as microcontrollers and sensors. We signed 22 licenses for Cortex-M processors in Q1, and these will bring advanced functionality to a huge variety of devices, including electric motor controllers, memory cards, motion detectors, and microphones. Our Mali processors bring advanced graphics and video capabilities to consumer devices.
In Q1, a major semiconductor company signed two licenses for our latest Mali products. Our Physical IP libraries help our partners build chips on the latest manufacturing process technology. During the quarter, we signed two major Physical IP agreements with leading foundries. One of those was announced with UMC last week. Both of these extend our relationships by developing new Physical IP products that chip companies can use when building products from 55 nanometers down to 14 nanometers. These contracts were backlog building, and there was a lower portion of revenue from terms this quarter in our Physical IP business. Therefore, this quarter, we are reporting lower Physical IP licensing revenues than the usual run rate. We expect that revenues will improve from Q2. I'll switch to the royalty side of the business. Arm's royalty revenues are reported one quarter in arrears.
Our royalty for Q1 was generated from chips sold by our licensees in Q4. Processor royalty revenue was up 15% year-on-year, outperforming the semiconductor market, which declined 3% in the relevant period. Processor unit volumes were up 10%, with strong growth seen in Arm-based microcontrollers and smart cards, which grew 20% on last year, and networking infrastructure chips, which grew 10% year-on-year. Volumes of Armv8-A chips more than doubled year-on-year to reach 280 million units. We estimate Armv8-A chips accounted for approximately 65% of the smartphone market in Q1. We saw growth in the number of chips containing multiple Armv8-A processors, such as the Octa-core chips sold into mobile phones. These amounted to 100 million units in Q1, roughly 25% of the smartphone market. Ultimately, we expect to see Armv8-A in all smartphones and many other devices too.
In the last few months, we've launched two processors that help expand the breadth of Armv8-A adoption. The Cortex-A35 is the smallest and lowest power processor that supports the full Armv8-A architecture, including 64-bit instructions. The Cortex-A32 processor reduces the size and power further by supporting only the 32-bit instruction set within Armv8-A. These new technologies from Arm will help drive higher royalties per device from billions of low-cost, smart, and connected devices. I'll now hand over to Chris, who'll provide further details on the numbers.
Good morning. Hopefully, many of you will have had the chance to have a quick look at our earnings release and the quarterly roadshow slides that are available on our website as usual. Q1 dollar revenues of $398 million were up 14% year-on-year, with 11% growth in technology licensing and 17% growth in technology royalties. As you know, quarterly license revenues are lumpy. Processor licensing had a particularly strong quarter, up 24% year-on-year to $135 million, whereas Physical IP license revenue was down 46% year-on-year to $13 million. This resulted in overall licensing growth of 11%, and we have a good pipeline of opportunities to underpin both processor and physical licensing growth for the balance of the year. Approximately 40% of our processor license revenue came from backlog in the quarter, and this is at the lower end of our typical 40%-60%.
Bookings were strong for Q1, and group backlog was down around 5% sequentially. Processor royalty revenue was $192 million, up 15% year-on-year, reflecting continued market share gains and the increased amount of Arm content being deployed in mobile computing and enterprise devices. Processor chip shipments grew 10% year-on-year, and revenue grew faster than unit shipments as Armv8, Mali, and Octa-core all gained share. For the smartphone chips reported in Q1, Armv8 had around a 65% share, Mali around 50%, and Octa around 25%. Physical IP royalty grew 38% year-on-year due to the increase in shipments of wafers using Arm's Physical IP at advanced nodes. These wafers are typically more expensive than older nodes and therefore yield a greater royalty per wafer. Normalized OpEx in Q1 was GBP 132.9 million, up 33% year-on-year, and adjusting for foreign exchange, broadly in line with our guidance.
There were two main drivers of the increase. Firstly, our accelerated investments have resulted in a 20% increase in headcount, the vast majority going into engineering and new business developments. Secondly, sterling has weakened significantly against all major currencies, particularly the U.S. dollar. As a reminder, around 40% of our OpEx is in sterling, 40% is in dollars, and around 20% in other currencies, whereas income is over 90% dollar denominated. A weaker sterling increases costs in sterling terms, but overall gives a net benefit to Arm. Excluding the impact of currency movements, OpEx is just 1% higher than in the previous quarter and at the upper end of the Q1 guidance given last quarter. Assuming exchange rates at current levels of around $1.42 to the pound, OpEx in Q2 is expected to be in the range of GBP 130 million-GBP 133 million.
This reflects the return to quarter-on-quarter OpEx increases in line with historical levels. The group's effective normalized tax rate was 15% in Q1, and we expect the full year tax rate to also be around 15%. As a result, EPS grew 15% year-on-year to GBP 0.082. Now onto the outlook. The licensing pipeline for the rest of the year is robust, with leading semiconductors companies and equipment manufacturers looking to license Arm's most advanced technology for the next generation products. Macroeconomic uncertainty remains and could influence consumer and enterprise spending on semiconductors in 2016. Based on current conditions in the semiconductor industry, we expect group U.S. dollar revenues for the full year to be in line with market expectations. Great. Thanks, Chris. Now we'll hand over to Q&A.
As usual, if you could ask just one question at a time, please, we will be able to get everybody's question in, and we can come round again if you have follow-up questions. Yeah. Operator, do you want to open the call for questions now, please?
As a reminder, if you wish to ask a question or make a comment, please press the star one on your phone keypad and wait for your name to be announced. If you wish to cancel that request, please press the hash key. Your first question today comes from the line of Matt Ramsay from Canaccord Genuity. Your line is open.
Yes. Thank you very much. Good morning, gentlemen. Simon, I guess a two-part question from me. Congratulations on the strong execution in a pretty choppy macro.
Thank you.
I guess with only a 25% or so penetration of Octa-core in the smartphone market so far, I guess where do you think that penetration can get through the year and how big of a driver of earnings could it be or royalty growth? I guess on the flip side, tons of moving parts last night with Intel, but one of the things that stood out to me in their results was a 60% growth in their networking business year-over-year. The numbers that you guys have put out is 10% growth. It strikes me that as Armv8 compares it a little bit tougher as we move through the year, that networking revenue might need to be a bit robust for you guys to keep royalty growth rates high.
I wonder, maybe you could just walk us through the differences in the growth rates you've seen in your networking business versus Intel. I know it's very different products and sort of what the growth rate of that business could be over the medium term. Thanks.
Okay. Thank you for ignoring my request to only ask one question, we'll cover that slide. First of all, so on the question about Octa-core, as we said in the first quarter, penetration was around 25%. That's grown fairly steadily over the last year. A year ago, Q1 2015, it was less than 5% penetrated. That grew through the year. Full year 2015 was about 10%. Full year 2016, obviously, it's hard to say. It's going to depend on the mix and what handsets get sold, et cetera. It wouldn't surprise us for full year to be about 30%. We do see room for growth of Octa-core for some time to come. If you were thinking where might Armv8-A penetration go in the full year, as I said a moment ago, about 65% for Q1 2016.
Could be 70%-80%, exit rate around 80%, maybe as much as 85% for the full year. Your second question was about networking. Networking volume for us grew about 10% year-on-year. A lot of that is still chips based on Armv7. As you know, we've spoken about a lot of licensing activity in the networking space over the last couple of years, and we're expecting Armv8-A based chips to come through. We think networking will be a driver of royalty in comparison to Intel that you mentioned in that space. Businesses are completely different. I'm not going to spend too long trying to draw a comparison between the two of them. We feel good about what's going on for us in networking.
The long-term driver, or medium-term driver rather, is about 5G and overall the increased sophistication of the equipment that's going to be required to deliver the kind of bandwidth and amounts of connectivity that 5G necessitates. I think we're in a very good position for the Arm partnership to service that.
Thanks so much, Simon. I appreciate the perspectives.
Thanks.
Question is from the line of Gareth Jenkins from UBS. Your line is open.
Thanks. I'll just keep it to one. You've done a great job in gaining market share in the last two years in graphics. I just wondered whether you can talk about some of the technological aspects of how you moat that business off a little bit and make it sticky through GPU, CPU coherency, or any other efforts that you see over the next two years that you can maybe make that more defendable against potential competition. Thank you.
Yeah. Really, a lot of that is about engagement with the ecosystem. We have a strong effort around the developer community, making it easy for people to write their applications that utilize the benefits of the Mali architecture. We do a lot of work on tools. We do a lot of work generally on enabling that ecosystem. That is one of the ways in which we do increase stickiness around our product. As you know, fundamentally, GPUs implement a third-party standard. OpenGL ES is controlled by the Khronos Group that we're a member of and very active in driving the standards forwards. Ultimately, it's got to come down to execution on building the best performance and the most power-efficient cores with the right mix of features for different markets. You mentioned GPU computing. There are going to be markets where that's important.
We're expanding our roadmap to have some GPUs with those features in for those markets and some GPUs with those features out, where there's more price sensitivity and power sensitivity. It's about the mix of all of those things. Having the right product, but crucially engaging in the right way with the ecosystem.
Can I get a follow-up on that, Simon?
Sure, mate. Just this once, Gareth.
Thanks. Do you feel that your GPU can scale to performance levels that are high enough to compete with high-end graphics companies that we see standalone, the likes of Nvidia, et cetera, on the compute side?
I guess the way we look at it is not so much can we compete with company X, it's more about market opportunity, and is there a large volume market opportunity that we think the Arm partnership wants to serve with IP licensed from us? We're looking at different markets and making sure that if we think that's a viable market that our business model serves, then making sure we've got the right products for it. There are always going to be some applications where the volumes are so low that they're not served best through an IP model, and there's no point in us making some enormous product that's only used by one or two people and the volumes are low. That just doesn't fit our business model.
We're looking at where the volume's going to be, we're looking at how the market's evolved, and driving our roadmaps accordingly.
That's great. Thank you.
Thanks.
From the line of Andrew Gardiner from Barclays, your line is open.
Good morning, guys. Thanks for taking the question. Just one on the mobile side. At the time of 4Q results, you'd highlighted that I think nine companies had licensed the next generation Cortex-A processor. I was wondering if there was, A, any more activity there in terms of lead licensing, and is this something that we're likely to hear more of perhaps later this year, or is it a little further out there into 2017, perhaps?
We haven't called out particularly which products were licensed and the licensing of advanced technology in this quarter. As ever, there are new products that we're developing. We will announce those in the fullness of time. In the last two quarters, we've launched at least three CPUs. Seeing very strong demand for that and very strong interest in the high end of our roadmap. We'll reveal more details of that as it comes and be explicit about the licensing as it goes.
Okay. Thank you.
Thanks.
Line of Kai-Uwe Kessel from Merrill Lynch. Your line is open.
Yeah, good morning. Thanks, gents, for taking my question. I had one on the revenue guidance at the Capital Markets Day in September last year. I think you guided to about GBP 40 million in incremental revenues this year. Do you have any visibility at this stage about the phasing? Also around which end markets and also which maybe segments we should expect those incremental revenues to come from? Just to confirm, that will be incremental revenues based on the OpEx investments that you have been carrying out in the last six months. Thank you.
The Capital Markets Day, we talked about GBP 50 million of incremental cost ramping up over the year, Sorry, GBP 40 million of cost ramping up over the year, and then generating $200 million of revenue in 2020. We didn't actually give an explicit revenue guidance for the first year. When we've given our guidance now, which says total revenue, we're comfortable that it's going to be in line with market expectations. That includes the incremental revenue that's coming from those investments. If you go back to the Capital Markets Day, revenue this year is around the continuing revenue from some of the acquisitions we made last year, such as Sansa Security, together with some early licensing wins in the other markets.
Actually, the big ramp-up in the additional $200 million will come in future years as those licenses turn into shipments by our partners to OEMs and generating royalties.
Okay, it would be mostly on the licensing side. Actually, I'm just looking at the slide where you have said GBP 40 million, which is slide 94. Just wanted to clarify that is mostly licensing and M&A, I guess.
I think the GBP 40 million you're referring to there is the GBP 40 million investment that we were talking about incrementally making through this year. As Chris said, that's products for the networking and server space and it's new product for IoT. We talked about our software platform around mbed. That development is going very well. We're seeing very strong interest and pickup across the industry in that. I think if your question's around how we're doing on those investments, as you'd expect, we do a fairly frequent look back at how the spend is going and how the revenues are going and it's exactly where we were planning it to be from the latest look back we've done.
Okay. Thank you.
Thanks.
Question is from the line of Francois Meunier from Morgan Stanley. Your line is open.
Hello. Yes, congratulations on the great quarter despite the weakness in smartphones. Really well done. I was wondering, what's your view on competition at the moment? There's a big competitor in Santa Clara, which is reducing headcount by 12,000, while your headcount is actually increasing by 20%. It's pretty clear that you're making an R&D push this year for a reason in servers, networking, and Internet of Things. What's the view, maybe it's a bit early, let us know, about 2017? Is it worth making an incremental push in 2017 to nail the competition, or basically is extra investment in 2016 enough?
We are planning on expanding generally over the next few years. The extra investments that we outlined last September aren't the only source of growth in terms of R&D in the business. You've seen us grow quite significantly over the last few years, we do expect over the coming years to keep expanding. What we saw explicitly last year was based on the success that we were having in those markets, the opportunity to push ahead. We view competition quite broadly. In different markets, there are different people who we compete with. In the microcontroller space, that's very fragmented, as an example. Some of the work we're doing is to help defragment that and create some standardization around the Arm architecture. In networking competition, there is different. There's incumbency that we need to address.
We've got to make sure we've got the right product, that we're working with the right ecosystem, and that's where those investments are going as well. When I think generally about how we've expanded over the last few years, it's been about deploying resources in a way that helps us sustain our very strong position that we've got in the mobile market whilst broadening the success we've had in other markets as well. It's incorrect, as some people do, I know, Francois, you know Arm very well, a lot of people still assume that we are completely dominated by what goes on in mobile. As you've seen from the numbers here, it's 45% of the units. We are having quite significant success outside of mobile. The investments that we spelled out are to really capitalize on that yet further.
If I translate that into the numbers for next year, we should see operating leverage come back to the numbers next year.
The operating leverage is going to be a function of where revenues go, obviously, and part of that is royalty that we can't control, at all. I was going to say very much, but at all. In terms of continued investment in headcount, we're expecting after the step up that we've gone through, for growth rates to return to more the kind of typical growth that you've seen on a quarter-to-quarter or year-to-year basis from us.
All right. Thank you so much.
Great. Thanks, Francois.
Question is from the line of Sandeep Deshpande from J.P. Morgan. Your line is open.
Yeah. Hi. Thank you. Simon, maybe two questions. Firstly, on the mobile market itself, do you think you've now seen the worst decline in the mobile market and now from here on, you can see the numbers reported by your customers that the biggest decline is over and there will be stable trends from there into the next quarter in the mobile market, or do you still expect, based on what you're seeing in industry reports, et cetera, that this may go down again in the following quarter? A follow-up on the server/networking part. Yes. You have got some wins with the data center guys. What about in the networking market itself for virtualized servers, for the networking market? Thank you.
Okay. To your first question, in terms of mobile growth, we're looking at the industry reports that are coming out at the moment. Haven't actually seen a number for unit growth in Q1, but the forecasts for the year are still in that 6%-7% growth rate range that we were talking about 90 days ago. 2016, it looks like it's going to pan out at the moment based on the forecast, from what everybody expected. From one quarter to the next, to be honest, I really do not spend a lot of time worrying about that. Totally outside of my control. It will be what it will be. We think in the long term, smartphones are going to continue to grow.
We're going to see increased functionality within those smartphones that's going to drive the demand for more computing performance and more sophisticated processes from Arm's partners. We're seeing core counts go up. We're seeing increased adoption of Octa. We've seen some deca-core products launched just in the last few weeks. MediaTek launched their X20 and X25 products, which have 10 cores in them. That is a trend we're expecting to continue. From one quarter to the next, you're going to get fluctuations within that. Really don't spend a lot of time trying to analyze it. Your second question was about supporting virtualization in network servers, I think. Was that right?
Yes. What I'm talking about is in a networking market, looking to virtualize altogether, and whether Arm is going to be positioned as for servers in a virtualized networking market.
We built features into Armv8 to support more virtualization. We talked a moment ago about the work we're doing in the OPNFV world to make sure that the software stack that is optimized to utilize those features and create efficient implementations of that. We're doing a lot of work, both directly in Arm, across our partners within Linaro, on making sure that all the open source software to support efficient running of virtualized networks on Arm is all there and all well supported. That work is a good example of where some of our increased investment is going. We're happy with progress on that, and see no reason why Arm technology can't be used to implement that class of networking product, which is going to be really important to deliver the performance that 5G requires.
Thank you.
Again, if you do wish to ask a question or make a comment, please press the star one on your phone keypad and wait for your name to be announced. The next question is from the line of Brett Simpson. Your line is open.
Thanks very much. Simon, just looking at consensus for PD royalties in the June quarter, I think it's about $185 million, which is coming off a $192 million base this quarter, is about 3% down Q1Q in dollar terms, which suggests a pretty healthy seasonal period. When I look at the March quarter for the industry, we've had an earthquake in Taiwan that took out capacity at TSMC, and I think we all know the iPhone builds have seen big cutbacks in the March quarter sequentially. When you say you're happy with consensus estimates, can you perhaps talk about why you see it being so strong despite some of these headwinds that's clearly impacted the sector? Thank you. Maybe just a follow-up as well.
Looking at Q1, can you maybe just talk about where we are with v8 and Mali and Octa as a percentage of smartphone units in the Q1 period? Thank you.
Okay. When we talk about in our outlook statement there about expecting group dollar revenues to be in line with market expectations, we are talking about the full year there. Again, one quarter is next. That split, as you know, we don't guide for the breakdown between licensing and royalty revenues. We're talking about the full year. As you say, there are things that have gone on. There was the earthquake that affected TSMC's capacity. I believe they've largely recovered that situation right now. There was, of course, a tragic earthquake the other day in Japan. Our understanding is that both Sony and Renesas are impacted to some degree around that. They are things that may well affect that.
That's why in our outlook statement, we talk about being comfortable with that guidance modulo the macroeconomic environment, that's the sort of thing that plays into it. In terms of overall demand for our products, we have a very strong pipeline of licensing opportunities and the big companies that we work with are investing for the future and investing to win. There's always going to be something that happens that is outside anyone's control. That's why the macro factor does play into it. Second part of your question was about breakdown of v8, Mali, and Octa in Q1 smartphone shipments. Was that how you wanted to look at it?
That's right. Yeah.
Yeah. Our estimate of that is v8A in about 65% of handsets, Mali in about 50%, and Octa in about 25%.
That's great. Thanks very much.
Great.
Line of Amit Harchandani, Citigroup. Your line is open.
Good morning, gentlemen. Amit Harchandani from Citigroup, a question, if I may, on your mobile royalties. We've seen over the years the move towards making silicon in-house by some of the larger players such as Apple and Samsung. In the past, for example, we have looked at mobile royalties as a % of ASPs, and there's been talk about ASPs declining or gradually coming down. My question is, as we move towards increased sourcing of content in-house as opposed to merchant silicon. How does that impact the way you drive your royalties from these customers? Is there any incentive to move away from % to maybe flat pricing that in turn would encourage greater uptake of Arm-based content? If you could just talk about the impact of moving away from merchant silicon to in-house silicon, that would be helpful. Thank you.
As you say, there are companies that do this. There are a lot of companies that don't do this because the merchant market is still supplying a lot of silicon into handsets. The concept of kind of OEMs building their own silicon is something that we've dealt with in our business model for many years. This isn't actually new to us. It's been a situation that's been around for a long time. We have ways of modeling the cost of chips, so that the royalty that we get from somebody who's building silicon via an in-house supplier is paying to a first approximation, the same royalty as somebody would if the silicon came from a merchant supplier. That is a very well-trodden path for us. We're used to dealing with it. We have models for that.
If I could just get a clarification around that. Nothing really changes as we move towards the greater adoption of v8, octa-core, and Mali? It still remains the same tried and tested approach that you've followed in the past.
Yeah. We're not seeing any big shift away from our conventional models. It's something that does evolve all the time. As you know, we've changed the rates around the newer technology, reflecting the value that it brings. Our business models do evolve as the industry evolves, but there's no big sort of step function that's going on or that we anticipate to go on.
Thank you.
Question is of Johannes Schaller of Deutsche Bank. Your line is open.
Hey, gentlemen. Thanks for taking my question. It looks like the adoption of v8 in mobile is maybe trending a little bit behind what you initially planned. Sounds like most of the low end stays on v7, at least for a while. Also for octa-core, maybe there's a little bit of less incremental adoption this year than we've seen last year. Just wondering if you could share with us some of your thoughts around the reasons for that. Is it costs from the client side? Is there technical reasons for that? If you could give us a bit of an update around that would be helpful. Thank you.
No, there's no technical reasons for that. I think that we're not a million miles away from where we thought we were going to be. The number of smartphones in a quarter that have got a v8 versus a v7 is down a mix shift fundamentally. What we've seen for quite a long time, probably over a year now, or probably about 18 months, is a shift to v8 based chips from the supply chain. Now at the very low end, where there is a supply of v7 based chips that are very low cost, you're going to see those get worked through the channel. I don't think there's any change in our view that in the long term, we expect all smartphones to move to v8. When we set that out, we said that was going to be a multi-year journey. The initial uptake was very rapid.
Uptake or mix shift slowed down maybe a little bit. We think through this year, we'll end the year with about 80%-85% of smartphones being v8 based. It's going to asymptote over the next couple of years to 100%. Again, one quarter to the next, things will speed up, things will slow down. Not something that I can control, not something I worry about too much. The numbers are what the numbers are. It's very much an output of what happens. The trend's unchanged.
That's clear. On Armv8, you seem to be quite confident. If you just look at octa-cores as a quick follow-up, targeting 30% maybe by the end of this year, 60% longer term. Is there any kind of view you have if that's a relatively gradual, steady progression towards the 60% over time? Or should we expect something that is a bit kind of out of line for next year to the positive or to the negative side?
I think there will be a fairly steady progression on there. Again, not expecting any big steps. As chips move to more advanced process geometries and it's economic to put more cores into each chip at same or small incremental price, then I think we're going to see functionality of smartphone chips go up at the entry level all the way up. The high core counts are going to trickle down. We think getting to 50%, 60% over the next few years is entirely feasible.
Okay. That's very helpful. Thanks.
Participants, once again, if you do wish to ask a question or make a comment, please press the star followed by one on your telephone keypad and wait for your name to be announced. Your next question is from the line of Euan Lamb from Liberum. Your line is open.
Hi there. It's Euan. Thanks for letting me ask the question. Accounts receivables increased, I think 54% year-on-year or GBP 18 million.
Accounts receivables is now growing at more than 2x revenue. Has there been a change in the revenue mix or accounting policy to lead to this faster growth in accounts receivable?
No change in policy. The increase is sort of a good news story and reflective really of the great bookings quarter that we had on the licensing side. It's just a function that if you close a number of deals in the last couple of months of the quarter, then you're going to have a big accounts receivable balance at the end of it.
Okay. We shouldn't expect any difference in the cash conversion or working capital outflow this year.
No, I mean.
with billed receivables?
No, is the short answer. I think when you look at our cash conversion, it's sort of somewhere between 90% and 110%, depending on license growth. When we've got high license growth period, and you can see that over the sort of the years from 2012, 2014, that's when cash conversion was a bit more than 100%. It trended down as we delivered on those licenses and the payment terms became due. It'll bounce around 100%. I don't see any reason why it won't, on average, over a multi-year time, be 100%.
Okay. Thank you.
Question is from the line of Vijay Anand. Your line is open.
Thank you. I had a question on the server market, specifically, your thoughts on competition from the Power Architecture. We've seen some statements from Google pretty recently that they've made much more progress in porting their software on the Power Architecture than on Arm. Just wanted to get your thoughts on how you see Power as a competitor, and whether you see them as a threat in terms of achieving your 25% market share target by 2020.
Well, I think the server market could well be much more fragmented in the future than it has been. Obviously IBM are a very long-term player in this market. Power Architecture's been around a long time. It's been in service since about 1995. It has something of a track record in this space. It was designed for that market in the first place. IBM created power.org to help push that into the future. I look at the Arm partnership and the number of silicon vendors and the coming together of the ecosystem that we have in the data center space, and there'll be competition in the future. I think what's been going on around other architectures doesn't change our view of where we could get to.
Okay, thank you.
The next question is from the line of Robert Lamb. Your line is open.
From Jefferies. My question, just following up on Euan's question about the trade receivables, and just the mix there in terms of the amounts recoverable on contract. It went up obviously at the tail end of last year, and it has come down this quarter, but it is still a bit high, I guess, compared to what we have seen over the last few years. I just wanted to understand what the business dynamic that has been going on that has caused this to increase. Is it fair that the reason it has come down is because it is now converted into trade receivables? If so, it just seems like the cash collection on this seems a little bit longer than normal. I just wanted to understand, is it linked to the new investment cycle? Just any color here would be great.
As you said, the AROC came down this quarter, and I know there were questions about it going up last quarter, and what I said then was it is really down to the timing of the quarter end and how that stacks with milestones in the license agreement. We were expecting it to come down. It has come down. That is what I said it would happen when we last reported. At the same time, we are signing new contracts that will serve to increase AROC. Again, I go back to the sort of the comments I made about accounts receivable. It is really to do with the timing of payment milestones, and that there is no change in the business model. There is no change in policy. It is the normal ebb and flow of the licensing process.
Okay, great. Thanks for the clarification.
Question is from the line of Lee Simpson from Jefferies. The line is open.
Great, thanks. Good morning. Thanks for letting me on. Just a quick one again on OpEx, if I could. Just trying to clarify what you mean by historic OpEx growth levels. Really just trying to understand what the underlying headcount implications might be of growth there before wage inflation, how does that sag with medium-term needs for servers investment and investment in the Mbed Incubator business?
Yeah, when you look back over time, it's been around sort of a 5%-7% increase in headcount, then you've got the wage inflation on top. It sort of translates into high single digit. That's what we're talking about for the investment. As we've said, we announced the big step up in September. We wanted to flag it. That's worked its way through now, we're returning to that more normal level of headcount and wage inflation. As you know, the majority of our cost is people. It's a headcount driven number.
Okay. That 5%-7% headcount increase, is it skewed to server space in particular, is it skewed to the Mbed Incubator? Is any color you could give in that?
It's quite broadly across the business. What we explained on the call just now was how we're investing in core roadmaps, so driving our processors, our GPU, our Physical IP to create the next products to enable server networking and mobile devices to be implemented. It's on the ecosystem around networking servers, particularly, and it's into our IoT team. They're looking at the Mbed platform. It's quite broad. I wouldn't say it's more in one area than the other. On those new fronts, that's where that incremental investment is mainly going.
Got you. Thanks so much.
Your next question is from the line of Anil Doradla from William Blair. Your line is open.
Hey, guys. Thanks for taking my question. Simon, one quick question on the software ecosystem for the data center market, for the infra server market. Obviously, you're seeing a lot of investment. Some of your chipset vendors are investing quite a bit. When we step back and looking at the big picture, how ready do you think the Arm server chips are ready to take on perhaps the whole ecosystem? Can you share which areas that still need a little bit more work on the software ecosystem? Thanks a lot.
It's hard to pinpoint a big area where we need to go work on. What we've been doing is working through with people as they are looking at particular workloads, looking at particular deployments, and seeing where the gap is and where we can optimize to get an even better implementation. There's been some broad kind of foundational work around the basic Linux infrastructure. The infrastructure that you need if you're a manager of a data center to install a server and manage it. Then it's down to particular workloads. What we're showing in the roadshow slides are different workloads that we're looking at from storage to web serving, going up through data analytics, et cetera. We're kind of knocking these off, and they're very much driven by our partners and where they see opportunity, where end customers want to use Arm-based servers.
We're getting quite focused on optimizing particular loads. It's not any one big thing. It's in fact, lots of little things. When you come down to it, this is often the case.
All right. Thanks a lot.
Next question is from the line of Martin O'Sullivan from Cenkos. Your line is open.
Yeah, thanks very much. Actually, my questions have been asked, but I was just wondering if you could give us a sense for how the recent acquisition and integration of Sansa is going, and whether there have been any notable milestones achieved with regards to Sansa, particularly with regards to IoT security.
Integration and working with that team, going very, very well. We've expanded that a little bit. We've put more heads into Israel. We have merged that team in with the rest of the guys who are working on the IoT area and within part of our business that's focused on systems IP. Some of the hardware security features that came from Sansa have been merged into that. Very happy with how that's going, and the integration of that team. I'm sorry, there was a second part of your question, I think.
There wasn't actually, no.
Right. Okay.
It's just really with regard to IoT security, how that's progressing.
Okay.
Okay. Thanks very much. Thanks.
Question is from the line of Jérôme Ramel. Your line is open.
Yeah. Good morning. I got a quick question, Simon, concerning your new agreement with TSMC on seven nanometer node, specifically the part on the High Performance Compute System on Chips. Should we read that you are maybe more aggressive targeting server networking with seven nanometer node? Should we expect a kind of acceleration of design wins due to the seven nanometer, or maybe just the 16 nanometer node will be enough to see the ramp-up in servers?
I don't think that our expectations of growth aren't dependent on any particular process node. We have for many years engaged with TSMC and others on advanced process technologies, on looking at how those process technologies are developed and the co-optimization of our processor with their process. The work that we have going on on 7 nanometer, 10 nanometer, and below is a continuation of that work. It's making sure that we kind of trailblaze with our foundry partners the implementation of Arm-based SOCs on their high-end processes, so that when our mutual customers then get to it, they know they can rely on the results and know that we have worked together to make it easier and make the result more deterministic. The work that we have going on on 7 nanometer is really an extension of that.
It's going to help with the high performance compute market, with servers, with high-end networking. It really is just a continuation of our sort of modus operandi of how we work with those partners who are developing advanced manufacturing processes.
Okay. Thank you very much.
Line of Alexander Duval. Your line is open.
Yes. Hi, everyone. Alex Duval from Goldman. I wondered, can you talk a little bit more about Arm's ability to benefit strategically from the buildup of the semis industry in China? We know there are significant plans to invest in the industry there perhaps catalyzed a bit by the issues ZTE has experienced in terms of sourcing components. Related to the China question, I wondered if you could give some more color on the consortium that was announced of Alibaba, Phytium, Baidu, et cetera, working with Arm on cloud computing, with some institutions in China. Many thanks.
Sure. Yeah. As you know, the Chinese government has an initiative around developing local semiconductor companies, and ensuring that more devices are sourced locally than are imported. I've seen data that says that China imports more semiconductors by value than they do oil. That has been a priority for some time. We've seen the growth of local companies like Spreadtrum, for example, Rockchip, Allwinner. We've seen the growth of within Huawei, their team HiSilicon are developing some of the most advanced chips in the world right now. That's a development that we've been close to for a long time, and we've been operating in China for 15 years or so by now. Have a team on the ground working very closely with those Chinese semiconductor companies, providing support and helping enable the innovation that they want to achieve locally.
Obviously, by using Arm technology, it prevents somebody from needing to reinvent the wheel because so many chips have microprocessors in them. There's so much software in the world that runs on Arm. It enables anyone, anywhere, to build chips that can then sell into a global market. Our business model is very supportive of the objectives of the Chinese semiconductor industry. We're seeing them invest a lot in manufacturing at the same time. How are those chips going to be used? Again, there is an initiative in China to ensure that their server infrastructure for the banking industries, for the local cloud companies, et cetera, is served locally. That led to Arm's participation in what's called the Green Server Alliance, which got announced last week. That again, is something that's been sometime in the making, sponsored by MIIT in China.
One of their Vice Minister, Vice Minister Huai gave the opening speech at that event last Friday. There are a number of companies involved. Arm is a platinum member of that. You mentioned some of the companies there, Alibaba, Baidu, Dell, the new joint venture company with Qualcomm is involved in that. H3C, HPE, Huawei, Lenovo, Phytium. Lots of companies, big names, are involved in that. It's an important initiative to develop local server infrastructure that is both energy efficient, and open by way of platform and participation.
Super helpful. Many thanks.
Thanks.
From the line of Douglas Smith. Your line is open.
Yes. Simon, there was a line in the press release about increasing R&D to develop next generation processor technology. Where do you think Arm processor technology will be in five years' time? What is the roadmap going forward for the next generation?
Well, our roadmap's always been driven really by maximizing the benefits of process technology and enabling next generation applications to be implemented as efficiently as possible. We take a systems view on that. It isn't just about raw instruction throughput or number of threads within the processor. We're looking at system design. We're looking at the applications that are going to be run in terms of workload, and what's the best way to partition that across a CPU or GPU, specific accelerators, how you optimize data flow around the chip, on and off chip, to minimize power consumption. As an overall theme driving our roadmap, they're the things that we worry about.
In terms of specifically where our roadmap's going, if you look back over time, when we introduced Cortex as a technology, we created specific architecture variants to enable the application processes, the real time processes, the microcontroller processes, to have the right feature set, yet have architectural consistency across them. More recently, we've seen at the high end of that, as we've achieved success in the enterprise space, at the high end, V8A, there are some processes with more enterprise level features in them, than the ones that we developed for the mobile market, which don't need those features. We've seen a kind of a split of the roadmap up the top there. I expect those to really be the trends over the next five years. As we're more successful in enterprise, I think that's an area where we'll push on performance.
We'll look at the different workloads that are being run on Arm, that will probably drive some evolution there. Meanwhile, the rest of the roadmap, again, just focused on that system efficiency, and delivering the best performance at the lowest possible power.
Great. Can I ask a follow-up on a previous question someone asked?
Sure.
Yeah. On the TSMC seven nanometer high performance computer on the press release. Obviously, data center is one of the few areas where you're not really there yet, I think there's still some skepticism out there. Is it thinking that it'll take until seven nanometers until the Arm architecture can really compete against the supplier who's dominant in that space?
Well, as I said earlier, we don't think that we need to wait for 7 nanometer processes to come along before we get anywhere in the server space. In our roadshow slides, if you have a look at slide 25.
25. Right.
The slide shows the growing number of markets which can be served as more performance and different chips come along. That shows growing from 2015, where we already addressed the storage and web serving market to over time. Through the combination of many functions and process technologies, one of them, more and more of the market being addressable by Arm technology. This is one of the factors in it, but it's not the case that nothing happens until 7 nanometer comes along.
Sure. Actually, that was a slide that led me to my question because you have HPC engineering as 2018, which is, I think, exactly 7 nanometer at TSMC.
That also has initial underway in the middle of this year as well.
Right. Okay.
Again, that's a kind of evolutionary thing.
Mm-hmm. Okay. All right. Thanks a lot.
Your next question today is from the line of David O'Connor from Exane. Your line is open.
Great. Thanks for squeezing me in, guys. Simon, question for you. When I look at the 2016 royalty drivers, I think they're well understood, and you've indicated that you're exiting the year at high levels of the Armv8. Then also you talked about, in the medium term, 5G will be a strong driver, particularly new segments such as networking and servers when those architectures are rolled out. That seems to be more like 2018 timeframe. My question is around what you see as the royalty drivers for 2017. Thanks.
Yeah. 5G deployment really starts around 2018, and will go for a number of years. It's going to be something that is a driver in that timeframe. In the meantime, there are many networking applications which are clearly being served by Arm chips today. We've seen 10% year-on-year growth. We've seen our royalties perform strongly driven by embedded, driven by mobile, driven by enterprise and home. We're expecting those trends to continue. We're going to gain share in many different markets. 5G is a new technology, but it's not like there is no evolution between 5G and now. Really what we're seeing with the development of advanced networking equipment, it is more evolutionary. There's not just a big bang that comes along every so often from 3G to 4G to 5G. In between, you're seeing the rollout of more and more advanced technologies.
That's going to be a driver over the next few years.
Great. Thank you.
Next question is from the line of Jaguar Bajwa. Your line is open.
Hi. Thanks for taking our question. On the licensing side, probably a bit more near term, but given the receivables grew a lot in the quarter, it would imply that a lot of deals were done late in the quarter. Also when I look at the last time your mix of turns versus backlog, in licensing was at 60% over the last four years, was in Q4 2014, and the following quarter was quite weak for licensing. We can also see the next six month composition of backlog is at a relatively low level. It's down year-on-year. This would imply, given your statements around robust licensing, that your turns visibility is very good. Can you just talk about that, given your historical trend that we saw in Q4 where the following quarter was quite weak? Just the confidence you have around that robust pipeline.
Just secondly, on the Mali penetration, I was just wondering how you see that exiting the year currently at 50%. Do you see that more runway to grow that? Also the royalty rates that you see trending for Mali. Do you see it staying at this rate or is there room to grow that? Thanks.
Just on the backlog. The backlog is actually flat year-on-year, we're not down. We're down Q-on-Q. I don't think you can read too much into the mechanics of turns versus backlog in terms of what that means perspectively, because prospective license income is all about the pipeline of deals that we have in our CRM system, and we have pretty good visibility of what that's going to be over the next 12 months. That's what's behind the guidance that we've given around full year revenues.
Okay. I was just talking about the six-month backlog. I think that's down year-on-year, okay.
The second question was?
Second question was about Mali. Our expectation is that
Mali gets to about 60% penetration in smartphones over the full year. Exit rate, I would expect to see about that level, maybe a little bit higher. Royalty rates, in Mali, the journey there, very similar to processors. As we add new capability, add new functionality, we expect to edge up royalties over time. I wouldn't brace yourself for any big step change in that in the near term.
Okay, great. Thanks. That's very helpful. Thanks.
Next question is from the line of François Meunier from Morgan Stanley. Your line is open.
I just wanted to ask about the Chinese server opportunity, someone else asked the question, thank you very much.
Okay. Perhaps we can just make this the last question as we are running out of time.
Your last question is from the line of Amit Harchandani. Your line is open.
Thank you. Just a quick follow-up around your current thoughts with regards to M&A. If you could please share how you are thinking about M&A, any changes in thought process there are more or less consistent to what you have said in the past, and whether it is still more centered around investing around the software, building up ecosystems. Any thoughts would be helpful. Thank you.
Yeah, there is no real change in our philosophy right now. We are looking at what IP fits our model, what are the IP building blocks that people are going to need for their chip devices to address the future generation of needs. What makes sense for us to do, what makes sense for the ecosystem to do from both a hardware and software perspective. No real change on that. You have seen us do a number of acquisitions over the last little while. Very happy with how that is going and the philosophy that we are taking to it.
Thank you.
Okay. Well, I'm afraid we're going to have to call it a day there. Thank you very much to everyone for dialing in today, and thank you for your questions. We will see you for Q2 results in July. Thank you and good morning.
That does conclude the conference for today. Thank you all for participating and you may now disconnect.