18 October 2007 SanDisk Q3 Conference Call
Eli Harari, Chairman of the Board, and CEO
Judy Bruner, Executive Vice President, and Administration & CFO
Sanjay Mehrotra, President, and COO
Lori Barker Padon, Senior Director of IR
Lori Barker: Thank you. Good afternoon and welcome to the financial teleconference for SanDisk Corporation for the third quarter of 2007. I’m Lori Barker, SanDisk’s Senior Director of Investor Relations. Joining me is Dr. Eli Harari, Chairman and CEO of SanDisk, Sanjay Mehrotra, President and COO, and Judy Bruner, Executive Vice President of Administration and CFO.
The agenda for today’s teleconference is as follows: Eli will start with an update on the industry and SanDisk’s overall strategy, Sanjay will follow with operational and business updates and Judy will end with our third quarter financial results and future guidance. We will conclude the teleconference with your questions for Eli, Judy and Sanjay.
Now I’d like to turn the call over to our CEO, Eli Harari.
Eli Harari: Thank you Lori.
Demand was remarkably strong in the third quarter, with high volumes straining our supply chain. We currently are selling more than one million cards per day, approximately half of which are for the mobile card market which barely existed three years ago and which we believe is only just beginning to take off. The star product performers in Q3 were mobile cards and USB flashdrives. Also, our international sales are growing rapidly and reached 67% of total product sales in the third quarter. The strong output from Fab 3 has allowed us to gain market share in key geographies where we were previously severely supply constrained and our financial performance is becoming much less dependent on U.S. consumer sentiment.
I want to share with you our current thinking on SanDisk’s demand and supply requirements in 2008 and beyond. We expect the #1 demand generator for Flash memory in 2008 to be storage in mobile handsets and we believe that this market is entering an explosive growth phase with a barrage of new multimedia handsets that will be voracious users of embedded Flash and Flash cards. We are very well positioned as a “one stop shop” supplier of both the mobile OEM and the mobile retail fronts. In the third quarter, more than 40% of our mobile revenues were generated through about 100,000 specialized mobile retail outlets that now sell our mobile products.
In emerging markets, we see the solid state drive (SSD) market as a key strategic long-term investment. This overall market is projected to grow nicely in 2008, accelerating in 2009, and by 2010 or 2011, we project that Flash storage in computing could become the second largest user of Flash memory after mobile handsets. We believe that our Flash systems expertise, and in particular MLC implementations as well as our extensive system IP, will become important differentiators in this market as it unfolds over the coming years.
Adding to the Flash demand equation are existing markets such as imaging, USB flashdrives, audio, video, gaming, navigation, and emerging markets such as enterprise, home entertainment, automotive and content distribution. We believe that our combined projected demand from these markets will put us among the top three consumers of Flash memory. In 2008 and beyond, we believe our projected demand will likely outstrip our total supply even with some reliance on non- captive sources.
We plan to address our 2008 supply requirements in the following ways. In Fab 3 we expect to maximize output through the following steps: first, complete the conversion to 56 nms (nm) in Q1 2008, second, convert a portion of the output of 56nm to X3, or three bits per cell, starting in March or April 2008, and third, begin the conversion to 43nm in Q2 08. In Fab 4 we plan a fast ramp on 56nm and we expect to begin conversion to 43nm as soon as 56nm yields have stabilized. In parallel, we expect to re-enter the non-captive market as a more significant customer in 2008.
In 2009, we expect to continue ramping Fab 4 to its maximum capacity. However, beyond 2009 we believe that we will need substantial new captive capacity if we are to retain our market leadership and competitively serve the very large markets that I just discussed. Given that a new captive fab takes 18-24 months to plan, to build, equip, qualify, and begin to ramp, we have commenced discussions with Toshiba for a potential joint Fab 5 for late 2009 or 2010 production starts. The timing for such a fab would be such that it would not require significant new capex outlays before the second half of 2009, at the earliest. Also, under a previously announced Memorandum of Understanding, we are in discussions with Hynix for a potential minority investment in one of their 300mm Flash fabs. If that should transpire, the Hynix investment would be on a much more modest scale than a potential Fab 5 with Toshiba.
As you can see the pace is picking up. We are addressing opportunities in Flash markets of substantial scale, and these entail substantial investments and risks. However, not all Flash competitors are the same, and I believe that SanDisk’s captive supply and retail presence, technology leadership and growing consumer brand position us very well in pursuing our vision for sustained growth and market leadership in the coming decade.
I will now pass it on to Sanjay.
Sanjay Mehrotra: Thank you, Eli.
Q3 was a very busy quarter for SanDisk with continued new product introductions and the new manufacturing plant openings of Fab 4 in Yokkaichi, Japan and our first captive assembly and test facility in Shanghai. We made excellent progress in achieving market share gains in Europe and Asia and further lowering our unit costs.
In the third quarter, overall demand for our products was strong with total worldwide unit sales in the retail channel doubling on a year-over-year basis, driven by strong unit sales increase in USB drives and microSDTM cards. Further, demand for our high-capacity 2 Gigabyte and 4 Gigabyte mobile cards was strong and the supply of these cards was constrained during the quarter as we could not ramp up backend assembly and test capacity fast enough to meet the accelerated demand. In the retail channel, during Q3 we shipped more mobile card units than we did in all of 2006. In fact, unit sales of the microSD card were at an all time record and it was the largest single product in the SanDisk portfolio. The mobile OEM channel grew nicely and continues to contribute more than half of our mobile revenue, however, revenue growth in the retail mobile channel accelerated and grew at an even faster rate. Going forward, we believe that video in mobile phones will become another strong driver for our high capacity products with shipments recently started of multimedia 5 mega pixel new phones like Nokia N95, Sony Walkman K850 and others.
USB flashdrive revenues have resumed strong growth as a result of combining, in the first quarter of 2007, our legacy SanDisk and msystems USB flashdrive teams under a single global organization. We realized solid sales of USB drives in the U.S. market during the back-to-school season and we also made excellent progress in overseas markets by expanding our channel reach, offering segmented products, and tailoring our products to each region.
According to GFK Europe, our overall retail market share in cards and USB flashdrives grew by more than 350 basis points in the first two months of Q3 compared to the first two months of Q2 and now our market share stands at about 28% – the highest share we have achieved in the Europe region for cards and USB flashdrives. I am pleased to report that for the first time ever, we achieved #1 European USB market share in August 2007. We also observed strong market share trends in the APAC region. As a result of our global sales push, our retail revenue in each of the regions of Europe, APAC, and Japan grew by more than 50% sequentially.
On the manufacturing front, Fab 3 continued to lead the industry in cutting-edge manufacturing technology. Fab 3 has now achieved its target capacity of 150,000 wafers per month and we are achieving excellent yields on our 56nm process technology on both 8 Gigabit and 16 Gigabit devices. We expect 56nm technology to account for more than two thirds of our fab bit production output by the end of 2007 and we are well underway to begin the transition to 43nm technology in the Q2 2008 timeframe. Further, work on 32nm technology development is proceeding well.
Industry wide assembly capacity for high capacity micro-sized mobile cards with multiple Flash memory stacked die is not sufficient to keep up with the rapid growth in demand. In that regard, our timing for our captive assembly and test facility in Shanghai is proving highly beneficial to us. The Shanghai facility is on schedule for a meaningful production ramp in 2008.
As we look to the seasonally strong fourth quarter, we believe demand will be strong for our products and the benefits of low-cost 56nm products should enable us to position and price our products attractively for holiday shoppers. We expect continued revenue growth in our Sansa line of media players with the addition of new products such as the Sansa Clip, a very small audio player with superior value, and the Sansa View, our first video-enabled media player. In addition, our high capacity and high performance products in imaging, mobile and computing markets will also continue to play an important role in our growth in Q4 and beyond.
With that, I will hand over the call to Judy.
Judy Bruner: Thank you Sanjay and Eli and good afternoon everyone.
Our strong third quarter performance is highlighted by exceptional megabyte growth in both retail and OEM channels and in particular for our mobile handset and USB businesses. Our 54% sequential megabyte increase in Q3 is an outstanding growth rate for the summer quarter. This growth reflects the continued development of the mobile handset mega-market for Flash memory, and market share gains in Europe and Asia in both cards and USB drives, the result of our strategic focus on growing share outside North America. The 16% sequential decline in our ASP per megabyte, a significant improvement from the 26% decline in Q2, stemmed primarily from Q3 price quotes provided to our OEM customers in Q2 and retail price reductions implemented in June which had a full quarter impact on Q3. We took very little pricing action during the third quarter.
Our retail revenue increased 25% sequentially, with megabytes sold up 46% and ASP per megabyte down 14%. Retail revenue was 66% of product revenue during Q3, a comparable mix to the second quarter. On a year-over-year basis, our retail revenue grew a strong 36% despite a 63% year-over-year decline in ASP per megabyte. Our year-over-year retail growth reflected an increase of nearly 100% in units sold and 267% in megabytes sold. Our strongest retail growth was in the mobile handset market, where retail unit sales grew more than 300% over the prior year. The average capacity of a mobile card sold in retail in Q3 was over 1.2 gigabytes, up 114% on a year-over-year basis. Our retail sales were also very strong in the USB market with both unit sales and average capacity up over 100% year-over-year. On an overall basis our average capacity in retail was 1.6 gigabytes, up 10% sequentially and 83% year-over-year, with the average growth rate influenced by the increasing mix of mobile cards which are at a lower average capacity than USB drives and camera cards, although growing at a very strong clip.
Our OEM revenue increased 32% sequentially, with megabytes sold up 82% and ASP per megabyte down 26%. On a year-over-year basis, OEM revenue grew 37% with megabytes sold up 206% and ASP per gigabyte down 57%. OEM growth was dominated by the mobile handset market, although we also had nice growth from bundled camera cards, cards for GPS devices, and a small contribution from solid state drive sales.
License and royalty revenue for the third quarter of $119 million increased 52% year-over-year and 11% sequentially. This growth was slightly higher than we had forecasted due to higher MLC mix and better SLC pricing in our licensees’ sales. Our Q3 non-GAAP product gross margin of 26.4% improved 7.4 points from the second quarter. The decline in our cost per megabyte was similar to the decline in our ASP per megabyte, with memory cost declines reflecting increased scale and the beginning of the transition to 56 nm product. The primary contributor to the improvement in our gross margin was a reduction in inventory charges. With inventory balances lower and price movement less aggressive than in the first half, inventory charges were significantly down, for both excess product issues and lower of cost or market adjustments. Another favorable factor in Q3 gross margin was the receipt of a one-time insurance payment for claims we had filed related to a fab power outage which occurred in the first quarter of 2006. Partially offsetting the favorable insurance payment was Fab 4 start-up costs. We began production in Fab 4 in September, ahead of schedule, resulting in start-up costs charged to cost of sales rather than R&D in the month of September. The non-captive memory mix in our shipments remained below 5%.
Non-GAAP operating expenses were $199 million, up $29 million from Q2. The growth in R&D came primarily from Fab 4 start-up costs in July and August and increased engineering materials, masks and NRE. In sales and marketing, the major contributors were variable co-operative selling costs, and higher spending on branding, merchandising and collateral materials. In addition, all expense line items were impacted by the reinstatement of salary increases which had been frozen from Q1, as well as higher spend on our IT infrastructure, and increased hiring. We are carefully increasing our headcount and absolute expense investment, while still reducing our operating expenses as a percentage of revenue, targeting to be in our long-term model of 15-18% in 2008.
Other Income came down sequentially, in line with our previous forecast, due to one-time gains in Q2 and a higher mix of tax-exempt investment income. Our non-GAAP tax rate came down to 35.1% on a year-to-date basis and to 32.3% for the quarter as our profits increased in low tax rate jurisdictions.
On the balance sheet, cash and investments, including long-term investments, remained unchanged at $3.2 billion. We generated $339 million in cash from operations and $43 million in cash from employee stock programs, and we invested $324 million in Fabs 3 and 4 and $82 million in property and equipment. Our Flash joint ventures drew down $131 million of operating leases bringing our net operating lease guarantees to slightly above $1 billion.
Our accounts receivable are up $126 million and our inventory is down $56 million compared to the second quarter, both due to significantly higher shipments in the third quarter. You can also see the impact of the higher shipment level in our deferred income in the liability section of our balance sheet, which is up $61 million. Given the increased rate of sell-through, channel inventory is at a lean level of 7 weeks, and inventory on our balance sheet stands at 73 days, which is a lower level than we would prefer heading into Q4.
Before turning to our outlook, I’d like to point out that we achieved a non-GAAP operating margin of 15.7% in the third quarter, moving into the 15-20% operating margin range that we established as our 2008 target model at our February 2007 Analyst Day. And on a year-to-date basis for 2007, we are at 10.7% non-GAAP operating margin, slightly above the 0-10% range that we predicted for 2007 at the Analyst Day. We believe we have executed exceptionally well through the pricing volatility of recent quarters, and the investments we have made in capacity, technology and human resources are paying off in market share gains and improving profitability.
I’ll now turn to our outlook. We expect strong demand for our product categories in the Q4 holiday season with growth primarily limited by petabyte availability given the completed expansion of Fab 3 and our low inventory levels at the end of Q3. We expect megabyte growth in the fourth quarter to be in the range of 45%-60% sequentially, bringing our megabyte growth for 2007 to approximately 200%. We expect price decline in Q4 to be moderately more than the 16% decline in Q3, driven primarily by Black Friday sales and other holiday promotions. The annual 2007 decline in ASP per megabyte is forecasted to be in the low 60 percent range. We expect our Q4 product revenue to be in the range of $1.05 to $1.2 billion, and we forecast Q4 license and royalty revenue to be about the same as Q3.
We expect fourth quarter gross margins to benefit from a higher mix of lower cost 56nm memory, with this impact partially offset by a full quarter of Fab 4 start-up costs and Q4 price declines. We forecast non-GAAP Q4 product gross margin to be in the range of 27% – 30%. We expect Q4 GAAP product gross margin to be lower than non-GAAP by approximately two percentage points due to the inclusion of stock compensation and acquisition-related purchase accounting charges.
We forecast Q4 non-GAAP operating expenses to be in the range of $240 million to $250 million with the growth from Q3 reflecting planned spending on TV and other advertising, increased holiday merchandising, higher investment in certain R&D areas, partially offset by the elimination of Fab 4 start-up costs from R&D, and increased spending on G&A infrastructure including the start up of a new ERP project. Q4 GAAP operating expenses are expected to be higher than non-GAAP by an amount similar to the $35 million delta in Q3.
Other Income for Q4 is expected to be approximately $5 million lower than Q3, due primarily to a continued shift to more tax-exempt securities which yield a better after-tax return. We expect the non-GAAP tax rate for the fourth quarter to be 34%-35%.
On the balance sheet, we forecast that Q4 inventory levels will be approximately flat compared to the Q3 level as we ship all we that can for the fourth quarter, but with work-in-progress increasing as we ramp Fab 4. We expect to make Q4 investments in Flash joint ventures of approximately $200 million, and we expect to guarantee new operating leases of approximately $130 million in Q4.
In summary, we are very pleased with the improved top line and bottom line performance in Q3 and we look forward to a similarly strong Q4.
We will now open the call for your questions.
Craig Ellis, Citigroup: Thank you and congratulations on the quarter. Judy, can you clarify how much second sourcing you’re assuming in the bit growth guidance of 45% to 60%?
Judy Bruner: At the moment, we are looking at the potential of getting some more non-captive supply in the fourth quarter. However, it is not clear that we will be able to get it in time to influence the fourth quarter. Let me turn that over to Sanjay and he can elaborate.
Sanjay Mehrotra: Yes, our non-captive purchases in Q4, as Judy said, if we are able to realize those, would very much depend on the timing of those purchases and of course the competitive pricing of those purchases. In all, we would expect the non-captive purchases in Q4 frankly to still remain below 5%.
Craig Ellis, Citigroup: Ok, and Judy, you’re saying that that isn’t included in the guidance, or it is included in the guidance?
Judy Bruner: That level is included in the guidance.
Craig Ellis, Citigroup: Ok, and secondly, just given the pricing mechanics in the third quarter, I would have expected royalty guidance to go up sequentially. Can you just elaborate on that?
Judy Bruner: For our primary licensee, they have not made any clear comments that we are aware of as to what their pricing was in the third quarter. We have tried to make our best estimate based on their bit guidance for the third quarter, their bit reports for the third quarter, as well as our estimates of their pricing and that’s included in the total overall license and royalty.
Craig Ellis, Citigroup: So are you saying that you are assuming that pricing didn’t go up for them in the third quarter?
Judy Bruner: We believe it may have gone up a little bit but we don’t have a specific number.
Craig Ellis, Citigroup: But you will have that when you close out the fourth quarter, obviously?
Judy Bruner: Obviously, yes,
Craig Ellis, Citigroup: Ok, and then a follow-up for Eli; Eli, can you just provide some commentary about how we should think about the ramp-up of x3 and how much of the mix could be x3 by the time we exit next year?
Eli Harari: For x3, it’s looking pretty good on 56nm and we don’t really see any issues with qualification and ramping it up. It is designed to be pretty much broadly based, in other words, general purpose, not limited to just benign applications, but I would say that by the end of 2008, the majority of our products, the majority of our megabytes will still be two bits per cell, partially because we will along with x3 on 56nm also be aggressively transforming the 56 to 43nm on two bits per cell and I believe at the 43nm, MLC is still more cost effective than the 56nm x3.
Craig Ellis, Citigroup: Ok, that’s helpful, and lastly for me, you sounded confident that the company would have to increase its second sourcing next year. Can you just elaborate on what you are seeing that leaves with you that degree of confidence?
Sanjay Mehrotra: If you look at our Fab 3 production, which has now in terms of wafer output, maxed out at 150,000 or so wafers, and Fab 4 is just starting to ramp up and will contribute meaningfully in 2008 timeframe. And you look at our total demand requirements next year, we believe that we will not be able to fulfill all of that demand through just captive supply. We will be needing some non-captive supply and the demand drivers that are really requiring that mix of captive and non-captive we believe in 2008 are primarily driven by mobile, the recent applications of video, et cetera in mobile phones we believe will drive high capacity storage in mobile phones, as well as all of our other markets, particularly supplying to international growth, we believe will drive strong demand in 2008.
Craig Ellis, Citigroup: Alright. Thanks everybody.
Jim Covello, Goldman Sachs: Great. Good evening. Thanks so much. A couple of quick questions; what are you guys estimating for a supply growth for the industry for 2008 that leads you to believe that you will need to source captive supply, especially after what Samsung did with the CapEx increase last week?
Eli Harari: I think that we believe that we are entering a new baseline level, particularly with the handsets. I think that there’s not enough realization in the market of how strong and how fast this market is moving in terms of requirements for storage. It may have been kicked off by the launch of iPhone, but really until today, a very small number, I’d say probably 30% of the phones that can use a card actually use it. We are seeing a very, very strong move now to retail, to after market purchases of cards, one-, two, and four gigabyte cards, and we think, as well as embedded. And the numbers there are so large that if you just double the number of cards and then double the megabytes per card, you very easily I believe can outstrip whatever Samsung is talking about or the other players.
And then on top of that, of course, the 200 millimeter supply that is being retired next year. That is very real. So I think that overall, 2008 could be a year where there’s a good balance between the managed supply, even with aggressive ramping of new capacity.
Now, there could be first quarter seasonality, weakness that we’ve seen in the past, but for us at SanDisk as opposed to our competitors, we see us being able to use everything that we can make and that’s not being sufficient to meet our forecasted requirements for next year.
Jim Covello, Goldman Sachs: And then just a couple of other quick follow-ups; first, on the solid state drive penetration for 2008, do you have any kind of percentage maybe of the on the notebooks, what you think solid state drives could represent of total notebook shipments in 2008? And could you offer some perspective on solid state drives in the data centers, when you think that might become a reality? And then my final question is, any update on the Samsung royalty negotiations? Thank you.
Eli Harari: Solid state disk, as I said, we believe is going to be a very, very substantial market and it definitely is very strategic. But until it moves to the pricing of MLC, it’s not really in my opinion going to take off. And so far, everybody’s solid state disk, including our own, is based on SLC, and as you know, SLC component pricing has been very high up to this point and in that regard, it is definitely not helping this market in this stage to go beyond the seeding stage.
I think in 2008, hopefully we will be able to move the market to MLC and I think at that point, the pricing becomes much more realistic and much more attractive.
As far as the enterprise space, I think you asked, that is definitely market that eventually would be, you know, a significant market for solid state disk. But in general, we still think that 2008, although it becomes a real market for solid state disk, is still relatively small compared to the growth in cell phones.
People are always looking, you know, where’s the catalyst for 2008, and I could tell you for SanDisk, the catalyst very simply is cell phones and international. All those other markets of course we are addressing, but in relative terms, the handset market is so immense. The growth there is so strong and again, international, because we started from a very low market share and very low presence, represents for us in 2008, a very major opportunity no matter what happens.
I don’t mean no matter what happens, but under reasonable conditions and of course, we don’t know about the Beijing Olympics and the U.S. elections and so on.
Jim Covello, Goldman Sachs: Thank you.
Eli Harari: As far as Samsung, there is really nothing new to report. Everything that I said last time, still applies. I think both companies would like to come to some kind of resolution before the current cross license expires and that still applies.
Ahmed Kapoor, Piper Jaffray: Thank you. Just turning again to the handset market, you mentioned that you sold about $42 million mobile units in Q3. I guess doing the math, that looks at about 15% penetration. How should we think of that penetration going forward and what are some of the key drivers you are looking at?
Eli Harari: I would not say 15% penetration. I would say that if the total number of handsets sold this year, if 1/2 billion cell phones have card slots, assuming let’s say 100 million handsets a quarter with a slot, 42 million units is not 15%. It is much, much higher than that.
Ahmed Kapoor, Piper Jaffray: Ok, so…
Eli Harari: Because not all cell phones have a card slot. We are, you know, we started a campaign that we call “wake up your phone” to educate the market about what that card slot is all about, because we think that most consumers today still don’t really know that they can in fact put four, eight gigabyte into the cell phone and convert it into a very attractive music player. So there is education that has to happen but I think the heavy lifting as far as standardization and slots and embedded is behind us. I think the iPhone is helping driving that awareness amongst consumers.
Ahmed Kapoor, Piper Jaffray: OK. That actually leads to my follow-up question in terms of most of the phones with memory card slots we see out there seem to be capped off at 2 gigabit densities. What do you think are some of the bottlenecks out there and what do you think is necessary to get those densities up?
Eli Harari: We had a bottleneck last year that most of these phones were designed not to take the SDHC, the high capacity SD, which caps at 32 gigabyte. I believe that the vast majority of cell phones today are already designed to handle the SDHC, which can go from 4 to 32 gigabyte.
Other than that, it’s the cost of the cards but that is coming down, so I think that, Apple in fact eliminating the 4 gigabyte version of the iPhone and just going with 8 gigabyte is in fact helping us go in the right direction.
I think at 3GSM in February in Europe, you are going to see a huge number of phones that have a lot of gigabytes.
Ahmed Kapoor, Piper Jaffray: Thank you.
Harlan Sur, Morgan Stanley: Thank you and great job on the quarter, the execution. Your guidance, megabyte shipments up 45% to 60% in Q4, given the constraints, obviously, but if you didn’t have these constraints, what do you think the megabyte growth would have been to support the true demand in the fourth quarter?
Judy Bruner: We’re not going to speculate on what the exact growth rate would have been. However, I will say that we believe it could have been somewhat higher if we had ample supply.
Harlan Sur, Morgan Stanley: Ok, and given the supply constraints, it sounds like you are pulling in the ramp of Fab 4. What is the barrier from motivating the team to accelerate the ramp of 43nm technology? Is it just too early days in the technology development phase?
Sanjay Mehrotra: Actually, we are making very good progress on 43 nm technology and as Eli mentioned, that we will have 43nm technology in Fab 3 starting to ramp up in the Q2 timeframe. So we are working as fast as we can in applying 43nm technology to the products, qualifying the products, and getting them ready for production ramp to begin in Q2.
Harlan Sur, Morgan Stanley: Ok, great, and then just one last question. The status of your MLC controller development for SSD laptop applications. Is the team still on track to bring that technology out in the first half of next year?
Sanjay Mehrotra: Yes, we are working aggressively on bringing, applying MLC to SSD and we plan to have products ready by mid next year with MLC SSD.
Harlan Sur, Morgan Stanley: Ok, great. Thank you.
Paul Koster, J.P. Morgan: Thank you. I have three related questions. The first one is, is the capacity constraint that you see uniform across the industry, both with a short-term perspective and a sort of ’08 perspective? In other words, are you going to lose market share to someone else during this period?
Sanjay Mehrotra: We do not believe that we will lose market share during this period. The capacity constraint, despite our strong output from our fab, is on the memory side as well as somewhat on the assembly and tests, and we are aggressively working on eliminating some of those bottlenecks and accelerating the output through our supply chain to really fulfill the strong demand that we are seeing for our products at this point.
Eli Harari: I think that there could be some temporary market share losses in some regions. We have the the U.S. usually around Black Friday, the demand really goes through the roof and we sometimes have to take care of that. But we ought to be able to recover that, if that happens.
Paul Koster, J.P. Morgan: I think a number of investors will probably be wondering why, given the supply constraints, the gross margins aren’t shooting up higher. Can you just explain why they are lower than this time last year when the circumstances were very different, perhaps over supply was occurring last time around?
Judy Bruner: Paul, I would say that the primary reason, if you compare gross margin expectations for Q407 to gross margins in Q406, which is I think what you are asking, is that pricing has come down 60% plus, and costs have not come down to that level in the last year.
Eli Harari: Year over year, I think we said pricing came down about 63% in retail and our costs, although came down very nicely, did not match that level and that’s the real difference.
Judy Bruner: Now, the gross margins we expect will go up from Q3 to Q4, as I said, primarily because there will be a higher mix of 56nm product in our Q4 cost of sales. But again, the cost reduction year over year is not enough to keep up with the price reduction in the low 60s range.
Paul Koster, J.P. Morgan: My last question is we would ordinarily expect a depletion of channel inventory in the fourth quarter, and it sounds like it could be even more the case this year, given the supply constraints. What does this imply for the beginning of ’08 when I think people are concerned that over supply rears its head?
Eli Harari: Yes, I mean this, you know we are just entering the real heavy selling season for the holiday sell season. I think the next two months will tell where things stand at the end of December, so I think it is premature to speculate on what it is going to be, and certainly premature to speculate about Q1. We know typically, as you know, that seasonally it is weaker than the fourth quarter, but how much we don’t know, because overlying the seasonality, there is also the underlying very strong growth that I mentioned.
Paul Koster, J.P. Morgan:Thank you.
Tristan Gerra, Robert W. Baird: Hi. Good afternoon. We see a trend where there the central OEMs are increasing, embedded NAND Flash content in some instances, de-emphasizing the bundling of the card. That’s the case for the Nokia 95, for example. The [?] version doesn’t even have a card slot. Could you give us a sense of the margin differential between your embedded business and your flash card business? And also, your view on the USF initiative and if that takes off, what is the potential licensing implication for you?
Sanjay Mehrotra: With respect to the embedded business, we are definitely looking forward to 2008 our embedded business in mobile phone increasing. We are seeing strong design wins for our iNAND products, as well as we have said before, in the beginning of 2008 we will start shipping multi-chip packages through our Qimonda joint venture.
The N95, 8 gigabyte phone with embedded flash that you indicated only gives a halo effect to increase awareness and in making the consumers aware that high capacities are needed to fully utilize the features of these phones. That is by the way, one example of an N95 model; there are other N95 models with card slots as well.
So while embedded in mobile phones, we expect it to increase for our business in 2008 the card slots as well as card bundling for OEM and the retail aspect of our business will continue to be strong drivers of mobile growth in 2008.
Eli Harari: On the UFS initiative, the spec is not out. The publicity, the advertising is that this is going to be the greatest thing since sliced bread. It will have phenomenal performance and it will do things that no other card has ever been able to come close. We’d like to see what happens.
As far as licensing or IP, I believe that we’ll have to of course see what the UFS does, but if it’s a device with flash trying to emulate a disk drive or a mass storage device, I think that it will more likely than not be covered by our very extensive licensing patent.
Judy Bruner: Tristan, let me also address the question you asked about the gross margin of embedded card products versus removable card products for the phones. The answer is that there is not a significant difference for us in gross margin whether we sell an embedded product or whether we sell a card that’s bundled with the phone.
Tristan Gerra, Robert W. Baird: Very helpful. Thank you.
Daniel Amir, Lazard Capital Markets: Thanks a lot. Congratulations on a good quarter. A couple of questions here. First of all, on the new testing and assembly operations in China, Sanjay, can you comment a bit about how it’s going and also what the potential margin impact this could have on your business in 2008?
Sanjay Mehrotra: So the assembly and test operation in Shanghai is going very well. We are very pleased with our ability to introduce new technologies for high capacity mobile cards in this factory.
Our goal is ultimately for this assembly and test factory to meet 30% of our unit requirements; of course, that will ramp over a period of time, over the next few quarters.
In terms of the factory’s ability to meet costs, it is very cost competitive but in addition to costs, this factory gives us the benefit of again being able to introduce fast to market, new advanced technology products as well as gives us the flexibility to meet the increases in demand that takes place in our business in times such as Q4.
But overall, we are pleased with the cost competitiveness as well as the production ramp of this facility.
Daniel Amir, Lazard Capital Markets: The margin impact?
Judy Bruner: The way that I would think about this is that producing a portion of our cards at this captive factory with very competitive costs is just one of the ways that enables us to reduce the non-memory portion of our products and try to keep the non-memory cost reduction in line with the memory cost reduction, which is achieved through the technology transition. So this is just one of the many ways, but as Sanjay said, the cost is very competitive.
Daniel Amir, Lazard Capital Markets: Ok. A follow-up question. It seems like now that you are starting production here in x3 next quarter, can you comment on where x4 stands in terms of potential production and how should we look at the future of both technologies as a major source of your future production?
Sanjay Mehrotra: So regarding x4, as we have said before, we are applying x4, we are implementing the designs of x4 on 43nm technology. We will have samples of x4 in 43 nm technology toward the end of the 2008 timeframe, and any production starts will really take place in the 2009 timeframe.
x3, Eli already mentioned that we will be ramping x3 in 56nm production in 2008. It will primarily be 56nm. In 2009 it will be 43nm, and x3 will become an increasing portion of our production output next year. The majority next year will still remain 2 bit memory, even by the end of next year, and as we ramp up in 43nm x3 in 2009, its percentage as part of our total production will increase.
x3 has been designed with a total focus on reliability and performance and we plan to apply x3 over time across all of our product lines. x4 technology will likely be applied first in certain products where its performance and specifications can be best utilized, such as possibly audio/video players.
Daniel Amir, Lazard Capital Markets: Ok, thanks a lot.
Vijay Rakesh, Oppenheimer: Just a couple of questions here. As you look at the non-captive mix for next year do you think it grows? It looks like it’s less than 5% range here in Q3? Do you think ’08 we go back to the 10% to 15% range?
Eli Harari: Yes, somewhere in that range in ’08 timeframe. We are not prepared to put a stake in the ground yet for this, but we do believe that we need non-captive, somewhere around that range in the 2008 timeframe.
Vijay Rakesh, Oppenheimer: As you look at x3 and 43nm here, will x3 get to 5% of output by 3Q08? Do you think you can get it to that level by 3Q08 or you still think it’s going to be a majority is 2 bit?
Sanjay Mehrotra: As we have said, by end of next year a majority of the output will be 2 bit, but x3 will continue to ramp during the course of the year.
Vijay Rakesh, Oppenheimer: Is 5%,10% a reasonable assumption for Q3?
Sanjay Mehrotra: I think in the Q3 timeframe we will likely have somewhat more than 10% in x3.
Vijay Rakesh, Oppenheimer: Ok great. Thanks a lot guys.
Gurinder Kalra, Bear Stearns: Thanks. As far as your x3 and x4 is concerned, how do you think about the licensing of the technology? Is there a part process which might suggest that let’s keep it to ourselves and Hynix, and not license it out to anybody else?
Eli Harari: x3 IP is somewhat different than x4 IP. x3 IP is very similar to MLC, 2 bits per cell. x4 is quite substantially more complex technology and has its own relevant IP. So I think I would look at x3 and x4 somewhat differently. But at this stage I’d rather not comment on that.
Gurinder Kalra, Bear Stearns: Ok. My second question, I know you can’t comment about Q1 and I hear you on the demand you are seeing, but we are looking at a lot of incremental supply coming on sequentially as we go into Q4 and then Q1, and clearly spot pricing has been declining so the chip level pricing doesn’t look that great. Obviously Q1 last year was pretty painful. What makes you so confident that Q108 will not have a similar level of pain?
Eli Harari: I’ve said it many times. Let me say it again. We believe that our markets are very young and have really tremendous growth opportunities for us over the next three to five years. We are really planning our future and our entire strategy is based on a fervent assumption that these markets are going to grow at a very fast pace and that we need to plan our destiny based on a very fast growth trajectory, a vector ramping up, a pretty steep vector. We cannot slow down in any one quarter, make quarterly course corrections that may look good for our quarterly results in any particular quarter if that means that we sacrifice the big opportunity that we see that we think is very, very unique.
We are very unique as a company in that we manufacture all our own flash. We’re in a pretty good position, I believe, to use 100% of our output throughout the next several years and we have the ability to use other people’s flash but we are not a component supplier [missed phrase]. We are definitely building new markets.
I just talked about the need eventually to have fab 5. There is no way that we can be a significant player in the solid state disk business in 2010 and 2011 based on the output of fab 3 plus fab 4 plus anything that we can get from the non-captive market. So these markets are huge, and it’s just a parallel front of these markets growing at a very fast pace, and we just have to plan for that based on our conviction in those markets.
So yes, we worry a great deal about what happens in December or in January or by the first quarter and we manage the business as best we can through these difficult quarters, but in the long term, the strategy has proven to be very successful for us, and we don’t plan to change that model.
Gurinder Kalra, Bear Stearns: Thanks Eli. Great execution.
Edwin Mok, Needham & Co: Ok. Thank you. First question is regarding fab 3. Since you guys reached 150,000 wafer starts is there any possibility that you guys can squeeze more efficiency out of that fab and actually increase your wafer starts more in the fourth quarter?
Sanjay Mehrotra: It may be possible to increase it a little bit, but really not by much.
Eli Harari: We’ve asked all operators to lose a few pounds so they can squeeze by in the aisles. I’ve never seen a Japanese fab. It’s so tight, it is just unbelievable.
Edwin Mok, Needham & Co: Ok. I see. Just a follow-up question also on the supply constraint. Am I understanding correctly that you have supply constraint because you can go and a buy non-captive supply, they’re not qualified within your product? Is that why you guys can’t go and buy supply from other suppliers? Because it seems to me that there’s additional supply coming online in the fourth quarter. Is that correct?
Eli Harari: Until very recently, non-captive supply was extremely expensive and would be very difficult for us to conduct our business purchasing at the price that was out there.
Secondly, I have to tell you without naming any names, the quality of flash MLC is really not the same and we do believe that the Toshiba SanDisk MLC quality is significantly ahead of what we can buy there, and we have to be very careful with what we buy. There’s a lot of inferior product that is dumped on the market at any price, and of course, consumers are not aware of that. We need to make sure that if we are in this market and buy substantial output of non-captive, that our controllers are sufficiently powerful, potent in terms of error correction and so on to make those up to our standard. It’s not that simple.
Edwin Mok, Needham & Co: Do you think that going forward though, in 2008, you guys might want to increase non-captive, but you might be limited by the same issue at all?
Eli Harari: First of all, yes. We’ve said that we are looking at increasing our non-captive. We think that the non-captive model is a very healthy model. It’s a good model. It probably will not be the 70/30 that we’ve had in the past, because I think frankly the numbers are so high looking forward, that 30% of our capacity being non-captive would be a huge amount of supply for somebody else to supply us.
But yes, moving more towards the 15%, 10% to 15% non-captive, I think it is something that will be very healthy for us.
Edwin Mok, Needham & Co: One final question. In terms of the mobile phone market, looks like Viper ST (?) was very strong. Can you put any color on the embedded side, any new design activity that you guys thought was seminal or in terms of growth in that area, what are you guys expecting for the fourth quarter?
Sanjay Mehrotra: I cannot really specifically call out the design wins, but I can tell you that for our iNAND products we do have design wins, and as I said earlier, we plan to increase shipments of our embedded memory in the mobile phone in the 2008 timeframe.
Edwin Mok, Needham & Co: Great. Thanks a lot or answering my questions.
Daniel Gelbtuch, CIBC World Markets: Congratulations on a great quarter. I was wondering if could you give some color about or maybe hazard a guess as to where you see ASP per megabyte going down or declining in 2008? Do you think it’s going to be hopefully much better than this year in the 60s? Can we get back to an equilibrium level of 40 to 45 or 50?
Eli Harari: Honestly, I have no idea. If you had asked me at the beginning of the year what would be the pricing for 2007, it went down really low in February/March. It went way up in the summer months. It’s now heading back to where it was in February/March. I’ve never seen that before in the market, and I would not be able to predict it. I have no idea.
Daniel Gelbtuch, CIBC World Markets: I’d say on a grand scale and a bigger scheme over the course of a year do you think the demand vectors are so strong, coupled with the supply shortages that you envision, you think that we could have a better, more stable year, I guess, in the bigger picture?
Eli Harari: Let me say that based on the technology, on the cost of the 50nm technology compared to the 63nm technology, clearly the cost structure of 50nm on 300mm in high volume will allow the kind of pricing that you are seeing in the market, as we speak, for those companies that have that kind of technology and production.
Daniel Gelbtuch, CIBC World Markets: Gotcha. OK well thank you very much.
[closing comments] Eli Harari: Ok, so, I believe this is a very exciting time for SanDisk. You’ve seen our thoughts of our growing markets, cards for mobile phones and industrial markets. Industrial sales. We are very happy where we are in the technology innovation. I’d like to thank everybody for listening in today. Have a nice evening.