My tentative plan for this next stretch is to work my way carefully through the material presented at SanDisk’s 25 February 2008 Analyst Day and post as seems appropriate. I will also probably put up transcriptions of selected presentations.
As a preamble, this week I’m posting some excerpts from three SanDisk Analyst Day Reports. I’m skipping commentary as that would be jumping the gun.
Both Bear Stearns and Lazard are guardedly optimistic with $40 price targets. Deutsche Bank is decidedly neutral with a $26 price target.
**** 26 February 2008 Bear Stearns Excerpts below *****
Target Price YE ’08: $40
Key Takeaways from Analyst Day
With regard to the current NAND market environment, SanDisk indicated that the price weakness it started to see in late 4Q07 is continuing into 1Q08, but similar to last year, they continue to expect supply-demand to improve as we progress through the year. SanDisk implemented price cuts in February which they had not planned for when they provided their guidance in January. As a result, we are lowering our 1Q08 revenue estimate from $825M to $805M, below the mid-point of the guidance range of $775-$875M, our product gross margin estimate from 24.5% to 23.0%, and our non-GAAP EPS from $0.28 to $0.24.
SanDisk did not provide an update on its royalty negotiations with Samsung, but indicated that they face two possible outcomes. The preferred outcome is a renewal of its agreement with Samsung at lower royalty rates, and if an agreement is not reached prior to August 2009, the alternative outcome is vigorous litigation. A renegotiation at lower rates is consistent with our expectation and should not come as a surprise to the market.
From a demand perspective, SanDisk expects the mobile end market to continue to be the fastest growth driver for the company in 2008. In 2007, bit shipments to the mobile segment grew 327% YoY, significantly exceeding the corporate average of 190% YoY. Demand is being driven in part by mobile operators up-selling cards to consumers in order to promote their services. Other future demand drivers that the company highlighted include video capture and SSDs. SanDisk plans to introduce MLC-based SSDs by year-end.
Despite weaker than expected pricing in 1Q08, mgmt reiterated their full-year guidance of revenue growth of 15-25% and product gross margin of 24-28% for 2008. Mgmt provided the additional color that they expect ASP and cost per bit to both decline by 50-55% YoY, and bit shipments to increase in the 150-170% YoY range.
Costs should benefit from the 43nm process transition in 2H08, and the initial ramp of 3 bit-per-cell production on 56nm. The company is also on track to sample 4 bit-per-cell NAND by the end of 2008. Based on a lower starting point in 1Q08, we are lowering our full-year 2008 and 2009 non-GAAP EPS from $1.96 and $2.39 to $1.78 and $2.22 respectively. We are adjusting our price target from $45 to $40. We believe the stock already discounts the lower estimates.
Although we expect supply-demand dynamics to remain challenging in the near-term, we believe at these levels, SanDisk stock is currently pricing in an overly pessimistic view with regards to the NAND outlook for 2008. We believe the NAND market should be balanced in 2H08, and that a bottoming in spot pricing in late 1Q08 or early 2Q08 should serve as a catalyst for the stock.
**** 25 February 2008 Deutsche Bank Excerpts below ****
25 February 2008
Price target: $26
Highlights from 2008 Analyst Day
Why buy a watch when all you want is the time?
With an uncertain outlook for NAND demand and growing concerns of oversupply in 2008 we are not surprised SanDisk chose to focus on the LT prospects at the Analyst Day. We believe SNDK is taking constructive steps but increased capital intensity and optimistic demand projections are only likely to extend industry oversupply. Although SNDK is trading at near trough valuations we believe a recovery remains uncertain and therefore maintain our Hold rating and $26 P/T.
1Q08 ASPs dropping quickly
NAND contract prices for 2H Feb were down 5% compared to flat in 1H Feb. If prices remained flat, then average contract prices would be down 25.4% q/q compared with -24% previously. Although SanDisk does not see US retail softness spreading to other geographies it was forced to take pricing actions in February, actions unforeseen in January. Not surprisingly the $/GB of retail flash cards in the US has dropped nearly 30% since the beginning of the year per our checks.
2008 revenue guidance too aggressive
SanDisk reiterated revenue guidance of +15-25% for 2008 on the basis of 50-55% ASP reduction and bit growth of +150-170%. Given our forecast of sharply slower industry growth (+9% in 2008, down from +27% in 2007) and SNDK growing slower than the industry in 2007 (+18% vs. +27%) we believe 2008 growth targets are too optimistic.
Leave P/T and estimates unchanged
Our $26 P/T equates to a P/E of 18x and P/S of 1.5x our 2008E, a slight premium to historical trough valuations to reflect our belief that the NAND flash market is near a trough and although recovery is uncertain we don’t expect further deterioration. Downside risks: excess NAND supply, share loss, and manufacturing excursions. Upside risks: market share gains, improved brand awareness of SanDisk, and development of new royalty and IP around X3 and X4 technologies.
Short-term Outlook Remains Poor
With an uncertain outlook for NAND demand and growing concerns of oversupply in 2008 we are not surprised SanDisk chose to focus on the long-term prospects for NAND flash at its Analyst Day. We believe over the long-term NAND flash revenue should outpace overall semiconductor industry growth, but we don’t believe industry supply/demand has yet bottomed and we continue to expect 2008 to be a difficult year. We note NAND contract prices for 2H Feb were down 5% compared to flat in 1H Feb. If prices remained flat average contract prices would be down 25.4% q/q compared to -24% previously. Although ASPs are likely negatively impacted by Chinese New Year we expect demand to remain tepid and industry growth to decelerate sharply from +27% in 2007 to +9% in 2008. Tepid demand forced SanDisk to take additional pricing actions in February, actions that were not foreseen when the company reported results in January. Not surprisingly the $/GB of retail flash cards has dropped nearly 30% since the beginning of the year.
SNDK achieved cost reductions of 55-56% per year since 2005, however over the last 2 years, ASP reduction of nearly 60% outpaced those cost reductions. SNDK doesn’t believe the industry can generate a ROIC if ASPs continue to drop 60% per year because the pace of cost reductions is likely to slow down as the industry completes the transition to MLC and 300mm wafers. The high cost of NAND fabs ($7-8bn for 200K WSPM) and the difficult of technology scaling will pose increasing challenges to incumbent vendors and likely dissuade new entrants. We agree with SNDK’s assertion that beyond Samsung, Toshiba/SDNK, Hynix, and Intel/Micron we are unlikely to be any new entrants in NAND flash, but even without new entrants the industry will remains very competitive as each vendor looks to grow market share and better leverage its fixed costs. SNDK is hopeful that either Intel/Micron or Hynix will slow its output growth but we think that is unlikely in 2008.
There was nothing new reported on the IP front. SNDK continues to negotiate with Samsung ahead of the August, 2009 deadline and the company remains hopeful that an agreement (with lower royalties) can be reached ahead of the deadline. If an agreement is not reached SNDK expects to eventually be victorious but only after a lengthy and expensive period of litigation. The company has had some success licensing its technology to white-label card vendors. We suspect several of the card vendors with high exposure to the US which would be negatively impacted by an ITC and/or Wisconsin decision are amendable to a settlement, but conversely those with low exposure to the US market have very little motivation to settle with SNDK.
Extending Technology Leadership
Fab output exceeding earlier expectations
SNDK continues to execute well on its manufacturing strategy as both Fab 3 and Fab 4 output is well ahead of schedule. Fab 4 reached 20K WSPM output in 4Q07 is on track to hit 110K in 2H08 while Fab 3 achieved its initial target of 150K WSPM. SNDK achieved captive bit growth of nearly +200% y/y in 2007, but it is expected to decline to +160-170% in 2008. SNDK believes that it will capacity constrained by 2010 even assuming modest bit growth of only 130% per year. As part of the its plan to grow market share SNDK and Toshiba have started plans for Fab 5 which is expected to be in production by 2010. However the investment case for Fab 5 assumes SNDK grows it market share to 20% in 2010. This suggests SNDK/Toshiba combined market share would be 40%, similar to market leader Samsung in 2007.
Fab 5 JV will be slightly different
SNDK will construct Fab 5 under a slightly modified agreement with Toshiba. SNDK will only contribute 25% of the capex for Fab 5. In exchange SNDK will receive 25% of the output at cost and another 25% of the output on a cost plus foundry margin paid to Toshiba. SNDK will have the option to convert the foundry model to a captive model at a future date, but it is unclear how much the option will cost SNDK. Clearly this option reduces the capex commitments for SNDK and adds some flexibility should industry conditions deteriorate. However the option comes at a cost and will likely negatively impact SNDK margins. The good news is SNDK should see improved costs from the steady in sourcing of B/E functions. SanDisk’s new assembly/test facility in Shanghai as expected to ramp quickly, from 10% of production in 4Q07 to 30% of production in 4Q08. In addition to lower costs the facility is expected to improve time to market, reduce cycle times, and improve flexibility.
In a tacit admission that the first generation of SSDs have not met with much success SNDK has reorganized its SSD organization. The company hired a new GM from AMD who better understand PC OEMs and the requirements of the PC platform. MLC SSDs remain on track to sample by the end of 2008 with widespread adoption expected in 2009. However it is clear that the PC is the most demanding application for NAND flash and overcoming the complexity of PC applications is likely to prove challenging. In addition to complexity, pricing remains a significant barrier as well and SNDK believes SSDs need to reach $1-1.50 per GB before PC OEMs begin to embrace the technology.
X3 expected to be 50% of bit output in 2009
SNDK is making steady progress on X3/X4 technology with 43nm X3 expected to sample by the end of 2008 and ramping quickly to 50% of bit output in 2009 (less by wafers). X4 technology on 43nm should sample in early 2009 but output will remain limited for much of 2009 as X3 and X2 continue to be the majority of bit output. SNDK continues work of 3D OTP and R/W technology but it is unclear when this technology will lead to successful commercial products. The primary competition appears to be optical media which at the moment offers more compelling density, performance, and reliability and continues to improve at a rapid pace.
MP3 market rapidly shifting to lower price points
According to SNDK the sub $150 MP3 player market is where all the unit growth is coming from and the market above $150 peaked in 2007. SNDK achieved a milestone in 4Q07, becoming the #1 brand in the sub $150 segment largely due to the Clip. The Clip is already lower priced than the Shuffle and more full featured so we do not need to cut our prices in response to Apple latest price cut. Clip is selling very well and SNDK has a strong roadmap in the sub $150 segment for 2008. SNDK believes it can be more competitive at the low-end of the market because brand matters a lot less at the low-end of the market and also less competitive on a relative basis. Above the $150 price point you have all handsets vendors, Apple, and Microsoft all competing for the same consumer, but at the low-end there are few established competitors. Despite all the promising talk we point out Sansa units were only up +12% y/y and revenue growth was likely even lower.
Financial Model Getting Tougher
Mobile phones or bust
SanDisk is now reliant entirely on the mobile phone segment to drive growth. In 2007 Mobile segment revenue was +64% y/y and represented 35% of total revenue, up from 25% of revenue in 2006. Despite unit growth of +35% and +154% bit growth (4 year high) Imaging segment revenue was down y/y. USB and A/V segment combined revenue was up slightly but revenue growth in these segments will become more difficult due to competition and market saturation. We point out that despite the success of the Sansa Clip, A/V units were up only 12% y/y. The “Other” segment shows tremendous promise but the industrial, PND, and SSD markets are still too small (3% of revs in 2007, up from 1%) to make a difference and barely offset the decline in Gaming (5% to 3% of revs) in 2007.
Capital intensity expected to increase
Capital intensity is expected to exceed 50% in 2008 and 2009, before declining to under 40% in 2010. SNDK expects to spend nearly $5.4bn in 2008/2009, up sharply from the 2007 Analyst Day forecast of $4.4bn. The increase is largely related to increasing output of Fab 4 (200K WSPM) and spending for 3D OTP memory ($400m). The higher capex comes on top of 2007 spending that was nearly $500m higher-than-earlier expectations ($1.9bn vs. $1.4bn) due to increased output at Fab 3 and a faster ramp of Fab 4. Of the $500m increase in capex nearly $350m was funded with cash. The higher capex in 2008 & 2009 is offset somewhat by lower spending in 2010 due to the revised structure of the Fab 5 agreement with Toshiba. However the revised capex will push SNDK into negative cash flow position in 2008 and 2009 and even in 2010 the company will only be modestly cash flow positive. Meanwhile SanDisk’s lease guarantees are expected to balloon from $1.1bn in 2007 to almost $2.4bn in 2010. It is doubtful that the company will be
2008 growth targets too optimistic
We believe SanDisk’s growth rate of 15-25% for 2008 is too aggressive. We point out that SNDK underperformed the industry in 2007, growing revenue +18% compared to the industry at +27%. Given our expectations of sharply lower industry growth in 2008 (+9%) we do not believe SNDK is likely to outperform the overall industry by such a large margin if at all. SNDK expects industry bit growth to slow in 2008, but expects its own output to outpace the industry. SanDisk’s forecast assumes market share gains, share gains which we don’t expect to materialize. Both Samsung and Intel/Micron should grow bits in excess of industry growth in 2008 and although Hynix may lag we don’t expect the company to lose 6 ppts of market share (implied by 3 ppt share gain for SNDK and Toshiba). Due to the continued oversupply situation we expect ASPs will be down another 60% in 2008 or perhaps higher.
Slowing retail growth a concern
We have already addressed the supply situation but given the uncertain demand environment especially in North American retail (~50% retail mix in 2007) we believe high density products will face demand headwinds in 2008 and therefore believe SNDK’s bit growth of +150-170% may prove to be too aggressive as well. We believe demand for low-density cards will show more resilience in a slowing macroeconomic environment than high-density products. Although OEM revenue continues to show strong growth (+35% in 2007) this channel typically carries lower margins (negotiating power of large customers) compared to retail and is subject to the sourcing decision of only a handful of customers.
**** 26 February 2008 Lazard Capital Markets****
Price Target: $40
SNDK: Analyst Day wrap-up — negative NAND environment
in the stock price; BUY Update
Mobile phones and SSD is the future. We expect that demand for removable memory in handsets, increasing demand for embedded and removable memory for video players, and SSD will be growth drivers for the company. We expect over 700M handsets with card slots to sell worldwide in 2008, up from 200M in 2006. Management also highlighted the SSD opportunity. We believe that up to 20% of notebooks (32M units) can be SSD-based by 2010, which is equivalent to ~40% of today’s NAND capacity. Nevertheless, we don’t see this becoming a real driver until 2009-2010 when 1GB=$1.50 in price.
Cost reductions on track. SanDisk shared with investors its view of the upcoming technologies for NAND. X3 remains on track to be a game changer in 2009 when the technology transitions to 43nm; current yields seem promising. X4 is targeted for production in 2009. Finally, its transition to 43nm in 2008 is following a similar track to the movement to 56nm in 2007. SanDisk’s cost reductions should be in the 50% range in 2008.
Financial outlook seems in line. For 2008, management indicated that they expect revenue growth of 15%-25% (unchanged from before) and OM in the range of 13%-16% of revenue. They guided for product GM in the range of 24%- 28%, despite their expectation for ASP per MB declines of at least 50% this year. For 2008, we expect revenue and EPS of $4.51B and $1.99, in line with guidance.
Valuation and risk. SanDisk is trading at 12x our forward 2008 pro forma EPS or 1.1x book. Our price target, based on sum-of-the-parts analysis, implies a multiple of 20x. The risk lies in steeper than expected NAND ASP declines. Reiterate BUY. While the short term remains challenging, we see better NAND pricing ahead in 2Q which is a tracker for SanDisk stock. As we expect a better year for SanDisk in 2008 than 2007, largely because of better cost reductions from 43nm, and Fab 4, we see the current long-term valuation as attractive.
Financial outlook seems in line. Management provided two-tiered guidance: one for 2008 and another for 2009-2010. For 2008 they indicated that they expect revenue growth of 15%-25% (unchanged from before) and operating margins in the range of 13%-16% of revenue. In addition, SanDisk guided for product gross margin in the range of 24%-28%, despite its expectations for ASP per MB declines of at least 50% this year. For 2008-2009, management indicated that they expect revenue growth and operating margins to be 15%-30% and 13-16% of revenue, respectively, with product GM of 24%-28% despite its expectations for ASP per MB declines of 40%-50% in 2009-2010. For 2008, we expect revenues and EPS of $4.51B and $1.99, in line with management guidance.
Mobile segment is the key to near-term growth. We believe that the mobile market will likely to grow north of 170% in 2008, higher than the company’s guidance of 150%-170%. The trend in mobile handsets is that many handsets are coming with high-capacity bundled SanDisk cards. In retail there is a higher attach rate with cards and phones and we see an increasing trend for mobile operators to push the sale of cards. Overall, we believe that the mobile handset segment could surpass 40% of total SanDisk sales in 2008.
Mobile phones and SSD is the future. We expect that demand for removable memory in handsets, the increasing demand for embedded and removable memory for video players, and SSD will provide growth drivers for the company. We expect over 700M mobile handset models with card slots to likely sell worldwide in 2008, up from 200M in 2006. Management also highlighted the SSD opportunity and its impact on the notebook market. We believe that up to 20% of notebooks (32M units) can be SSD-based by 2010, which is equivalent to ~40% of today’s NAND industry capacity. SanDisk believes that the revenue opportunity in the industry for SSD could be north of $6B in revenues. Nevertheless, we don’t see SSD becoming a real driver until 2009-2010 when 1GB=$1.50 in price. In addition, this will occur only when the industry will be shipping MLC SSD, which SanDisk plans to introduce later this year.
Fab 5 terms are positive for SanDisk. Management highlighted the outlook for Fab 5 that was announced last week. SanDisk elaborated that Toshiba sees NAND flash as an important strategic division and that Toshiba would like to have more control over it; therefore, they were flexible with SanDisk in Fab 5 where they own 75% of the fab. This has suited SanDisk well, as the company wanted the flexibility to drive the business in terms of capacity matching demand. Therefore, Fab 5 investment is earmarked to be $1B in 2010 rather than $2B. In addition, this gives the company an opportunity to get equal wafer capacity at cost plus, despite the fact that the company owns only 25%. Considering the heavy investment in a new 300mm 32nm 200K wafer fab of $8B, we see this deal as positive for SanDisk.
NAND down-cycle continues, though we have an optimistic outlook. While we are currently in the third year of a NAND flash down-cycle — in which ASP/MB is declining more than 50% per year — cost reductions at the major NAND players are not declining at the same rate. Therefore, we don’t see 50%+ price drops as sustainable in the long term. Nevertheless, for 2008 we believe that investors should largely focus on when NAND pricing is likely to turn around, which we believe should occur in mid-2Q.
Short term is challenging but risk-reward appears compelling. While the short term remains challenging in NAND pricing, we see better NAND pricing ahead in 2Q which is a tracker for SanDisk stock. In addition, we see the long- term NAND price declines of 50%+ as unsustainable, and we believe SanDisk is well positioned for the time when the industry shifts to more rational pricing. As we expect a better year for SanDisk in 2008 than 2007, largely because of better cost reductions from 43nm, and Fab 4, we see the current long-term valuation as attractive.