Good morning, ladies and gentlemen. Thank you for standing by, and welcome to Arm Q3 analyst conference call. At this time, all participants are in a listen-only mode. There'll be a presentation followed by a question and answer session, at which time, if you wish to ask a question, you'll need to press star one on your telephone. I must advise you the conference is being recorded today, Tuesday, the 23rd of October, 2012. I would now like to hand the conference over to your first speaker today, Ian Thornton. Please go ahead, sir.
Thank you, Jenny. Good morning, everybody. This is Ian Thornton, Head of Investor Relations at Arm. On today's Q3 results conference call, we have Warren East, Chief Executive Officer, and Tim Score, the Chief Financial Officer. On today's call, Warren and Tim 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 earnings release can be found on the Arm Investor Relations website at www.arm.com/ir. Before I hand over to the team, 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 investments, and the information communicated on this call is being made available only to investment professionals.
Any persons present on this call who does not have professional experience in matters relating to investments should not act or rely on the contents of this call. The following conference call will contain forward-looking statements which are other than statements of historical fact. The company's actual results for future periods may differ materially from these statements, as they are based on current expectations and are subject to a number of risks and uncertainties. On this note, I'll hand over to Warren.
Thank you, Ian. Good morning, everybody, and thank you for joining our call. I'm going to run through the highlights as usual, then hand over to Tim, who will provide some more detail on the numbers. I hope we'll cover most of the content in Q&A. It's been a quarter with very strong momentum. Royalty revenues are at record levels, volume shipments at record levels, licensing at record levels, and our already record backlog at the half year up by a further 6% by the end. It's been a very good quarter.
We highlighted three months ago how various market leaders are licensing Arm technology and introducing new products into a wide range of end markets, and that has continued during Q3, with several important design wins being announced in enterprise networking, new products launched in mobile computers, Arm-based servers, wireless sensors for the Internet of Things, right across the application spectrum. The fourth quarter has also started very well. We've recently seen Google announce its new Chromebook from Samsung, and later this week, Arm-based mobile computers running Microsoft's latest PC operating system will start shipping. Generally, the demand for high performance, low cost computing is driving our license revenues. As these products are coming to market, our royalty revenues are growing as well. This quarter, our royalty revenue increased by 27% year-on-year, and that's significantly outpacing the semiconductor industry.
This revenue growth has enabled us to continue to invest further in our R&D capability. It's enhancing our ability to create new products. In fact, since the beginning of the year, we've added 252 people, with 115 of those joining us in Q3. At the same time, we've been able to grow earnings by more than 20% and it's been a quarter of record cash generation. We're going into the final quarter of 2012 with a record order backlog. The opportunity pipeline is robust, and that points to another strong quarter for licensing revenue in Q4. Our Q4 royalty revenue is generated from Q3 shipments, as you will recall. The data that we have from our customers so far suggests that there will be a moderate sequential increase in Arm royalty in Q4.
Overall, we'd expect group dollar revenue for the fourth quarter to be in line with current expectations. Now I'll discuss revenue drivers in a little bit more detail. Starting with the processor division and license quarter. For a very broad range of end applications from embedded microcontrollers for Internet of Things, right through to image processes in digital cameras and mobile phones. About a third of the licenses were signed with companies taking their first Arm license. Many of these new licensees are located in Asia, and we continue to see strong demand from this region. These new companies are licensed across the full spectrum of Arm products from Cortex-M, for use in embedded applications, through to Mali graphics and Cortex-A in mobile and mobile computing. At the high end, we signed two Armv8 licenses for enterprise and networking applications.
This is including one Armv8 architecture license and one Armv8 processor license. We've seen increasing demand for the use of our technology in high-end networking applications, and these licenses are continuing to build on the momentum that we've seen in that application space generally over the last few quarters. In addition to the Armv8 licenses, we sold six Cortex-A processor licenses, mostly into consumer and to server applications. That included another license for big.LITTLE technology for use in mobile computing. That brings the total number of partners now enabled with big.LITTLE to 14. big.LITTLE, as a reminder, incorporates our most performance-intensive processor, the Cortex-A15, coupled with our most power-efficient processor, the Cortex-A7. We've seen very exciting feedback from the partners in relation to this, and we expect to see the first big.LITTLE implementation shipping next year.
In fact, we have our own silicon back now, early test silicon is confirming our initial expectations on big.LITTLE, delivering on the promises that we talked about a year ago when we launched this technology. The early silicon is suggesting that big.LITTLE implementations can produce a 2x improvement in performance at the same time as halving the energy required. That's compared with a typical smartphone apps processor that's shipping right now. Enough of big.LITTLE. At the other end of the spectrum, we also signed licenses for our Cortex-M processors, 13 licenses for Cortex-M, mainly for microcontrollers and storage applications. Two of these licenses were for Cortex-M0+, which is our smallest, lowest power processor. That's the one with the potential to control something like an embedded sensor for up to 10 years on a watch battery.
This new technology enables our customers to have Arm software compatibility right across their entire product range, including chips that might previously have been 8-bit. In fact, Freescale this quarter announced a range of Cortex-M0+ MCUs designed to help migrate their consumer and industrial application base that are currently using their legacy 8-bit and 16-bit microcontrollers to Arm. Finally, five of the processor licenses were for our Mali graphics processor for applications ranging from digital TVs, car entertainment, imaging, three of those licenses were for companies taking their first ever Mali license. Arm's Mali is now the most widely licensed graphics IP. We've sold 69 licenses to over 50 companies. Switching to royalty. Our royalty revenues are reported one quarter in arrears, Q3 royalty was generated from Q2 device shipments.
Processor royalty revenue was up 27% year-on-year compared with the relevant industry revenues increasing by 4%. Of that, 10% was due to market share, 8% from Cortex, 5% from Mali. All told, there were 2.2 billion Arm processor chips reported during the quarter. The mobile non-mobile split was 50/50. It's a 16% year-on-year increase in volume, and that was mainly driven by growth in non-mobile markets. Arm is continuing to gain share in markets like digital TVs and set-top boxes, as well as microcontrollers and smart cards. Digital TVs and set-top boxes actually were particularly strong this quarter, more than doubling year-on-year compared with an industry that was down 5%-10% over that period. The growth in functionality of consumer products like smartphones, tablets, and so on, is also having a positive impact on our royalty.
During the quarter, we saw a doubling of Cortex-A class shipments, a nearly fivefold increase in the number of Mali processor shipments compared with a year ago. Typically, as you know, Arm receives a higher royalty percentage for Cortex-A and an additional royalty percentage for chips containing Mali. The growth in Cortex-A and Mali has helped increase the average royalty per chip from $0.044 a year ago to $0.049 now. It's worth noting, when we're looking at the quarter's royalties, that on a year-on-year basis, this quarter's compared with a relatively low base because the royalties in Q3 were adversely impacted by a number of effects last year, also including the Japanese tsunami. When you look to Q4 2011, royalties were unseasonably high.
That's pointing to a relatively lower year-on-year growth rate in Q4 this year than that which we've just seen in Q3. Moving on to physical IP now. Total physical IP revenue was up 9% at GBP 28 million. We signed 3 new royalty-bearing platform licenses, including 2 new platforms at 20 nanometers and 14 nanometers with GlobalFoundries. We'll be working with GlobalFoundries to deliver our optimized Arm-based system-on-chip designs, including their 14 nanometer FinFET process. We continue to see demand for the processor optimization packages comprising our physical IP optimized for use with Cortex-A. We signed 4 more PoPs in the quarter, bringing the total to 36. Just after the quarter closed, we announced that we're making available a processor optimization pack for the full range of Mali graphics processors as well.
Underlying physical IP royalties in Q3 were $13.9 million, that's up 28% year-on-year. One of the drivers behind that is the royalty revenue from our physical IP at advanced nodes, 45 nanometers and newer. That's continuing to increase and now accounts for a third of the physical IP royalty revenues. In other parts of the business, sales of development systems were $12.1 million. That was a decrease of 3% year-on-year. As you know, we're transitioning that business to focus on microcontroller tools and premium toolkits for multi-core systems. At the same time as generally in the market, there is a rise in the availability of open source tools. The transition for that business remains on track, and we expect our development systems business to achieve our stated target of being broadly flat year-on-year. Just a few operational points. It was a busy quarter for marketing.
We had record numbers of people attending our partner meeting in Cambridge. We share each year with our partners our roadmap plans, we had 500 representatives from our partners and key ecosystem companies. Next week, in Santa Clara, we'll be holding our annual TechCon, we're expecting 3,000 visitors for 3 days, we'll be presenting new technologies, have exhibits, demonstrations from our engineering teams and from companies across the ecosystem. Meanwhile, back at Arm, we've continued our investment in R&D. We've grown the engineering teams. We've hired 250 people so far this year, we expect to continue that investment through Q4. With that, I'll hand over to Tim for some further detail on the numbers.
Thanks, Warren. Good morning, everyone. As you've heard, Warren's covered quite a lot of the financial headlines, I'll keep my comments relatively brief. Just a reminder of the headlines in the third quarter. Dollar revenue is up 18% year-on-year. Normalized PBT up 22%, earnings up 22%. That revenue growth of 18%, strong in processor licensing up 17% year-on-year, as Warren said, processor royalty up 27% year-on-year. Going into the third quarter, we reported that July that the backlog was at record levels, approximately 60% of PD license revenues in the third quarter were generated from backlog. We still exited the quarter with backlog up 6% sequentially.
The usual analysis of backlog maturity and composition is included in the quarterly slide set on the website, that shows that approximately 30% of total backlog is expected to be recognizable as revenue over the next two quarters. As we stated in the earnings release, prospects for order backlog in the fourth quarter look promising, and this combined with a healthy licensing opportunity pipeline points to another strong quarter for license revenues in Q4. Processor royalties again outperformed the industry, up 27% year-on-year. As Warren mentioned, the percentage growth year-on-year has been flatter to some extent as royalties in Q3 2011 were impacted by the Japanese tsunami. A very strong performance nevertheless. In previous years, normal seasonality has resulted in about a $10 million sequential increase in royalty revenues for Arm from Q3 to Q4.
However, the macroeconomic environment remains uncertain, and industry data for the third quarter indicates that the seasonal uptick typically seen in Q3 was lower than normal, hence our guidance on royalties for the fourth quarter. Normalized OpEx in Q3 headline was £72.3 million compared to the guidance given in July of range £68 million-£70 million. OpEx in Q3 included a net charge of approximately £2 million due to the impact of a weaker dollar on the accounting for derivative instruments, and this compares to a £2 million credit in the second quarter, which you may recall from the half year release. Stripping out these effects from both Q2 and Q3 indicates that underlying costs have moved from £68 million in Q2 to £70 million in Q3. As noted in the earnings release, the increase in OpEx is due to increased investment in our research and development teams.
Normalized OpEx for Q4 is expected to be in the range of £71 million-£73 million, assuming current exchange rates. Before moving to outlook, a couple of comments on cash flow and balance sheet. At £88 million, net cash generation in Q3 was at a record level, reflecting the increasing profitability of the business, notwithstanding the ongoing investment which Warren referred to in our research and development teams and in the organization infrastructure that supports the company's growth. You will note that there is an other debtor on the balance sheet of just over £100 million at the end of September. We regularly evaluate strategic opportunities, and in relation to one of these opportunities, we had conditionally committed this amount at the end of the third quarter. This represents the entire investment being contemplated, and in the event the transaction is not completed, the amount will be returned to Arm.
This investment has no impact on our stated progressive dividend policy, which means that over time, as we have said before, we expect to grow dividends faster than earnings. Now closing on outlook. Arm entered the final quarter of 2012 with record order backlog and a robust opportunity pipeline for licensing, pointing to another strong quarter for licensing in Q4. Q4 royalty revenue generated from third-quarter shipments, as you know, and data from our customers and from the industry at large suggests a moderate sequential increase in Arm's royalty revenue in the fourth quarter. We therefore expect group dollar revenues for the fourth quarter to be in line with current market expectations. With that, throw it open to questions.
Thank you very much. As a reminder, please press star one to ask a question. Your first question from François Meunier from Morgan Stanley. Please ask your question.
Yeah. It's actually François Meunier from Morgan Stanley. On the cost, Tim, I understand that the budget is going a bit higher, I think it would be good to reassure us that this will remain under control into next year. Also, in Q4, the GBP 72 million you're guiding to, is there any bonus payment or anything? I'm just trying to see what's going to be the new run rate for next year. That's going to be my first question. The second question obviously is the ASP, which is very strong this quarter, a very big increase, and thank you for providing the bridge on page 13 of the presentation. What's going to be the blended royalty rate for this year? Last year it was 1.2. Is it going to be around 1.3, 1.4 already this year?
On OpEx, François, as I said, the underlying cost in Q3 were at GBP 70 million, just at the top end of the range we guided. GBP 72 million for Q4, it would be broadly in line with consensus estimates for this year. There is no change to our medium-term view of the expenses that we need to incur to get the revenue growth. No change in the shape and costs absolutely are under control. As Warren said, we are investing. We have a lot of opportunity to take Arm technology into a very broad range of new markets, and we're obviously invested in the R&D capability to execute on that opportunity. There's no change to our medium-term guidance in terms of operating leverage that we expect to generate through this business.
François, with regard to your second question about ASPs, well, we haven't got a number for you today on what that's going to look like for the full year. We'll wait until we see the full year on that. I think if you look at the trend in Cortex-A shipments, you can see that that's now up to 9% of the total. You'll see Mali volumes increasing significantly. As you know, both of those have a positive impact on both the royalty rate, and they're tending to go into chips of higher value. Therefore, you have a higher percentage multiplied by a higher value, and that is driving this average take per chip up. At the level that it's at today, we expect that trend to continue at the same time as we expect the trend in growth of microcontrollers to continue as well.
We'll wait and see what the answer is when we get to the end of the year.
Okay. Thank you very much, and well done again for a very good quarter.
Thank you.
Thank you, sir. From Merrill Lynch, you have a question from Didier Scemama. Please ask your question.
Yeah. Good morning, gentlemen. Congratulations also on my end on the ASP licensing. Very well done. Great execution there. Two quick questions, if I may. First, on licensing, I think there was a lot of confusion in the last three or six months on that particular business where people, I think, expected that to slow down dramatically. Can you explain a bit the underlying dynamics in licensing and, in particular, whether you've seen any changes in the underlying ASPs you charge or perhaps also the contribution from subscription licenses? I've got a quick follow-up. Thank you.
I'm not exactly sure, Didier, where that confusion would have come from. Obviously, my sort of guidance on quarterly license revenue has been gradually increasing over the last few quarters. The backlog has continued to go up. What is driving that really is the broadening applicability of Arm technology to the full computing spectrum. That is encouraging existing licensees to license Arm more widely, and it's bringing a lot of new licensees to Arm for the first time. I think the trend that we've seen in licensing has been fairly clear. As I say, I'm not quite sure where the confusion would have come from.
No, I think what I meant is skepticism.
Skepticism. Well, that's a different word.
Sorry, my English is poor.
Oh. You're very harsh on yourself.
Always. Second follow-up is basically on the ASPs. If I've done the math correctly, you've disclosed for the first time that the Cortex-A class of processor was 35% of royalty revenues. If I've done the math correctly, it looks like the underlying ASPs on Cortex-A chips are GBP 0.189. Versus GBP 0.035 on non-Cortex-A type chips. Would you assume that the non-Cortex-A type chips, effectively ASPs would go up in the longer run as you start to see more Mali contribution as well as maybe also higher royalty rates in the non-Cortex-A type of chips in the medium to long term?
It's possible, but these things really do depend on where you see the proportion of straight Cortex-M microcontrollers compared with the Cortex-A. You're right with the arithmetic on Cortex-A, that's going to continue. Mali tends to go into the same sort of chip that you would find a Cortex-A. We're not really seeing Mali sitting alongside something like a Cortex-M processor. I would expect the dynamic that's been established, where you have higher priced chips with higher value Arm microprocessor cores, that's going to continue as more phones get to be smarter phones, as mobile computing turns into reality, as we get into servers. We're also seeing these types of technologies deployed in the networking infrastructure product as well.
At the other end of the spectrum, we do see a huge volume opportunity for microcontrollers, and typically, these chips will contain a single Arm microprocessor core. It will be a lower priced chip, and it will be a core that is commanding a royalty rate that is at the lower end of the spectrum.
Got it. Then the final one on the Renesas announcement, I think this morning on their microprocessor product line based on Arm. That has been historically sort of proprietary CPU cores for that pool of products from Renesas. How significant is it for the company?
Well, obviously, we've been working with the companies that currently comprise Renesas, Hitachi, NEC, and Mitsubishi, for many years. They've all had Arm in the stable. We do think this morning's announcement is quite a milestone, because, for microcontrollers per se, it has been a bit of a no-go area for Arm. It's a great milestone, and I would liken it to a couple of years ago when Freescale came out with their first Arm microcontroller announcement, and we've seen that sort of become more pervasive across the Freescale microcontroller line. This quarter, as I mentioned a few moments ago in the narrative on the call here, even more announcements from Freescale. It's the start of a journey.
Thanks so much. Congrats.
Thanks.
Thank you. From Citigroup, you have a question from Amit Harchandani. Please ask your question, sir.
Good morning. Amit Harchandani from Citigroup, thanks for taking my question. The first one is following up on the previous question on processor royalties. You've talked about outperformance of 10%-15% relative to the industry. At the same time, you have this dynamic of increasing proportion of Cortex royalties, which in turn have a higher royalty rate. Basically a multiplier effect. Would it be fair to say that over the next, say, three to four quarters till this plays out, that the outperformance in terms of processor royalties would be closer to the higher end of the 10%-15% range or maybe even higher? Then I have a follow-up.
What we've just laid out there is totally logical. There is a double effect with the Cortex-A and the Mali processors. The only modification I'd make to what you've said is that this is an effect which is going to play out over several years rather than several quarters.
In that case, would it be fair to say then we are looking at a range of, say, 15% or even higher going forward in terms of outperformance?
Well, obviously, what happens on a quarter-by-quarter basis depends on the sort of particular dynamics in that quarter. As a trend for several years, I think we'd be comfortable that that positive effect from the high-end chips with the high-end cores is going to continue to be a major driver of Arm's overall outperformance.
Okay. As a follow-up, if I may, could you provide us with some insight into the licensing of V8 by various end markets other than networking? In particular, approximately what proportion of these would be customers that one would normally associate with the mobile computing or the smartphone space? Also maybe what proportion of your backlog currently consists of architecture license revenue? Thank you.
Right. Well, let me start with the V8 first while Tim's having a think about the backlog one. It's very early days for V8 at the moment. We've seen, as we mentioned, V8 providing a hook for things like networking and things like servers. A key differentiator about V8 is 64-bit, and a lot of networking applications find 64-bit very useful. A lot of server software assumes that there is a 64-bit processor there. These applications really hitherto, there have been significant barriers to the Arm architecture, and V8 overcomes those barriers with 64-bit. That's why you're seeing some presence of V8 licensing there. In due course, already, we are seeing interest, and that interest will turn into reality. We'll see the V8 products finding their way into computing and high-end mobile applications as well. I think that's an inevitability.
It's something which is going to happen over the next several years rather than any time very soon.
We don't break out by type of license, the composition of the backlog. Just a reminder, the backlog is a combination of longer-term licensing arrangements, like subscription deals which would include architecture similarly. The majority relates to new technology or technology to be finalized that has been licensed, where we're not yet recognizing the revenue because there are still engineering milestones. Architecture is a component, but it's a relatively small component of the overall backlog.
All right. Thank you, and congrats for a good quarter.
Thank you.
Thank you, sir. From Goldman Sachs, you have a question from Simon Schafer. Please ask your question, sir.
Yes. Thanks very much. Actually wanted to ask a question on royalty units in the wireless segment. Just because of this disaggregation specifically on the connectivity side, looks like your units actually year to date in wireless haven't really grown. Obviously significantly less than your company average and also much less than smartphone growth. When are we completing this sort of disaggregation cycle of components that has been depressing that unit growth number in wireless? I'm just trying to get a sense as to when you would expect some sort of re-acceleration for unit potential on the mobile phone side.
Well, it's true, Simon. We've seen only a 6% growth in volume, 22% in value. The thing about mobile is we've got two trends. We've got increasing proportion of phones being smarter phones and containing more functionality and typically more chips. Against that, we have the integration trend of, in particular on the connectivity chips, things like Wi-Fi and Bluetooth coming together. That's why there isn't a simple relationship between the volumes of phones actually shipping and the volumes of Arm chips within phones. The two just are a bit disconnected because you've got these two trends happening, and they're happening at different rates with different drivers, and it depends on different suppliers and market share changes, and it's very difficult for us to help you model that other than highlight the underlying trends.
When you talk to individual suppliers, maybe you can refine the modeling. I don't think we're really able to help much more than that.
Got it. Okay. Thanks, Warren. Thanks for that. My second question on the cost side, as you said, another 250 or so employees since the beginning of the year added. As you look into next year, what's the sort of preliminary budget plan? Specifically, I think you almost have, well, still have almost twice as many employees in the U.K. as compared to what you have in the U.S. Over time, are you going to continue to leverage your U.S. headcount and therefore have more of a natural hedge? I guess I'm trying to understand just the avenue of cost growth as we go into next year. Thank you.
Well, we are depending obviously on the business climate and our views of revenue prospects for the next year and the years going forward. We will continue to increase the resource, particularly as we get into the new market areas that I talked about. In particular, the enterprise space does require us to do different things. The functionality in mobile consumer and mobile computing, we also need to do different things today than we did five years ago, and that does need more resource. We are continuing to hire, as we said, through to the end of this year, and we have plans that are sort of fairly similar in absolute terms, but less in proportionate terms, probably for next year, again, subject to how the revenue plays out.
I think as Tim mentioned, in addition to that, from a cost point of view, when you're growing people, which in our case is mainly engineers, then they need tools and infrastructure to work with. It isn't just about headcount. In terms of where we put the people, we are hiring and growing our resources in the U.S., but we're going to continue growing them in the U.K. as well. Although you talk about natural hedges with the currency, there is an element of that. It's really about putting engineers where the talent is, where the customers are, and bearing in mind overall costs as well. It costs us a lot less to employ somebody in the U.K. than it does to employ them in Austin, Texas. At the same time, Austin, Texas is one of the global centers of expertise for microprocessor design.
Arm has a strong presence there, and so does several of our very major customers. I think you're going to see the pattern that we've established continue, and we don't really see a rationale for significant changes.
Great. Thanks, Warren.
Thank you.
Thank you. Your next question comes from Justus McAvilly from Credit Suisse. Please go ahead.
Hi. Good morning. My question is around Mali. I see you've had several chip makers taking both Mali and Imagination PowerVR licenses, and I'm just wondering if you can provide some color on maybe the two key factors that impact their selection process, especially in terms of smartphones and tablets.
Right. Well, obviously you need to talk to Imagination to get the full color on this as well. We believe that our technology offers the best combination of power and power efficiency together with performance. This is a sort of selling feature, if you like, of the Arm Mali graphics processors. Against that, of course, when one is considering the design of a new chip, you look at what you had before in the chip. The cost involved in switching between different flavors of graphics processor is an input to that decision. Different customers will trade off the cost of change versus the technical benefits of better power consumption, smaller area, and those sorts of things. In the world of graphics processors, a lot of the software is written for open standards. OpenGL ES 2 type is one example of the standard API that programmers use.
It is relatively, compared with, say, a microprocessor, it is relatively easy to deploy both types of graphics processor. I think what we're seeing is something fairly normal. It's rather like several years ago, if you go back 5 to 10 years, some of our semiconductor partners had Arm microprocessors in some products and MIPS microprocessors in other products. Gradually over the 5 to 10 years, we've seen our market share increase versus MIPS. Now it's shifted altogether in many of the application spaces. Maybe we'll see something similar happen in graphics, don't know.
Okay, great. Thank you.
Thank you. Your next question comes from Nick James from Numis. Please go ahead.
Morning. Just a couple of questions. One was on the Q4 guidance on royalties. I think last year, the sequential growth in Q4 was about 19%, and I think at the time it didn't really feel that was necessarily a snapback from the Japan tsunami, but it was more about build for iPhone 4S. I'm just trying to understand how the dynamic is different this year from last year, given the macro-wise are both quite similar. The second question, I think you said on the big.LITTLE that you're expecting chip shipments next year, which I guess suggests that the time span from license to chip shipments is narrowing somewhat. Is that a trend that we're seeing across the business?
Nick. Tim, on the royalty issue, I'll try not to overcomplicate this. If you look at Arm's royalties in Q3, Q4, and Q1 last year, I think I'm right in saying they went from sort of GBP 84 to GBP 100 to GBP 93. I think we commented when we reported GBP 100 in Q4 last year, that there was probably some element of the weighting of shipments between Q3 and Q4, i.e., our Q4 and Q1 being more weighted to Q3. Because if you look back at Arm over the last 5, 6 years, it's actually very unusual for Q1 royalties to be lower than Q4. Okay. I think in two senses, Q4's result last year was quite unusual.
One, because of the snapback, as you say, from the tsunami, and the second thing is there was probably a little bit of draw forward from what turned out to be our Q1 royalties. I think that's one fact. The other fact is purely if you look at the messaging coming out of the big semiconductor companies for their third quarter, it is in many cases lower than normal seasonality. I think we had Marvell overnight -6%, I think Qualcomm down sequentially. You've got TI obviously down in their wireless business. I think there are a number of companies who their third quarter was flat to down and not up. Similarly, I'm well aware of the fact there are a number of companies who are strongly up, and so it's a mixed picture. That is the context in which we are guiding our Q4 royalties.
Okay. With regards to your second question about big.LITTLE and product launches and so on. Yes, we expect to see our first big.LITTLE shipments next year. Bear in mind, this is very much the vanguard of big.LITTLE. As is normal, we work with lead partners ahead of completing the designs, and we've done that with the big.LITTLE processors as well. It's always been the case that one of the benefits of being a lead partner with Arm is that you get a time to market advantage. Arguably, some of the semiconductor partners have been getting very competitive and really pushing hard to narrow that time difference or the time gap between when we first have a design and ship it to them and when they're able to get chips out the door.
It might be a sort of minor effect of that, I think it's a combination of that and the fact that we're talking about first silicon, and that's a lead partner thing. I don't think that it's the start of a serious trend to narrow the typical time delay between licensing and appreciable volumes of products.
Great. Thank you very much.
Thank you. From Deutsche Bank. Your next question comes from Kai Korschelt. Please ask your question.
Thanks for taking my question. I had a couple. The first one was just on operating margins. Looks like they were flat year-over-year despite meaningful revenue growth. I'm just wondering if your assumption is right and there will be a relative slowdown in revenue growth. Could we see a contraction in margins on a year-over-year basis on earnings, in fact? My first question, and the other one is, sorry to follow up on OpEx, but maybe to make it a bit easier for you in terms of the historic context. I think you were previously indicating that OpEx growth typically tracks around half of the overall revenue growth. Is that rule of thumb still applicable, maybe as a little bit more tangible sign for next year? Then maybe just one brief question.
It looks like the die sizes from an average perspective at Apple, looks like they're getting smaller with the shrink to 32 nanometer on a like-for-like basis. I'm just wondering, does that impact you in any regard? Thank you.
Thanks, Kai. You're right. At a headline level, operating margin in Q3 2012, 44.6%, which compares to 44.6% in Q3 2011. I think the key thing to take into account there is that, as I mentioned earlier, there's a GBP 2 million mark-to-market charge in the third quarter this year. Actually, if you look back, you'll find that there's a GBP 2 million credit in the mark-to-market in Q3 2011, as well as, by the way, a GBP 2 million credit in Q2 this year, which I was just referring to in the sequential movement. Basically, again, to get your underlying costs, you need to make there effectively a GBP 4 million swing on FX. Actually, the underlying OpEx operating margin is quite well up this year versus last. In terms of the overall trajectory, as I said earlier, operating leverage is still expected to come through in this business.
Our operating margin has increased broadly from 30% to 45% over the last two or three years. We have multiple conversations with investors about what the trajectory is going to be out into the future. I think when you look out five years, and I think we covered this again at the Analyst Day, this margin is going to be higher. How much higher? Is it going to be 50? Is it going to be 55? Is it going to be 60? Will depend on as much as what we're investing in the out years for opportunity to develop further technology for licensing royalty, as well as the penetration into the markets that we currently see. Very difficult to give a specific trajectory, but certainly expect ongoing operating leverage over multiple years.
With regard to your second question, Kai, about the smaller die sizes. I think that's quite a good trend because what it really reflects is use of more leading-edge manufacturing technology. Typically, wafer prices are higher. They're significantly higher there, and that's quite a positive for our physical IP royalty. As I mentioned in the commentary, we've seen good growth in physical IP royalty, and one of the drivers there is the use of newer leading-edge technologies. Secondly, these leading-edge technologies are indicative of people pushing the technology further right across the board, and that means using more sophisticated Arm processors. As we've mentioned, the more sophisticated Arm processor, the higher the functionality, the higher the royalty rate. I think underlying, it's a symptom of a very positive trend.
Okay. Thank you.
Thank you very much. Now from JPM. You have a question from Sandeep Deshpande. Please ask your question.
Yeah. Hi. Thank you. I have a couple of questions. Firstly, could I talk about PIPD? If you look at PIPD, you've seen some recent growth. You've talked about processor POPs driving growth. When you look at the PIPD revenue, it has essentially grown just about 10% since you acquired the business. When do we see this acceleration in PIPD growth associated with the processor POPs, which will actually make this a growth business that we can value as part of the Arm story? Secondly, can we talk about the processor royalties that you're getting for the non-Cortex-A series processors? In Cortex-A, you have talked about raising the royalty rate from the historic 1% towards the 2% level. How is that move happening in the Cortex-M, for instance, or the Cortex-R series?
Finally, when we look at the percentage of royalties that you have got from Cortex-A, clearly very strong growth over the last year. At the same time, when you look at the non-Cortex-A royalties, they have actually declined from fourth quarter of last year to this quarter, for instance, more than 10%. I'm trying to understand how we should be modeling the non-Cortex-A royalties versus the Cortex-A royalties, because Cortex-A clearly growing very fast, why is the other business not growing at this point? Thanks.
Okay. Let's start off with the physical IP one. I don't think, Sandeep, you're going to see a massive sort of switch. What we are seeing is a continuous trend. As I mentioned in the commentary, the leading-edge technology is now accounting for just over 30% of the physical IP royalty. I think the POPs are exclusively at those leading-edge technologies. As you get the leading-edge technologies occupying a greater and greater proportion, as more of that 70% that is still trailing technology turns into leading-edge technology, then that will be more of the POPs chipping and driving our physical IP royalty growth. As for the Cortex-M and the Cortex-R versus the Cortex-A, well, Cortex-M and Cortex-R don't command higher royalty rates.
Typically, these royalty rates are going to be in the 1%-1.5% range rather than pushing up towards 2%. We're not going to see a big change in royalty rates from those sorts of products. What we're seeing overall, and as a contributor to Arm's royalty rates, we're seeing the royalties driven by products incorporating things like Cortex-M having a negative impact on the average royalty per chip, because typically these chip prices are lower than the basket of Arm chip prices. This is what we said in answer to one of the earlier questions. You have a positive effect from the Cortex-A, the higher value chips with the higher royalty-bearing components within them.
At the same time, we have growth in Internet of Things type applications, microcontrollers and so on, which the royalty percentage per chip is constant at the lower end of the percentage spectrum, 1%-1.5%, and that's multiplied by a lower chip price. There's no news in that. That's been a trend for a long time, and that is going to continue, but that's incremental profit for us to have a greater share of those microcontroller spaces and is overall good news for our absolute USD royalties.
Great. Warren, just a follow-up to that. I understand that point, that any royalties is great because it's money which comes to your bottom line. What I'm trying to understand in terms of modeling it as such is that clearly Cortex-A grows exponentially over the next 5 years, we need the rest of the royalties to also grow over the next 5 years, which we haven't seen, say, over the last 3, 4 quarters. Is this because of where we are in the semiconductor cycle, or is it because some mix has shifted in this non-Cortex-A which is causing this to happen? How should we be looking at that rest of the Arm business to trend over a 3, 4-year period as such?
Sandeep, it's Tim. I think the best advice we can give really is to focus on the segment penetration slides that we publish on a regular basis. We give our estimates of the total opportunity. We give details of where our penetration is, how it's tracked in recent years, and provide commentary on where we think it's going to go. I think from a modeling standpoint, it's a volume opportunity. As Warren said, the ASPs of the chips tend to be broadly flat. The percentage we take is in the range that Warren said, and the modeling of it is really a question of how fast we think that Arm is going to penetrate those markets over time. No change there, really. In terms of very short term, these things are much harder to map out and relate to each other.
Longer term, I think the trend is clear. Arm is growing in each market, but in all markets that we're targeting. Those markets are characterized by different chip prices and different penetration levels at this point, but they're all growing.
Okay. Thank you.
Thank you very much. From UBS, your next question comes from Gareth Jenkins. Please ask your question, sir.
Thanks. I have one follow-up question from an earlier question, Warren. I just wondered if you could highlight within mobile what your average royalty revenue per chip has done in the quarter. I think last quarter you cited it was growing around 12% year-over-year. I just wondered whether you could give us a sense of what it's done in the current quarter. Away from mobile, just in terms of networking, I wonder whether your biggest areas of opportunity here are penetrating companies like Broadcom in home Wi-Fi, et cetera, or whether it's actually into totally new areas. If so, can you maybe detail a few of those?
Finally, just in terms of strategic investment, obviously I know you don't want to provide much detail around that, but I just wondered, flipping some of the earlier questions on the head, why you need to externally invest when your ROIC is so high that it would make sense to actually invest internally more than external. Is it to seed new companies like Calxeda, or is it in totally other areas? Thank you.
Okay. Righto. Let's start off with the quick one on mobile. Yes, last quarter we indicated that mobile ASPs were growing by about 12%. This quarter, we'd say the corresponding number is about 17%. That, as I mentioned in the earlier commentary, that's sort of Cortex-A and Mali driven. In the current quarter, on a year-over-year basis, we've seen significant increase in the Mali. Networking. Well, the answer is both. Yes, in the home networking products where Arm has been present for many years. We are seeing gradually an increase in our penetration there, and undoubtedly moves by companies like Broadcom help. I think the new opportunity in networking is more in the infrastructure, in the big infrastructure, in switches and routers, particularly in mobile infrastructure as companies roll out LTE implementations and beyond.
We're seeing a tremendous amount of change in that marketplace, evolution in form factors. People basically trying to address the serious issue of how on earth are they going to power all this equipment, Arm has a very power efficient solution to that. As I mentioned, the lack of a 64-bit processor has held us back from some of those sockets. Now people can see that 64-bit processor coming on the horizon. We are seeing design-ins for the earlier 32-bit implementations in many of these applications. We're at the design-in stage at the moment. Very little of this is actually product shipping. It's going to be a significant driver over the sort of medium three to five year time period.
Yeah, Gareth, on your third question related to the other debtor that I referred to in my introductory remarks, just over GBP 100 million at the end of September. You're also right, we're not in a position to provide any further details on that. I don't think there's any point in discussing our rationale when we're not in a position to discuss what we're actually contemplating. I think should the transaction close, we'll obviously be in a position to discuss that in detail in due course, and if it doesn't, that other debtor will revert to cash.
Thanks.
Your next question from the line of Sumant Wati from Redburn Partners. Please ask your question.
Hi, guys. Thanks so much for taking my question. Congratulations on good quarter results. I think a little bit of it is a bit of a follow-up from the previous question on the strategic investment. I understand that you can't talk a lot about the detail of where that investment is going. Could you possibly just tell us that if it actually does go through, whether the investment would be paid by your cash on the balance sheet, or would you actually be looking at any other sources of financing this particular project? My follow-up is essentially, I think this is for the first time ever that your non-mobile units now represent about 50% of your royalty bearing units, essentially.
Although from a unit perspective we understand that, if it is possible, could you just give us a rough idea of what the value composition of these two areas is, essentially? Thanks.
Yeah. Sumant, in the paragraph in relation in note four to the statements, you'll see the reference that these funds which are being conditionally committed, were actually financed in the short term by a facility because we didn't want to break existing deposits that were in place in the normal course of business. There is a short-term facility which transaction or not, we would expect to be repaid in short order.
To your second question about the non-mobile, mobile split in value terms. We don't intend to do a lot of detail on this because we don't want our customers reverse engineering these sorts of statements and playing it into the commercial discussions we have with them. 50/50 in volume terms today translates into about 60/40 mobile, non-mobile.
Okay. Thanks very much. I think, sorry if I can just ask one very quick follow-up. On OpEx guidance for Q4, you've given that, but I think a couple of people have asked this. For 2013, if we are going to start modeling it already, do you think using the base case of what the OpEx would look like for this year would be a fair view into budgeting for next year, essentially, at this point?
Well, I think the things to take into account, Sumant, are obviously, as Warren said, we would in a sort of normal business environment, continue to make investments in our resources. I think the other thing you need to do when you think about next year is factor in the full year effect of activities in this year, as well as things like inflation. Those are the factors that need to be taken into account. Obviously, we are in the fairly early stages of our budgeting for next year, and we'll give a lot more guidance on our 2012 cost base at the end of the year.
Okay. Thanks very much, and congratulations.
Your next question from the line of Andrew Gardiner of Barclays. Please ask your question.
Good morning, guys. Thank you. I had a question around tax. In particular, now that we're starting to get a bit more information from the government regarding the Patent Box, I was wondering if you had determined with any more accuracy how your tax rate's going to look going forward as the Patent Box is phased in over the next few years.
Yeah. Hi, Andrew. Similar message, really, to the one I gave at the analyst day. As you say, the Patent Box has now been legislated, and it becomes effective April 2013, and is phased in over five years with 60% of the benefit in year one, i.e., tax year April to April 2013 to 2014. My expectation would be that all other things being equal, Arm's tax rate will be below 20% at the back end of that period. Obviously, if you look at the line between where we are today and that lower rate, a reasonable kicker to that is in year one, which straddles the financial years 2013 and 2014.
Okay. Thanks very much.
Next question from line of Pierre Ferragu from Bernstein. Please ask your question.
Good morning. Thank you for taking my question. I have a question on architectural licensing again. It seems that type of licensing is gaining share amongst your mobile clients. You have now two very big clients using them. I was just wondering how you see that impacting your average royalty rate as of today and going forward if you think it can change the trajectory of your royalty rate in percentage terms. Do you think that it opens the door for some of your clients, for instance, taking longer to one architecture before moving to the next one because they have more flexibility on how far they can exploit the architecture they're working on at the moment?
Well, the answer is there's no fundamental change there. Obviously, the royalty arrangements that incorporate into our licenses are individual and confidential to the arrangement. As a rule of thumb, in modeling Arm, you shouldn't consider royalty arrangements related to architecture licenses to be any different from our perpetual implementation licenses.
What about how this could affect the behavior of your clients? It could encourage maybe some of them to stick longer to an architecture before moving to the next generation?
Not really. I think we've always had architecture licensing as an option for some of our semiconductor partners who want to take advantage of it. Historically, we haven't seen people with architecture licenses take any longer over a particular generation than people who use Arm implementation. I don't see any reason for a change in behavior now.
Don't forget that many of our customers use a combination of both of architecture licenses and implementation licenses. In doing that, they're obviously making their decisions, but there isn't a particularly different timeline to market in either case.
Okay, thanks a lot.
I think we better make this next one the last question. I'm afraid we're running out of time.
Next question from the line of Jerome Ramel from Exane. Please ask your question.
Yeah, good morning. Just two quick question. What is the current attach rate of the graphic Mali with Cortex-A? Second question concerning the royalties coming from the smart card business. You still have a major design win which are not shipping yet and which could easily boost your market share from 10% to roughly 30%. When do you expect the first major shipment to happen?
Well, in answer to the first question, Jerome, I'm afraid we're not disclosing that. As far as information provided is concerned, we've said that Cortex-A shipments account for 9% of the total volume. We've also said that our target for this year is to exceed 100 million units for Mali shipments. When we were at the half year, we said we're on track for that. At this point in the year, I'd say we're certainly on track for that. If anything, we probably could say we are running ahead of that run rate right now. We're not running ahead of it to such an extent that we want to come up with a sort of revised target. I think you can say that we're very confident we're going to exceed 100 million units by some margin.
I think that's about all we can really say on attach rate for Mali and Cortex-A. As for the smart card design-ins and when these are going to take off, I'm afraid we can't really comment. It's such a small space that we can't really comment on particular designs because there's only a few players, and we're not able to talk about specific customer design wins like that.
Okay, fair enough. Thanks a lot.
Okay, thank you. Congratulations to Jerome for getting the last question in. With that, I'm afraid we're going to have to draw it to a close. As we said at the start, it's been a great quarter. Very strong set of results, lots of momentum in the design wins, lots of excitement in the new applications and spaces we're getting into. Record licensing, record order backlog. We're saying that the opportunity pipeline's looking pretty robust at the moment, and prospects for the backlog for Q4 look very promising. We'll be back at the beginning of February to tell you how we got on in Q4. Thank you very much.
That concludes conference for today. Thank you for participating.