NVIDIA F2Q09 (Qtr End 7/27/08) Earnings Call Transcript

NVIDIA Corporation (NASDAQ:NVDA) F2Q09 Earnings Call August 12, 2008 5:00 PM ET

Executives

Michael Hara - Investor Relations

Jen-Hsun Huang - President, Chief Executive Officer and Director

Marvin D. Burkett - Chief Financial Officer

Analysts

Daniel Berenbaum - Cowen & Company

Joanne Feeney - FTN Midwest

Patrick Wang - Wedbush Morgan

Arnab Chanda - Deutsche Bank Securities

Daniel Ernst - Hudson Square Research

Hans Mosesmann - Raymond James

Doug Freedman - American Technology Research

David Wu - Global Crown Capital

Operator

Good afternoon, ladies and gentlemen. Thank you for holding. I would now like to turn the call over to Michael Hara, Vice President, Investor Relations. Thank you, sir. You may begin.

Michael Hara

Thanks, Kevin. Good afternoon and welcome to NVIDIA's conference call for the second quarter ended July 27, 2008. On the call today for NVIDIA are Jen-Hsun Huang, NVIDIA's President and Chief Executive Officer; and Marv Burkett, NVIDIA's Chief Financial Officer.

Before we begin today’s call, I would like to take care of some general administrative items. Your lines have been placed on a listen-only mode until the question-and-answer segment of today’s call.

During this call, we will discuss some non-GAAP financial measures about diluted net income per share, tax rate, and gross margin when talking about our results. You can find a reconciliation of these non-GAAP financial measures to GAAP financial measures in our financial release, which is posted on the investor relations page of our website at www.nvidia.com.

Unless otherwise noted, all references to market research and market share numbers throughout the call come from Mercury Research or John Petty Research.

This call is being recorded. If you have any objections, you may disconnect at this time.

Please be aware that if you decide to ask a question, it will be included in both our live transmission as well as any future use of the recording. Also, shareholders can listen to a live webcast of today’s call via the investor relations page of our website. The webcast will be available for replay until the company’s conference call to discuss its financial results for its third quarter fiscal 2009.

The content of today’s conference is NVIDIA's property and cannot be reproduced or transcribed without our prior written consent.

During the course of this conference call, we may make forward-looking statements based on current expectations. These forward-looking statements are subject to a number of significant risks and uncertainties, including statements as to our financial outlook and projections, visual computing, the GPU and CUDA, the impact, performance and availability of and demand for our products and technologies, our competitive position and our growth and strategies. Our actual results may differ materially from results discussed in any forward-looking statements.

For a complete discussion of factors that could affect our future financial results and business, please refer to our Form 10-K for the fiscal year ended January 27, 2008 and the reports on Form 8-K filed with the Securities and Exchange Commission.

All forward-looking statements are made as of the date hereof based on information available to us today and, except as required by law, we assume no obligation to update any such statements. The content of the webcast contains time-sensitive information that is accurate only as of August 12, 2008.

Consistent with the requirements under Regulation FD, we will be providing public guidance directly in the conference call and will be unable to provide significantly more information in offline conversations or during the quarter. Therefore, questions around our financial expectations should be asked during this call.

At the end of our remarks, there will be time for your questions. In order to allow more people to ask questions, please limit yourself to one question. After our response, we will allow one follow-up question.

With that, I will now hand the call over to Jen-Hsun.

Jen-Hsun Huang

Thanks, Mike. Good afternoon and thank you for joining us. Q2 was a challenge for NVIDIA. We reported revenue of $893 million and non-GAAP diluted earnings per share of $0.13, significantly below our original expectations. The simultaneous impact of desktop PC end market weakness around the world, adjustments of our desktop GPU ASPs in response to competitive pricing, and the delayed ramp of our next generation MCP substantially contributed to our revenue shortfall.

We also noted that we would be taking a non-recurring charge against cost of revenue to cover anticipated customer warranty, repair/return replacement and other associated costs resulting from a weak die/packaging material set in certain previous generation GPU and MCP products using notebooks. Although the failures are only seen in a small percentage of all the chips we shipped with this material set, the repair cost of a notebook can be expensive. In total, we took a charge of $196 million. We will continue to support on OEM partners on responding and resolving end customer issues.

We underestimated the price performance of our competitor’s most recent GPU, which led us to mis-position our fall lineup. The first step of our response was to reset our price to reflect competitive realities. Our action put us again in a strong competitive position but we took hard hits with respect to our overall GPU ASPs and ultimately to our gross margins. The price action was particularly difficult since we are just ramping 55-nanometer and the weak market resulted in taking longer than expected to work through our 65-nanometer inventory.

Throughout the second quarter, we observed a general weakness in the desktop GPU segment, consistent with the results published in Mercury Research Q208 Graphics Report, desktop standalone GPU shipments declined over 23% from the first quarter, mostly in the low-end GPU segments. We believe the decline was primarily due to a weak economic environment, driving a mix shift to lower priced PCs, as well as a continued mix shift to notebook PCs.

Until we can offer motherboard graphics for Intel platforms, our penetration of notebooks and low cost PCs will remain low. We hope to start production this quarter.

There were several bright spots in our business. We see continued strength in our professional solutions business. The Quadro business grew 38% over last year. The mobile workstation marketplace is growing at 30% a year and over the last 18 months, we’ve gone from around a 40% share of this market to over 90% share.

Despite the challenges we faced in the second quarter, we remain optimistic about our growth opportunities. There are a number of steps we are taking now and in the next several quarters to regain our growth momentum, which I will outline later in the call.

Let me now turn the call over to Marv to discuss our financial results.

Marvin D. Burkett

Thanks, Jen-Hsun. My remarks today will cover the second quarter of fiscal ’09, including comments regarding both GAAP and non-GAAP results for the quarter. Included in the quarterly results is a charge of $196 million to cost of revenue for a warranty issue, which we previously disclosed. The difference between GAAP and non-GAAP is this warranty charge and the charges for stock-based compensation. A full reconciliation is included in our press release and on our website.

As Jen-Hsun stated earlier, revenue for the quarter was $893 million, which is down 23% from Q1 and down 5% from the prior year. The issues with regard to revenue in the quarter are almost entirely focused on the desktop GPU segment. Desktop GPU revenue was down 40% quarter to quarter and down 25% year to year. The remainder of the business combined was down 10% quarter to quarter and up 11% year to year.

Desktop GPU units declined by 20% but more importantly, desktop ASPs declined by 25% quarter to quarter. Jen-Hsun has already gone over the issues here, so I won’t repeat them.

Notebook GPU revenue was actually up quarter to quarter by 8% and up 43% year over year. Professional solutions revenue decreased by 12% quarter to quarter as expected after an exceptionally strong Q1.

Memory in the quarter was down by 41% to $21 million.

The decline in both desktop unit volume and ASPs created significant pressure on margins. The 25% decline in ASPs falls directly to the margins. As a result, overall corporate margin decreased to 16.8% for GAAP and 39.1% for non-GAAP. The decline in desktop ASPs did not carry over into notebook. Notebook ASPs were relatively flat quarter to quarter.

GAAP operating expenses were $305 million, down almost $7 million quarter to quarter. We were able to start the process of limiting OpEx growth. During the quarter, we added 195 new employees and ended the quarter at 5,553 employees. Much of the growth in the quarter was new college graduates.

Depreciation in the quarter was $46 million, up about $5 million from Q1. Capital expenditures were $63 million, which is down $139 million from Q1. However, the Q1 capital additions included $150 million for the purchase of property in Santa Clara. The Q2 additions of $63 million included $25 million of test equipment.

Despite all of the issues, cash flow from operations was still positive in the amount of $90 million for the quarter. Interest income and other was $9 million, which is down about $1 million from the first quarter, primarily due to lower than average cash balances and lower interest rates during the quarter.

The non-GAAP tax rate for the quarter was 17% while the GAAP tax rate was a benefit. Going forward, we expect the GAAP tax rate to be 23% and the non-GAAP tax rate to remain at 17% for Q3 unless the R&D tax credit is renewed in the quarter.

GAAP EPS for the quarter was a negative $0.22 per diluted share and non-GAAP EPS for the quarter was $0.13 per diluted share.

There were no stock repurchases during the quarter. However, this week, our board of directors approved an increase in our stock repurchase program from $1.7 billion to $2.7 billion. Through July 27, 2008, we have repurchased 68 million shares under our existing stock purchase program, for a total cost of $1.16 billion.

On the balance sheet, cash and marketable securities were $1.66 billion, which is up $36 million quarter to quarter. Accounts receivable were $679 million, up $28 million, and DSO was at 69 days. This reflects the profile of shipments during the quarter. We have not experienced any problems with receivables.

Inventory was $432 million, up only $12 million quarter to quarter as we did a good job moderating production volumes.

Accrued liabilities, which is up $247 million in the quarter, includes the accrual for the warranty charge. There are no other highlights on the balance sheet.

On the outlook, revenue will be the most difficult to forecast. We are expecting our market for desktop PCs to grow less than seasonal. While we expect growth in MCP, Sony, and desktop, other business units may be flat to slightly down. Therefore, we expect total revenue to grow slightly.

For gross margins, we would expect small improvements until we can make the transition to 55-nanometer. This is in process but we have inventory of the 65-nanometer parts which will have to be sold first.

Operating expenses will be up from the second quarter approximately 5% but we will try to reduce that growth as much as possible.

The GAAP tax rate will be 23% and the non-GAAP tax rate will be 17%.

With that, I will turn it back to Jen-Hsun.

Jen-Hsun Huang

Thanks, Marv. The visual computing universe is expanding and is driving exciting growth opportunities for the GPU. Visual computing is no longer just about graphics. We are driving revolutionary advances in four distinct areas -- photorealistic rendering, image processing, such as video processing and computational photography, parallel processing for the simulation of complex algorithms like real world physical properties, and rich displays, such as 3D touch U.I. and 3D stereoscopic displays.

We have been innovating and investing in these four areas for some time and plan to launch products and initiatives soon.

At the core of these growth strategies is CUDA, our general purpose, massively parallel computing architecture. CUDA is a new mode in all of our GPUs, starting with the GeForce 8 family. CUDA is not GP/GPU. The programming model is not triangles and texture maps. When in the CUDA mode, our GeForce Quadro Tesla processors behave like a general purpose processor and is programmed using C and a C compiler. Nearly 100 million CUDA compatible processors have already shipped in desktop PCs, notebook PCs, and workstations. Our GPUs are CUDA compatible from top to bottom.

And the many years of products on our development pipeline are all CUDA compatible. CUDA is already the world’s highest volume parallel computing architecture and the adoption of CUDA is accelerating. Developers all over the world are using CUDA to dramatically accelerate their applications, ranging from nano-molecular dynamics, computational photography, end body simulation for astrophysics, computational fluid dynamics, computational finance, such as options risk analysis, and tomographic reconstruction.

Developers are seeing accelerations ranging from 10X to 300X. CUDA is clearly revolutionizing super-computing around the world and dramatically accelerating the time to discovery from life sciences research to the search for alternative energy sources.

Over 100,000 copies of the CUDA C compiler have already been downloaded since we launched CUDA almost two years ago, and nearly 52,000 copies of the CUDA C compiler were downloaded in Q2 alone, and the rate continues to grow rapidly.

HPC Wire readers chose CUDA as the most significant HPC, which stands for high performance computing, software of 2007. Our website links to over 100 published papers that reference CUDA, but we estimate many times that have been published worldwide.

Over 50 schools and universities worldwide now include CUDA as part of their computer science curriculum or in their research. We recently sponsored a new parallel computing lab at Stanford University and appointed the first two CUDA centers of excellence, the University of Utah and the University of Illinois at Urbana-Champaign. Both universities are leaders in parallel computing and like many others, have adopted CUDA. We are honored by their investment around CUDA and thrilled by the work that they are doing in parallel computing.

CUDA has opened up a new and exciting market to us. With our Tesla product line, we are poised to revolutionize a new class of personal super computers. Through CUDA, we have discovered a pent-up demand of researchers around the world desperate for super-computing performance but cannot afford the enormous expense to run and maintain a super computing cluster. We will even revolutionize super computing clusters. Because of the dramatic acceleration using CUDA, our Tesla product line will reduce the space and power consumption of clusters by 10 to 20 times. This translates to either more performance possible in any server room or millions in energy savings each year.

Linux Magazine likened the introduction of GPU computing to the great cluster disruption that transformed data centers in the 80s. They added that we could be at the beginning of another paradigm shift.

We will start to ship our Tesla processors this quarter. The Tesla C1060 computing is the world’s first teraflops processor, while the S1070 teraflops -- a few missing words there -- delivers four teraflops per one new configuration. Tesla and CUDA look to be change agents for a multi-billion dollar market.

One of the most important applications of CUDA is the simulation of physical properties of objects and environments. The most popular demo at the press launch of the GeForce GTX 280 was the natural motions backbreaker. Backbreaker uses NVIDIA's PhysX engine to simulate the motion of players.

PhsyX is the world’s most pervasive physics engine available on the Xbox 360, PlayStation 3, Nintendo Wii, and PC. PhsyX allows game developers to use physical world simulation on the GPU to create content automatically, making the game feel and behave like the real world.

We have registered over 26,000 developers using PhsyX and are seeing over 4,000 PhsyX SDK downloads per month. Nearly every game in development today will incorporate PhsyX simulation to enhance realism while moderating the escalating cost of game art.

The PhsyX engine runs on any PC; however, when accelerated by a CUDA compatible GPU, effects like cloth, soft body, particle, and fluid simulations can speed up 10 times over a quad-core CPU alone. PhsyX adoption is accelerating rapidly. Game engines, including Unreal engine 3, Gambrio, Trinigy use it. Middleware developers such as Natural Motion, AISeek, Speed Tree are licensees. There are over 150 titles shipping and well over 200 titles currently in development.

The visual computing revolution is happening and it is all around us. On August 25th to 27th, along with companies worldwide contributing to the visual computing universe, we are hosting NVision 2008, the world’s first mega even dedicated to the visual computing movement. This will be one of the largest events ever held in San Jose and will be attended by game developers, designers, artists, programmers, engineers, scientists, researchers, system OEMs, and gamers and technology enthusiasts from all over the world.

For gamers, NVision ’08 includes a massive LAN party and the finals of the Electronic Sports World Cup.

For artist, we’re hosting Demo Scene, Machinima using games to create movies, and digital art events.

We’ll have an exclusive showing of the new 3D movie, Fly me to the Moon, hosted by Buzz Aldrin and its director, Ben Stassen. We’ll have technical tracks for content creators, developers, artists, and scientists.

We are also hosting the emerging company summit in which 60 new companies meet the financial community to discuss breakthrough products using the power of visual computing.

We got tripped up in Q2. Some things we can’t control but there are many things we can. We’ve dusted ourselves off and we are now intensely focused on changes and improvements that will enhance our performance in the future.

For all who have followed NVIDIA over the years, you know what we can achieve when challenged. We will surely overcome this challenge as well, meanwhile continuing to drive the visual computing revolution we started.

We’d be happy to take your questions now.

Question-and-Answer Session

Operator

(Operator Instructions) Our first question comes from the line of Daniel Berenbaum with Cowen & Company. Please proceed.

Daniel Berenbaum - Cowen & Company

Thanks very much. Jen-Hsun, when you talk about gross margin improvement and you talk about improving slightly until you fully get to 55-nanometer, how should we think about that trajectory over the next couple of quarters? And then, how much of the gross margin improvement is going to be because pricing has now stabilized because you have readjusted your products to the right price points and how much of that comes from cost improvements?

Jen-Hsun Huang

First of all, with respect to the price adjustment, our price going into Computext, which was the June timeframe, was just wrong and we miscalculated the competition. Now that we’ve properly calculated and gauged the competition, we’ve reset our price to it and we have a strong position going into the fall.

So my expectation is that pricing is going to be stable going forward. From this point out, I guess the real focus is gross margin improvement.

There’s all kinds of levers we could pull and we ought to be working on all of them. You’ve seen us improve our gross margins in the past and we will improve our gross margins going forward, so the company, as you can imagine, is very, very focused in this area now and even more focused in the past, and so we just have to go figure out how to get our costs in line and our gross margins back up.

It’s not a singular thing that explains it all away. Our situation was particularly aggravated because we were the first to 65. We had such a large momentum of 65-nanometer products and when the market slowed down unexpectedly for us, we ended up with a lot more 65-nanometer capacity than -- or inventory than we like.

All of our products are now 55-nanometer and all of them are ready to ramp and many of them have ramped but until we consume the 65-nanometer capacity, or inventory, excuse me, we’re not really going to be able to benefit from the cost reductions of the 55.

But that’s not the only thing that we need to do. I mean, the company needs to do a lot more and I expect the various groups around the company to be doing a lot more in gross margin improvements.

Daniel Berenbaum - Cowen & Company

Okay. Thanks very much.

Operator

Our next question comes from the line of Joanne Feeney with FTN Midwest. Please proceed.

Joanne Feeney - FTN Midwest

I had a question about the charges and your view of what might still be out there. There are some concerns that there may be more product failure than in your opening remarks. Jen-Hsun, you remarked that you felt like you were going to continue to support the partners. Do you see that support leading to potential more write-downs in the future?

Jen-Hsun Huang

We’re not expecting more write-downs in the future. When we scoped out the problem, we had -- we felt we had enough data to project out the anticipated failures from the various platforms that are out there. This doesn’t happen to all of our chips and it doesn’t happen to most of the notebooks that are out there. There are only a few examples of them and of all the notebooks that have shipped. So we think we have a pretty good handle on the situation but -- and we thought that we were relatively conservative but we’ll see how it goes.

Joanne Feeney - FTN Midwest

Okay, and then if I could, just a question on the motherboard GPU product. You remarked that you were ramping production. When do you think that becomes -- could you give us a timeframe for when that becomes a more significant share of your revenues, and perhaps give us some numerical targets that we should expect in terms of the share of output?

Jen-Hsun Huang

I can tell you that we are building a lot of them right now. I can tell you that from a technology perspective, it is a solid year to maybe year-and-a-half ahead of what the competitor has in this space. In terms of GPU computing and -- you know, many core programmable processing on the motherboard, it is the only solution for from what I can tell several years out. So we have a very significant advantage relative to the competition.

We also have some very exciting design wins -- nothing that we can announce today but hopefully we’ll learn about them in the near future. So this is an important project for us. We’ve put a lot of engineering into it and I am expecting it to be a really exciting product.

Operator

Our next question comes from the line of Patrick Wang with Wedbush Morgan. Please proceed.

Patrick Wang - Wedbush Morgan

I just wanted to talk real quickly about desktop GPU again. Can you help us better understand what you are expecting in terms of timeline for the desktop GPU to stabilize?

Jen-Hsun Huang

Well, I wish I could tell you. Part of it has to do with the economy. Some of it has to do with the U.S. currency and what we found is that there was a pretty strong shift from mid-range desktop PC segments or products to entry PC and some of it to notebook. And so -- and we also saw that the U.S. OEMs were stronger because of the weak dollar relative to the European system builders. And our European business tends to be substantially larger in terms of GPU attach.

And so there’s a lot of things that conspired against us but at the core of it, I think that the weak economy played a pretty significant role. We didn’t really -- we didn’t lose any share. The marketplace just -- the market just got soft on us and so it’s hard to say exactly when it’s going to recover but I think that time will tell.

Patrick Wang - Wedbush Morgan

And then just a quick question on gross margins, and I guess two parts here; first off, I understand you guys have some 65-nanometer inventory to work through. How much is there and how long does it take? And then second, if Marv, you could talk about some of the moving parts on gross margins in Q3, fiscal Q3.

Marvin D. Burkett

Well, how long it takes us to work through it depends on how many we sell in each quarter. I think we’ll see some improvement in Q3 and some improvement in Q4 in the transition to 55-nanometer. We’ll have to see how quickly we work through that inventory.

Some of the moving parts on gross margin going forward is that I think as Jen-Hsun said, we’ve seen prices stabilize in the desktop GPU area, so the improvement in gross margin there is going to have to come from cost reduction. In MCT, which will become we think an increasing portion of our business, obviously it is not at the gross margins of the rest of the company, so that could be a slight drag. Offsetting that, we expect nice growth from the Sony arrangement, which is obviously very, very high gross margins, so that would be a benefit. So there’s lots of moving parts. When you put it all together, we expect some slight improvement in gross margins.

Operator

Our next question comes from the line of Arnab Chanda with Deutsche Bank Securities. Please proceed.

Arnab Chanda - Deutsche Bank Securities

A couple of questions; one, could you talk a little bit about your handheld and Tesla and Tegra product lines, what sort of revenue expectations you have, either for later this year or maybe possibly next year? And I have a follow-up, please.

Jen-Hsun Huang

Tegra is doing very, very well. The design wins are -- I think everybody would be excited about the design wins when they come out, and the engineering teams are heads down trying to ship these products. They won’t ship until the early part of next year and so I’m not expecting any revenue contribution from Tegra this year but hopefully next year it becomes a real business and hopefully a real substantial business.

Tesla, you know, this is a brand new category for the world and this is not a -- this is using GPUs for super-computing and it’s not GP/GPUs as I explained earlier but it’s the CUDA mode of GPUs. It’s an array of general purpose processors with a modern CPU instruction set that’s compilable by C and C++ in the future and this is a revolutionary product. I mean, if you look at the speed-ups, it took us nearly 20 years to improve the performance of computers by 1,000 times and yet, over the course of one year, researchers are discovering they can speed things up by 20, 50, 300 times using one GPU. And when you put more GPUs in the system, you get linear speed-up. And so this is revolutionary stuff and we are getting just an enormous amount of enthusiasm around the world. And despite the availability of Tesla, people are building homemade super computers in universities all over the world. I mean, as I traveled the world this last summer, every university I went to has a home-made super computer based on NVIDIA's technology.

So I think that this is a market that is eager for disruptive product and as such, it’s kind of hard to predict exactly how big the market is going to be. But I think that there are quite a few researchers around the world and they are all desperate for more super computing horsepower to advance their research.

And so I think my back-of-the-envelope would suggest this is a multi-billion dollar business opportunity and if you look at the number of CPUs that are sold into clusters, we have an opportunity to really, really dramatically capture a lot of that share.

Arnab Chanda - Deutsche Bank Securities

I have one follow-up on your [TT200] series. I think you said that you haven’t loss share and it’s been a pricing issue. It seems like your competitor is getting reasonably good reviews, so do you think that you are going to maintain your share but your gross margin is going to be lower because of having to lower prices? Or does 55-nanometer kind of correct that, where you can -- your margins will go up and your share will be maintained? What sort of dynamics do you expect there?

Jen-Hsun Huang

Our competition has built a nice product but nobody ever said that our product isn’t awesome and the nice things that people write about their product is that it’s well-priced. But if you take a look at our products, the GeForce GTX, it’s not only a great graphics product. Because of CUDA, we’re going to be able to accelerate physics in the hundreds of games that are out there and you saw early review, or you should be seeing early reviews of our physics processing and all of the reviews are fabulous. I mean, what’s not to like about increasing performance by 10 times? And so there’s a lot of exciting things that we are doing around physics.

CUDA also enables transcoding and video processing and we know that video transcoding is probably one of the killer apps around the world today. I don’t know how many million PCs are running all night long transcoding a move to put it onto an iPod. If we could just make it super real-time, just think about the amount of energy we are going to save for the world.

And so millions and millions and millions of PCs are spending all their nights cranking on video transcoding and we are going to make it so that you can transcode faster than real-time. And that’s a capability that we have, and we invested in CUDA and so our GPU is slightly larger.

Our computing performance, if you look at using folding at home as a surrogate, which is a Stanford approaching folding application that -- it’s the only application, in fact, that runs on a CPU, a cell processor, and GPUs. And if you compared our performance to that of our competition, it’s not just 50% faster -- it’s four times faster. And so our computing performance is really, really good and it’s an area that we’ve invested a great deal in, and it makes our die size a little larger but I also believe that this is how we are going to revolutionize visual computing.

We can’t just keep selling chips that make graphics run faster and cheaper. I mean, that’s all very nice and it’s all good but we need to advance the visual computing field in some remarkable and important way, and parallel computing is one of the most important investments that we are making.

So take a look at the stuff that we are rolling out now. People are really excited about the price points that they see and then on top of that, you get all the benefits of CUDA and GPU processing that we bring, and physics processing that we bring.

Operator

Our next question comes from the line of Daniel Ernst with Hudson Square Research. Please proceed.

Daniel Ernst - Hudson Square Research

Three quick questions, if I might -- do you see any second order impacts from the chip packaging issue on notebooks? Is there any loss of design wins that could come out of that issue with your OEM partners?

Two, with the rise of the ASUS EEPCs and all the many look-alikes and the [inaudible], the new Dell network out today, do you see that having an impact on your overall notebook market share or do you think you have a roll in those ultra-light portables? And then I have a quick follow-up. Thanks.

Jen-Hsun Huang

Frankly, on the work that we are doing supporting our OEMs to help them repair and to support their end users, frankly all of our engagement with all of our OEMs, they have been just delighted by the work that we are doing. Obviously this isn’t something we absolutely need to do but we stepped up to do it because we think it’s the right thing to do. And so each case is a little different, so we have to look at each case carefully but our open-minded approach and our good partnership approach is welcomed by all the OEMs. And so if this is going to be anything at all, it should be a positive.

I missed some of the early parts of your second question but I think it is related to the ultra mobile PC. We don’t really have any offerings in that space. We have offerings in the space that I think is probably even more important, which is the mobile Internet device. The single most important mobile Internet device today is the iPhone and that is much, much smaller and has a much more rigorous thermal and computing requirement than what is possible for a new MPC, if you will. And I know exactly how to use an iPhone and I know exactly what kind of enthusiasm the market has for it, but a UMPC, I haven’t figured out exactly how to use. So I think that the MID is a real important area. We are doing to see hundreds and hundreds of millions of MID/smartphones that come out over the course of the next several years and I think this is going to be the second personal computer revolution and Tegra is designed to hit that sweet spot, and so that’s our focus.

Daniel Ernst - Hudson Square Research

Understood -- to follow-up on both those, maybe just a few more, on the back-end, is there an opportunity to recoup some of the costs that you are having on fulfilling the OEM [inaudible] program, either warranty or recoupment from your packaging partners, where the issue derived from?

And then a follow-up on the MID question -- so Jen-Hsun, are you saying that you don’t believe the ultra mobile, UMPC market will be large enough to address or you just don’t think your technology addresses that type of computer?

Jen-Hsun Huang

We have no comments about the recouping. We don’t know and it’s too early to tell.

With respect to UMPC, that’s just not our target. I mean, there are a lot of things that are not our targets and that just happens to be one of the things that is not our target. We are much more targeted at the MID and the smartphone marketplace where we think that the segment is desperate for a well-designed computer on a chip, and that’s what Tegra is, and we also think that’s an area that is where growth could be really significant. So that’s our target and that’s our focus and frankly, from our initial engagements with the marketplace, we’re delighted by that.

Operator

Our next question comes from the line of Hans Mosesmann with Raymond James. Please proceed.

Hans Mosesmann - Raymond James

Jen-Hsun, a couple of questions; the way NVIDIA introduced GPUs into the marketplace, a single GPU at the high-end and you cascade going forward into higher volume segments -- has that strategy changed in relation to what AMD has done here with their strategy?

Jen-Hsun Huang

Hans, you might recall, we were actually the world’s first GX2 and I forget which chip we did it with -- was it 7900? I think it was 7900 we introduced the world’s first GX2.

We’ve got nothing against GX2s and recently, we just had another GX2 with the 9800 GX2. It has its advantages and disadvantages and so I don’t know that there’s any particular philosophical approach that we take here. We just have to look at the market and build the right product.

From a scalability perspective, single-chip is always better than two. I mean, if you take a look at the scalability of crossfire and you compare that to SLI, if you take your top 200 most popular games, SLI or GeForce GTX 280, the scalability across those 200 games is 100%. Every one of those 200 games are going to scale when you put a GTX on it. When you put SLI on it, those 200 games, the vast majority of them will scale, maybe 90% of them will scale. And with crossfire, a lot less will scale. And so when you use an X2 with crossfire, you are just going to see a lot less scalability. There are many games that don’t scale and there are some games that even get worse in scaling.

So you just have to take it case by case and -- but we think our approach is the right approach. The best approach is to do both. If we could offer a single chip solution at 399, it certainly doesn’t preclude us from building a two-chip solution at something higher. So I think that having the right price, right product at each price point and the best-performing product at each price point is the most important thing.

Hans Mosesmann - Raymond James

Okay, and a quick follow-up then -- thanks, Jen-Hsun -- the current headcount, are you guys comfortable with that, given the current environment? A lot of silicon valley companies have cut heads and I’m just wondering if that’s something that you guys are thinking about.

Jen-Hsun Huang

Our OpEx is too high and we need to moderate that and the management team is very focused on moderating that. The first thing to do is just stop growing the OpEx. Now, we had a lot of things that were scheduled for the second half of the year that we are very committed to and we think it’s the right thing to do, one of which was NVision 2008. This is a very, very important event for the visual computing industry to have. It’s the first one and it’s a recognition that the visual computing industry, the visual computing ecosystem or the visual computing universe or movement is here, that there are hundreds of companies that can participate in visual computing and come to a conference that’s dedicated to visual computing. So I think that’s a really, really important investment to make and we are going to continue to do that.

There are other initiatives that we are really committed to do. For example, the launch of all of the CUDA applications and products that are derived from CUDA are just coming out. You know, although researchers around the world have been benefiting from CUDA, most consumers have not and we are going to launch exciting products this second half related to that, and so that’s really, really important for us to do and it’s going to really significantly differentiate our product stream. Instead of just graphics processing, we’re going to do parallel processing and physics processing and image processing and video processing and we are just going to keep rolling those things out because CUDA makes it possible. And so those efforts need marketing support and we are committed to do that.

So there are some preplanned OpEx increases already but aside from that, we need to make sure that we do our best to moderate or improve our OpEx and bring it back in line relative to our business.

Operator

Our next question comes from the line of Doug Freedman with American Technology Partners. Please proceed.

Doug Freedman - American Technology Research

Jen-Hsun, just building off of sort of that OpEx question, you talked also about reducing costs. Are there some cost savings to be done in the cost of goods sold side for the gross margin improvement going forward? Other than just straight shrinks, are there architectural things here that can be done?

Jen-Hsun Huang

There are all kinds of ways to improve costs. You know, the -- how I start, holy cow. As I’m trying to answer your question, a list of 2,000 things just popped up in my head. So the answer is yes.

Improving gross margins is not just my job or Marv’s job. This is the job of 5,553 employees and anybody who is on our payroll needs to be figuring out how to improve our gross margins, and people who aren’t on our payroll ought to figure out how to improve our gross margin.

So the answer is, Doug, it’s an enormous issue and we are frustrated as hell that we got ourselves in this position. Some of it we can control, some of it we couldn’t but we should have done the things and we could have done the things we could have controlled, and so we are going to go do it. And so the beautiful thing about a disruptive quarter like this, from a company’s position perspective and the type of work that we are doing, nothing has changed, and our opportunity hasn’t changed. But this is a wonderful wake-up call for all of us and it’s a -- you know, no greater time to heighten the awareness of gross margin improvements at NVIDIA.

Doug Freedman - American Technology Research

In recent analyst days, you’ve put up some slides showing how increasing GPU and taking down the CPU actually improves the user experience. Have you seen any change from the system configuration side that that’s starting to happen? Any way that we can track that or anything you can do to help us understand the marketplace?

Jen-Hsun Huang

Yeah, it’s possible. If you take a look at our performance segment, I think the performance segment -- hey, guys, help me if -- I think I’m right here, but our performance segment, the performance segment increased year-over-year by about 40%. So the performance segment is growing.

Now, this last quarter, the mid range PC market really declined and the mid range desktop PC market happens to be our low end GPU market. And that part of the marketplace really, really collapsed during the second quarter. But the high-end marketplace was healthy. There we had to adjust our ASPs but the high-end PC marketplace -- not the high-end PC marketplace, the high-end GPU marketplace, the performance GPU marketplace from $149 and above, is robust and it grew a lot year over year.

So I think -- I don’t know if it’s related to shifting, nor is it that important. The most important thing is that our initiative is really about helping people understand that there are a lot of different ways to design computers and a lot of different ways to improve your user experience, and that’s really ultimately our focus, is to enhance user experience.

Operator

Our next question comes from the line of David Wu with Global Crown Capital. Please proceed.

David Wu - Global Crown Capital

On the transition to -- I know this is not the only problem that you have but let’s say by the end of your fiscal ’09 year, from what I piece together, actually the whole company top to bottom on the GPU side would be on 55-nanometer. So the die shrink part of the problem should be behind you. And I was wondering about this testing problem that you have had in the last six months. Would that be behind you as well?

Jen-Hsun Huang

Testing problem?

David Wu - Global Crown Capital

Well, I think you weren’t able to test enough good chips and some of those chips sort of got thrown away.

Jen-Hsun Huang

Oh, that’s improved a lot. The ops team is very focused on that, David, and I think that that has improved a great deal. In fact, I looked at the metrics just this morning and it’s improved a lot quarter over quarter. So I think there is still a lot of work to do and that’s one of the areas where gross margins can improve. We need to dramatically reduce the amount of scrapage that we have as a company and those are great opportunities for gross margin improvements that doesn’t affect price and it doesn’t affect our market position and only improves our earnings.

And with respect to your first comment, hopefully by the end of the year, we should be top to bottom 55.

David Wu - Global Crown Capital

Can I follow-up with one quick one? I understand from various websites that the tape-out has already occurred for the 40-nanometer part that you have been designing out of your B foundries. I assume that means you will be going to production in the new fiscal year fairly early. Is that more or less correct?

Jen-Hsun Huang

Well, we made the mistake of having a bad quarter -- let’s not make the mistake of pre-announcing our products.

David Wu - Global Crown Capital

Okay. Marv --

Marvin D. Burkett

Yes?

Operator

There are no more questions at this time. I will now turn the call over to Jen-Hsun. Please go ahead.

Jen-Hsun Huang

Thank you all for joining us this quarter and we look forward to talking to you again next quarter.

Operator

Ladies and gentlemen, that does conclude the conference call for today. We thank you all for your participation and we ask that you please disconnect your lines. Have a great day, everyone.