Good morning, everyone, and welcome to Arm's Q2 and H1 results. Thank you for joining us this morning. Hopefully, you've had a chance to read the press release and are primed with some questions. We'll come on to questions later. I'm going to give a bit of an overview of the business first. I'll then hand over to Tim, who will talk through some of the numbers, and then we'll do Q&A up here in the chairs. Before we get going, I'll just refer you to the usual cautionary statements. I assume you're all very familiar with this, and we'll take it as read. Let me start by giving a bit of an update into what's going on across the business. As you have seen from the release, our revenues in dollar terms were up 17% year-on-year.
In this period, the dollar sterling exchange rate moved against us. In U.K. terms, in sterling terms, our revenues were up 9%. The performance of our business was driven very heavily by very strong licensing, up 42% year-on-year. Royalty in this period was up 2%, and I'll come back to that later. In terms of licensing, this has been one of our strongest quarters ever, with 41 licenses signed across a very broad range of technology, including two licenses for products we haven't actually announced yet. I'll come on to those a bit later on. Despite the exchange rate headwind, our earnings per share was up 11%, and we've declared today the interim dividend up 20% year-on-year. Let me come back to royalty.
As you'll have seen, our royalty grew 2% year-on-year, and that's lower than we've experienced in recent quarters. That's been driven by a number of factors, many of which have been anticipated. There's been seasonality in our numbers. Q2 is ordinarily a weaker quarter. There's been some industry cyclicality and actually a tough comp year-on-year. 12 months ago, we had a very strong quarter in royalty performance. Industry cycles do happen, especially in consumer markets. When I talk about consumer, I mean phones, TVs, DVD players, digital cameras, and the like. When we look at the composition of Arm's royalties, about 80% of the royalty value comes from these consumer-driven markets. When there is a cycle there, our royalty revenues are disproportionately impacted by that. You can see that cycles are nothing new to Arm.
We're showing here on the slide, in the mobile sector alone, some of the peaks and valleys that we've experienced over the last few years, and you can see that some of the peaks are higher and some of the valleys are, in fact, lower. Over the years, as we've been through these cycles, there have been a number of factors that drive the industry up and down. We've seen different product mixes. We've seen technical discontinuities come along. We've seen the anticipation of new devices drive the markets up and drive the markets down. There have been a number of these kind of factors that have gone on over the last few quarters. In the last quarter, one such example is in China, where the operators are using subsidies to heavily drive the sell-through of 3G handsets in anticipation of moving to 4G later on.
What that means is there's a delay in the uptake of the most advanced chips from our customers by the OEMs, and that is one of the factors that read through into our royalties. There are a number of things going on. The real question, I guess, is, are we at the nadir? Are we at the bottom of the valley? Are we still on a downward trend? From what we see looking into the industry, talking to our customers, looking at overall trends, we believe that our licensees have already experienced a stronger Q2 for them that will result in stronger Q3 royalties for Arm, and there is a momentum building that will carry into Q4 and beyond. We think we are at or certainly near the bottom of this cycle.
One factor that gives us confidence and underpins our belief in further strong royalty growth for Arm is the strength of licensing of our advanced technology, our Version 8 architecture products. We've now licensed Version 8 products 50 times to 28 different companies. When we look at some of the key segments for Arm, we see that all of the top 10 chip companies that build chips for smartphones have licensed Version 8. Nine out of 10 of the companies that build application processors for tablets have licensed Version 8. In consumer electronics, TVs, DVD players, set-top boxes, it's four out of five. In enterprise networking, where we're growing market share, again, four out of five of the top guys who sell into this market have licensed Version 8 of the Arm architecture. Version 8, as you know, commands a higher royalty rate than its predecessors.
As these devices come to market, as they end up in OEM products and deployed into data centers, this will drive higher royalties for Arm. In Q2, as I mentioned, we started licensing new technology roadmap products that we haven't actually announced yet. We have two products that have been licensed, and today these go by the code names of Maya and Artemis. We're not going to give a lot of detail about those products today. It's a bit of a teaser. I'll talk about that more openly later on. In the meantime, if you want to know how to spell Maya, come and see Ian afterwards, and we'll give you that. More on those products another time. Artemis and Maya are results, fruits of the investment that we've made in R&D over the years.
If you look at our headcount over the last 12 months, we've increased by about 500. Most of the people that we've hired are engineers working on our next generation of technology. They're working on multiple products. Artemis and Maya are just two of the things that are coming out now. We have many investments going on in many different product lines to help us drive the business forwards. The products that they're creating are going to help us sustain market share in markets where we're traditionally strong. In mobile, we see huge growth ahead of us. We believe that the products that we're building will enable Arm to maintain a very high market share there.
It's going to help us gain share in some of our non-traditional markets like automotive, like the data center, as cloud develops, and in the networking, as I said, where we are starting to grow share. It's also going to help us capture a large share in new and emerging markets, in the Internet of Things, in wearables. These are new markets which have much growth ahead of them, we believe, and many changes ahead of them. We believe the products that we're building will allow us to take a very large share there. This investment, this continued investment in R&D is really important to us. Now, if you came to our Analyst Day in May, you'll have seen this slide.
Just to give a recap here, all of those different technology areas we grouped broadly into three different pillars: mobile application processors, enterprise infrastructure, and embedded intelligence. I'll just touch now on the progress in all three. When we look at our progress in mobile, 12 months ago, I stood here outlining how we see the mobile market, the smartphone market, evolving over the next five years, how the different categories of smartphone will experience different growth rates over the next five years, and how the kind of different semiconductor devices our partners are building will power those devices. 12 months ago, we talked about our expectations for the adoption of Version 8 of the Arm architecture, for the adoption of our graphics technology and our video technology.
The progress that we've seen in the last 12 months really does underpin our belief in the growth and in the model that we spelt out, leading to somewhere between 15%-25% CAGR in our royalty revenue stream from mobile. In the last quarter, the very strong licensing further gives us confidence in that model. I'm more confident about that today than I was 90 days ago and I was 12 months ago. Version 8 licensing, as I said, has been very strong. We're expecting to see many new devices based upon silicon built on v8 later this year and into next year. When we look at progress on servers and in enterprise infrastructure, the last quarter has been really interesting and a really positive thing, I believe, for Arm-powered server chips.
We've seen three of our licensees, AMD, Applied Micro, and Cavium, all do some fairly major announcements about what they're doing and the products that they're building. Early in the quarter, AMD demoed their first product, they talked about their roadmap using their architecture license from Arm, and they talked about devices which are pin compatible, enabling and accelerating a migration to Arm-based servers. At Computex, we saw Cavium launch a range of devices tailored for different workloads and delivering very high compute performance with multiple very high core count devices. Most recently, we saw Applied Micro launch their products tailored at high-performance computing, but in a very low power envelope. We've seen a lot of silicon devices, and we expect those to ship in real servers towards the end of this year as those devices start to get delivered to our customer's customer.
It's been a very strong quarter for progress in the enterprise. In embedded intelligence, nearly half the licenses that we signed in Q2 were for our M-class processors. These are the products that we've designed specifically to target microcontrollers and very small and very power-efficient systems on chip that will be in wearable devices, Internet of Things devices, and in embedded products everywhere. Those licenses, we've done that 20 licenses for Cortex-M, we expect to come through into silicon over the next few years and contribute to the strong growth we're seeing in unit shipments of Cortex-M. It's about 900 million units in Q2 of Cortex-M-class Arm-based microcontrollers. As well as on the silicon side, we've seen some evolution in the ecosystem around the embedded space as well in the quarter.
You may not be familiar with it, the board in the middle is an Arduino board. Arduino is a development platform with a very large ecosystem around it, which has historically been based on very simple eight-bit microcontrollers. In the last quarter, we saw new products from Arduino based on 32-bit Arm processors, helping that developer community move over to higher performance devices. Similarly, Atmel, who've been a big driver behind Arduino as well, are evolving their ecosystem called Atmel | SMART to make it easier for their customers to move from very simple devices to more sophisticated, more higher performance 32-bit Arm-based devices. The silicon side is evolving. Lots of customers coming to us looking to innovate in this space. We all see, many of us see the growth of embedded as a real driver for the semiconductor industry.
In summary, we've made great strategic progress in the last 90 days. A strong licensing performance, developments in our ecosystem, really do underpin the long-term growth opportunity for Arm. Licensing is a precursor to royalty. Licenses that we've signed today will come through in volume, typically in a two to three-year time period. 41 licenses signed in a quarter is really significant, one of our best licensing performances ever, and that does act as a very strong precursor to future royalty growth. Royalty in the short term has been impacted by various market conditions. When I look at the activity that's going on, Arm and our licensees have not lost market share anywhere. This is about the dynamics of the end market.
It isn't about market share changes we're continuing to see, as is evidenced by the 41 licenses, strong uptake of our newest technology, and that will drive us into the future. Based on the conditions we can see today, based on what is going on in the market, we do expect better royalty performance in Q3 and that momentum to accelerate into Q4. With that, let me hand over to Tim to talk through some of the numbers, and then we'll get back into Q&A.
Thanks, Simon. Morning, everybody. I think Simon has given us a good overview of the licensing and royalty dynamics that have impacted both Q2 and what we should expect down the road in terms of royalty growth. This morning, I'll keep it fairly brief, and just focus in on a bit more color on the P&L. Some comments and reminders about balance sheet, cash, capital structure. Sort of revisit a little bit the backlog dynamics, because obviously, it's an interesting licensing quarter in terms of the number of licenses, the movement in the backlog, the outlook for the rest of the year. We'll give a little bit of color on that. Just to reiterate the outlook, how we see the balance of the year. As Simon said, 17% dollar growth. Quite a tough headwind at the moment with regard to the currency.
The Q2 effective translation rate was 165 in 2014 compared with 154 last year. That's a relatively painful move. That translated, as Simon says, into 9% sterling revenue growth. PBT, therefore, more closely aligned with that, again up 9%. Now that is after another, if you sense headwind from currency, of the mark-to-market when we revalue our balance sheet, revalue our forward contracts at the quarter end. The FX actually moved from about 166 and some at the end of Q1 to 171 at the end of Q2, that typically would result in us having a mark-to-market charge in that environment. When the dollar strengthens quarter-on-quarter, you tend to get a credit. Yeah, in most quarters it's ±GBP 1 million. In this particular quarter, with quite a significant weakening through Q2 of the dollar, it's GBP 3 million.
On an underlying basis, OpEx GBP 85 million in Q2 compares to our previous guidance and consensus of about GBP 87 million. Running a little bit under on an underlying basis. You'll see there the guidance for the third quarter, a little bit higher in the sort of GBP 90 million-GBP 92 million range. We have been continuing, as Simon said, to hire, to invest in the R&D opportunity. We're also investing in, if you like, the back office, the infrastructure, the IT, all of the things that underpin the future growth of this business. Q3 also carries some costs that don't necessarily appear in all quarters. We have our big annual partner meeting in the third quarter, probably our biggest single marketing event. That gets charged in the third quarter.
We also closed a sort of bolt-on acquisition at the end of May called Duolog, which you'll see referred to in the back of the release. Relatively small acquisition, but 60+ people. The cost of those people will also be reflected in the Q3 numbers. Somewhere in the sort of early GBP 90 million we would expect as a base case at current exchange rates. That has driven a normalized EPS up 11%. The items in between there, I think as we all know, we're all getting less return on our cash at the moment. There are one or two bankers in the room who are looking a bit sheepish in terms of how much yield they're giving us on our cash. That suffered a little bit.
On the other side of the equation, the tax rate is running at about 18%, just over, compared to just over 20% last year. This is the dual benefit of the ongoing implementation of Patent Box, which, as most of you know, is being implemented over 5 years, and also the U.K. corporation tax rate continues to be legislated down. So the tax tailwind continues, earnings up 11%. Also, in Q2, not only are we investing in our new R&D capability and innovation, we have also been reviewing, as we look forward, the mix of sort of skills and capability generally that we need in Arm as we grow. So it is not all about going out and identifying the new. We are also looking at the rump and the mix that we have. And that has given rise to around about 130 people leaving Arm in Q2.
Net head count obviously up, on a continuous basis, if you like, reviewing the skills mix, and this was a particularly focused exercise around doing that. So there is a restructuring charge of GBP 8 million in the quarter. Moving on to the cash and the balance sheet. Strong quarter. Cash generation, GBP 87 million. You can see on the right there the sort of cash journey that we have been on. Without boring you with ancient history, the way Arm has managed its cash is we were cash collectors until 2003, 2004. We then managed cash down, as you can see there, down to about GBP 50 million just before the world went into meltdown going into 2008. And we did that via a 3 and a half year rolling buyback, where we bought back about 16% of the share capital during that period.
When we went into the downturn, we sort of deliberately let the cash ride up, and as you can see, have broadly continued to do so, such that we find ourselves now with GBP 750 million net. So I think, with our cash balance, we have a lot of opportunity to invest. We have opportunity to do further M&A if we need to. And we have opportunity to continue to return cash to shareholders, probably on an increasing basis as we look forward. You can see the graph in the middle on the share count. Looking back again over a nine-year view, the share count has been flat. And we said at the beginning of this year that we continued to maintain a share count. What that will mean from here is a limited buyback. Current numbers, probably about 10 million shares in a year.
We did a little bit of that in Q2, 1.4 million shares, GBP 12 million. And you can expect to see more of that as we maintain that flat share count. And on the dividend front, since we introduced the dividend back in 2004, we doubled it at the beginning, and then have consistently grown it in the sort of 20% area. Some years plus, some years minus, depending what is going on in the market. And today, we are saying 20% in the context of earnings up 11%, consensus earnings for the full year up about 15%, and our future outlook up, we think, around those numbers at least. So we are increasing the interim dividend by 20%. So as I say, backlog I think is just worth spending a little bit of time on. Because lots of moving parts on license revenue.
Here's a quarter where we've beaten expectations by about GBP 15 million. We've signed, as Simon says, 41 licenses. I think that's the second time we've been through 40 licenses. In a sense yet, the backlog down 10% sequentially. It was down about 5% sequentially last quarter. What should we be reading into all that, I think is what I'll seek to address. I think it's worth reminding ourselves over the last five years, backlog has grown broadly 5x. What's been driving that? Even though it's been down the last couple of quarters, you can see it's pretty close to its historic high levels. I guess, the key drivers from the launch of version 8 of the architecture, a combination of things that really drive the backlog up. Architecture licenses, subscription licenses.
A subscription license, just to remind those who's not fresh in the mind, is a multi-year license. Often three years, sometimes five years, that a company that uses Arm widely across multiple divisions and targeting multiple end markets often finds it attractive to take a sort of corporate-wide license to a defined suite of our technology. They don't have to be negotiating individual licenses for individual processors all of the time. The way the accounting works for those licenses, because they incorporate technology to be developed during the period of the subscription, you can't recognize revenue in the normal percentage of completion way. You have to recognize it on a linear ratable basis. When you sign a subscription license, you get five years' worth or three years' worth of contract value in the backlog on day one.
That backlog pertaining to that particular license will reduce on a quarterly basis through to the end of that license. In most cases, our subscription licensees will then renew at the end of that subscription. Most of them stay with that model once they've embarked on it. After three or five years, the backlog will shoot back up again as they renew that subscription, and then it will wind down again. That's been quite a powerful dynamic in that growth. Also, the other key component of backlog is when you sign licenses with customers before you're in a position to deliver all of the things that they're signing up for. Clearly, you can't recognize revenue before you deliver products, or you still have costs to incur to complete those products.
When you're an early or a lead licensee, which is a model we've used throughout Arm's history, the value of the license will go into backlog. It will be transitioned from backlog to revenue as engineering milestones are met, contract deliverables are made. Therefore, inevitably, you go through a period of, as I say, recognizing ratable from the subscription. It just so happens that in the last couple of quarters, engineering milestones relating to the initial processes from the version 8 architecture have been met. Multiple licensees, multiple deliverables. Obviously, money goes from backlog into revenue as that occurs. You look forward, you look at the new subscriptions, you look at the subscription renewals. In a sense, the cycle then begins again as we introduce new technology.
Simon's just touched on a couple of the next wave of ARMv8 processors that are going to be coming to the market in due course. Lead licensees are already beginning to engage. That's driving out the backlog. What we say in the release is that we expect looking into the crystal ball of licensing in the second half, if you look at the mix of licenses between licenses that generate revenue in the short term, licenses that generate backlog, we expect the backlog to be rising in the second half from where we are, which I think by the time we get to the end of year, I think will be a strong support for target license revenue next year. You've heard me say many times that we don't expect to be growing license revenue 40% or 30% deep into the future.
I grant you that this accelerated growth is lasting longer and actually higher than I have been signaling. This is very good news. Don't be disappointed if in Q3 or Q4, we're not growing licensing at 40%. If you look at the comparisons in the same way that Simon said that the royalty compares in Q1 and Q2 have been tough. The licensing compares have been easier. You remember Q3 last year, we jumped through GBP 100 million of license revenue for the first time. Therefore, in Q3 and Q4, license revenue compares will get tougher. We see a world of accelerating royalty revenue growth year-over-year, but license revenue growth will be slower relative to the 42. Anyway, these are the dynamics. Backlog will go up some quarters, it will go down in others.
We are very confident it is supportive of the licensing outlook that's out there in the market, in 2015 and beyond. Reiterating the outlook. Licensing prospects in the second half look good. They look good for turns business, as we call it, that yields revenue in the short term. We also see the pipeline of opportunities generating outward pressure on the backlog. Simon's touched on royalty, industry data, guidance from our biggest shippers looks positive for a pickup, which we would expect to benefit from in Q3, in Q4, in Q1, because obviously we are one quarter in arrears. Putting all that together, we see our full year USD revenues being in line with market expectations, which are in the sort of $1.295 billion area at the moment, $1.295 billion. With that, we'll go to Q&A.
Perhaps I could ask in a Q&A, as we have been doing of late, if you could ask one question so we can get around the room and perhaps state your name and affiliation. Right. Oh.
Thanks. Achal Sutaria from Credit Suisse. Simon, you mentioned about 15%-25% growth in your smartphone royalties long term. We looked at that chart which you showed about mobile, like the impact from mobile inventory correction. Shouldn't we expect a significant rebound like we've seen in the past again this year in your royalty revenue growth, when you come out of this inventory correction in the second half of this year? 15%-25% growth is what you expect from smartphones, and then you've got additional share gains in enterprise and networking. I'm just trying to understand why can't your royalty revenues grow, like 25%-30%, in the next few years going forward?
Yeah, there may well be a strong rebound. That 15%-25% is a five-year CAGR for just for smartphones. As you say, when you layer on other things, potentially there's much higher royalty growth to come from that. It's all going to come down to the timing of when products come out. When you look at a server or networking equipment, are driven by very different market dynamics, very different time frames than consumer devices. Adding it all up, they're all layers of royalty, which contribute to long-term growth. The point I was making was specifically around long term for smartphones.
I think one thing to bear in mind on that is one of the assumptions in that 15%-25% is that by the end of that period, all smartphones, entry, mid, and premium, are incorporating version 8 of our technology, and that, as Simon said, is the one that yields the higher royalties. Actually today, that's a minimal contribution. Today we're just starting to, but we're not really seeing the benefit yet of the version 8 being designed into smartphones. That is a longer-term comment on that one sector.
Would it be fair to say that you can outgrow the semi market by more than 15%, which has been your guidance long term? That's where I was coming from.
Right. Well, certainly we have in the past, and I think we've been developing a strategy for broader market penetration, broader adoption of Arm technology, which potentially can lead to those kind of levels of outperformance in the future. We'll see.
Thank you.
For the next question. Gentleman behind you.
Yeah. Aditya Misra from Bank of America, Merrill Lynch. Simon, your guidance basically seems to imply 10% PD royalty growth for this year. When do you see your outperformance getting back to the 15 percentage points that you've talked about? Can we see that in 2015, considering your comps are going to be easier this year? Secondly, do you see your year-over-year
PD royalty growth rate accelerating into the first quarter of next year, considering this, your fourth quarter this year seems to have been affected by some of the push-outs on the 4G space.
The question is about royalty growth. Yes, certainly, I think the re-acceleration isn't something that just ends on December 31st. I think as we work through the inventory issues, one would expect to get back to something more normal. At the same time, I would expect the regular seasonality. You can see from the shape of that graph that these cycles do come and go on an 18-month to two-year kind of cadence. At some point in the future, I'm sure we'll be impacted by some of that again. The key for me is long-term, are we delivering the right technology into these markets? Are we continuing to maintain share in very high-share markets and grow new share? All of that is happening. I think that supports the kind of long-term model that we've outlined in the past.
Thank you.
Yeah. Maybe just behind you.
It's Janardan Menon from Liberum. I'm just trying to paint a picture of your royalty rate, your percentage of royalty, as you've transitioned out of ARM11. ARM11 has now fallen by 50% according to press release, and you're down to 3%. That's taken over by the Cortex-A class. That's going to start transitioning to the v8 architecture going forward. If you were to paint a picture of royalty rate going from, say, 1.1% when this journey sort of started and rising to somewhere in the 2%-3% range, is this transition now, the ARM11 finishing and the v8 starting off, is that an upward inflection moment in that graph, or is it a steady state, you're sort of ticking on 1.2, 1.3, 1.4, I mean, whatever the number is today?
If you could just give us a qualitative description of how you would see those transitions on your royalty rate.
I think in the mobile space, in the smartphone space, I wouldn't necessarily inflection point because I think whilst we talk about v8 being pervasive, yeah, it is going to transition in. I think it'll be faster than. There are two things. There's the smartphone mobile space, and then there's the overall blended. We get-
Sure. The question is more on the smartphone mobile.
Yeah, I think in the smartphone, it will obviously depend very closely on the sort of v8 RAM. I think it's going to actually be quite noticeable within the mobile space. When you look at the overall Arm royalties, it gets a little bit more submerged under the volumes of microcontrollers and IoT type stuff. I actually think it will become quite noticeable in the mobile space as we move to v8.
You're saying that whatever the trajectory of growth we've seen in smartphones over the last, say, two or three years in the ARM11 to Cortex-A transition, you'll probably see a slightly more accelerated rate of growth going forward?
I think what you're going to see is where ARM11 is used is in kind of feature phones right now. Feature phones are being replaced by very low-cost smartphones based on single-core Cortex-A5, Cortex-A9. It isn't moving to v8 yet. I think there's a generation maybe or two, or maybe more actually, of very low-cost devices based on ARMv7 architecture, where there's a less marked delta in the royalty rate. Maybe they move to multi-core over a couple of generations before going version 8. That is an ultra cost-sensitive market. To make a whole smartphone unsubsidized for GBP 30, you've really got to work at that. It will be a while before all the latest, greatest technology does ripple down. Probably going to happen over time, I think the first transition in that feature phone space is from ARM11-based feature phones to Cortex-A5, Cortex-A9-based single-core smartphones.
Maybe keep working back through that way, next.
Good morning. Sandeep Deshpande from JPMorgan. Just coming back to this royalty cycles that you've shown in the graph. Historically, your royalty cycles tended to go along with what happened in the industry, whereas the semiconductor industry, if you have noticed from the first quarters, had a very strong performance in the semiconductor industry. Why has Arm performance been slightly deviating from the semiconductor industry? Following on from the earlier question, overall, if you do 10% royalty growth this year, I mean, you're not guiding to that number, but 10%, 12%, you're going to be in single-digit outperformance versus the industry, which hasn't been the case for the last five, seven, eight years as such. Why is this year so different from past years?
I think when you look at the semiconductor industry as a whole, all of it, through this year, with the retirement of Windows XP, that has driven, more recently, a bit of a resurgence in PCs, perversely, which obviously we are much less exposed to. Certainly when I talk to CEOs of semiconductor companies, no one's exactly high-fiving about the growth rate of the semiconductor industry for them over the last year or, in fact, multiple years. It is slow growth, and share is being moved around between companies. There's a lot of consolidation going on. There's definitely been some bright spots there. But for the markets we're exposed to, as I said earlier, we haven't been losing share. Therefore, this is about end unit shipped which just ripples through for us. Now, we're not in control of that.
The only thing we can govern is the rate at which we're licensing new technology and the adoption of it into the market and the build of the ecosystem to make that easier over time, which is where our focus has been.
You're not saying this is because of this trend of the slowdown in the high-end, mid-end handset market versus the low end. Since you can actually see where the royalty growth is coming for or lack of royalty growth within that mix? Can you say that because the low end has been growing very fast and you're very highly exposed in that segment of the market, whereas some of the high-end players have been much slower. Are you seeing that trend impacting this?
As you say, we're in all of those devices, so we move with that market. There's no magic that disconnects us from those end market trends to our royalties. That's highly correlated. We're in such a high market share position. Brett.
Thanks. Brett Simpson, Arete Research. Just a quick question for Simon. There's been a big debate in the industry over the last year or so about Moore's Law no longer giving the economics it once did. I guess if I look at 20 nanometer, which is coming up for ramp up fairly soon, the wafer price increases we're hearing is pretty dramatic. I just wanted to get a sense because the die sizes we're seeing at 20 nanometer are pretty similar to 28nm high-K, which would mean there's going to be cost pressures for chip makers. What does this really mean for Arm's business and royalties, particularly for mobile chip prices? Do you think we're going to go through a period of inflation for high-end mobile chips as we go into 20 nanometer and then other nodes beyond that?
There's some fundamental economics here, right? 20 nanometer is a more complex process. There are more process steps. The scaling that 20 nanometer has achieved isn't what the historic trends have been, the price per transistor hasn't been on the same curve that we've enjoyed for decades. There is no getting away from that, somewhere along the supply chain, someone's got to pay for it. The equipment that the fabs put in, more and more expensive, the R&D cost of the chip, more and more expensive. It's got to be funded somehow. I think this isn't new news. People have been talking about this for a number of years with 20 nanometer. I think what we're going to see is the number of people that adopt 20 nanometer is relatively low compared to most process geometry transition points.
You'll see more people stay on 28, innovate more around design. You'll see the foundries create variants of 28 nanometer to make sure that improvements can be delivered, we've already seen some of that. We've seen for a number of years the big investment that's been going on into next generation transistor technology to provide that big jump up in performance and area scaling and voltage scaling to really give you the benefit for the cost. It's all about, fundamentally, transistors are about what benefit can I get for how much do I have to pay. If that's in the right direction, you can sell your product for more. I think we're in a transition period. I think the industry generally is looking to get to FinFET's next generation technology as quickly as possible, you'll see fewer people on 20.
The inference for your business is obviously the chip ASPs, the mobile prices, the chip prices, particularly the high end, because they've been fairly stable over the last couple of years. Do you think that's going to change as part of this transition?
Well, if the cost is absorbed in the chip price through, then on those very high-end devices, you should expect them to get more expensive. At the same time, with 28 nanometer becoming more mature, more volume may be run on that than, again, in an ordinary mode. You might see downward cost pressure on that through the regular kind of economies of scale that the semiconductor industry delivers. Exactly how the net of that plays out is kind of TBD.
If I can just have a follow on for Tim. Tim, if I go back six months ago, your Cortex-A royalty units were growing triple digits year-on-year. We've seen quite a big decline just in the growth the last couple of quarters. I'm just checking, is there any one-off events here, inventory events that you think might have caused such a sharp decline in the growth? Or in your view, when you look at the landscape, was inventory a factor at all in such a marked decline in the Cortex unit growth?
Well, I think inventory is a factor. I think also probably integration is a factor. There's no kind of one-off strange thing that we're not sharing with you. As Simon said, to some extent, with the market share we have, we're just very closely correlated to what's happening in the market overall. Clearly, why are we growing at 2% not 20%? I think a lot of that is around the inventory and what we've seen in that top end.
Tim, another one.
Hi, thanks. This is Sumant from Redburn. Just two quick questions. One is could you talk a little bit about your value of your server and enterprise royalties? How big are they as a mix of your overall royalty revenue at this point? How much could that grow to in the next two years? Do you expect it to reach to about 25%-30% of overall royalty mix in this near term basis or not?
In a one to two-year period?
Yeah.
No. In terms of volume, the volumes of those chips are dwarfed by microcontroller. They're dwarfed by smartphone chips. Now they're large, valuable chips, multi-core Version 8 architecture, so they're valuable on an individual basis. I think over the next couple of years, as volumes start to grow, it's not going to create an enormous additional layer of royalty for us. Out in time, our expectation is we grow 10%-15% market share here. It should, at that point, it should start to be more meaningful. I think in the next couple of years.
Measure the success on the design wins and the overall ecosystem development, as opposed to really trying to spot that in the numbers.
Could you give us a bit of color on what the value of that is for you at this point in time as a percentage of your royalty?
Last year, we said that enterprise networking was 5% in unit terms. I think on average, the yield from that is higher than the blended. I think we have also said that if you look at Q4, the annualized is probably closer to 10% rather than 5%. You saw, again, strong growth in Q1 in unit terms. That gives you a feel for where we're heading. We're not going to give you values and units by sector.
Okay. One quick question on the cash capital structure. This is a very useful slide looking at the net cash share count sector. I'm just wondering, obviously, your cash will probably grow from current levels as your earnings grow. If you maintain your share count, is the only way you return cash to shareholders is via dividend going forward in a material way, or do we expect a lot more M&A happening in the medium term?
Two headline comments there. One is, we've stated repeatedly that over time, we see the dividend payout ratio increasing. Okay. That continues to be the plan of record. We're non-committal on timing and quantum, but we do see the payout ratio we think is consistent with this business model, with the operating leverage, the expansion. Secondly, we confirmed in February that we intend to maintain a flat share count. That doesn't mean that's the limit of our buyback, in a sense, ambition or option forever. If we think it's the right thing to do to manage the cash via buyback out in the future, we will do like we did between 2005 and 2008, when we bought back 16% of the shares.
We're not saying that's the maximum we'll ever do, but we did confirm in February this year that we have a specific plan to maintain a flat share count using the buyback. That doesn't mean to say we're not going to do more buyback in due course.
In other words, you'll be returning the cash. Even if you maintain a flat share count, you will have an excess amount of cash which will keep developing over the next three to four years. Do you keep that on your balance sheet or do you return it, or do you buy some stuff with that?
I think you're going to see a combination of an increasing payout ratio, at least a buyback to maintain a flat share count. You're going to see a lot of investment in the business. You're probably going to see some level of continued M&A, as you have done. In terms of precisely what level of cash we're going to run with over the next two, three, four, five years, we're not being very specific on that. I've said to investors multiple times, we don't plan to build a cash pile for the sake of a cash pile. We can debate what a cash pile is in Arm terms. The conversation there says we don't see a need to have materially more cash on our balance sheet than we currently have.
Thanks.
Good morning, gentlemen. Amit Ajmani from Citigroup. My first question pertains to your backlog evolution. I think at the end of Q1, we said backlog would be flat till the end of the year. Now we have talked about rising backlog. I would just maybe like to understand your thoughts on where do you see backlog ending up at the end of the year. More importantly, as you look longer term, obviously backlog has outgrown licensing significantly, and at some point of time, you would expect the trend to stabilize, maybe even reverse. I'm just trying to look out as we look towards the coming years with licensing growth maybe tapering off, would we expect more of a stabilization in backlog, or would you expect backlog to sort of continue, maybe gradually coming down over the next couple of years? What's your thought on backlog evolution?
Well, I think in time, backlog is a more lumpy concept because of the revenue recognition rules. Okay? It's a more lumpy concept than revenue recognition itself. If you look through all of the accounting, in reality, backlog has to broadly grow at the same rate as license revenue out in time. Because backlog represents typically somewhere between 40% and 60% of our license revenue. There is inevitably a close correlation. You've seen in this period of high revenue growth, that backlog's grown much faster because of the reasons that I went through. It's very hard to call backlog on this quarter, that quarter.
I think what we're saying is, we expect, given the amount of revenue recognition based on product deliverables that have come out of the backlog in the first half, and based on the pipeline that we see of licenses, the mix between revenue generators short term and backlog builders. When we look at that pipeline in the second half, we expect the backlog to be higher at the end of this year than it is now. Will it be precisely where it was at Q1? We'll have to see. Obviously, when we spoke about that in April, we wouldn't particularly have been predicting that the backlog will be 10% down in June. What happened in practice was if you look at the mix of deals that were in flight in June, the revenue-bearing ones kind of signed, and some of the backlog-building ones have moved into the second half.
They haven't gone away. We expect them to happen this year. That's going to be good for backlog. Do I know what backlog's going to be at the end of the year? Not precisely, but it looks positive.
Over the next few years, would you expect backlog to be stable?
Well, I think, as you say, it's going to normalize as license revenue normalizes as well, but there'll be a different timeline.
Maybe as a quick follow-up, if I could. FX has been painful, clearly. Could you maybe remind us if there are any tools available at your disposal to mitigate the effect? I know it's very difficult, something outside, but how should we think about it?
Am I a magician?
You're not a magician. Clearly, you are not, and that's why I said there are factors outside your control. Is there anything in terms of-
We had a non-exec on the board, actually, who just unfortunately retired after nine years. He was a magician. A member of the Magic Circle. I still don't think that he could eradicate FX movements. I think, hedging this and that, you can smooth the impact. The fact is, Arm is a less profitable business when there's a $2 to the pound than when there's a $1.50 to the pound. That's never going to change unless we moved all of our people out of the U.K. and went into dollar currencies. There isn't any magic. I think we kind of have to, in a way, suck it up. As you know, we do a sort of rolling hedging program. We are going to be less profitable.
I'm sitting here at 171 thinking I'm not a currency forecaster, but the future is bright for Arm's currency. I probably thought that at 165 as well, so we'll see.
It swings at roundabouts. We lose one way, we gain the other. Seven years ago, it was two to one. We went down to about 145, and the FX adjustments were in our favor. We don't win prizes for that. We don't lose-
In terms of whether you'd be reconsidering your rolling FX strategy, is there something you're thinking in terms of hedging a bit more longer term, or you just continue with the waiting side?
There isn't really anything structural. Obviously, you can make long-term bets on hedging. What that means is you're going to make a big bet, which you may be very right or you may be very wrong. I don't really think that's the game we're in. That doesn't really help us. I think we just have to accept the fact that we're more profitable if the dollar's stronger. Go dollar, basically.
Thank you.
Can we keep moving along here? Maybe.
Morning. It's Nick James from Numis. Just a question on licensing and the engagement in new products. I guess we used to think about new products from Arm being bigger, faster processors, and I guess it maybe feels like that's kind of different now, and that the opportunity looking forward is for wider pieces of semiconductor IP. If you can just kind of help us understand the market opportunity for licensing looking forward.
Yeah. As you say, the bigger, faster processors kind of get the limelight. When we launched version 8 products, the Cortex-A50 series, that was a big kind of, "Here are our new big up into the right products." In reality, we're developing products across our entire roadmap all the time. It's processors, it's graphics cores, it's physical IP, it's system interconnect components. The acquisition that we've just done is about technology that helps people put these very complex chips together, and we have IP for that as well. The sophistication of those products is going up all the time, to enable people to build very high-performance SoCs in a very area and power-efficient way. As over the last few years, the pervasiveness of the Arm architecture has broadened. It isn't just about the next generation processor for mobile.
We're looking at the next generation processor for enterprise networking, for cloud servers, for mobile, for tablets, for IoT devices, for automotive. There's just a broad range of markets that we're looking at. We've been building the business to capitalize on that opportunity. Over the last couple of years, we've stood here and talked about the increase in headcount. It's to enable us to create those roadmap of products, to service those markets, and further broaden the royalty opportunity, and some of that is now starting to come out.
I guess there's kind of new customers coming in and licensing Arm. Is there a widening opportunity within existing customers for dollars of IP licenses with the new products coming through?
There's certainly a category of customer, and this has always been the case for Arm, who either has multiple business divisions and they use Arm in one place, then they start using Arm in other places. That gives us greater opportunity. Customers who, having been successful in one place, are looking to see where can they leverage the technology that they've built into other markets. Again, that does broaden our opportunity. On the other side of that, you've got consolidation going on, and these things kind of balance out against each other. Broadly, the use of Arm is getting wider. We're targeting more markets. That's why we believe there's great growth ahead.
Thanks.
Thanks. Andrew.
Thank you. Andrew Gardner from Barclays. You spent quite a bit of time on the enterprise slide talking about servers and the latest product launches. I was just wondering if you could help us with a bit more detail around the networking side as well. If we go back to 2012, I think you had maybe one, maybe two partners start to ship. Last year, it built even further, up to, I think 5% market share is what you've had on your slides. Just as we've come through the first half of this year, can you give us any sort of update around perhaps number of partners now shipping or number of subsegments within the enterprise networking space that's shipping, just to get an idea of how that ramp is going relative to the licensing activity we saw a few years ago?
Yeah. Nothing's fundamentally changed since we were here talking in detail about that at the Analyst Day in May, just a couple of months ago. I don't think you were here, actually. We talked a lot.
Caught me.
Sorry?
Caught me. Yes.
Turn up next time. Yeah, we talked a lot about that, and our prospects for growth and our anticipated market share. Over the last quarter, progress kind of expected to plan. As you said, over the last little while, we've done some of the licensing. I mentioned on the slide four out of five of the top chip companies providing this market have licensed version eight of the Arm architecture. We expect to see those market shares that we laid out actually one year ago, come to fruition over time.
Perhaps, Tim, just a quick follow-up on the FX question. Given that you're recognizing, as you say, sort of 60%-ish, in the most recent quarter of licensing out of backlog, can you give us a sense as to where the average USD rate is for the current backlog, just to get an idea of how that phases in over the next few quarters?
I think the trend that you've seen in the last couple of quarters, effectively, the license revenue effective FX rate on translation, better than or stronger than the weighted average of the FX through the quarter, I think will continue for a good while. Obviously, stuff that's going into backlog now is going in at 170. Obviously, most of the stuff that's going to backlog has gone in in the 150s and 160s. You're still going to see a translation rate that's a little bit stronger than market. I think it's Gareth at the back.
A lot of my questions have been answered. Just wondered on OpEx, if you could talk about next year in terms of budgeting. It's probably a bit early. You've historically talked about 50% of your revenue growth into OpEx, and I suspect given that you're bouncing back quite aggressively on royalty revenues, we should expect a lower growth rate than that 50% of growth. Could you help us around overhead and personnel increases for next year? I've got a follow on for Simon, sort of separate.
Well, the 2015 budget. As you say, we are quite early in the 2015 budgeting cycle, as in we haven't really started it in detail. What we are in the middle of actually, is looking at our five-year annual update of our sort of five-year plan. All things being equal, we would expect to continue to be investing, both in the R&D capability of this business and as I describe it, the business infrastructure. I think as we sort of think of our models for next year, I don't see a particular step change. Obviously, we're going through a year of quite strong net increase in headcount, which will obviously flow through into next year. I don't see any fundamental change.
Now, precisely the trajectory of hiring and investment that we make will inevitably be influenced by how we see the market and the opportunity at the time. That's probably all I can say at this point of the cycle.
Yeah. Just to say that, we do see huge opportunity here, and we are building the company for the future. Some of the infrastructure development that Tim mentioned, I don't know if that, Gareth, is what you meant by overhead, or maybe you meant Tim and I. That is a really important part of the company. We have a lot of people working very hard and being able to enable them to get their engineering work done as quickly as possible and deliver on time to our customers with the highest quality is really important. We are looking at, as we grow in a five- and ten-year period, what do we need to do now to make sure we're successful there? That obviously requires some investment on the assumption that we grow, and we believe those growth opportunities are big.
Now, by that, I don't want to scare you into thinking, "Oh, my God, the OpEx is going to go up massively." We will do that in a measured, sensible way, based on how the top line is going to grow.
We're very focused on the sort of effectiveness and efficiency and productivity of the people. If you look at the way Arm's OpEx develops, we tend to go through phases of investment and then phases of more sort of digestion and productivity. We've obviously been through quite a long period now of investment in seizing this growth opportunity. I think we are very focused as we look at this sort of five-year plan period of how we can actually make ourselves more effective and more efficient in everything we do. When you have time to sort of sit back and think about how you do stuff, there's always opportunities to be better and more effective.
That, I think that's going to be a key theme, which in financial terms will offset, if you like, some of the opportunity and the need to invest in seizing the opportunity.
a completely different question, just on IoT. Going forwards, can you talk about some of the steps that the industry's taking to maybe become a bit more harmonized, so things like the Thread Community and so on, and what Arm's place will be within that? Or whether we'll see a more fragmented world where Imagination gets some market share, Intel gets some market share, and Arm gets some market share, hopefully the lion's share.
Yeah. Fragmentation in IoT is clearly a risk. When you look at how devices are going to sense data, share that data, and make their data available to services that might utilize that data, the easiest thing to do is to build a closed system. We think that's a bad outcome. We want to see accelerated adoption of IoT because it drives our volume, and to remove some of these barriers for data to be shared, because we think there's a greatest return on all of this if that happens. We are looking at various activities for how Standardization, appropriate standardization can occur around some of the technologies required for IoT, and Thread is one of them.
It's about how do products from different manufacturers actually work with each other, instead of requiring their own completely closed system, which is obviously very inefficient, and likely to lead to more cost and lower deployment at the end of the day. Thread's one of those. We are talking to lots of people up and down the supply chain about these issues. We're looking at security, we're looking at encryption. We're looking at the standards that are going to revolve around this, and what the role of government is in setting off in the right direction as opposed to the Wild West followed by massive control, which again, I think will lead to a suboptimal solution. There's lots of this. It's very early days for IoT.
Given our desire to see a rapid acceleration of the use of Arm processes here, we are investing to try and lower some of these barriers. Okay. Any other questions? Right.
Just a follow-up. Your mobile customers, licensees, can you perhaps talk about how many of those are double-digit % of sales customers? We're seeing a lot of consolidation in mobile semiconductor. I'm just wondering, how concentrated your business now is with some of those folks getting bigger and bigger.
In terms of units, you can see the data there. You know who's shipping what, and broadly speaking, the dollar isn't a million miles away from that. There is some concentration going on, but there's a lot of competition there as well. I wouldn't take the situation that we're in today as an indication of what's going to happen for all time. You roll the clock back 10 years, the people who were dominant then have changed. Question down the front.
Just a quick follow-up in terms of the kind of licenses. For example, your perpetual licenses, sometimes licenses are for per use, maybe for a term period of time.
We've seen 41 licenses signed right now, despite that, the backlog went down simply because there was recognition from the backlog. I'm just trying to understand whether it's essential not only to look at the number of licenses, but also in terms of the quality, because I guess perpetual is more expensive compared to say, a per use compared to a term use. Just trying to understand, how are you seeing the adoption in terms of these different flavors or variants of licensing? Is it more precise per use licenses? Is it more perpetual? If you could shed some color on that. Thank you.
I don't think there's been any big shift there. Typically, I think for the entire time of Arm, that our newest products, we don't license on a per use basis until they have been matured in the market for a while. Part of that is because we want to engage with people who are going to commit to our latest generation products and really work with them to make it successful. Your measure of quality, I don't think of quality of our licensing. Every one of those license deals has an opportunity to turn into something big. By your measure in terms of timeframe of license, the newer technology is going to be on a much longer term basis than perhaps some of the older technology. It's not to say that the older technology is always licensed on a per use basis either.
There's a real mix of that, and that can change big time from one quarter to the other. Every one of those Typically, I guess, new companies, maybe first company to Arm, or a company that takes a first license to Arm technology might engage on a smaller term just because of uncertainty. You never know who's going to turn into a big company.
Nothing's changed there. It's like normalized mix that we have seen so far.
Yeah.
Thank you.
Okay. All right. Well, if there are no more questions, thank you all very much for coming, and we'll see you out on the road. Thanks.