Arm Holdings - Earnings Call - Q1 2011
April 27, 2011
Transcript
Operator (participant)
Standing by and welcome to the 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 that time, if you wish to ask a question, you will need to press Star, one on your telephone. I must advise you that this conference is being recorded today, Wednesday, the 27th of April, 2011. I would now like to hand the conference over to your first speaker today, Mr. Ian Thornton, Head of Investor Relations at Arm Holdings. Please go ahead, sir.
Ian Thornton (VP of Investor Relations)
Thank you very much indeed. Good morning, everybody. This is Ian Thornton, Vice President of Investor Relations at Arm Holdings. On today's Q1 results conference call, we have with me Warren East, the Chief Executive Officer, and Tim Score, Chief Financial Officer. On today's call, Warren and Tim will take us through the highlights and comments from the quarter's results, and then we'll open this up to a Q&A session. As a reminder, the presentation and press release can be found on the Arm Investor Relations website at www.arm.com/ir. Before I hand over to them, I just have to read out a few words with respect to this conference call and what we're about to discuss.
The contents of this conference call are being directed only to those of you who have professional experience in matters related to investments, and the information communicated on this call is being made available only to investment professionals. Any persons present on this call who do not have professional experience in matters related to investment should not act or rely on the content of this call. The following conference call will contain forward-looking statements, which are other than statements of historical fact. The company's actual results of 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.
Warren East (CEO)
Thank you, Ian. Good morning, everyone, and thank you for joining us this morning. I'll go through the business highlights and then hand over to Tim to provide some more detail. As usual, we expect that we'll cover most of the content during the question and answers. 2010 was an excellent year for Arm in this morning we're pleased to announce that that momentum has continued through into the first quarter of 2011. As we look forward, we can see further healthy indicators for our growing strength in the medium and longer term. We came into the year with a record backlog. We had several new products scheduled for launch during the year and a strong competitive position as we're making further gains in these long-term structural growth markets. What's happened is that we've continued to outperform the semiconductor industry with strong growth both in licensing and royalties.
We're benefiting from electronics companies developing products enabling a more pervasive consumer-oriented Internet, and that's the main driver of licensing, with 39 processor licenses sold during the quarter. Shipments of Arm-based chips used in these sorts of products grew 32% year-on-year, and that's about 3x the rate of overall semiconductor industry growth, which drove those Arm royalties. The revenue growth is enabling us to continue to invest further in R&D, whilst at the same time growing earnings by 34% and delivering near record levels of cash generation. We are looking forward, therefore, to the rest of the year. We anticipate normal seasonality for royalty revenues that we tend to see during the second quarter, and Tim will comment on that in a few moments. There is some economic uncertainty as we're all looking for clarity on the medium-term impact of the Japanese earthquake and its aftermath.
Nevertheless, we expect group dollar revenues for the full year to be at least in line with market expectations. Now, I'll just discuss drivers for revenue in the different parts of the business in more detail, and we'll start with the processor division. We signed 39 licenses in the quarter. That's actually our highest number ever in terms of number of licenses. Most of them were as a result of current trends for increased Internet connectivity and also for smarter microcontrollers. There were eight licenses sold for Cortex-A family processors and seven Mali graphics processor licenses. 17 licenses in total were for Cortex-M family processors, and that's very good news, further increasing our penetration in microcontrollers and smart cards. We're particularly pleased that several major companies who are developing chips for digital TVs and set-top boxes have decided to start using Arm processors in the future.
This is a market we've been focusing on for a long time, so it's excellent to see years of technology development, business management, marketing, and design activity actually coming through and delivering some strategic wins. Two major companies signed subscription licenses: Broadcom and LG Electronics. These subscription licenses are long-term commitments where a customer pays a fixed annual fee and then has access to a range of our processor portfolio. Companies who've signed that type of agreement are able to readily deploy Arm technology more quickly across multiple divisions. After a few years, these sorts of companies often appear much higher up the table of royalty payers, so that's very encouraging news. Switching to royalty, I'll note royalty is reported one quarter in arrears, so the Q1 number that we're just reporting is as a result of Arm-based devices sold during Q4 2010. Processor royalty revenue was up 32% overall.
The total volume was a record at 1.85 billion units, and that was driven by continued growth and increasing sophistication in smartphones and continued growth in microcontrollers. There's a more complete breakdown of volume shipments provided in our earnings release and in the presentation, which has gone up on the IR section of our website. If we look at the individual segments, then there's continued gain in market share in non-mobile applications, which is very encouraging. We've seen a 55% increase in microcontroller shipments against a backdrop in industry where microcontroller shipments growing less than 20% overall. In DTVs and set-top boxes, we've seen a 15% rise in the number of Arm-based chips, and that's into an end market which has been broadly flat overall. Good news there in non-mobile.
In mobile, we are clearly benefiting from an increase in Cortex-A family chips used in smartphones and tablets and those sorts of products. Shipments of Cortex-A family are now up to 4% of Arm's total unit shipments. That was 3% last quarter and 4% now. As discussed back in February with our full-year 2010 results, we're often able to achieve higher royalty rates for Cortex-A-based chips as these A-based processors deliver more value to our semiconductor partners. These chips also tend to be the more expensive devices, and we benefit from a higher royalty rate on a more expensive device. These factors contributed to the average royalty per chip increasing sequentially from $0.046 to $0.048 in spite of the strong growth in microcontrollers. At the microcontroller end, the Cortex-M-based products are now accounting for 13% of the total volume. Cortex overall is 17% of total volume.
Now I'll switch to our physical IP division. In physical IP, revenue was up 19% overall, $23 million, and that was helped by licensing growing strongly, 43% year-on-year. Much of that growth has been from the strategic agreements that we've signed over recent years with foundry companies to develop advanced physical IP at 28 nm and below. As our engineering teams are developing and delivering that IP, we're recognizing the revenue, and that's coming through and driving that licensing growth. We've continued to see demand for the Processor Optimization Packages that we talked quite a bit about back in February. These enable our Cortex-A9 customers to readily achieve high-performance and low-power processor implementations using our Processor Optimization Packages.
For every chip that's implemented using one of those, let's refer to them as POPs, Arm receives royalty for both the processor that's in the chip and for the physical IP that's there. During the quarter, we signed three more licenses for POPs, and in addition, we also signed agreements to create new Processor Optimization Packs for Cortex-A family processors at 28 nm and at 20 nm. That's looking good for the future. Underlying physical IP royalties were actually down 1% year-on-year, but that was in line with the foundry industry. We're beginning to see benefits as high-volume production moves to the more advanced nodes of 45 nm and below, and we look forward to that effect going forward.
If I just touch on a few other operational and marketing points, we've continued with our recruitment that we grew fairly heavily in number last year, and we've continued in the first quarter of this year. We've added 33 people in total. This is growing our engineering teams working on advanced processors and graphics products, and that brings the total Arm headcount to 1,922. We intend to continue to invest in the team during 2011. The big marketing event during the first quarter was Mobile World Congress, which was probably our best Mobile World Congress yet. We had many OEMs showing off new high-end smartphones and tablets, mostly based on dual-core Cortex-A9 chips. In addition, several semiconductor partners showed off their next chips for future generations of mobile computers, and these included NVIDIA, the quad-core Cortex-A9, STMicroelectronics, and Texas Instruments, both announcing Cortex-A15-based chips. With that, I'll hand over to Tim to provide some further detail on the numbers.
Ian Thornton (VP of Investor Relations)
Thanks, Warren. Good morning. As Warren's given a fairly clear overview of the result there, I would just provide a little color on some of the numbers and some of the outlook. As you've read in the release, Q1 dollar revenue is $185.5 million, above consensus, which was around $174 million, driven primarily by a very strong performance in license revenue, which is up 50% in the processor division year-on-year and up over 40% in the physical IP division. Just over $51 million PD processor division license revenues were ahead of the normal range of $40 million-$45 million per quarter that I referred to at the full year 2010 analyst presentation at the beginning of February. Prospects for licensing continue to look promising with a high backlog and a strong opportunity pipeline.
We do not necessarily expect the number of licenses that we've seen in the last two quarters, i.e. 39 this quarter and 35 last, to be the norm, nor would we expect licensing revenues to be in excess of $50 million every quarter. In addition to the high number of licenses yielding short-term revenue in Q1, several partners have entered into longer-term commitments to use Arm technology, where much of the revenue associated with these agreements will be recognized in future quarters. As a result, the group backlog at the end of Q1 is up about 15% sequentially and is more than double the level it was at this time last year.
The usual analysis of backlog in terms of maturity and composition is included in the slide set on the website that Warren referred to, and that shows that 24% of total backlog is expected to be recognizable as revenue over the next two quarters, and approximately 50% of PD license revenue in Q1 was generated from the backlog. Processor royalty also grew strongly in the quarter, up 32% year-on-year and up 7% sequentially to just under $88 million. Typically, seasonality results in about a 5%-10% sequential decline in Arm's royalty revenues for the second quarter based on Q1 shipments. This has been the pattern over the last four years. In 2009, the sequential decline was more than that at 18%, but that, of course, was a time when the industry was going into downturn.
I think 5% to 10% sequential decline in line with normal seasonality is the way to look at it. Based on the industry data that we've seen for the first quarter, which, of course, is the relevant period for our Q2 royalties, we would expect to see normal seasonality again this year. When you look at the models and compare Q2 2011 forecast royalties to Q2 2010, it's probably worth remembering that last year we had a $9 million catch-up in processor royalties. Moving on to costs, normalized OpEx in the first quarter was $60.3 million compared to the guidance given in February of $57 million-$59 million.
As mentioned in the earnings release, the sequential increase to a level just above the top of the guidance range is largely due to the impact of a weaker dollar in Q1 on the accounting for derivative instruments, which has resulted in a net charge of just under £2 million being included in G&A expenses in Q1. As a result of the strong revenue performance, the group is reporting normalized operating margins in excess of 42%, with net cash generation in the first quarter well up on the same quarter last year. This is at the same time as we have been continuing to increase our investment in new technology by growing our R&D capability. Looking into the next quarter, we expect normalized operating expenses, assuming effective exchange rates broadly similar to current levels, to be in the range of £58 million-£60 million.
Reiterating the outlook that Warren referred to, we do anticipate normal seasonality for royalty revenues in Q2. Notwithstanding the current uncertainty as to the economic impact of the earthquake in Japan on the semiconductor industry supply chain and end product markets later in the year, we expect that group dollar revenues for 2011 full year will be at least in line with current market expectations of about $730 million. With that, we'll throw it open to questions.
Operator (participant)
Okay. We will now begin the question-and-answer session. If you wish to ask a question, please press Star, one on your telephone and wait for your name to be announced. If you wish to cancel your request, please press the Hash key. Your first question comes from Sandeep Deshpande from JPMorgan. Please ask your question now.
Sandeep Deshpande (Analyst)
Yeah. Hi. Thanks for the question. Just a first question on PIPD. I mean, despite the wins you're seeing in PIPD, and PIPD royalty revenues seem to be declining year-on-year despite an increase in the semiconductor industry. Is there an explanation for that?
Warren East (CEO)
I think there is a very small decline year on year. When we look at how that's made up, it's mainly due to physical IP users switching their production down to the 65 nm node. If you go back and look at the licensing that we were doing a few years ago, then actually our physical IP at 65 nm is much less penetrated than at 90 nm and indeed at 45 nm and 32 nm and below. That's why I'm actually quite encouraged and said in the opening commentary, encouraged to see 45 nm and more advanced nodes starting to get to production because we do have a little bit of a soft patch in 65 nm. That's what's coming through and affecting royalty today.
Sandeep Deshpande (Analyst)
Do you expect that to continue for a few quarters? I mean, because 45 nm is, I mean, more second half loaded in terms of shipments.
Warren East (CEO)
There will be downward pressure from 65 nm and upward pressure from 45 nm. It's quite hard to call exactly which is going to dominate over the next couple of quarters. If we look through several quarters, we can clearly see the world moving to 45 nm and below fairly quickly.
Sandeep Deshpande (Analyst)
Warren, would you also like to comment on what you're seeing in terms of the reported royalty rates on the Cortex processors at this point as opposed to what you're signing and negotiating, what the reported royalty rates are?
Warren East (CEO)
Yeah. Obviously, we're not going to comment in specific cases. The Cortex-A products, which are shipping at the moment, accounting for 4% of shipments, are commanding the higher royalty rate that we talked about back in February. That, combined with the typically higher ASPs, is what has driven the sequential increase in absolute dollars and $0.01 per chip in spite of obvious downward pressure from the strong growth in microcontrollers.
Sandeep Deshpande (Analyst)
You're not commenting on what this actual, is it a 1%, 2%, or 2.5%, or what that figure is?
Warren East (CEO)
No. We won't really comment on particularly the percentage there, Sandeep. There's a range. We've said by the time you get through to Cortex-A9s and Cortex-A15s, then you're at the higher end of the 1%-2% range. The reality of most of what's shipping today, though, there's very few other than Cortex-A8 products, and that's a little bit earlier on in that journey.
Sandeep Deshpande (Analyst)
Thank you very much.
Warren East (CEO)
Thank you.
Operator (participant)
Your next question comes from Gunnar Plagge from Nomura. Please ask your question now.
Gunnar Plagge (Analyst)
Yes. Hello. I was wondering whether you could speak a little bit more about the Mali design wins. I think you also mentioned recently that compared to your main competitors, the ecosystem here is not as developed. Maybe you could talk a little bit what we can expect over the course of the year here. As a follow-up, just a quick comment on the share-based compensation, which seemed to have jumped quite substantially from Q4 to Q1. Thank you.
Warren East (CEO)
Okay. I'll pick up on the Mali question, and Tim will come in on the stock-based compensation in a few moments. We were pleased to see seven Mali licenses during the quarter. That is for a mix of mobile computing design wins and DTV set-top box type design wins. Mali is going into our processors for both those end products. Obviously, that's licensing activity. We'll have to see how that translates into shipments over the coming years. In terms of what we expect for this year, we continue to expect, as we said in February, shipments of Mali-based products in the tens of millions this year. That is going to be in a mixture of smartphones and in DTV-type products. It's correct what we said in February about the maturity of the Mali ecosystem being less mature than the Imagination ecosystem, which is essentially the incumbent in the mobile space.
We expect to see an acceleration in the maturity of the Mali-based ecosystem as those products start shipping this year in, as I say, the tens of millions of units. I think that's it on Mali.
Ian Thornton (VP of Investor Relations)
Yeah. Gunnar, Tim, on share-based compensation, there are some accounting effects which are one quarter only in Q1 based on the very significant rise in the share price during the quarter. We ended the year at £4.23, and by the time we did our RSU grant in early February, we were at £5.75. By the end of the quarter, we're a little bit higher than that. What I expect to see happening in the subsequent quarters of this year is the share-based compensation quarterly charge will return to the levels that we saw in Q3 and Q4 last year. I think the other point to note is that our grants to employees are broadly based on proportion of salary. Therefore, what we are seeing is the number of shares that are granted to employees is going down as the share price goes up.
In terms of the overall forward dilution, in terms of number of shares, that's going down.
Gunnar Plagge (Analyst)
Thank you.
Operator (participant)
Your next question comes from François Meunier from Morgan Stanley. Please ask your question now.
François Meunier (Analyst)
Yes. It's François. Congratulations on the results again. I've got a question first on the deal signed with LG yesterday. It looks like it's a pretty broad type of licensing deal. I don't think LG has made chips before. It's a bit of a new thing. If you could please comment on this deal with LG and maybe if you expect more OEMs going forward to use this type of strategy. Maybe I'm looking at Sony as well in terms of their new gaming consoles type product. Thank you.
Warren East (CEO)
Good. Thanks, François. There's nothing particular to note about the LGs deal. It's a subscription agreement, as noted. It's targeted at a range of products, including phones, potentially some mobile computing, also DTVs and set-top boxes. It includes normal Arm microprocessors and Mali graphics. I think that's all there is to note on it for the time being. Arm has worked with LG for some time. Arm processors have been designed into LG's DTVs for some time, and this is just an extension of that relationship.
François Meunier (Analyst)
Okay. That's a good question I would have. It's about the ASPs, which are continuing to trend upward, so that's very good. Obviously, we've seen recent tablets being launched with die size much bigger than the ones which were launched last year. Obviously, almost all tablets being launched this year are dual-core. Shall we expect the ASP for Arm on those new tablets to be much higher than what they were last year?
Warren East (CEO)
I think you probably can. I mean, these are new tablets launched around new chips. Of course, the phenomenon that we usually observe is that when a new applications processor is first launched, it's released at a higher price than the generation of chips that it's replacing. Then a normal ASP decline takes over a period of a couple of years. Yes, we do expect to see higher ASPs from those particular chips. I'd like to caveat, before everybody runs away and revises their models upwards, caveat that with normal ASP decline over a couple of years. In spite of it's great to see all these tablets being launched. Don't forget, we're still talking about a market of 40 million-60 million tablets for the year. We just shipped nearly 2 billion Arm microprocessors in the first quarter. You have to put that into perspective.
François Meunier (Analyst)
Thank you.
Warren East (CEO)
Thanks.
Operator (participant)
Your next question comes from Didier Scemama from RBS. Please ask your question now.
Didier Scemama (Analyst)
Yeah. Good morning, gents. Congratulations on the results. A couple of questions, if I may. Just on Francois's question, I just wanted to add one more element. When we look at the diary of what's available on the Internet of tablet processors, roughly what we are seeing is that the CPU, dual-core CPUs, takes about 20%-25% of the diary. If I do the math right, that would be anywhere between $3-$4, if you want, of the cost of sales, which is materially above what you charge. What I'm wondering is if in the longer term, I'm talking the next 5-10 years, we could see the royalty that Arm Holdings collects on the CPU side for those processors going into PCs and so on could trend towards those numbers. That's my first question.
My second question would be, and I apologize if you've been misquoted, but I can see on Bloomberg a comment that was made earlier this morning saying that AMD could do a lot more with Arm Holdings in the future. I was wondering if you could elaborate on that, please.
Warren East (CEO)
Okay. First question about the royalty rates for chips. As we said in February, if we are supplying more value to a semiconductor company, then we are replacing work that that semiconductor company would otherwise have to do. We're replacing costs that that semiconductor company would otherwise have to incur. Therefore, we are able to strike an agreement with the semiconductor company that in exchange for additional value, we will take home a higher royalty rate. That trend continues. We try very hard to steer away from discussions about silicon area when we're discussing royalty with semiconductor companies. The point of the Arm business model is that we supply the microprocessor, and different semiconductor companies differentiate around that microprocessor in different ways. In some instances, that will mean a greater amount of silicon area is in the semiconductor partners' differentiation piece.
In other cases, there'll be a lesser area in the differentiation piece. Because of that, it's a very difficult commercial discussion to have for both sides. What we tend to do is look at the value of the Arm microprocessor in terms of the functionality that it delivers rather than the area. I think that's all we really need to say about the royalty rates and chip areas. On AMD, we were asked a question this morning about AMD. We simply said that AMD is a big semiconductor company. They're in the business of shipping microprocessors. We're in the business of licensing microprocessor designs. It's a perfectly normal course of business for Arm to be trying. You would hope that we would be trying to sell microprocessor designs to AMD.
Clearly, AMD has signaled that they're going through a bit of a rethink of their strategy at the moment. Therefore, that presents, as far as we're concerned, standard commercial common sense will tell you that that presents a heightened opportunity. At the moment, they don't use Arm microprocessors. As far as we can see, there is an opportunity there where they might use Arm microprocessors in the future. You've got to expect that we'll be trying to persuade them of that.
Didier Scemama (Analyst)
Brilliant. One final follow-up, if I may. What we've seen from the OMAP 5 from TI is a dual Cortex-A15 and some slave, I believe, Cortex-R5 or Cortex-R3M. I'm sorry, I forgot exactly which. On those chips that have dual CPUs plus sort of slave CPUs, how much higher would be the royalty rate on those? I mean, the sort of quantum of improvement relative to just a dual Cortex-A9, say.
Warren East (CEO)
As you know, I'm very sensitive to commenting on particular semiconductor companies. In the general case, where there are multiple Arm microprocessors per chip, each microprocessor is commanding a royalty of its own. In the general case, a dual A15 will command a certain level of royalty, not because it's dual but because it's A15. The M4, such as it is in the OMAP device, will be commanding another royalty for M4. Typically, there's a discount for the second and subsequent cores, and that discount might be quite significant, like about 50%.
Didier Scemama (Analyst)
Perfect. All right, thanks so much.
Warren East (CEO)
Thank you.
Operator (participant)
Your next question comes from Gareth Jenkins from UBS. Please ask your question now.
Gareth Jenkins (Analyst)
Yeah. Thanks. A couple of, if I may. Just firstly, I wondered what you make of Google's intentions to reduce fragmentation and also an update on Leonardo from your side. Secondly, I just wondered whether there's anything that you've seen from Intel that's particularly surprising, the sort of acceleration of TikTok to 14 nm and how that positions you. Finally, just the timing on 64-bit, please.
Warren East (CEO)
Okay. On Android, there has been in the quarter some speculation about and discussion about Google's intention to standardize. I think really to talk about Google's intentions, you have to talk to Google, not Arm. As far as we can see, Android is an open-source platform. There are benefits in the standard, but because it's open-source, customers are able to take the starting point and make modifications to suit their own products. At one level, if those modifications are not done in a controlled way, that can lead to fragmentation, and you can end up with applications that are written for one particular end product. For example, an end product that has a small screen. Maybe a particular product that's got a large screen, you'll get a different type of application performance on that. It's very much down to how sophisticated is the modification performed by the OEM.
What this open standard does is enable Google's customers to differentiate. Exactly how well they do that differentiation is up to them. Exactly how much Google wants to control that differentiation is up to Google. We are keen to support Google and its customers in bringing a range of diverse products to market, and we'll continue to do that. Efforts such as Leonardo do help those OEM customers to work across multiple system-on-chip devices and multiple system-on-chip devices based on the Arm architecture. That's why we started Leonardo a year ago. Leonardo's continued pretty much as planned. The number of engineers deployed across Leonardo has continued to grow as planned. I think they're up to about 120 people at the moment. You have to talk specifically with Leonardo about that. Arm is now an ordinary member represented on the Leonardo board.
We no longer have a sort of controlling interest or steering role in Leonardo that is unique any more than any of the other Arm partners that's involved in it. I think that's all I can really say on Android and Leonardo. On Intel accelerating TikTok down to 14 nm. Yes, Intel has been quite explicit about manufacturing technology being their key area of expertise. We have yet to see how that develops. The Arm partnership has access to leading-edge foundry capability from the leading foundries. In our physical IP division, we're striking agreements with leading-edge foundries to collaborate to 14 nm and below as well. You saw announcements from Arm and IBM. You saw announcements from Arm and TSMC at the TSMC event the other day as well. It is very much normal progress in the semiconductor industry. Your last question was about 64-bit.
All we've said on 64-bit is that clearly it's something that's been done in the world of computing. Everything we've done so far in the Arm architecture has simply followed what's been done in the world of computing and taken that and implemented it on a system-on-chip type scale. Yes, people are talking about 64-bit versions of Arm, but we're not in a position to make product announcements. We'll make the product announcements as and when we're ready to do that.
Gareth Jenkins (Analyst)
Thanks, Warren.
Warren East (CEO)
Yeah, that's it. Thank you.
Operator (participant)
Your next question comes from Kai Korschelt from Deutsche Bank. Please ask your question now.
Kai Korschelt (Analyst)
Yes. Good morning. I have two, actually. The first one was on ASPs. They moved up a bit sequentially. Given this is the sort of strong Q4 holiday season, that's reflected in there with higher smartphone and tablet percentage. Should this sort of trend down in Q1, which is your Q2 again, or should ASPs stay at around that level? My second question is if I could maybe be a bit more pushy on stock options. Clearly, it's a large part of your OpEx base. If we treated it as an expense, your margins would be materially lower now. It's obviously a good thing to have employees participate in the success of the company. I was wondering if, given your strong balance sheet and cash generation, are you planning to at least offset the dilution to the existing shareholders from this via a buyback, for example, at any stage?
Gunnar Plagge (Analyst)
Thank you.
Warren East (CEO)
Okay. I'll deal with the first one, and Tim's going to answer the second question. On ASPs, yes, you're right. The upward pressure on ASPs has been driven by smartphones, mobile computers, DTVs, these sorts of consumer-oriented products which tend to be more seasonal. As we look at the next quarter, typically, we'll see market sectors such as industrial, which is much less seasonal. We'll see microcontrollers probably dominate more, and that could result in a flat ASP for the quarter or even a decline as we've got these two sort of offsetting forces, one pushing the average rate up, one pushing the average rate down. As we've discussed many times, this is simply the result. A good result for Arm may well be a lower average number overall because that will mean that growth in microcontrollers has been stronger.
Therefore, a lower average number could actually be a good result for Arm, and you shouldn't necessarily interpret that as a bad result for Arm. Q2, we expect normal seasonality. We expect overall royalty revenues to be down 5%-10%, as we've said, through normal seasonality. If you look at the mix there, we would expect smartphones and tablets and things to have less of an impact than microcontrollers.
Ian Thornton (VP of Investor Relations)
Kai, you raised two issues there. One is, in a sense, remuneration policies, with particular reference to share-based remuneration. You are touching on, if you like, the wider issue of capital structure management and balance sheet efficiency. Of course, there could potentially be linkage between them. Our approach to share-based remuneration has been consistent over many years. As I said earlier on the call, each of our employees gets some proportion of their salary in share-based remuneration. What that means today is that we are issuing somewhere around 0.5% or 0.6% per annum of our share capital to employees. That policy has been consistent. As to whether we do share buybacks to specifically offset dilution or for wider balance sheet efficiency management is something that we will wait and see. We have obviously done that before.
Between 2005 and 2008, we bought back about 16% of the shares and returned about £260 million of cash. Our current plan of record is progressive dividend. We have been growing the dividend through the cycles, and that remains our plan of record. Share buybacks remain an option. We have the authority to do buybacks up to 10%, and we routinely renew that. It remains an option. There is nothing to announce on that today.
Kai Korschelt (Analyst)
Okay, thank you.
Operator (participant)
Your next question comes from Raj Seth from Cowen & Company. Please ask your question now.
Simran Brar (Analyst)
Good morning. Thank you. This is Simran Brar calling in for Raj Seth. My first question was for Warren. For Mali, you continue to see strong momentum for the second quarter. I was wondering, when do you expect revenues from Mali to sort of become material for your model? In terms of gaining share versus your dominant competitor, how much of a deterrent are switching costs for potential customers?
Warren East (CEO)
Yeah. Okay. We are pleased to see continued momentum, seven licenses in the quarter. As far as license revenue is concerned, then a Mali license is fairly similar to a generic Arm microprocessor. Typically, the price would be less than something like a Cortex-A9, but it's more like a high-end Arm 11 or something like that in terms of pricing. That's the sort of contribution that's being made from a licensing revenue perspective. Out of a total of 39 licenses in the quarter, seven of them were Mali licenses. You can gauge the contribution to license revenue from that. In terms of when it's going to become significant for royalty, I think 2011, as I said on the call, we expect to see tens of millions of units of Mali. That's obviously a small contributor compared with the run rates on this quarter of 1.85 billion units in the quarter.
Tens of millions of units is relatively small. Mali does tend to go into the more expensive devices sat alongside a high-end Cortex processor in an applications processor. It should be compared with the Cortex-A products, which in the last quarter were about 4% of total shipments. In terms of that comparison, tens of millions of units this year will look like a reasonable contribution. It's those shipments of units this year which make an impact to the switching cost because at the moment, the key issue about switching is that the maturity of the Mali ecosystem is clearly less developed than that of the incumbent, which tends to be imagination-based devices. With tens of millions of units shipping this year, we would expect to see that maturity develop and start to develop quite rapidly.
That will bring down the switching cost so that by the end of 2011, the switching cost will be a lot less substantial than it is today.
Simran Brar (Analyst)
Thank you. One more question for Thorn, if I may. In terms of your model, Tim, could you remind me what typical seasonality is for the processor royalty side of your business in the second half? Assuming a constant exchange rate, how should we think about expense growth in the second half of the year? Thank you.
Ian Thornton (VP of Investor Relations)
If you typically look at the shape of processor royalties through the year, as we've been saying this morning, we tend to see a 5%-10% decline in the second quarter versus the first based on Q1 shipments. What we tend to see then is an uplift in Q3 and then a bigger uplift in Q4. I think the reference today to the uncertainty around Japan, I just think, needs to be considered when looking at the forecast for royalties in the second half. I think it's too early and it's very unclear as to what impact there may be on consumer behavior in Japan at the back end of the year. We do expect some impact. Normal seasonality is for an uptick in Q3 and a bigger uptick in Q4. Assuming that similar level of we've guided for the second quarter, we are on OpEx.
We're in investment mode. We expect to see an increase in headcount in the second half and therefore probably a gradual increase in OpEx. Consensus OpEx coming into these results is about £236 million for the full year 2011. That may edge up a little bit, but I don't think substantially based on today's results and the guidance for Q2.
Operator (participant)
Great. Thank you. Your next question comes from Jérôme Ramel from Exane BNP Paribas. Please ask your question now.
Jérôme Ramel (Analyst)
Yeah. Good morning. Just to come back to the full-year guidance where you say you expect revenues to be at least in line with consensus. If I look at what Broadcom recently said on the wireless, Nokia units, iPad units, it seems that the wireless is a little bit weaker than what people were expecting going forward. If you feel confident with the consensus, I just want to check if actually you feel more confident on the licensing revenues rather than on the royalties for the full year.
Ian Thornton (VP of Investor Relations)
You're right. I think the recent announcements of some of the semiconductor companies who are Arm licensees about their first quarter and our second quarter does underpin what we've been saying about normal seasonality. Although there are clearly mixed guidance out there, some companies are flat to growing, others are in that range, and some are below that range. I think in terms of the full year, our comment is looking at our overall revenues. We do expect to see growth in royalties in the second half. We're just noting the fact that Japan is likely to have an impact on the shape of that. We have a record order backlog that underpins licensing in the second half, and we have a strong opportunity pipeline. Our comments are really based on the business overall.
Jérôme Ramel (Analyst)
Thank you. On the backlog, any idea of where it should be flat or still continue to grow this quarter?
Ian Thornton (VP of Investor Relations)
Yeah. I think it is very difficult for us to call backlog on a quarter-by-quarter basis. I mean, when we look into our opportunity pipeline for licensing, there is the normal mix of deals between those that yield short-term revenue because they're for more mature technology and licenses for technology currently under development or future technology and indeed potential longer-term deals such as that we've seen this quarter. It really depends which of the pipeline closes in the second quarter as to what the backlog is at the 30th of June. Looking overall into that pipeline, it does look pretty promising.
Jérôme Ramel (Analyst)
Okay, it's okay. Thank you very much.
Operator (participant)
Your next question comes from Nick Hyslop from RBC Capital Markets. Please ask your question now.
Nick Hyslop (Analyst)
Thank you. Good morning, guys. I think most questions have been asked, but I had a couple. Obviously, in the light of what Broadcom has said overnight, when you look at the kind of royalty checks coming in, does the comments from Broadcom on weakness in mobile suggest that the maybe -5% to -10% might be closer to the -10%? Do you have any sympathy with what they've said?
Ian Thornton (VP of Investor Relations)
That one sounds like a crack record, Nick. What we're seeing out there is a range of reported results for Q1 for our customers and indeed for their guidance. There are companies that have seen sequential growth, there are companies that are flat, and there are companies that are down more than 10%. I think in the round, 5%-10%, I think, is the range. We have actually received a very small proportion of royalties to date. I think to be more specific than that would be unwise at this particular point. Like you, we note very carefully what all of our shippers are saying in their public result to the extent they're public.
Nick Hyslop (Analyst)
Okay. That's fair enough. Maybe just could you tell me what you're expecting for the full-year amortization charge of intangibles? It just seems to be very low. I know it's not an important one, but I'd be keen to know.
Ian Thornton (VP of Investor Relations)
I think the run rate will be rather consistent. What's actually been occurring is that the intangibles that arose on the acquisition of Artisan at the end of 2004, the great majority of them have now been fully amortized. You're seeing a much lower run rate relating to largely more recent acquisitions.
Nick Hyslop (Analyst)
Therefore, for the full year, that is now four times that number. That is the right number.
Ian Thornton (VP of Investor Relations)
That's a good way to think about it.
Nick Hyslop (Analyst)
Okay, that's great. Thank you.
Operator (participant)
Your next question comes from Nick James from Numis. Please ask your question now.
Nick James (Analyst)
Morning. I just had a question on the physical IP POPs. I wondered if you could remind us when you started selling these and how many of your customers have adopted them so far, and also the timeframe within the chip development cycle when people license the POPs, and in terms of how long the delay is until you start to see royalty revenues from these products.
Warren East (CEO)
Yeah. We started selling these as products at the very back end of 2009. So far, we've sold 13 POPs. If we analyze where those are going, people are using them for Cortex-A9. Very roughly, the sort of attach rate of a POP to Cortex-A9 is about 50%. You can look at our A9 licensing. Very roughly, half of those licenses, POPs have been licensed. Exactly the extent to which the POPs are used in the chip implementations, of course, we don't know at the moment. People are paying good money for them, and we anticipate they are being used as part of the chip design process. The length of time to develop the device is no different from normal. We're not seeing a lot of people licensing a POP after they've licensed an A9 a long time afterwards. We'd say typically about four years to meaningful royalties.
We started effectively at the beginning of 2010. You might see, because people who are buying A9, typically, they'll be fairly experienced chip designers, you might see a little bit of an improvement on the four years. Initial royalty, we'd expect to see probably more like three years. Any appreciable volume that you'll actually notice in terms of royalty revenue, then I expect that to be about four years.
Nick James (Analyst)
Great. That's very helpful. Thank you.
Warren East (CEO)
Thank you.
Operator (participant)
Your next question comes from Brett Simpson from Arete. Please ask your question now.
Brett Simpson (Analyst)
Yeah. Thanks very much. I have some questions on the PD licensing business. Revenue growth in this area clearly coming through much faster than historically has been the case. The last few quarters up 50% year-on-year, and you've had record license intake in each of the last three quarters. Warren, you've spoken in the past about the relationship between licensing activity and the sort of semiconductor design cycle where we see peaks and troughs in license activity, which has proven to be lumpy in the past. Is this dynamic becoming less relevant in your view? To what extent is this level of license intake sustainable over the next 12 months or so?
Warren East (CEO)
Okay. The answer is that yes, as Arm's sort of market becomes much bigger and much broader, then the lumpiness that is associated with licensing does tend to be smoothed out. However, we have said back in February, and we're sort of repeating it today, the license revenue that we would expect for a quarter run rate we would expect to be in the sort of $40 million-$45 million. We don't want people to think that north of $50 million is the norm. We are seeing a period of strong licensing at the moment as people are licensing Cortex-A9s, Cortex-A15s, not really sort of contributing in any way to license revenue at the moment. That's a brand new product. We are in the sort of sweet spot of A9 licensing at the moment. You would expect, therefore, it to be a little bit product cycle related.
That's why even though we've just reported two quarters of north of $50 million of licensing, we're still saying $40 million-$45 million is the normal range. That doesn't mean that there won't be another quarter of $50 million plus. There might be. If there's something more like 20-30 licenses in a quarter instead of 40 licenses in a quarter and license revenue in the $40 million-$45 million range, that's perfectly normal as well. You're right. There's a little bit of lumpiness, but it's probably a lot less lumpy than it used to be, simply lure of bigger numbers instead of lure of small numbers.
Brett Simpson (Analyst)
Sure. Just a follow-up there, Warren, if I may. In the presentation pack, it says you've signed cumulatively 68 licenses for Cortex-A. Can you talk about specifically how many licensees that translates to specifically for multi-core Cortex-A? I'd like to know how many chipmakers are developing these types of application processors as there's a concern that only a small number of chipmakers are actually scaling up in this area, particularly with Apple taking a disproportionate share at this stage.
Warren East (CEO)
Yeah. I mean, it's true that the Cortex-A on the chart refers to A15s, A5s, A9s, etc., etc. If you look at it from a number of semiconductor companies' point of view, there's about 30 of them involved in that activity.
Brett Simpson (Analyst)
Okay. Great. How many do you think, when you look at the total addressable market, how many chipmakers do you think can the ecosystem that can sustain this sort of development given the growth outlook here?
Warren East (CEO)
It is true that as a general rule, bigger investments mean fewer companies are able to make those big investments. However, we have been talking at these sort of conference calls and the meetings about the decline of Arm Holdings' licensing opportunity in terms of the number of customers who are going to buy licenses. We've been talking about that for the last decade, as long as I've been doing this job. In that time, the number of licenses has consistently grown. The number of licensees has consistently grown. Indeed, in the last quarter, we had 10 new licensees, including a couple coming to the party for the first time with Cortex-A-based products. Yes, that general trend must be out there. Common sense tells you that it must be out there.
Actually, we're continuing to find new customers who are ready to buy into the Arm architecture and existing customers who are ready to broaden their relationship and continue to invest in Cortex-A-based products. I don't want to forecast the demise of this trend too early.
Brett Simpson (Analyst)
Sure. Okay, thanks very much.
Warren East (CEO)
Thank you.
Operator (participant)
Your next question comes from Alex Shah from Davidson. Please ask your question now.
Alex Shah (Analyst)
Good morning. Warren, if I could, I just wanted clarification on your opening comment about LG and Broadcom, the licenses there. You mentioned that Broadcom and LG were licensing for set-top boxes and digital TV. Is that true for Broadcom as well, or were you referring just to LG at that point?
Warren East (CEO)
In the answer to the question, I was referring to LG specifically. The subscription licenses signed by both companies are broad licenses for a broad range of products. Broadcom will make their intentions clear when they're ready to do so. It is not for us to comment.
Alex Shah (Analyst)
Great. Thank you.
Warren East (CEO)
Thank you.
Ian Thornton (VP of Investor Relations)
Operator, can we go for one more question, please?
Operator (participant)
We can do. Your next question comes from Andrew Gardiner from Barclays Capital. Please ask your question now.
Andrew Gardiner (Analyst)
Thanks very much. Just a very quick one. Just another clarification on the POPs within physical IP. Can you give us any sense as to how significant they are from a revenue point of view in terms of licensing relative to perhaps sort of the norm?
Warren East (CEO)
Yeah. I mean, these are low-value licenses of typically a few hundred thousand dollars a piece. Our intention with the POPs is to enable semiconductor companies who've invested in a Cortex-A9 or above type product to be able to bring that product to market quickly and accelerate times of royalty. The purpose of the POPs is not really to generate lots and lots of revenue out of licensing them. We do like people to make a commitment, and that's why it's at about that level.
Andrew Gardiner (Analyst)
I see. I presume most of it at the moment is high-end smartphone mobile computing related. Is there any likelihood that you could expand this into other areas as well?
Warren East (CEO)
It's for Cortex-A9s at the moment. Actually, they're selling to people who are building apps processors for DTVs and set-top boxes as well as for smartphones and mobile computing type applications. It's literally just a way of getting a deterministic result out of your Cortex-A9 or A15 implementation. In that way, that's one problem in the design process out of the way, and you can get your chip to market faster.
Andrew Gardiner (Analyst)
Okay. Very clear. Thanks very much.
Warren East (CEO)
Okay. With that, we'll conclude. Thank you very much for dialing in and taking part. We're enjoying a good start to the year, a healthy pipeline looking forward. We look forward to talking with you again in July when we come back with our half-year results. Thanks.