Earnings Call: Q2 2010

Jul 27, 2010

Okay. Good morning, everybody, and welcome to, Q2 and first half results presentation and those of you on the line on the call. It's great to see everybody. Those of you who to whom I've spoken close-up. I'm, fermenting myself the champagne bottle tide this morning because we're quite pleased with our results for the first half, and we'll go into a bit of detail on that right now as soon as I can advance the slides. So, as I say, very, very good set of results, obviously trading conditions in the first half of twenty ten for everybody in our industry, have improved compared with 2009. But what's really good about the ARM position halfway through 2010 is that we feel we're very well positioned for long term growth the outperformance of the industry right throughout the cycle has continued. Our revenues profitability and earnings that I'm sure we'll talk about this morning all reflect that. And that's because in these key market sectors with long term structural growth that we've been talking about for some time. We're continuing to gain market share. And during the quarter, we've also had some very high profile announcements with industry leaders adopting arm technology, some fun for the first time, but generally for deeper use of of ARM technology, and that's all very good and bodes very well for the future. Our physical IP, I'll talk about in a few moments, but we've had some great progress there as well. And on a day to day basis, business is in good shape. Orders coming in, which position our backlog at absolutely record levels after a strong quarter. So we're feeling good. We have continued with a regime of pretty strict financial discipline, and that's led to healthy margins, which have in turn led to healthy cash generation as well as permitting increased investment in R&D through the quarter and indeed as we move into the second half of the year, you'll see us do more of that. The slide is a standard format reminding everybody of axes of strategic growth, and, we're reporting progress in all of those three areas. This this quarter. So very good. Now I'll just go straight into the detail down one level, and we'll start with a process of business. And and Processor Licensing, it was a very good quarter, actually record booking this quarter which has driven the backlog to a record as well. But by some margin, we've continued to grow our base of processor licenses at a steady rates, continuing to grow the installed base. And there are, several announcements, which have happened during the quarter, obviously, last Friday, high profile announcement from Microsoft, but the more normal licensing, which sits behind that, is in very good shape, too, with had 5 licenses for Cortex A Class Processes, and we've now got a full lineup of lead partners for Eagle. Our new, Cortex A class product, which will be launched a bit later in the year. At the other end of this room, the microcontroller end, we've had additional licenses, the Cortex MFAS products, and we've had some microcontroller announcements as well. Looking at our graphics portfolio, then, we're very pleased with further commitments from partners for Mali and the licensing the licenses with potential to generate future royalties there now incremented up to 29. The pie chart on the bottom is a familiar pie chart to you now. We've used this format before, but we're putting it there to continue to remind people that non mobile is still a strong driver for our new licensing. So people taking licenses for digital televisions for microcontrollers and some of those microcontrollers being used in new applications, such as smart meters. We've also seen in the mobile space, some new customers who've come to the armed party for mobile in the last couple of years starting now to create their chips and get their chips out into the marketplace and come back for more licenses as well. So that's pretty encouraging. Now a little deviation from the standard format of the presentation, I said last week, we had a high profile announcement with Microsoft or rather that's a low profile announcement with a high profile company. And this slide is here because we could spend a lot of the Q and announcement. We can't talk much about it, and this is as much as we are able to say this morning know, the facts are Microsoft have bought an architecture license from Arm. Architecture licenses allow customers to build their own implementations of arm microprocessors, which are fully compatible with the arm instruction set. It is a multi year agreement, which enables them to work more closely with us. Microsoft have actually been collaborating with Arm for many years, it says on the slide 13 years, my arithmetic is not too good, but I think it's probably getting on for 14 years now, actually, and Microsoft have brought several of their operating systems to the ARM architecture in that time. And those are the facts, and there's really not much more to say when pushed, I have obviously said that working with a leading technology company like Microsoft and then working more closely with ARM is obviously we're taking that as good news. But as far as that goes, I don't think we have anything further to add on Microsoft. So You won't get much more in the Q And A, I'm afraid. Back to the standard format of the presentation, if seen this slide before, but here's an update on our licensing activity and how we're growing the licensing base. When you pull these numbers apart, you'll see a very good mix of new and existing customers when you look at the the earnings release in some detail. We're pleased with the mix of products and also the mix of the type of licenses that are in here. Last year, during a bit of a semiconductor downturn, we did see a shift in the type of licenses that we've done. And that was talked about, and we explained that at our Analyst Day in 2009. And we talked at our Analyst Day a few couple of months ago about how that situation was changing, and that's reflected in in this licensing that's happened. A reminder of our new products for 2010 that's, aligned with the 3 different ProPro as the applications profile, the real time profile and the microcontroller profile in our architecture. So we have a third lead licensee signed up for Eagle. And as I say, we'll launch that product fairly soon, Heron, which is the new real time processor. We've lead licensees signed up there. And, the new one at the microcontroller end, we've already launched that product. It's called CorteX M4, and it brings DSP capability to microcontrollers. And the big announcement there in the quarter was with with Freescale, where they've announced the family of microcontroller is based on core XM4. And as a reminder, of how the 29,000,000,000 unit opportunity that we see in 2014 is broken down, categories targeting about 29,000,000,000 unit opportunity. So now for licensing, Just a reminder of the theory of how licensing leads on to royalty growth chart at the bottom left hand side of the slide. This is our installed base of licenses. You'll see in the first half of twenty ten, it's ticked up again at about the normal rate 34 licenses in the first half. So we're now over 690 licenses deployed in the field with the potential for generating royalty in future. And as a reminder, chart on the right hand side shows that these licenses yield royalties over a very long period. And our royalty revenue grows, as we get new licensees starting to come to the armed party ship, their 1st ARM based ships, the quarter we're This quarter, we're reporting 5 new shippers, starting to ship on products, or it could be when an existing customer start deploy arm in a new division. And a couple of quarters ago, we had a great example of that with ST. They starting to get a lot closer to shipping arm in their TV products. Arm based chips becoming a standard in new market sectors are another way for driving royalty growth. And certainly, the activity that we're seeing in microcontrollers is the great example of that happening as we speak. And we're also seeing new types of customer adopting ARM Technology working more closely with technology and the announcement from Microsoft last week is a great example of that. So that's the theory of licensing and driving royalty growth. How does that theory play out in practice, for this quarter. I want to see the histogram at the top there. It says a couple of things. It says that our royalty is back to pre downturn levels. It also highlights the fact that arm is not immune to seasonality, and we're seeing the effects of that in the Q1, Q2 transition that you can see on histogram there. Which is why we tend not to look at these things on a quarter by quarter basis. If you look at things on a non quarter by quarter on a sort of yearly basis, you can see our arms penetration coming through and increased royalty, underlying royalties up 54% versus an industry up less than 40% in the corresponding period, and that that's driven by unit shipments, unit shipments, up to 1,000,000,000 units this quarter. That's nearly double where we were a year ago. And it's right across the piece of our but we're seeing particularly strong growth now in the early stages of core tax deployment. So core tax products that is across the whole of the different core tax profiles, AR and M, applications, real time and microcontroller, up tenfold. So cortex is now up to 6% of unit shipments. And, the point to note there is that Port FA products, which we've been saying some time typically, at a higher royalty percentage per chip, those Cortex A products are just beginning to ship in volumes that start to move the dial. So that's encouraging. That said, for the quarter, as a whole, we've seen very, very strong growth in things like Bluetooth microcontrollers and other very low cost devices, which is why the average royalty when you look at that average royalty per chip is down 4.5%. This is actually good news. It's reflecting very strong growth in applications like Bluetooth. And just to note and remind everybody on the chart, there's a little characteristic there on the far right hand side, reminding everyone that these Q2 royalties, we're talking about are underlying royalties ignoring the catch up royalty that we reported during this period. Those royalties are outperforming the industry growth. So there's a continuation of a chart that you've seen before with the ARM growth rate and the industry growth rate overlaid on top of the actual arm royalties. And you can see the gap continues and, in fact, widens in the current quarter. This is a reflection of, arms growth in market share in the sectors in which we're already penetrated and a reflection of increasing penetration. And you can see on the right hand side of the slide, we've broken down in the usual way the key sectors that are driving that, obviously, the absolute numbers are very strong. They're strong because we're talking about gamers on top of markets, which are themselves growing very strongly on an annual basis compared with this time last year, but arm sheet, arm numbers are actually growing more strongly in all cases than the underlying industry growth. And I'll just highlight that is even the case in mobile, where particularly we're getting used more in mobile phones. So the average number of microprocesses per mobile phone now up from 2.4to2.6 cores per phone, and Bluetooth is a big driver there. So that's licensing and royalties, one of the ways in which, of course, we encourage licensing to turn into royalty revenue is providing the environment for our licensees to be successful. And one of the ways in which we do that is by investing in our ecosystem And that's the general principle. We invest in this ecosystem, which creates this massive community of third party developers around the armed processes That means that customers have a fantastic choice of associated technology that they can use when they're deploying ARM based solutions. Also means that all those 3rd party developers are tied up working on the ARM architecture rather than other architecture. So that's why we invest in the ecosystem. And in the quarter, we had a significant announcement at the beginning of June when we kicked off the Lenaro initiative Vinaro is, a separate organization, which, for the time being, is controlled by Honor, but will disappear as a separate organization or disappear from arm, I mean, become a separate organization And that is all about optimizing the Linux kernel for ARM, so and producing optimized tools for Arm in the Linux world so that users can deploy Cortex A class products from multiple silicon suppliers, in an easy and straightforward way. So it's, it's unifying that, that community out there. It's also, as well as matching multiple SAC devices It's working with the multiple Linux distributions that you see there. So Android, Livo, Migo, Ubuntu, WebOS, etcetera. And making life easy for people ecosystem isn't just about Linux. Linux and the Lynaro initiative is obviously the highlight of our quarter in this part of our business. But as a reminder, we're investing right across the software ecosystem. So Once again, a reminder, on Microsoft, we've been doing that for a long time, Adobe Flash widely deploy in mobile computing and smart phones in the internet world. And Adobe Flash 10.1 alarm is now shipping. And just a reminder that it's not all about high end application processes and mobile computing. At the microcontroller end, we're also investing in the software in software to make it easy for microcontroller users to deploy ARM based microcontrollers with our CMC cortex microcontroller software interface standard for cortex M based microcontrollers as well, and that's also been happening. So right across the software ecosystem, some investment from Arm there. And that investment goes on to help create, armpowered products. And here's a reminder of some of the new, onpowered products. At the top of the slide, you can see eBooks, TVs, Most of the internet connected TVs are now based on Arm, all Android announced TVs, are based on Arm, and the majority of the tablets, even some of the net books you can see net book there, based on ARM. By 2014, we expect 1,000,000,000 of these screens in the interconnected internet connected screen area. By that time, we expect at least 75% of them, to be based on on. And sat behind those internet connected screens are the infrastructure real time products with real time processes. So most of hard disk drives are based on arm today, nearly all the baseband modems that you see out there, allowing these things to communicate half the Wi Fi processes are based on ARM. And by 2014, we estimate the total to be about 1,000,000,000 devices with real time processes, and we expect that a great number of these will be based on on. And at the embedded end, this is where microcontrollers are typically deployed. We're seeing huge numbers of design wins today with companies increasingly adopting the ARM architecture for these types of products, from things like smart meters to cars to motor control in washing machines and those sorts of things. It's a huge growth opportunity for Arm. And, and so a reminder of the sort of activity that we're doing in terms of driving design wins, which is what ultimately drives royalty. Now I'm going to change gears and switch to physical IP, It was a great quarter for our physical IP business, in terms of activity with Leading Edge customers. And really influential customers. We sold 3 new platform licenses during the quarter, not all at Leading Edge, our platform that we sold at 100 and 13 nanometers as well. And in fact, these older technologies will live for many, many years because it's very economical for people to build products on some of those older technologies. And those mature nodes are actually quite good for us in terms of ability as well because we have increasing reuse. And so the licensing activity that we do there is inherently more profitable. Early this quarter, though lots of the activity took place last quarter and actually in quarters before, we announced a license with TSMC at advanced nodes at 28 nanometers and also 20 nanometers. And so at the 28 nanometer nodes, we are now working with TSMC in our physical IP division for both low power and high performance process variance. And that really completes the set of key semiconductor foundries for the 2832 nanometer nodes. So now if you're a customer and you want to work at those nodes, you can use on physical IP and you can go and get your devices made with any of the leading foundries in the world. So we're very pleased to see TSMC coming to that party with Arm. We're even more pleased to see them sign up with us for collaboration on 20 nanometer technology as well. And, again, at both a low power and high performance variance. So physical IP, customer activity is in pretty good shape. In terms of perhaps a backward looking indicator, in some ways, our Q2 Royalty in physical IP, underlying royalties are up hugely with the rising market, of course, on the year before if we strip out the catch up. So bit of context on the physical IP development, you've seen a variance on this slide several times before. So Here, we are updating it for the middle of 2010, is basically saying that our leading edge technology is on track At the right hand side of this slide, we're pushing our R and D towards, towards 20 nanometers, 32 nanometer common platform technology is all now available and tape outs are happening. We're actually seeing our very, very first 32 nanometer royalties coming through from the very, very first customers. But in terms of much wider deployment, is out there and ready for designing right now, which will generate royalties in a few years. And as I said on the previous slide, Tier 10 see is now, come to that party at the far right hand side of the slide. So that program is very much on track. Stepping back and looking at how our physical IP business is developing and will develop, It's about growing market share. The histogram on the top of the slide shows that we are little by little growing our market share, this is an analysis of, the number of wafers or 200 millimeter equivalent wafers at any rate, and those wafers which deploy on physical IP. And so you can see that over a period of a few years, that's iterating upwards. We're growing market share. But we're growing market share of a market which is growing hugely. And what we expect over the next 5 years, the, as IDMs around the world seek to outsource their silicon production more, we expect the foundry market to grow substantially. So we're gaining share in a market, which is itself growing substantially. And then when you peel back another layer and look at the histogram at the bottom of the slide and look at where that growth is driven from, you can see superior growth at the leading edge nodes, and it's those leading edge nodes where Arm now has a full house in terms of physical IP technology deployed across the leading foundries. So that's really what's going to drive our physical IP business over the next several years. So let me summarize, it's a great quarter. There's a lot happening and the results reflect progress in all parts of our business. We're very pleased to see influential market leaders coming to on and his 3 significant announcements over the last couple of weeks. We've talked about market with long term structural growth, mobile devices, digital TVs through the connected internet, drives because the internet means lots of data and people love storing the data, and microcontrollers as intelligence gets deployed more and more wild, widely, in the products around us in our everyday lives. So aren't gaining share, again, in all of those areas. At our Analyst Day, I said, we have the best business model technology and ecosystem for now. We absolutely do. These results are showing that showing us outperforming the industry with our royalty up over 50% versus an industry growing in the 30s, and that's driving superior profitability, cash generation so that we can increase the dividends. And at the same time, invest in R&D to drive that ecosystem and our product portfolio for the future. So with that, I'll hand over to Tim. Morning? So the good news is well, very good news. I'm not going to say anything about that, but the big values, I've got three slides. So let's get to it in quite quickly because you've got a lot of information in your packs and you're dying to get on to the questions. And we'll see what they will relate to. So as Warren said, and Warren has covered quite a lot of the detail here. So I'm going to go through it pretty quickly. I mean, what a good strong licensing quarter, not just for revenue, I mean, 36,600,000 processor was a good recovery from the lower levels we saw in 2009. Probably, more importantly, from our standpoint, looking forward, is that there are several licenses that occurred that build backlog in the quarter. The Microsoft deal, which we've touched on, Warren referred to a handful of other A class processor licenses, which help backlog as well as short term revenue and the Eagle license, which predominantly is a backlog builder. And what that's meant is that we're more than 20% sequentially and about 50% higher than we were a year ago, and it will come on to to what that means for the outlook a little bit later. I think we touched on royalty, You may note, that's the first time that we've reported catch up royalty in the Processor division since Q1 2006. So this is a rare event. It doesn't in any way undermine my confidence in the sort of the general ability of our licenses to pay us royalty on time. I think this generally is a one off event where predominantly one particular part, which shipped for multiple quarters. In fact, about 3 years got missed off, and was identified in preparation for one of our normal audits. So I wouldn't expect to see anything like this very often at all, but it obviously did affect the numbers. So we tried to make very clear that that was a effect be a one off. On the cost side, Q2 OpEx a shade over 52 Consensus was about 513. I mean, about half of that difference is a translation issue consensus. Was about 152. The actual cost rate for Q2 was about 149. That makes up about half the difference. The balance obviously, we're talking about small numbers in and out, but the incentive payments are running a little bit higher, as the revenue and booking trajectories are running at an elevated level. So there's a little bit more sort of bonus type accrual in there as well. I think 4 For next quarter, we say in the release, assuming the FX is around similar level, somewhere in this sort of 53 to 55 range. 4 OpEx. And I think if the currency stayed at where it is right now at around 154, 155, it'd probably be at the lower end of that. Low end of that range. So what, of course, this has been yielding, is marginal expansion, but not at the cost of lack of investment. I mean, obviously, we've been managing, as Warren said, cost base pretty carefully during the downturn. Headcount has been relatively flat. We said we're going to grow it this year. That's what we're doing. It's up 65 at the half year, and it'll be up again at the end of the year. By how much we will see. I mean, we, we obviously need to get the right people into the organization and the pace of recruiting them will to some degree relate to what's going on outside of the big, wide world. But this margin expansion, 39% in Q2, if you strip up the catch up royalty, it's still much higher than we've seen prior to coming into this year. This has been throwing off a lot of cash. I mean, the net cash is about three times the level in the first half than it was last year. And having increased the dividend by 10% for the last 2 years sort of through the downturn, we're now increasing the interim by 20%, which is more a reflection of how we see earnings growth developing over multiple periods. Rather than obviously what's going to happen in the balance year of 2010, where current consensus has our earnings growing in north of 50%. This is a slide that usually sits in the back, but I thought I'd just bring it forward. Mean, what we and I think, yeah, when you're thinking about licensing going forward, looking at that maturity profile is quite interesting. I mean, the revenue to be recognized over the next couple of quarters is a little bit lower than when you saw this last time. But it's actually only much higher number. So the absolute coverage of target license revenue already in backlog is higher. And when we look out into the second half of the year and look at the the mix of licenses that sit in the opportunity pipeline, that gives us quite a high degree of confidence that for 2011 the coverage of target license revenue that will sit in backlog at the end of this year will be higher than we normally enjoy. Which I think is an encouraging sign. I mean, we're obviously way incent way out, but it's encouraging sign as we look forward into 2011. So again, just wrap up before questions. We're continuing to execute the strategy. Clearly, broadly, visible in royalties and visible in licensing agreements, and market share gains in all of our target markets. Think we're very well positioned for the 2nd half. We have record order backlog, which is good for short term rebel revenue. It's also good for long term recurring revenue. The opportunity pipeline, as I say, is pretty robust and it's a good mix of deals. Which are going to yield revenue in the second half and yield backlog for 2011. And we see no reason why the March share gains that we've been enjoying in royalties do not continue. We say in the outlook, as we have said, in recent outlooks, that we still see, uncertainty in a macro environment, and I think it would be remiss not to remind ourselves of, I guess, at the very best, you can say currently mixed messages that are out there. But who knows how the austerity packages are going to play into consumer demand and the global economy over the next few quarters? I think, yeah, we just flagged that. Now, whether that has an impact on ARM in 2010, early in 2011, not at all, we will see. But it would seem to us to be careless to assume we could be immune from what may be coming down the road in that regard. And I said on the some of you picked up, I said on the wireless call this morning that it is possible that this could manifest itself in a lower seasonal uptick going into Q4. But that is given how much guidance we've seen for Q3 from our shippers, that's largely a theoretical comment I mean, there are 1 or 2 who have guided already sub seasonal. But whether that would be enough to impact, I mean, normally, you'll note that arms royalties are often up somewhere between $5,000,000 $10,000,000 Q3 into Q4. That might happen. It might not. I mean, as you know, there's not a lot we can do about it other than report it, but we just giving the notion that it seems to me that that could be a possibility, maybe it's more likely to impact really next year along the end of this, given that we're reporting arrears, but we will see. I think in terms of licensing, almost irrespective of what happens in the macro environment, I don't really see any material impact on the licensing behavior between now and the end of the year, frankly, given the backlog, given the pipeline. So I think the licensing outlook is pretty encouraging. And lastly, I guess some of you will recognize the tone, if not the wording of that guidance. And the slight complexity this time is we wanted to stress that we are confirming current consensus effectively plus 9, not current consensus, including 9. If it was including 9, some of you would assume that we were downscaling our guidance, which we're not. Whether we will, as we have done in recent multiple quarters, and do better than that, only time will tell. But that's what we are confirming today. Good morning gents. Couple of questions. First of all, looking at the licensing, the de licensing outlook, A couple of questions on that. First of all, have you started recognizing have you started recognizing revenues from the Microsoft license in Q2 or that's coming up in the second half. And second, we've seen over the last 3 years, number of it's a non traditional semiconductor companies or nontraditional customers outside of the semiconductor world licensing. Microsoft and other company I have in mind. Is there any reason why other software companies that may have made semiconductor investments in the recent months may also be interested in taking architecture, architectural license And the second question is on Mali. Can you just frame a little bit Mali opportunity in terms of royalty rate with regards to GPU and the video processing, IP, as it is combined with a CPU, you say, how much royalty percentage would you get if you get the 3 IP blocks? The first question is there's no revenue in Q2 from Microsoft. We would expect to start recognizing that revenue, which will be on a subscription accounting basis. Over the and about, the known traditional companies who have, licensed on, whilst we cell licenses, primarily to Semiconductor Companies, we work with at a technical level a number of other companies, either in the ecosystem companies like Microsoft that are providing complimentary technology to somebody who's building a product or with somebody who's like an OEM who's actually building a product. And in some instances, those companies will feel that they can do something a little bit special, if they have a little bit more control, over the implementation of the microprocessor. And so under those circumstances, we enter into discussion about how they could work more closely with ARM. And in a very few instances, those discussions head off in the direction of an architecture license, and that has happened in this case. It's by no means a normal thing and an architecture license, if you're going to deploy it effectively, you still need access to a very good microprocessor implementation team, and we'd like to think that we have some very good micro our implementation teams. And so the chances are, that we can do something very, very good for you as well. But we stick to our roadmap and other people sometimes want to do something a little bit different. And our goal is to enable all the players. And that's why we have flexible licensing models, and we're prepared to enter into those sorts of agreements. But we're very, very strict about maintaining compatibility of the ARM architecture. So you write a bit of software for an ARM 70, the MI in 1996. And it'll run on a Court XAA. It'll run on a Court X A9. It'll run on an architecture license product from any one of our architectural licensees out in the future. Moi, royalties are broadly similar to general purpose microprocessor royalties. We pull Mali out. It's a different type of processor. The video process has a different type of processor again, and it makes engineering sense to have a different type of processor to implement those functions. From a commercial point of view, it looks exactly like an ordinary on microprocessor. And if you have multiple alarm microprocesses or a single piece of silicon, then we enjoy multiple royalties. So I think we're going to move on. Couple of questions. The first question I have is, I mean, again, on architecture licenses. Is there any correlation between architecture licensees and the volumes issue because historically, you have signed a lot of architectural I mean, not a lot, but, I mean, a handful of architectural licenses and they really haven't really contributed most of those companies who have signed architectural licenses haven't very much contributed to your royalty units. I mean mostly your own processes have done better than the architectural licensees. That's the first point. The second question I have is on the PRPD business. I mean, PRPD clearly, I mean, on the licensing front, you seem to be showing some progress there, but on the a royalty front, I mean, if you look at TSMC's revenue in the first quarter, it actually was flattish to slightly up versus the fourth quarter, whereas your royalty revenues were down by a few percentage points or more in a PIPD. So what is that reflecting on the PIPD royalty units? I mean, what is exactly given that you're talking about the market share gains. And finally, I mean, there has been lots of speculation, I mean, which you have addressed before about acquisition of ARM, etcetera, I mean, what protection will armed licensees have in the event of an acquisition of Arm given that they are basing their own roadmap on the back of, so if, for instance, Arm was not able to provide them with IP in the future, what happens to those licensees? Okay. Let's, let's deal with that. So architecture licensees end of their contribution to ours royalty. Is no correlation between our architecture licensees and high volume. The architectural license, as I say, is for somebody who wants the microprocessor that believes that in their implementation, they can, in some way, do something different, I say different, not necessarily better, different, from the standard arm implementation. We target our implementations at a wide range of Semiconductor Companies and their requirements. Some people believe they want something that was a little bit more special. That's what the architectural license is for. Whether or not they derived significantly more volume from that is a question you have to talk to them about the economics of their business. We earn the same royalty for the chip, whether it's an architecture licensee or whether it's an implementation licensee, and we're interested in enabling our partners to do what they want with the arm architecture. Our architecture licensing. Architect license is typically taking implementation licenses as well. So they often take an architecture license as an option for a product in a particular market, but alongside that, they're also licensing our own implementations. So again, it's but in terms of the ultimate royalties, it's I wouldn't say there's a correlation that we're saying. So PRPD Royalty was your other question. And for those of you in the room, perhaps slightly repeat the question. The point was that, if you look at TSMC revenues on a sequential basis, they increased, and we talked about gaining market share and actually our physical IP underlying royalty didn't increase as much as TSMC's royalty. And this is simply because we have a market share at the moment as put on the slide, it's about 20 percent of the number of wafers. Most of the wafers are generated by the big shippers the Qualcomm's of this world, who deploy their own physical IP at the moment. And TSMC enjoyed a large growth in the period on the basis of those types of customers, where armed physical IP is currently not deployed. So I think it's an artifact of that arithmetic and nothing more. In the sort of non Qualcomm type portion of foundry business, as shown in the histograms, we're increasing our market share. I didn't share within the telecoms and Broadcoms, which historically haven't been on the farm license. Well, as and when, we have an announcement to make with a company like Qualcomm, and I'm sure we'll seek to make it. The other question was a bit more academic, really, about acquisition. We're a public company. And, consequently, if somebody wants to buy arm enough, then that's what they'll do. We believe that, it's in the best interest of our customers and shareholders, for our, to be independent. We are explaining to you how were exposed to areas of big structural growth and, on prospect as an independent company are very, very positive. Do customers have, as Tim's just reminding question is do customers have protection? Customers, particularly in the light of recent rumors, ask us the question. And in you have an ARM license. Typically, it's a perpetual license. As long as you continue to pay royalties, you can continue to do designs with those products. Future roadmaps, we're working on future roadmaps. Those customers that have architecture licenses can can do what they want with them. So there is an element of built in protection anyway. Nick? We're working from front to back down outside of the room. Yeah. Two questions. Could you, the PD catch up wonder if you could expand a bit on how that came about. $9,000,000 seems quite a big miss or misunderstanding with one of your customers over a number of quarters. I'm wondering having had that misunderstanding whether you feel there might be scope or a few other misunderstandings among some of the other customers, on royalties. That's the first question. Well, as I said in the well, I was on the podium there. That's the first material. That's the first process of royalty catch up. We've had to port since Q1 'six. And there wasn't any before that either. So effectively it's happened twice. In this particular case, it was just one very high volume part that was, for whatever reason, systematic didn't find its way into the royalty reporting. Yeah. In a sense, that can happen. I mean, these customers are building multiple on these devices. And, in this case, it was missed. And it's very, very unusual. So I don't as I said, I don't think you said anything structural, but we continue to have our ongoing cyclical royalty audit program that we use all the big 4, and they're basically circling around the world all the time doing these order Anya, this 9,000,000 was identified by the customer, not the auditors. So it's not prompting you to have an extra review. It was identified by the customer in advance of an audit, which is now taking place. Okay. That's right. The second question, unrelated. You talked about, yourselves being happy with consensus for 2010 it currently is, perhaps you can let us know what you think consensus is for 2010? Before today or today, it's 584. And what we're saying in here is, we assume the 9 will be added. And then you will do what you do. When you look at the result against what you had in for Q2 and what we're saying about Q3 and Q4, So I, you know, I wouldn't be at all surprised if it doesn't do what it's done in recent quarters, which is edge up. So if you had a follow-up question about PIPD, a lot of the volume obviously at some point will cover about 3x, no, that's your customers. But what how should investors measure the portion that you actually address? It's very difficult for us to see as to how many programs would be done on libraries that the foundries are using themselves and how many how much of their wafer output is based on yours. So how should we how should we get a better understanding of that? Well, we're going to have to work on that. From a sort of presentation point of view. I agree it's difficult to separate out that, which is addressable market and that, which is not really practically addressable. I mean, ultimately, we think it is very much all addressable. And so we don't really see any structural barriers to that 20% continuing to increase. And the picture that we were trying to paint was a lot of the 80% at the moment is Cat 1,000,000 house. A lot of it is historic, process nodes we are signing platforms at some of the older, more mature nodes, but where we're concentrating our efforts as indeed, we have been talking about for the last several years, we've been concentrating our efforts at the leading edge so that when the market eventually migrates to that leading edge, we have it completely covered and a much larger proportion becomes addressable. But now I appreciate that it's difficult to separate out those 32 chunks. And you could make some approximations, how much is the top 5, 10 fabless players who typically use in house, and you could make a first approximation excluding those and then look at our share in the remainder and look at our share tracks in the remainder. As I say, long term, I'm not for separating them. I see it all as opportunity. My second question is just on the balance sheet, just I asked this every quarter, the last couple of quarters, net cash up 3x. So as a percentage of your loan capitalization is sort of the same situation that we had in 2000 and 1, and eventually, of course, it led to a rather large acquisition. So I was wondering how you felt about the balance sheet today, how much room there really is to redeploy more. I appreciate you've taken dividend already, but, what are the other avenues to speak? Well, I mean, Yes. Your memory of the 2 1001 banks, it's probably better than mine. But I mean, right now, we have 200,000,000 net cash and, yeah, You guys know what the market cap is. I mean, in that sense, it's not sort of an egregious amount of cash burning and holding our pocket mean, having said that, we have a history of doing bolt on acquisitions. I mean, I accept that we did a major one at the end of 2004, But that the size of that, I think, showed you that it was nothing to do. There was no sort of in advance building of a cash balance, so we could do it mean, it was something that we felt strategically right for the long term. So the plan of record is in progressive dividend policy. You've seen us do other things we've returned 1,000,000 of cash via buyback between 1,000,000 and 1,000,000. The buyback remains available to us. So yeah, I think over time, it's highly likely that there are going to be other capital structure management. Issues to be communicated to you in addition to the dividend. But we don't see a pressing need to materially reduce Thanks, Catherine, on TBS. A couple if I could. Just wondering, you mentioned last quarter about integration being effectively accretive to business model. I just wondered whether you're seeing the first signs of that sort of accretion coming through. Secondly, is it just on Eagle? Announcements of 3 lead licensees. I just wondered in which areas those licensees are looking to roll. Is it smartphones, mobile computing, which are the specific areas. And then finally, it's a kind of more holistic one just on the chart that you put up about PD royalties against the industry growth. You've seen that spread widen. So your rate of pace of market share increasing against the industry. And I just wonder how you see that playing out over the next 12 to 18 months do you see your market share gains accelerating from here given all the licensing activity we've seen or just keeping pace, basically? Martijn, three very different questions. I think by the integration and the accretive nature of integration, you're referring to multiple arm processes on the chip eventually leading to a more royal from that chip than when there are 2 chips. As per our Analyst Day presentation, We're showing you the mechanics of how that works. I think it is too complex to to read much of that happening into these numbers right now. It's an interesting question that I'm sorry, I'm just not prepared with a canned answer for other than what we showed you at the Analyst Day. Maybe we can have a think about trying to look for some examples of that and try to draw out how it is contributing. I'm looking at Ian at time. As I've seen that, and we have a thumbs up from Ian in the back of the room. So we'll have a look at that for you before the next time. Where is Eagle going to be used? Well, Eagle is a high performance A class processor, There's no reason why Eagle can't be used in a mobile computer or some other computing platform. Probably the first deployments, though, will be in high end smartphones. The first deployments will not be pushing the top end frequency of the Eagle, they will be aiming very much at the power envelope of smartphones. And that's where you'll see. So it's more capability for performance, but fundamentally, it's more efficiency. And that's, the first ones. And then, PD royalties versus the industry, while we're painting a picture of exposure to for growth markets, things like digital TVs, things like microcontrollers, which in and of themselves, are sectors which are set for superior growth to the bulk of the market over the next 18 months to 3, 4 years. And you can see within those sectors, I'm very heavily exposed. And so right now, I would say you might see a modest acceleration. Thanks. I just had two questions. First was on gross margins, which seemed to have been coming out a little ahead of expectations. If you could give us some guidance, as to where those are going forward? And secondly, just again on Marley, if you could just give us an update on the kind of run rate annual volumes for Marley today, And what are the material segments where volumes are coming from at this time? I think at the last time I stood up here, I gave a slide that sort of broke down the component parts of gross margin. The context of painting a picture that is going to increase over time. And really, there are a few variables, but the by far, the overriding one is what is the proportion of royalty revenue as a total of the whole, because royalty revenue is 100% margin business, it's essentially that that's driving up the gross margin. But I think when you're modeling out forward, then something around these current levels is the way to think about it. I mean, yeah, the headline number this quarter will be somewhat flattered by the catch up. So it's not quite as elevated, but it's certainly operating. And I think I said in a fairly medium mouth way 6 months ago that it's more likely to be north of 93 in South going forward. And I think that's still a good way to think about it. On the Moly in volume shipments, right now, Moly is not really moving the dial. In terms of the royalties that you see on a quarterly basis. It's been very much a story of licensing and design ins. Typically, these ships are 2 to 4 years from licensing to royalty. We kicked off with Mali at the back end of 2000 and 6. Therefore, we expect to see royalties starting to happen around about now. And we expect that First, I mean, there are shipments happening in smartphones, and that's where you're going to see the shipments predominantly. There's a little bit of design in also in digital TVs, and you'll see some more of that. And ST talked about that at the back end of last year, I think if you look at apps processes in smartphones and apps processes in digital televisions, that's the target market. And we see our share in that growing as further licensing that's been done over the last few years, there will be a significant portion of those that will switch to Marni. And then we had a question at the front. Sorry, yes. I seem to remember I may recollect incorrectly. The delay year also ago, you gave 1,000,000 unit run rate for Marley as the numbers being shipped I don't know if that's probably right. Did we? 15,000,000. No. It wouldn't have been for Mali. Mean, obviously, we have some legacy shipments of In imagination? But Marley, we only acquired Marley, right at the back end of those. Yes, I think maybe a couple of years ago, we would have been talking about graphics' processes and probably smartphone volumes at that time, were about 50 million units, and that's probably where the number was coming from. That remains the target market that and the apps processes in digital TVs, as we look forward, that 1,000,000 has grown apps processes in smartphones to several 100 of 1,000,000 today on a per annum basis, and Marley will take a share over the coming years. Again, it's an interesting question that we should probably, as and when Mali starts shipping, then we should look to a little more disclosure on that. But right now, it's so immaterial to our royalties that complicates things to talk about it. Okay. Thanks. We do have a question at the front, sometime. Morning. The questions about the Linaro related charges. You said that about 4,500,000 investment over the 1st 12 months roughly the current run rate from Q2. Two questions. 1, is that essentially just covering the running costs of the of the operation as it set up, and B, what happens after that? Are you expecting the same sort of rate of investment, or do you inject some capital and let it loose? I mean, in the so in these Q2 results, you see 1,000,000 costs incurred. And most of that is actually sort of one off launch costs of getting this entity into play. And some of it is the very initial because this was launched on 3rd June, very initial sort of running cost of the entity, which as Warren said, is actually currently on a temporary basis owned by arm. And because of that, it's sort of pulled out separately. In terms of once that organization is settled down, what arms ongoing contribution to it will be will depend on, the number of contributing participants. When that settles down. What we're disclosing in here is what we expect the cost to be in the 1st 12 months from June to June. But I think probably back end of this year very early next, I'll be able to update you more that all probability, we would have lost control of the entities that are no longer be a subsidiary and B, what the ongoing, armed commitment is likely to be. Ahead. So I had a question on the mobile royalty business and the unit assumptions you're using If we just look at the quarter, you said 1,400,000,000 cores and 62% were mobile. And then the course performed were 2.6. So that backs out at about 335,000,000 handsets is your assumption for Q1. For Q1 shipments globally for handsets. My question is, how do you account for gray market players? So we talked to MediaTek and there are publicly saying they're going to ship 500,000,000 units this year, spread from MStar. All these players that are your licensees, are they included in your unit assumptions? Maybe you can just clarify Well, Ian behind you is going to tell you how he came up with a number. So I think the formula I use is I take what's reported by customers and then I use the, traditionally use the Gartner mobile phone numbers and have done for a number of years. And so as those evolve to include to better include the China numbers, then maybe we'll get a better better representation, but at the moment, I just want to be using a standard sort of numbers and be consistent. So that's the numbers that I'm using. Yes. It's Kai Kosch with Deutsche Bank. Just a quick question on OpEx. I think your Q3 guidance is pretty clear. But on Q4, if we assume that consensus revenue estimates on that should we think about a sort of sequential increase again in Q4 in OpEx or a stablish run rate? I mean, I think it's I think I mentioned, when we did the last results, it's kind of likely to just gradually edge up, which is kind of what the consensus, I mean, coming into these results, consensus was 53.3 for Q3 and 54.3 for Q4. And that feels fairly sensible. I mean, we're in sort of net investment in people mode, and it's sort of gradual And so I think that's probably a good shape of the OpEx as I would guide you to at the moment. Good morning, Andrew Gardiner from Barclays. Just a couple of clarifying questions. First, on the auditing of the royalties, Can you tell us how often you're auditing the customers? Some of these seem to be extending a back of fairway. You mentioned sort of 2000 and 7 to 2010 is the as the overall period. So I was just wondering whether this had been missed in one of the prior audits or not. And also in terms of the back Clogue. You'd announced the Microsoft deal last night, but you seem to be indicating sorry, last week, but you seem to be indicating that that's in the 2Q backlog, is a TSMC deal also in the 2Q backlog, or is that going to drive 3Q up yet again? Every 3 years, broadly, and no, and yes. Hi. It's Jonathan Mennan from Liberum Capital. There was an announcement, I think it was yesterday or today morning from an automobile manufacturer consortium. I think it's called LVI or something like that, not to show up, it consists of all the major automotive companies and the component suppliers. And they basically announced that, they have chosen Migo as their operating system of choice. For all in card telematics, which includes GPS, entertainment, etcetera, etcetera. And you guys have invested in Lynaro to try and push on implementation on Linux versions. My question is, looking forward, when you talk about these kinds of devices, which is non mainstream mobile phone and also you're going into automotive, television, etcetera, is the choice of microprocessor going to be more based on the choice of operating system, which is is the operating system taking preference over the microprocessor or will the microphone be chosen on its own merits? And I completely grant that you will see Migo implementations on on course. I don't dispute that for a minute, but generally one would assume that the majority of Migo implementations will be on Intel Processors. Rather than on processes. The second question is just on PIP, just a clarification. Market share has been running at about 18, 20, as your chart suggested. But I thought I saw something saying that I think only 15% of your royalties is sub-ninety nanometer or something like that. So now that you've sort of and you've got a lot of leading edge process nodes going. Does that automatically lead to a rising market share in PIPD? Or are we back to waiting for Qualcomm and those types to sign up? I mean, are they both is that a mutually exclusive decision? Or does your existing customer base itself give you a big leg up on the back of these leading edge processes? Okay. First question, Migo and the Automotive announcement. Migo happens to be a Linux stack in which Intel has, put their weight behind all the first implementations of MIGA will be ARM based, I suspect, in volume terms, the highest volumes of our products running Migo, will be on for quite some years to come. And so, I think your question was about sort of microprocess as an operating system. If you look to the user, the user doesn't care what microprocess microprocess is in there. User, the operating system is very much about that. Now we don't know whether Migo is going to win in the Linux world or whether it's going to be Android or LEMO or Ubuntu or any of them, what we do know we want to do is enable most of those to be onpowered, so that the people who are building the products, who are the ones that do care about the microprocessor, have the simplest least resistance, lowest cost route to developing their products around the ARM market processor. And that's what Menaro is all about. It's about optimizing the Linux kernel and the tools for the range of arm based system ownership devices so that those people building the products are not limited in any way when they choose 1 of the several Linux operating system distributions. And some of those people will choose MIGO, and we hope they'll be very successful with their MIGO based products running on microprocesses. For PIPD Royalty Growth, then it's true that the volumes from the leading edge products at the moment are very small. We've concentrated on the leading agent. So certainly, we're opening up the potential for significant growth. And what I've tried to show in the presentation was the enhanced growth in the foundry space coming from those leading edge processes is how Arm is exposed to those high growth areas within the foundry business for physical IP. In terms of design wins with companies like Qualcomm and like Broadcom, they can only accelerate, be, we're working with them to, to try to bring that day forward. Think we're running out of questions. So let's just have the last one or maybe 2, if there are any. Excellent. Well, thank you very much, everybody, and we'll see you again in early February. Well, I'm sure we'll see many of you at all. So thank you very much.