Greetings, and welcome to the Advanced Micro Devices Second Quarter 2016 Earnings Conference Call. [Operator Instructions] As a reminder, this conference is being recorded. I would now like to turn the conference over to your host, Ruth Cotter, Chief Human Resources Officer and Senior Vice President of Corporate Communications and Investor Relations. Thank you. You may begin.

Thank you, and welcome to AMD's second quarter conference call. By now, you should have had the opportunity to review a copy of our earnings release and the CFO commentary and slides. If you've not reviewed these documents, they can be found on AMD's website at ir.amd.com. Participants on today's conference call are Lisa Su, our President and Chief Executive Officer; and Devinder Kumar, our Senior Vice President, Chief Financial Officer and Treasurer. This is a live call and will be replayed via webcast on AMD.com. I'd like to take the opportunity to highlight a few dates for you. Lisa Su will present at the Pacific Crest Global Technology Leadership forum on the 9th of August in Colorado; Raja Koduri, Senior Vice President and Chief Architect, Radeon Technologies Group, will attend the Jefferies Semiconductor, Hardware & Communications Infrastructure Summit in Chicago on August 30; Mark Papermaster, Senior Vice President and Chief Technology Officer, will present at Deutsche Bank Technology Conference on September 13 in Las Vegas; and our third quarter quiet time will begin at the close of business on Friday, September 16, 2016. Before we begin, let me remind everyone that the second quarter 2016 was a 13-week-quarter, and we expect to record our extra week in the fourth quarter of 2016. Today's discussion contains forward-looking statements based on the environment as we currently see it. Those statements are based on current beliefs, assumptions and expectations, speak only as of the current date and, as such, involve risks and uncertainties that could cause actual results to differ materially from our current expectations. Additionally, please note, we will be referring to non-GAAP figures during this call, except for revenue, which is on a GAAP basis. The non-GAAP financial measures referenced are reconciled to their most directly comparable GAAP financial measure in the press release and CFO commentary that we've posted on our website at quarterlyearnings.amd.com. Please refer to the cautionary statements in today's earnings press release and CFO commentary for more information. You'll also find detailed discussions about our risk factors in our filings with the SEC and, in particular, AMD's quarterly report on Form 10-Q for the quarter ended March 26, 2016. Now with that, I'd like to hand the call over to Lisa. Lisa?

Thank you, Ruth, and good afternoon to all those listening in today. Our ongoing focus on diversifying our business model and delivering great products is creating solid market and financial momentum. Looking at the second quarter specifically, strong semi-custom demand and better-than-seasonal graphic sales drove a 23% sequential revenue increase and our return to non-GAAP operating profitability. We also recorded our first full quarter licensing gain from our China server JV with THATIC, and improved our balance sheet with cash proceeds from the formation of our ATMP joint venture with NFME. Looking at our Computing and Graphics segment, revenue decreased 5% sequentially, as stronger than seasonal GPU sales were offset by a decline in desktop processor sales. First half 2016 GPU sales increased by a double-digit percentage from a year ago as the investments we have been making to develop leadership graphics hardware, software and drivers, combined with new marketing campaigns, are paying off. Our strong second quarter graphics performance was capped by the launch of our new Polaris-based RX 480 GPUs at the end of June, which helped contribute to our highest desktop channel GPU shipments since the fourth quarter of 2014. The Radeon RX family resets expectations around the experiences and features gamers now want in a mainstream GPU. We are pleased with the Polaris launch, initial channel sales and OEM design wins. We expect this strong demand to continue and help drive revenue growth in the third quarter with the launch of the RX 470 and RX 460 desktop GPUs and Polaris-based notebooks from our OEM customers. We also delivered our third straight quarter of sequential professional graphics revenue growth and believe we gained share driven by increased adoption of FirePro graphics by OEMs as well as several cloud data center GPU compute wins. In Client, mobile…

Thank you, Lisa, and good afternoon, everyone. The second quarter was a strong quarter punctuated by better-than-expected financial performance, driven by demand for our semi-custom SoCs and the closure of our ATMP joint venture transaction with Nantong Fujitsu Microelectronics, which bolstered our cash position. Second quarter revenue was $1.027 billion, up 23% sequentially, driven by higher sales of semi-custom SoCs. The year-over-year revenue increase was 9%, with higher sales in both reportable segments. Gross margin was 31%, down 1 percentage point from the prior quarter, primarily due to a higher mix of semi-custom SoC sales. Operating expenses were $342 million, up $10 million from the prior quarter, primarily due to increased marketing investments. We also recognized a $26 million licensing gain associated with our JV with THATIC and restructuring credits of $7 million, primarily related to facilities. We are pleased to report operating income of $3 million this quarter. Net loss was $40 million, with loss per share of $0.05 calculated using 794 million basic shares in the quarter. We recorded a pretax gain of $150 million related to the ATMP JV transaction, an equity loss of $3 million based on our 15% ownership stake and taxes of $27 million related to the JV transaction. Adjusted EBITDA was a positive $36 million compared to negative $22 million in the prior quarter. Now turning to the business segments. Computing and Graphics revenue was $435 million, down 5% from the prior quarter, primarily due to decreased sales of client desktop processors and chipsets. Revenue was up 15% year-over-year, largely driven by higher client load per processor and graphics sales. Computing and Graphics segment operating loss was $81 million compared to $70 million the prior quarter, primarily due to lower revenue. Enterprise, Embedded and Semi-Custom revenue was $592 million, up 59% from the…

Thank you, Devinder. Operator, if you could poll the audience please for our question-and-answer session.

[Operator Instructions] Our first question today is coming from the line of David Wong from Wells Fargo.

My first question, can you give us any detail on the ramp, in dollar terms, of revenues from new semi-custom products in the September and December quarters? And what the seasonal pattern of these products might be?

Sure, David. This is Lisa. Let me start with that. So our Semi-Custom business overall, we see as strong for the year. I think the predominant business is the current generation game consoles and as they ramp into the second half and the stronger holiday season. We expect the seasonality of semi-custom to be similar to prior years, where we'll see a peak in the third quarter and a decline in the fourth quarter. And we will see the new business layer in starting in the third quarter, but I don't expect that to change the seasonality pattern. So we're building up for a strong holiday season overall.

Okay, great. And since the launch of the Radeon 480 graphics product end of last month, have you been able to supply the demand for the cards? Or if not, when do you expect demand will rise to match supply?

Yes. So, David, we're very pleased with the launch of the Radeon RX 480. We had good supply at major retailers on launch day. Since then, the demand has continued to be strong, and so some of the retailers are out of supply. We do see that the 14LPP yields are good, and we're ramping up production steeply. So we expect that, that will equalize as we go through the quarter. We're also, very soon, going to launch the rest of the Radeon RX family and so you'll see 3 products in the third quarter in terms of overall product momentum.

Our next question today is coming from Mark Lipacis from Jefferies.

First one, Lisa, maybe for you. The China licensing deal was -- I thought it was creative and fascinating. Can you characterize kind of the pipeline of IP deals that you have or technology deals? Do you -- Is this it? Or do you have more in the pipe? Can you talk about timing or the types of deals you might be working on?

Sure, Mark. So the China JV, as you said, we announced it last quarter and it's now in operation. It's actually going well, so we're on track for some of the important milestone deliveries that we had. In terms of the overall pipeline, I would say, we have a nice set of interesting opportunities. They're very much -- as you know, they take a while for them to come to fruition, but we feel good about the model, which is partnering with folks that need high-performance technology, and we'll continue to work those deals as we go through the next couple of quarters.

Okay, great. And a follow-up, if I may. I was hoping that you could just help me with the accounting of the semi-custom design wins. So if I remember properly, there's 2 that you expect in the back half of this year and then 1 I thought it was the first half of next year. So Scorpio is 1 of those 3, but that's next year. So that's the 1 for next year?

Yes. Let me help you out with the accounting. I know that we've had several different pieces of information on the semi-custom new design wins. So what we said is we'll have a total of 3 semi-custom new design wins that will account for about, let's call it, $1.5 billion of revenue approximately over the next 3 to 4 years. We're starting the ramp of new business this coming quarter, this third quarter, and that will be one of the semi-custom design wins. Scorpio is also a design win, and that, as our customer said, will be in 2017. And we'll give you more information about the third one as we have more visibility.

Do you have more design wins that you just haven't announced or mentioned timing of?

Well, we're trying to get out a little bit of the counting game, but I think, overall, we're pleased with the semi-custom pipeline. Yes, I think some of the questions that we've been asked are do we believe we have design wins outside of game consoles? And the answer is yes, we have design wins outside of game consoles. I think we view the pipeline as good, and it's a business model that works well with our high-performance technology slant and our SoC capabilities.

Our next question today is coming from Matt Ramsay from Canaccord Genuity.

Lisa, I have a couple of questions on Zen. You gave some color in your prepared remarks about server opportunities for Zen starting in the first half of next year. Maybe you could give us a little color on potentially what markets you're going after there, where you see the opportunities? It seems to me that a lot of different moving parts in the cloud market with open power and ARM and then custom CPUs from Intel, but a pretty big enterprise space that's x86 captive for you guys to go after. So I'd just be interested to see what the strategy is going forward there and where the design wins might come from.

Sure, Matt. So, yes, let me give you a little bit of color of how we see it. First of all, look, that is a brand-new CPU design and we had very ambitious goals for it. And given where we are today, it's actually coming up very well. We've demoed both desktop systems and now we have server systems working in both our labs and our customers' labs. I would say, as the progress of Zen is -- goes along in terms of validation, the customer interest has increased, and so we did close a number of design wins in the second quarter and we have a number more in the pipeline as we go into the third quarter in more widespread availability. Our view is Zen is a general-purpose server architecture that can play in many different places. So you mentioned the cloud, I think that's certainly one target market given the growth there. We also see there are workloads that would be particularly -- fit well for sort of the performance areas that we're targeting. I think Enterprise is also, again, very much a target area for Zen. What we're looking for, again, is ensuring that we have a very high-quality launch, and so we're working very closely with customers to make sure that we run their workloads and demonstrate the performance improvements that we expect. But as I said, overall, it's going well, and we continue to work with customers to ensure that they see the benefits of Zen and working with AMD.

It's good to hear. I guess another question on Zen, more in the PC market because I think your prepared remarks focused a bit more on server, but maybe you could give us a little bit of an update in the timing of desktop and notebook potential launches? It just seems to me going into the holiday season, it's still a little unclear as to where Zen is going to land relative to holiday ramps in the Western markets and to Chinese New Year. So any color around that would be really helpful.

Yes, no, that's a fair question, Matt. So we have been very focused on the server launch for the first half of 2017. Desktop should launch before that. In terms of true volume availability, I believe it will be in the first quarter of '17. We may shift some limited volume towards the end of the fourth quarter based on how bring-up goes and the customer readiness. But again, if I look overall at what we're trying to do, I think the desktop product is very well positioned for that high-end desktop segment, that enthusiast segment in both channel and OEM, which is very much a segment that AMD knows well, and so that's where we would focus on desktop. You should expect a notebook version of Zen with integrated graphics in 2017, and that development is going on as well. And so, I think it's just a time of a lot of activity around the Zen and the different Zen product families.

Your next question today is coming from the line of Stacy Rasgon from Bernstein Research.

I was looking at the implied guidance for Q4. You said the -- I guess, the full year, up low single digits so, I mean, call that 3%. But it would imply Q4 down 16%, 17% sequentially and actually down on an absolute basis lower than I would've thought. I think Q4 also has an extra week in it. I was wondering if you could give us, I guess, some color on how you see the drivers, I guess, for seasonality going from Q3 to Q4 across both of the businesses, given the, I guess, the trajectory of the different project launches that we have in the back half. Like how do you come to that number?

Sure, Stacy. Maybe I'll start and then see if Devinder would like to add up -- add to it. So, look, our -- when we started the year, our expectation is that we would grow revenue in 2016 versus 2015, but we were coming off of a very low base in the first quarter. So we've been pleased with how it played out. Certainly, our second quarter revenue and the third quarter revenue guidance. Overall, the businesses are performing well, so we do expect both Computing and Graphics and EESC to both grow for the year. I think the Semi-Custom business is the large driver of the fourth quarter in terms of just how we see the overall business playing out. But the Computing and Graphics business is playing out as you might expect. So the second half should be seasonally higher, certainly with Polaris and as we launch broader availability across the product line as well as the 7th Generation APUs as they go into back-to-school and holiday. So that's the way we should think about it.

Okay. For my follow-up, I just wanted to take a look at your notebook shipments in the quarter. So they were up again sequentially. I think they were up double digits in Q1. I think they were up in Q4 as well. So 3 quarters in a row of sequential increase. Obviously, off of a low base as well. But how confident are you? Like, what's driving that? Is this actually sort of sell-out that's actually driving the demand? Or is this sell-in into the channel? And how should we think about that potentially normalizing?

Yes, so if you're talking about the notebook computing business, actually, we are kind of pleased with how it's performed. So if you look at it overall, the -- our OEM customers have adopted our technologies. So we see that across both notebook and desktop, but presumably notebook. We are also making progress in commercial, and that was a very important initiative over the past 5 or 6 quarters, so that's been nice to stabilize the business. And again, I don't believe it's a sell-in phenomenon. Actually, it's consumption share that we see increasing. And I think we have a reasonable opportunity to add to that in the second half of the year, but it's really around our products and the platforms that we're putting together with the OEMs.

Got it. If I could ask one more really quick. I was just a little surprised at how much your wafer purchases at GLOBALFOUNDRIES came down quarter-over-quarter, given the increase in notebooks as well as the timing of the Polaris launch. Any -- what -- should we read anything into that?

I think, basically, the purchases of the wafers are in line with product demand and mix of business. And as I said in the prepared remarks, year-to-date, we purchased about $260 million of wafers and we are getting into the back half of the year, which as you can see, with our revenue guidance in Q3, is pretty strong.

Our next question today is coming from John Pitzer from Crédit Suisse Group.

I guess, Lisa, my first question. I'd kind of like to get your view on the news earlier this week around SoftBank and ARM. You're sort of in a unique position because you're both kind of a partner with ARM as an ARM licensee, but if you look at your custom silicon business, especially on the x86 architecture, you could make the argument that you're a little bit of a competitor as well. So I'm kind of curious how you see kind of the reaction to that from a customer perspective? And does that change your view of how you might be able to monetize your own IP in the future?

John, thanks for the question. Well, I think it's a very interesting deal, the ARM SoftBank deal. I think we have a lot of respect for ARM. They are a close partner. I'm not sure we would call them a competitor. Overall, they are a partner. As it relates to how we think about our IP, we really believe that our IP, particularly the high-performance element of it on the microprocessor and the graphics side, is very unique. And in some sense, there are very few places in the industry where you have access to it. I think the applications that need it, whether you're talking about consumer applications or enterprise cloud applications, are growing, and it's so it's an opportunity for us to look for larger markets to monetize beyond just our own products. And that's, in general, the way we think about IP monetization is we have a set of products that is very important to our business model but our IP can go further than our product -- than our products themselves. And so we would continue to look for opportunities to monetize our IP.

And then, guys, I apologize if I missed this. You commented on sort of your expectations for GPUs going into Q3 and for semi-custom. Kind of how do you view the PC outlook within your total outlook above [indiscernible]? And I've got a quick follow-on.

Sure. Let me talk about the PC market overall. I think our view of the PC market is pretty similar to others in the industry. We would say, overall for 2016, high single-digit decline. We have seen some positive signals and we saw some positive data points from IDC earlier the last week. And then we saw a little bit of pickup in consumption in June from our OEM customers. But again, that was compared to, I'd call it a soft April, May. Our view is OEM business looks okay. The channel for us is still weaker than we would like, and that's our view of the PC business. Even in that framework, we believe we can grow our Computing and Graphics business on the strength of the products -- of the graphics products we've talked about as they ramp in the second half of that year. And on the computing side as well, I think we have a number of new platforms, and back-to-school and holiday are important for us. So that's the way we would characterize PCs.

So Lisa, not to put words in your mouth, but for Q3, PC is up, but perhaps not up as much as seasonal, is that the best way to think about it within your guide?

I would say roughly seasonal. I mean, roughly seasonal.

That's helpful. Then lastly, I apologize, but Devinder, now that you've got revenue growth behind you, and you've got some momentum here, how do we think about the gross margin progression from here and kind of the puts and takes? And I know there's a lot of different dynamics there, including the mix in semi-custom, but from these levels, how should we think about gross margin progression over the next, call it, 2 to 4 quarters?

Yes, John. I think the way you think about it is it is the mix of business, so that's a good observation. But also, you see us making continuing investments in our road map, in graphics and other areas, and we believe that will help us improve the gross margin. And then as you look out longer term, as you observe, get beyond the couple of quarters and get into 2017, so we do have the Zen product introduction coming up. We also have a positive [ph] in the pro graphics area, and that should drive even higher gross margins compared to where we are today, given the mix will get better with those products.

Our next question today is coming from Ross Seymore from Deutsche Bank.

This is Sidney Ho calling in for Ross. Just follow-up with the last question by focusing on third quarter. You guided third quarter gross margin to be flat, but the profile seem -- of revenue growth seems to be -- would suggest gross margins should decline on a mix adjusted basis. How should we think about gross margin by segment in Q3? What are some of the moving parts? And maybe follow-up to that, do you expect the gross margin of the new custom design wins ramping this quarter to have better or worse margins than your game console business?

Yes, I think that's well put. I mean if you look overall, when we have a higher mix of Semi-Custom business, typically, the gross margin is lower since the business on the semi-custom side affords a lower gross margin just the way the model is constructed. So keeping it flat at 31% with the guide in Q3, which is where we were in Q2. We're pretty pleased with that, and I think you're seeing some of the strength underlying the non-semi-custom business is allowing us how to keep it flat quarter-on-quarter.

That will be great. And then my follow-up question is you talked -- you guys talked about the Zen products at COMPUTEX and it is on track for launch in the first half of 2017. Can you talk about the impact on R&D expense and maybe on gross margin on the preparation of this launch over the next few quarters? And if you hit your revenue plans that you have -- you have thought about, at what point do you need to start increasing OpEx from the $330 million to $350 million level?

The OpEx side of it, if I look at it from that standpoint, you have seen our OpEx actually go up. In Q3, we're guiding at $350 million level, with specific targeted investments we are making in key R&D areas and products, in particular, the high-performance road map that you've laid out. But we're also making some specific investments in the marketing area, given the new products that are coming out on the PC side as well as the GPU side, trying to attract end users back to the AMD brand. And I think you'll continue to see us do that and that helps obviously with these new products underlying the better margins that you get with the fresh cycle, in particular, with the new technology areas that we are putting our products in.

Great. Maybe just one quick follow-up. For Q4 with an extra week in the fiscal quarter, should we expect the OpEx to be at the high end of levels, that, that will be enough of that $350 million...

I'm not prepared. One of things about OpEx is, as I always say, even if you look at Q3, the OpEx is up on a guided basis at $350 million, but very much and as I always say, we manage the OpEx pretty tightly. We've done that over the last several years, and obviously, it gets modulated somewhat by revenue. Q3 revenue is up 18%, OpEx is at $350 million, and we'll talk about Q4 when we gather here in about 90 days.

Our next question today is coming from Joe Moore from Morgan Stanley.

I was also curious about GLOBALFOUNDRIES being so low in Q2. And I guess, have you guys finalized the 2016 Wafer Supply Agreement? And can you talk more generally about which -- how you're deciding which products are allocated to GLOBAL and TSMC and is there anything that's exclusive to one or the other?

Yes, several parts to it. I mean, the working relationship, as Lisa said, with GLOBALFOUNDRIES is very good. We continue to work through the 2016 WSA and that's not finalized. We are in the process of negotiating that. In the meantime, we continue to get delivery of wafers for the products that we need in line with the product demand and mix of business. And relative to your -- which products from which foundry, we typically do not share the source of foundry for any of our wafers or products.

Okay. Okay, that's fair. And then with regards to kind of thinking about next year's OpEx, when you look at the importance of a CPU that can attack the server market and can attack the high end of the sort of enthusiast CPU market, I guess, I don't want you necessarily be constrained by the R&D requirements of the last couple of years. I mean, so how do you think about that if you start to show some revenue traction? I mean, can you -- how much flexibility do you think there is on spending? And just -- I'm not asking for a number, but just qualitatively, how are you thinking about profitability versus investing in these opportunities next year?

Joe, I would say, my engineers would be very happy to hear you say that. Look, I appreciate the question. I think you've seen, and Devinder said this, we're going to be disciplined in OpEx. And as -- this quarter was an important turning point for us to return to operating profitability, I think we have a nice sort of a view into Q3. We will look for opportunities to ensure that we are strategically placing the big bets in R&D, both on the CPU and the GPU side. For example, this year, we have ramped up our investments on the graphics side with the formation of the Radeon Technologies Group and what we're doing in both hardware and software, very significant investments, and similarly, on the CPU side as well. So I think we take very seriously our commitments in terms of profitability, but we will invest in the future, and we'll continue to look for opportunities to balance those in the right way. I don't know, Devinder, if you want to add.

I think that's well said. In 2017 and obviously as we get to the back half of the year, Q3, we have guided, we're getting to that -- get through Q3 and get into Q4, and at that point, we'll start thinking about where we want to place our bets as these are put in, in terms of the longer-term investments in 2017. Right now, we're really focused on getting -- we got to non-GAAP operating profitability this quarter. We want to maintain that in Q3 and then we'll see what happens from then.

Our next question today is coming from Christopher Danely from Citigroup.

This is Marco Chen calling on behalf of Chris Danely. My first question is, could you please update us on your expectations on the THATIC x86 server JV in China? Are there any potential legal risks with it, potentially Intel?

Yes. So relative to the server JV that we have with THATIC, as we stated before, we believe that our joint venture is operational. I think it's well underway. The technology that we're licensing is AMD technology, and so we don't have any issues relative to licensing. I think...

Yes. And if I may add, I mean, one thing on the update, I guess, if you look at the prepared remarks and the commentary for Q3 is we now expect, based on the technology transfer milestones completion, to recognize $22 million of a licensing gain in Q3 and approximately $75 million for the year, which is higher than what we had said previously when we said it was $52 million for the year. So it's incremental and, therefore, as Lisa said, things are operational, they're on track, and we continue to meet our technology transfer milestones from a date standpoint.

That's very helpful. And then my follow-up, I know you guys mentioned PCs and Graphics segment should be up year-over-year. Could you guys get into more detail, maybe talk about each segment individually, say, your trend for PCs versus GPU?

Yes, we don't usually go into granularity about the sub-segments of the business. I think, it's fair to say that our PC business has stabilized, and what we're seeing now is the opportunity as we go into the seasonally stronger half of the year on both the PC side and the graphics side. On top of that with some new product launches, we believe that we will end up growing year-over-year. But again, we don't usually guide to the details within the segment.

Our next question today is coming from Ambrish Srivastava from BMO.

This is Gabriel Ho calling in from Ambrish. I have a follow-up on your Project Scorpio comment. I think it's going to be large in 2017. And I think you have significantly higher performance and support for KVR [ph] So how should we think about maybe the pricing or ASP that -- compared to what you have indicated on the current generation game console platform?

So again, we don't -- we wouldn't disclose anything that's incrementally more than what our customer has disclosed. So I think Microsoft has talked about their goals with Project Scorpio. I think we are in support of those goals. Given the performance levels, you would imagine that there's more capability on-chip. But I wouldn't want to go more than that.

Okay. As a follow-up, I think your competitor has launched a comparable mainstream product, I think it's GeForce 1060. So -- and given the absence of your maybe a higher performance part launching maybe the second half, so how should we think about the rest of the year on the demand on your graphics side?

Yes, we are feeling very good about our Polaris launch. This was our strategy from the beginning. I mean, we laid out a strategy where we were going after the mainstream and trying to create really a new experience in terms of both capability and price point, and I think we've done that. We're also very, very focused on DX12 and our performance on new APIs, like Vulkan. So we are happy with our competitiveness. I think it's played out as we thought it would to. Certainly, our aspirations in GPU are to, certainly, have very competitive products across the entire product line, and so we've talked about working on Vega, which is the next-generation, high-end architecture. But in terms of our competitiveness, again, we've executed what we thought we were going to execute. And it seems like from both customer reviews and analyst reviews, that it's pretty well received by the market.

Our next question is coming from Ian Ing from MKM Partners.

So EESC, looks like you achieved operating income, same level as the third quarter of last year, but with $45 million lower revenue. So just wondering how you got more profitable in this segment, given these are both strong game console quarters and also you've got annual price declines in game consoles, I assume.

I think, if you look at the profitability, you've got to remember that in the EESC segment, we also have other businesses embedded, for one, and also the investments we are making for the Enterprise side of it with the data center product that's coming out in 2017. So the profitability, you're right about the observation. But profitability depends upon, obviously, the largest business we have in there right now is the Semi-Custom business. But also, the investments on the OpEx standpoint that we make year-on-year, and we talked about the targeted investments we are making in some of the product areas.

Okay, great. But these are relatively older products then, Embedded and Enterprise products at the moment, then, it sounds like. Okay, great. And then, my follow-on is, you had a 1-year cadence for GPUs now. Can you talk about the expectations on how last year's R9 300 series winds down and how that plays out? It looks like in Q2, there was really just a lack of a pause ahead of a known refresh and do you expect sort of a sharp wind down at some point? Or is it more gradual?

Actually, Ian, that's a good observation. The 390 Series or the 300 Series, overall, actually did do okay in the second quarter. We were also wondering whether there would be a pause prior to a new generation. I think we see it as a very orderly transition. It's actually one of our better product transitions as we go from the 300 Series over to the 400 Series. So again, nothing very spectacular to report other than the sell-through has continued -- the sell-through was good in the second quarter, and we believe that it's an orderly transition of the inventory.

Your final question today is coming from Jaguar Bajwa from Arete Research.

Just a question on Q3. Could you just lay out which would be the fastest-growing sequential segments in Q3? Will that be the GPU business, CPU or Semi-Custom?

Jaguar, I think what we would say is that our Q3 guidance is both Semi-Custom and Graphics will be the drivers of the sequential revenue growth.

Okay. And then, could you just give a bit more clarity around Vega timing? And also, what do you expect to be your differentiation here versus what NVIDIA has done with Pascal?

Again, I think we'll talk more about the details of the Vega architecture in time. But certainly, Vega is a high-performance GPU that will use high-bandwidth memory as part of it. So I would leave the details for a more -- as we go -- get closer to the Vega launch date.

Okay. Well, maybe just one final one, then. When we get into the server CPU cycle with Zen, should we expect any kind of contribution from an APU format with CPU plus Vega? Because we're seeing a lot of acceleration now in the data center. Do you think that could be a key product for you guys, which maybe your competitors may not have, bringing an APU to the server market for high-performance servers?

Yes, I think it's fair to say that we do believe we can combine a higher-performance CPU with a high-performance GPU. And as we look at our GPU compute in general, sort of our both professional graphics and server markets for GPU, I think that will increasingly become an area of focus for us as we continue to grow the Graphics business, so I think the answer is yes. I mean, obviously, it will come in time, but it's an area where combining the 2 technologies makes a lot of sense.

Operator, that concludes the question-and-answer session. If you wouldn't mind wrapping up the call, please?

Certainly. That concludes today's teleconference. You may disconnect your lines at this time, and have a wonderful day. We thank you for your participation today.