Advanced Micro Devices, Inc. (AMD) Q3 2012 Earnings Report, Transcript and Summary

Good afternoon. My name is Marvin, and I will be your conference operator. At this time, I would like to welcome everyone to the Xilinx Third Quarter Fiscal Year 2012 Earnings Release Conference Call. [Operator Instructions] I would now like to turn the call over to Rick Muscha. Thank you. Mr. Muscha, you may begin your conference.

Thank you, and good afternoon. With me are Moshe Gavrielov, CEO; and Jon Olson, CFO. We'll provide a financial and business review of the December quarter, and then we'll open the call for questions. Let me remind everyone that during our conference call today, we may make projections or other forward-looking statements regarding future events or the future financial performance of the company. We wish to caution you that such statements are predictions based on information that is currently available, and that actual results may differ materially. We refer you to the documents the company files with the SEC, including our 10-Ks, 10-Qs and 8-Ks. These documents contain and identify important factors that could cause the actual results to differ materially from those contained in our projections or forward-looking statements. This conference call is open to all and is being webcast live. It can be accessed from our Xilinx Investor Relations website. Let me now turn the call over to Jon Olson.

Thank you, Rick. Xilinx sales were $511.1 million in the December quarter, a decrease of 8% sequentially and 10% on a year-over-year basis. This was slightly better than the revised guidance we provided on December 19, due primarily to increased business from large communication customers in Asia-Pacific and North America during the last 2 weeks of the quarter. Gross margin was 65.8%, up from 63.9% in the prior quarter and better than anticipated, driven primarily by favorable customer mix and continued overall cost reduction. Operating expenses were $199 million, flat with the prior quarter and a little less than guided, primarily due to lower variable expenses associated with lower revenue and increased attention to fiscal discipline. Operating margin was 26.8% for the quarter. New product sales decreased 13% sequentially and increased 5% on a year-over-year basis. Virtex-6, Spartan-6 and our 28-nanometer product families all posted solid sales growth during the quarter. The new product category, however, was impacted by declines from our Virtex-5 customers in the wireless segment. Mainstream products declined 1% sequentially, and base products were flat. Let me now turn to a discussion of end markets. In the face of a continued weak wireless demand environment, communication sales declined 10% sequentially, more than we were anticipating heading into the quarter. Wired communications, on the other hand, increased during the quarter with particular strength from enterprise networking. The industrial and other category was down, as expected, with sales decreasing 7% sequentially, driven by industrial, scientific and medical and test and measurement applications. Consumer and automotive sales were weaker than anticipated during the quarter, driven by declines from audio/video broadcast, consumer and automotive applications. Data processing sales were approximately flat. Net income for the quarter was $127 million or $0.47 per diluted share. Other income and expense was a net expense of $7.2 million. Third quarter net income included a tax benefit of $15.3 million or $0.06 per diluted share related to one-time items, including the lapse of the statute of limitations with respect to certain previously unrecognized tax positions. Operating cash flow for the September quarter was $181 million before $19 million in CapEx. For the past 4 quarters, cash generation from operations was $863 million, a new company record. We paid $50 million in cash dividends and repurchased 1.1 million shares of stock during the quarter for $32.5 million. The tax rate in the December quarter was 2%. Diluted shares for the quarter were 268 million. There was a 2.5 million share dilutive effect from our convertible notes due to an increase in the average stock price during the quarter. For questions relating to the dilution associated with our convertibles, please visit our Investor Relations website at www.investor.xilinx.com. Let me now comment on the balance sheet. Cash and investments increased $95 million to approximately $3.0 billion. We have approximately $1.3 billion in convertible debt, and our net cash position is approximately $1.7 billion. Days sales outstanding increased 3 days in December quarter to 38 days. Inventory dollars at Xilinx declined by $2.2 million sequentially during the quarter. Combined inventory days, both at Xilinx and distribution were 142 days, up from 126 days in the prior quarter. In the March quarter, we expect both dollars and days of inventory to decline significantly as we realign our supply with the demand trends. Let me now turn to a discussion of guidance for the March quarter of fiscal year '12. Our backlog heading into the quarter is up sequentially. We are expecting strong growth from our 28-nanometer and 40-nanometer product families. From an end-market perspective, we are expecting sales from communications to be up slightly sequentially, with wireless up modestly after a weak December quarter and wireline sales approximately flat. We expect industrial and other sales to increase as normal spending patterns return to many of our customers following 2 quarters of inventory drain. Additionally, we are expecting defense sales to increase during the quarter, driven in part by last time buy activity, but more significantly by the commencement of a large program. Consumer sales are expected to be approximately flat, and data processing sales are expected to decrease sequentially. As a result, we are expecting total sales to be up 2% to up 6% sequentially, with sales from Europe expected to increase, sales from Asia-Pacific expected to decrease and sales from North America and Japan to be approximately flat. The midpoint of our sales guidance is predicated on a turns rate of approximately 55%. Gross margin is expected to be between 64% and 65% as we absorb new product ramp costs and customer mix returns to normal historical levels. Operating expenses in the March quarter are expected to be approximately $207 million, including approximately $2 million of amortization of acquisition-related intangibles. The majority of the growth in spending compared with the December quarter is in R&D and related to the 28-nanometer ramp as our tape-out activity remains high. There's also some growth in SG&A related primarily to variable spending increases as result of higher revenue and growth in some key marketing and sales programs. The full year spending number is approximately $810 million, lower than our previous estimate of $820 million to $830 million. We will provide fiscal year 2013 guidance at our upcoming Analyst Meeting on February 15 in San Francisco. Other income and expense is expected to be a net expense of approximately $10 million. The share count is expected to be approximately 270 million shares, and the tax rate is expected to be approximately 14%. Let me now turn the call over to Moshe.

Thank you, Jon, and good afternoon to you all. While we obviously have not been immune to the weakness in the current economic environment, I am pleased with the nearly 20% combined sequential growth from our Virtex-6 40-nanometer and Spartan-6 45-nanometer families. Virtex-6, in particular, experienced very strong growth from wired comms and storage applications. We expect these families to show continued growth in the March quarter based on the beginning backlog and customer forecasts. As we begin 2012, I've grown ever more confident about the competitiveness and customer value of our next-generation 28-nanometer product portfolio. This is because our top customers around the world are consistently telling us we're delivering indisputable leadership in terms of 28-nanometer technology. I'm extremely pleased with the execution of the 28-nanometer rollout, which is without question the most aggressive and complete new product rollout in our history. We have now delivered more than 10 tape-outs and are confident, by the middle of this year, to be sampling members from our entire 28-nanometer family. Additionally, the industry's first 28-nanometer product, the Kintex-7 K325T, recently passed rigorous qualification criteria, would release for production shipments, as planned, in the June quarter. 28-nanometer success would not have been achieved without our exemplary technology partnership and collaboration with TSMC. TSMC's fully qualified HPL process optimized for both performance and power has provided us significant time-to-market advantage and is providing yields that are in line with our aggressive roadmap and better than what we had experienced in previous generations. This past quarter, we achieved 2 very significant 28-nanometer milestones. First, we began shipping the industry's first extensible processing platform, the Zynq-7000 device. This family enables a new level of hardware-software systems integration with a complete ARM Cortex processing subsystem and low-power programmable logic. With a very significant time-to-market advantage over the competition, design wins already totaling well over $100 million, this new class of product is creating an unprecedented level of excitement value in our customer base. We also began shipping the first device from our SSIT, Stacked Silicon Interconnect Technology family. The Virtex-7 2000T device is the industry's highest capacity programmable logic device, with 2 million logic sales, delivering twice or more of the capacity of competitive devices through single-chip FPGA integration. This innovation is absolutely game changing, conservatively providing us with at least a 2-year time-to-market advantage. Design win momentum has been exceptionally strong with over 30 wins valued at more than $100 million. Our only challenge at this point is our ability to more quickly ramp supply to meet all of the demand for this product. We're very actively working through all constraints and are confident that we'll be able to align our supply with the demand for this product line. We now have four of the five 28-nanometer product family shipping in the market. The Zynq and Stacked Silicon Interconnect initial shipments were preceded by the shipment of key Kintex-7 and Virtex-7, both mid- and high-end family members. With 28-nanometer design wins now totaling nearly 350 and well over $1 billion in terms of design win value, we have approximately 90 discrete customers currently designing in members of the 28-nanometer product family, demonstrating the broad-based acceptance of it. In the March quarter, we are anticipating continued strong sales growth and by the June quarter, expect to comfortably exceed $10 million quarterly sales for this family. Finally, after reviewing input from over 120 customers on our 28-nanometer leadership position, I can tell you that in my more than 30 years in the high tech industry, I have never seen a single product generation which has the level and breadth of customer impact as our existing 28-nanometer portfolio. I look forward to providing you with an update on our 28-nanometer momentum at our annual Analyst Meeting, which is scheduled February 15 in San Francisco. Let me turn -- let me now turn the call back to the operator to open it up for the Q&A session.

[Operator Instructions] Our first question comes from the line of James Schneider with Goldman Sachs.

I was wondering on the wireless communication side, could you talk a little bit about the visibility you're seeing in terms of your customer programs there? It sounds like you're expecting that to increase slightly in the March quarter. When do you think that might bounce back more materially? And are you seeing that bounce back happening more in the 2G, 3G or 4G areas?

Jim, this is Jon. The visibility on wireless is really not very good for us. And while we still have, I would say, a reasonable level of business broadly across older generation technology, the uptake, particularly in North America, has been much weaker than the trend has been over the last 4 quarters. And then I think it's also pretty widely known that the next wave in China has yet to materialize in the same way that's being talked about. So we're kind of, in some respects, we're in a bit of a holding pattern waiting because the information we're getting from our customers doesn't really give us a precise quarter that they think the capital spend is going to start increasing at all those various providers. So it would be premature for me to speculate on when that might be. I still strongly believe it's just a matter of when, not if, just because of the strong need for bandwidth and the move towards LTE more globally, where we have an extremely strong position. So I'm not being very helpful, I know, relative to telling you exactly which quarter, but we think it is coming.

And then maybe as a follow-up, I'll ask you maybe a similar thing on the wireline side. It sounds like that really hasn't even bounced back cyclically at all at this point, and we heard anecdotally from a lot of carriers that they were postponing a bunch of new wireline programs into sometime in 2012. Any more visibility on -- I know you have limited visibility across the board, but do you think that wireline comes back later than wireless when it does?

I don't know exactly what specific programs you're referring to but -- and if you can tell, I probably couldn't give you specific customer data anyway. But we've seen a recent, I'll say, mini-surge off a low inventory point at several key customers. And so we're trying to read that ourselves to figure out whether it's an inventory replenishment above all others or it's got more legs. I think if you listen to the network guys, what they're saying, it's a little patchy as to whether they're highly confident or less confident going forward. So I think we're still in choppy waters there.

Our next question comes from the line of Uche Orji with UBS.

Moshe, let me just follow by asking you about some of the comments you made on 28-nanometers. First of all, congratulations on the very strong design wins. But can you help me understand your 28-nanometer design wins today compare that with 40. And over the next year, my assumption is that 40 will still be the main revenue driver even though 28 design activities going on. Where do you see your 40-nanometer market share versus 28? And how do you see the growth of 28 ramping between now and, say, the end of 2012?

Okay. So let me first talk about 28. We have an extremely broad product line at 28, 5 distinct families: Altix, Kintex, Virtex, SSIT, which is a high-end and unparalleled in terms of density so it actually opens up a new market for us, and then there's the Zynq product line, which again is -- opens up a whole new set of market. When we look at our competitive position, our expectation is that we will have 70% of the market at least, and that's based on the fact that in 2 of the 5 areas, there is no competition now, and we expect that competition to take a long time to come. And on the other 3 areas, we actually feel we have an incredibly strong hand. Now in terms of how quickly we expect it to come out of the gate, it is coming out very quickly, much faster than the previous 2 generations of product. And because the yields are also very high and the demand is there because it either can enable capabilities that aren't possible with the previous generation of technology or if they are possible, then it's a lower power point and over time, a lower cost point, we expect that to grow at a very fast rate. So if you look at Virtex-5, which was our previous generation, a previous more successful generation, then that took 3 to 5 years to reach its peak. Here, we expect that the peak A will be much higher because it opens up significant new markets, and we actually feel that we have a much better competitive hand. So we would expect, as we indicated, $10 million or more in what is the April, May, June quarter, and we expect it to grow rapidly from that point forward. So we aren't giving predictions, but that's much faster than we've seen previous generations of product. The 40-nanometer and 45-nanometer, we clearly came out later than the competition. Our strength there is driven by the fact that we have a low-end product in the 45 and that is doing well. Overall, our goal on the 40 and 45 is to close in, over time, on 50% market share. We're clearly not there now, and it's going to take us some time to get to that point. But in parallel, we expect the 28-nanometer to totally flip the way the market share has grown over the past few years and enable us to win back significant market share. When that will become extremely visible? It becomes visible when the 28-nanometer moves into significant volume. If you look at the numbers, given that we're -- the market in of itself is between $4 billion to $5 billion now, then it requires more than $10 million per quarter to make a significant dent there. But at the rate that it's growing, by the end of the year, this should be very visible.

I'd like to add one more thing. When we looked at this last of how we're doing versus the ramp on 40-nanometer, it actually wasn't this quarter. It was a quarter ago when we talked about having accreted $600 million of the design wins. We were 60% ahead of where we were on -- we were on 28 versus where we were at 40. And I actually didn't have a look at it at the end of this quarter, it could even be a bigger percentage. We'll try to give a little bit more information about that at the Analyst Day.

Can I just ask you one question? Of the $10 million revenue you have now, how much of that is the SSIT product? Because from the numbers you provided, Moshe, it is a very high ASP product, and I don't know whether that is something that could be a sustained growth beyond the accumulation phase. So I'm not sure how to think about the sustainability, what the mix of that is and what the sustainability of the SSIT product would be.

Yes, so the projection we made for $10 million is in the April through June quarter, and that does not require a huge amount of SSIT revenue. And so you shouldn't assume that SSIT is necessarily a huge percentage of that. And the reason that we believe that it has longevity is that if you look at the capabilities there, there's no way to get to those capabilities unless you design a high-end ASIC, which obviously is becoming less and less viable. So we actually are seeing 2 -- no, sorry, 3 very large markets using it, one is prototyping and you can say that's the one which tends to surge and then decline, but the 2 which have longevity are the wired communications, where it replaces ASIC and the Mil/Aero where there's a huge demand for very high-end devices. And both of those, obviously wired communications happens at a faster rate typically than the military market, but both of those tend to have long life cycles.

Yes. I also want to point out a couple of other things about the SSIT family, is that we have more than one member of that family, there's 3 members. And some of those members have a very high capability and total bandwidth for transceivers as well. And so this will get used a lot in communications design. So as those get taped out and through the sampling phase, et cetera, we think there's going to be a lot of legs in this particular family.

Our next question comes from Ambrish Srivastava with BMO.

Just on gross margin, guys. You guys have done a really good job in committing to the Street and then executing on that. I was just curious, despite the big comp customers coming back, you had a slight upside. How should we think about margins? What are the puts and takes around the 64%, 65% target that you guys have set? And then a quick follow-up, on inventory, Jon, is this a continued buildup from the end-of-line legacy that caused the total inventory to go up in the last quarter as well?

Yes. So on the gross margin, Ambrish, it spiked up in part by, we talked about customer mix, which essentially means we had a lower percentage contribution from large communications customers. And that had an impact of somewhere less than a point, but in that general neighborhood. So that brings us back down into more in the normal range that we have. And then as we tape out, we have many, many tape-outs going on right now, all going on at various stages in parallel, and that causes us to produce units that engineering has to test, et cetera, and some of those are salable, some of those aren't. So some of those things end up impacting our margin, but we've been able to hold ourselves at least within our model even to the start-up phase. So those are some of the bigger impacts. One other impact is that as we are cutting wafers pretty significantly, we have been the last 2 quarters, and that will kind of, I'll say, overrun our revenue decline now and therefore we'll see a precipitous drop in inventory dollars next quarter. And that has some high costs falling out that had some downward pressure to our gross margins and puts us in that 64% to 65% range. So those are the factors. So I would expect as we continue to have the ramp and get rid of the start-up costs, I suspect that we'll have, out in time, upward pressure going on in the gross margin number. We're kind of at a low point, if you will, at this point in time. From an inventory perspective, look, as you know, dollars went down very slightly, but let's say inventory was flat and days just shot up, and that's just a function of the arithmetic because revenue is down, and you use the cost of sales number to create the days number, so it was in absolute dollar, a lower number than in previous quarters. So days happened to shoot up even though inventory dollars really didn't go up. There is some of the last time buy inventory that is in there. It stays relatively constant for the next few quarters, just like our last time buy revenue will stay relatively constant for the next several quarters, so we're not seeing any real revenue blips because of last time buy or any significant declines in revenue coming because of the last time buy. It just takes a -- we've had to build so much of it because of the fab line that's being discontinued. We're just trying to manage that to a flat dollar level over the next 4 quarters.

Our next question comes from the line of Vivek Arya with BoA Merrill Lynch.

I'm trying to compare your current visibility versus how it was, I think, in December or so when you gave your outlook. What end markets are you more confident about now? What end markets are still sort of uncertain? Where I'm going with that is your March guidance is suddenly getting closer to normal seasonality, but I'm still sensing some uncertainty in your comments about end markets. I'm just trying to read the lines here on what the underlying turns are? What end markets do you like, what are still sort of uncertain right now?

I gave some comments earlier about the communications market. I think by our tone in the words, you'd probably come to conclusion that we think wireless has bottomed out for us, and we're -- now we're kind of waiting for whether there's additional CapEx spends by our customer's customer or not, even though we do see some modest increases going on there. And in the wireless segment, still choppy, not easy to predict where that's going or where the IT spend is really going to go or where the cable and satellite provider spend is really going to go. So I would say there, it's not really whether we like it or we don't like that. I'd say that the confidence level is -- we feel like we're pretty close to bouncing on the bottom, and we're not really sure when things are going to, I'll say, more fully recover to where they were 3 or 4 quarters ago. Almost the same statement on industrial and other. We do feel that that's more of an inventory drain situation, and that people have corrected their inventory and we're seeing a lot more orders coming. One of the reasons that revenue was a little higher than we thought, both primarily communications customers, but distribution also ended up being stronger in December than we had anticipated. Overall, we've pretty much counted on the last week or 2 to be very sparse with factories being closed because people didn't have a need to build stuff so they gave time off, and that was some of the word that we were getting through distribution. But then that actually didn't happen in the same way that we had anticipated. So a lot of our industrial customers come through our distribution networks, so that gives me a very positive feeling around. Again, we feel like we've reached the bottom on industrial. When is it going to come back? Consumer typically a little weaker in the March quarter anyway, so that's pretty much in data processing. That doesn't move the needle one way or the other that much for us. So I hope that's helpful.

Very helpful. And just a follow-up clarification and a question. What was the 40-nanometer as percentage of sales mix? And then the bigger question is very impressive design wins at 28-nanometer. I'm curious; when do you expect that to be, I don't know, say, take 10% of industry sales? Is it more a 2013 or 2014 time frame?

I'll deal with the second one first. So the 28-nanometer design wins are a myriad of end markets, end customers in various stages of rolling out their products. So you can't really project the size of the ramp, if you will, or the speed of the ramp. But because we have so many different customers and so many designs, just our experience is that ramp is going to manifest itself faster than we've seen in the last couple of generations. So in some respects, we're going on some very specific information from specific customers. In other respects, where it's a very broad set of customers, we're kind of going on the way trends have worked in the past. So again, we try to give the best information we have, not overcommit on things and we think we've given you a pretty fair statement on that. Your question on 40-nanometer was around percent of sales. We really did not provide that this quarter. I don't think we're going to provide that number.

Our next question comes from the line of C.J. Muse with Barclays Capital.

I guess first question, and I realize visibility is somewhat limited here, but curious how you see the shape of the cycle off of the trough here?

Yes. I'll give you my two cents. I don't know if Moshe has anything to add to that. But having lived through some of these before and see that there's certainly some end markets here where this is more -- it's a demand driven in addition to inventory correction, I view this generally as kind of a slow movement off the bottom over the next couple of quarters and then a better second half. But that's just an experiential thing, I don't think we have any data that necessarily supports that, because I think we look at the same reports that you guys do.

That's helpful. And as a quick -- sorry, go ahead.

My prospective, this is Moshe here, is that the one market which unquestionably is on fire relates to the smartphones. And if you look at what is happening there and the demands that they're putting on the underlying infrastructure, at a minimum, as those -- when people are moving away even in emerging markets and lower-cost markets, they're moving away from what I call feature phones, these smartphones. These phones generate huge, huge, huge bandwidth demands and as such, definitely, the systems need to be upgraded to support those. So on communications, both in particular, wireless communications infrastructure, coming back to Jon's comment at the very beginning, that can't continue. It has to be upgraded; otherwise, people will revolt against the fact that their phones, while flashy, can't be used for anything. And it's a huge business opportunity, and I'm very convinced that that's a question, as Jon put it, of when as opposed to will it happen. It definitely needs to happen soon. And we're seeing similar trends in other markets, which has similar characteristics but maybe they're not as visible because it's not driven by consumer smartphones which everyone knows about. So the other thing I'm hearing is in the whole host of ecosystem plays, there's been a huge bounce back from a very significant trough to now everyone building up to support the imminent requirements. And so I feel that good things are happening and timing might still be a question mark but just generally, I'm very bullish about the secular trends which are driving all of these needs.

That's very helpful. And if I could just ask a quick follow-up. When you think about 28-nanometer growing as a percentage of your mix, I'm curious, given the increased capital intensity at the foundries growing roughly 50%, we're in the move from 4x to 28-nanometer, what impact that will have on your gross margin? So how should we think about that? I know it's down the road a bit, but we'd love to hear your thoughts there.

Well, every generation becomes significantly more expensive and 28-nanometer, vis-a-vis 20, if you just look at the wafer prices, there's no doubt that they're going up there, there's less suppliers and the equipment is more expensive and the number of customers who are driving at the leading edge is much smaller so as a result, you don't see the normal, more slow, that's one trend. The other -- on the other side, to balance that, there's huge amount of engineering, which is being put in place in the product design, in the product design for manufacturability and in the operations flow, which makes us very confident that we can, to some extent, compensate for those. So in terms of margins long term, we're shooting for the same margins, and in the short term, as Jon mentioned, we're actually being able to absorb this new technology, which is a lot more expensive for us, better than we have before, a lot of it due to these elements plus obviously, there's the higher yields we're seeing, which are helping out. So we're not, at this point at 28, expecting negative impact on margins for our business.

Our next question comes from the line of Glen Yeung with Citi.

So you guys are guiding -- if I look at the numbers, kind of in line with what I would view as normal seasonal for the first quarter. I wonder if -- I guess just some questions about the relative expectations for the year. And 3 things: One, do you think normal seasonal is a pattern that is sort of the benchmark you're going to use for the rest of the year? Two, do you think that your business will outgrow the overall chip market? And thirdly, do you think that you should outgrow the end markets because you'll benefit from an inventory replenishment on top of just demand growth?

So, no, even though it does appear March is in much seasonal line with this, I'm not sure I feel like we're -- the whole year ends up being seasonal, because I think there's some recovery that comes on this year, and so that kind of wipes out the seasonal -- the underlying seasonal activity that would normally go on end market-by-end market. So I don't see that seasonal affect naturally holding on this year. So I guess maybe I'm a little more bullish this year than maybe what seasonal or 5-year trends would say -- averages would say for us. I do think we are going to outgrow the rest of the semiconductor industry. I think we've been saying that for some time, I still believe that. We tend to talk about this in longer-term CAGRs, so any given year could look one way but we do think on the whole, it will be better. And then the intensity we have in terms of penetration at our large customers is growing, and the breadth of our new technology is helping us get there just as Moshe talked about today. So I guess that's a no, yes, yes.

And then let me ask you a question just by bringing it closer to now, I suppose. Can you just talk a little bit about how your order patterns looked as you exited the year, and then how that may have changed, if at all, in January? And then you talk about sort of an uncertainty as to whether or not this will stick. I just wonder what are the kind of things you'll be looking for from your customers. What do you need to hear from them to give you confidence this is a recovery or it's not?

Yes, so the order patterns, again, we left -- we ended the quarter stronger than we thought. We ended with backlog higher than we entered the previous quarter. So these were all things that we viewed as very positive signs, and we were probably a little more bearish during the December month than we should have been, obviously. And so I think we -- we're bolstered by what we see in the backlog and how things are going so far in the first, very short 2 weeks of January. But quite frankly, we need the large customers to order parts and not push out what's on the backlog today, and that will give us some good signs. To me, that's the biggest sign, when the top 25 are in the game and they start putting orders on us and in some cases even expediting, that's what's going to make me start feeling a lot better.

Our next question comes from the line of Shawn Webster with Macquarie.

Well, it sounds like you'll give us an update on the trajectory of OpEx next year. But I was wondering if you could talk about maybe when you're -- when you expect the peak quarter to be in your 28-nanometer tape-outs? I believe in the past, you thought OpEx could decline next year. I'm just trying to figure out what the trajectory could look like at this point.

So I have a hard time answering that question without actually answering the question I said I'm not going to answer, which is talk about spending number for next year. And we -- part of lowering our, and I've said this on the investor conference, or a good part of the lowering of our spending engineering number this year, our OpEx number for this year lower than we had anticipated was because we've moved some things from one year to the next. We've re-calendarized some of our tape-outs in addition to taking out discretionary spending, and that's what's caused our number to be going down from the $850 million or so kind of number down to -- $850 million or $860 million number. We're talking about million dollars of operating spending, now down to $810 million for this fiscal year. So that does imply that there is some of that going on. But because we haven't talked about what the full year is going to look like, I'd be hesitant to say exactly what the profile of engineering spending is going to be next year. That's what we're going to talk about in 3 weeks.

You can't talk about when you think the peak tape-outs will be, if they will be in the first half or the second half?

Well, the peak, yes. The peak will be, of 28-nanometer tape-outs, is more likely in the first half, but you have to actually know the profile to know whether that -- whether there's a flat roof or a peaked if you go back to my Analyst Day presentation. [indiscernible]

Okay. And then a follow-up on the...

I'm not sure that's helpful. I'm afraid you're going to take -- you're going to draw conclusions from that, that may not be accurate, so just be careful with that.

My conclusion is I need to wait for next month. So if I could ask a follow-up question on the inventory side. Can you give us some color on how channel inventory -- just the inventory declines stay flat or go up? And what's your view on the OEM side of the inventory equation?

Yes. Our channel inventory has been pretty flat from a days perspective, so it's kind of tracked -- it's tracked our revenue numbers; been pretty flat around 15 total days for the company, plus or minus a couple, quarter-by-quarter for the last 4 quarters. And essentially we've again had short lead times. The distributors are quite confident they can get parts very rapidly, so they haven't been stockpiling parts nor have they been necessarily overly leaned out with respect to Xilinx. I can't speak for other lines that they're carrying. So for me, the inventory number isn't particularly a concern. Inventory at customers, we -- throughout the December quarter, we definitely had an increase of pull-ins by customers, meaning numbers of customers who were asking us to pull things in, which is a good sign that they have leaned up their inventory and they need things. These are all sizes of customers. I did point out that the last 2 weeks we had unexpected labels from large communication customers. One more in addition to our distribution customers. So to me, that's a good sign that we are bouncing along the bottom because while we still have some push-outs and reschedules, we also have some pull-ins. So it feels to me like that inventory at customers has worked its way through, for the most part with some exceptions, and we're now kind of more, I'll call it, hand-to-mouth, depending on what their end demand is.

Our next question comes from the line of Ruben Roy with Mizuho Securities.

Moshe, I have a question on the 28-nanometer products and the comments around $10 million in revenue through the April to June quarter. If you think about the ramp and how you guys were thinking about 28-nanometer maybe 6 months ago or even last year at your Analyst Day, would you say that the revenue ramp is moving in line with your expectations? Or do you think the macro -- the challenging macro has had some impact and perhaps some delays in some of the programs that you had won? And if so, do you think that there will be potentially an inflection point or a step function increase in 28-nanometer revenue later this year?

So that's another great question. So intuitively, it feels that it's happening more or less a quarter later than we had expected. Part of those are issues in the market, part of those are making sure that the products are fully qualified. As we have this scalable architecture and now that we have the first product basically having gone through -- both the processes qualified, that's happened a long time ago at TSMC and now the product is qualified to extend, that enables us to apply the pedal to the metal in terms of getting all of the other products out in an accelerated mode. And I do believe that, that will enable faster growth from this point on. And so the numbers we've projected are just the beginning of what we consider to be a significant ramp. And given that we have 5 product families, 4 of which already are sampling and the fifth is actually the simplest one, which already actually has been demonstrated, it's a subset of the previous ones, we feel very, very, very confident in our ability to bring this to market. And we do expect it to accelerate. Now the different products do tend to have different profiles and because we've started with the midrange product, which is Kintex, and we're now moving, and the Virtex product is catching up with that, that should provide some acceleration, too, because that tends to have a much higher ASP as does the SSIT, which was one of the questions we were asked about previously. So I know this is a supposed answer, but the answer -- if you summarize all of that, it should grow very rapidly through the second half of the year due to the breadth of the product offering.

And just quickly around Zynq and given that you've got some early returns from customers, would you say that the initial activity is in areas that would be incremental to the FPGA TAM? Or how are design wins working out? Is it kind of FPGA late replacement sockets or new sockets that you wouldn't traditionally hit?

So the markets we're addressing definitely -- we have tens of thousands of customers in nearly every market. So this is less opening a new market, but it enables us to significantly increase our penetration into markets which are not necessarily FPGA-centric. And an example on Zynq, which was targeted at 3 general markets: automotive, ISM and to a lesser degree, communications, we're seeing a totally new set of applications in automotive, for example, where truthfully, our percentage of the BOM before would be minuscule, and the percentage of the BOM now or the bill of materials is very, very significant. So no, it basically enables us to replace the solutions which used external processors, external DSPs, external analog mixed signal capabilities and external FPGAs to -- so it is a market expansion because it is -- it does have an FPGA element, there is an element of FPGA design, the tools we are providing to enable a larger, less skilled set of customers will be the key to enabling us to address that larger market in terms of designers, and that will enable us to reach a larger TAM.

Our next question comes from the line of Mark Lipacis with Jefferies. Mark Lipacis - Jefferies & Company, Inc., Research Division: It seems that both Altera and Xilinx are doing FPGA and microprocessor integrations, and it's not clear to me what the difference is, and I'm hoping that you can help me understand what's the difference between what you guys are doing, what your competition is doing.

Okay. Well, we've been focused on this area for a very, very, very long time, and you can check comments made by the competitor on this area as to when they got into it. As a result, we believe we have customer engagements for years now that are now reaching fruition, and we have a very, very significant time-to-market advantage. At this point in time, what we see from the competition is very early and if we compare it, it's probably equivalent to things which we showed customers over 2 years ago. And so we feel very comfortable with our product offering. We were there first. We have a very broad product offering. We have sampled, it works. It's in customer systems. The customer systems are working. We feel very good at our -- with our competitive position. We obviously need to continue running, right, because if you stop running, then the competition will catch up, but we definitely have a very significant early lead there.

Yes, Mark, one more thing. I don't think we can actually say much about their products since it only exists on the paper and a block diagram, which happens to look close to silicon block diagram, looks very close to our block diagram. But there isn't any unit out there, there isn't any evaluation boards out there that I'm aware of, and so it's hard to know exactly what they will do with it. And I think, as you know, it isn't just about a piece of silicon, it's around all the software ecosystem, et cetera, that you build around a processor subsystem. And as Moshe said, we've been working on this for quite a few years and with real customers engaging on this. So it's actually difficult to make a real competitive comparison at this juncture. Mark Lipacis - Jefferies & Company, Inc., Research Division: That's helpful. Can you just say what are the biggest end-market applications?

They are automotive, industrial and communications.

Our next question comes from the line of Christopher Danely with JPMorgan. Christopher B. Danely - JP Morgan Chase & Co, Research Division: When your sales started to pick up towards the end of the year and into this year, what did your customers say or distributors say as to the reason that they were ordering more? Was it that their demand was better or was it just that they'd kind of gone through the inventory burn phase, and they were taking some up or a little of both?

I would say it's about 50/50, based on what I can tell. So I tried to go through the end markets and the customers because I was trying to understand that and get people to explain that to me as well. And the best information I have today from the distribution side were it's just a lot of different customers involved in it. It's really kind of been a little of both. The networking side, et cetera, I don't have as clear of an answer on that question. Christopher B. Danely - JP Morgan Chase & Co, Research Division: And then just as a follow up on, it sounds like things are bottoming here. You guys have plenty of cash. I'm just curious as to why not get a little more aggressive on the buyback?

Yes. Well, this is an annual process that we go through in our discussions with the board, and we certainly will be discussing all of the alternatives. As we have been committed historically to both and we continue to be committed to both dividends and cash repurchases, so I think you can expect us to continue to look at both of those as opportunities.

Our next question comes from the line of Srini Pajjuri with CLSA Securities. Srini Pajjuri - Credit Agricole Securities (USA) Inc., Research Division: Jon, you talked about, last time, a program in defense. I'm just wondering if you can quantify it and also said that it's not going to go away, at least this quarter. I'm just wondering when it goes away.

Yes, there are 2 things going on in defense for next quarter. There is some last time buy which is, I'll say, a relatively small portion in terms of billions of dollars, really small, if you will, but it's just needing some additional -- some parts, I guess, the last time buy program that we have. The bigger growth is really this new buy for a relatively new program or I should say an increase in the program of one of the defense contractors. So that's really where the biggest growth amount is coming from. Overall, if you look overall at the last time buy from this fab line obsolescence thing that we have going, it's a very small increase overall for the company next quarter, and it remains kind of in this low-single digit bouncing around tolerance band of revenue. It was last quarter and it will for the next 3 or 4 quarters. So we're not really getting a lot incremental is what I'm trying to tell you from the last time buy. Srini Pajjuri - Credit Agricole Securities (USA) Inc., Research Division: And then just one clarification. On the enterprise side, you said that, that business grew in December quarter. Just wondering if you think that is just demand being a little better? And also if you could give us what's your expectation for that market is in Q1, that would be great. I mean, I'm talking the March quarter.

Yes, well, I would read that as a positive sign for us in the network area from the standpoint of just the overall ordering patterns have been low and choppy, and I think the increase was a positive sign relative to what's going on in the potential IT spend area going into the new year, so our least expectation is that people are going to continue to buy equipment, so I think that's a pretty good sign. Srini Pajjuri - Credit Agricole Securities (USA) Inc., Research Division: And your expectation is that it will grow in Q1?

Yes. Again, that's back to my answer to one of the earlier questions. I think wireline is still pretty choppy, and it's not clear to me that it's sustainable yet.

Our next question comes from David Wong with Wells Fargo.

A couple of clarifications on what you said about 40, 45-nanometers. If I understand, you said you want to close in on 50% share at this node. But that includes your low-end product on 40, 45, so you would have lower than 50% market share at the high-end product, is that correct? And can you also give us some sense of what percentage your 40, 45-nanometer reps now at high end versus low end?

So traditionally, the high end has been 2/3 of the market, and the high volume or lower end has been 1/3. And what we expect is for the sum of those to enable us to get to 50%, and part of that is driven by the fact that there is no competitive -- significant competitive product at the 45-nanometer node that we're competing with. So we think we'll have pretty plain sailing and we have a lot of high volume but more lower ASP design wins there, which benefit from what is a very strong profile in terms of low cost and actually very low power, too. And that would be our projection. It requires acceleration of our growth because we're definitely starting from behind these families. We're several months behind the competition, which did a good job and came out first, which is by the way why we made sure that at 28-nanometer that did not repeat and if anything, we feel we're well ahead on 28-nanometer.

We didn't really comment on the components. We're not going to, but the 45-nanometer, our high-volume product, didn't grow as well quarter-to-quarter and as well as the high volume and we said there was 20% growth for that category quarter-to-quarter.

Our next question comes from Sujee De Silva.

Maybe can you try to, from your anecdotal conversations, handicap the likelihood of wireless infrastructure spending snaps back some time in 2012 or can it stay at this low level?

Well, my position is that it can't basically because in Palo Alto, I can't get cell coverage, right? And then in China, which I was there last week, it used to have pretty good cell coverage and now seeing a lot of dropped calls in China, and this is based on one person's anecdotal information. I fundamentally believe that what is happening is there's a gross mismatch between the infrastructure and the consumer purchases, and that this will sooner or later generate a huge problem that will need to be addressed. And if you look at North America, the big carriers are trying to address this by limiting bandwidth, throttling back. There's a whole host of techniques that are being used. But at the end of the day, the number of smart cellphones is growing at such a fast rate that, in my mind, it's going to totally overwhelm the system. Having -- and this is not a North America-only phenomenon. It's clearly happening in China, too, in a big way as the number of smartphones is growing at a very rapid rate, and we expect it to have worldwide ramification. So I just don't see how you can go through another year of not investing in bandwidth unless there's a year of not selling cellphones, right? It's just at an imbalance already at this point and it's getting worse. And the number of applications and the usage models are just more and more bandwidth-intensive. So I don't think it's sustainable and it will happen. It will happen in different rates where it's driven by government mandate as opposed to where it's driven maybe only by commercial interests, but I think in both cases, there's absolutely a huge demand.

And then on 28-nanometer, thanks for giving us the $10 million forecast. I imagine that's concentrated around the Kintex and Virtex products initially. A year out, will that be spread between the 5 product families or still concentrated among one or two 28-nanometer families?

It will be spread for sure on all 5, but it won't be evenly spread on all 5. I would expect Kintex and Virtex to be the 2 largest ones, then SSIT to be the third and then Zynq and Altix to be the fourth. And there's just an intuitive response as to how it's expected to be. And some of the significant growth in Kintex is driven by the fact that we believe it has sustainable competitive advantage in the midrange, particularly for wireless applications.

Well, thanks for joining us today. We have a playback of this call beginning at 5 p.m. Eastern Time -- 5 p.m. Pacific Time, 8 p.m. Eastern Time today. For a copy of our earnings release, please visit our IR website. Our earnings release date for the fourth quarter of Fiscal Year '12 will be Wednesday, April 25, after the market close. This quarter, we'll be holding our Annual Analyst Meeting at February 15 in San Francisco. In addition, we'll be presenting at the Goldman Sachs Technology and Internet Conference on February 16, and the Morgan Stanley Technology, Media and Telecom Conference on February 28. This completes our call. Thank you very much for your participation.

Ladies and gentlemen, this does conclude today's conference call. You may now disconnect.