I thought it would be entertaining to put up some excerpts from UBS as we roll into Q3 earnings.
Uche Orji, the UBS analyst, makes his best efforts to rationalize his recent downgrade to Sell.
Apparently its bad news that NAND prices are strong (negative price elasticity), while at the same time its bad news that prices could weaken (seasonal demand peak limiting further ASP increases).
Not surprisingly, Uche neglects to mention Sanjay’s 100% bit demand growth for 2009 Deutsche Bank comments. Neither does Uche mention SanDisk’s change of heart regarding the selling of wafers.
Tuesday will tell how the quarter played out. FWIW, my take is that the news will be good- much to Uche’s chagrin- but, it will take Q1 2010 to get a read on how 2010 is going to play out.
UBS has SanDisk’s Q3 top line coming in at $776M; L&R at $113M; Product GMs at 18%; and Pro Forma EPS at $0.15.
No consideration was given to a potential inventory charge reverse.
FWIW, my take is that top line has a good shot of exceeding $900M. The top up driven by wafer sales from (excess) inventory. Although we’ll probably never know, my guess is that the majority of these wafers are destined for Apple, probably via Toshiba.
If lots of wafers have been sold at good prices, SNDK will take an inventory charge reverse.
UBS probably has product GMs right at around 18%.
Non-GAAP EPS looks likely to be the show-stopper, easily exceeding Q2’s $0.36. Probably over $0.50. In the best case, if the stars align, over $0.60.
The bigger question is 2010. Should Q3 be outstanding, the bears such as UBS will argue that recent strength is simply attributable to one time events and the bottom will fall out in Q1 2010.
UBS has product GMs for Q1 2010 coming in at 17%. UBS also is estimating L&Rs for Q1 2010 to come in under $70M. Both these numbers look low to me.
The number to watch for 2010 is product GMs. UBS expects SNDK product GMs for 2010 to be under 20%. I think UBS estimates will prove really really low, but-
Time will tell.
*** UBS excerpts below ****
12 October 2009
Uche Orji, Analyst
Nearing Seasonal Peak. Downgrade to Sell
Expecting peak demand by mid-C4Q to limit further ASP appreciation
With NAND spot/contract pricing up 24%/9% in Q3 and holiday product builds typically complete by mid-C4Q, we downgrade SNDK to a Sell rating based on: 1) negative price elasticity that could limit bit sales, 2) a seasonal demand peak in Oct/Nov that could limit further ASP increases, a key catalyst for shares, 3) GM downside risk from new, lower Samsung royalty rates taking effect. Downgrade to Sell rating with an unchanged $18 PT.
Greater exposure to retail could limit bit shipment upside
Whilst we do expect SNDK margins to improve meaningfully in 2H09, retail card sales could be negatively impacted by higher prices or fewer promotions. Relative to peers, SNDK has low embedded NAND exposure in high capacity smartphones (16GB or higher), a product that we believe is driving much of the demand strength in Q3 and similar to late Q1/Q2. Though industry supply growth is moderate, Q1 demand seasonality is likely to be unsupportive for ASPs.
GM upside could be limited by new royalty rate, non-captive mix growth
We believe GM upside offered by higher prices and cost reductions from the 32nm migration could be partially offset by the effects of lower Samsung royalty rates, which have a partial quarter’s effect in Q4, and the shift to a more balanced 30% non-captive bit supply mix in coming quarters.
Valuation: $18 12-month PT; Downgrade to Sell Rating
Our DCF-based PT of $18 is equiv to 1.3x P/BV. Sell rating. SNDK trades at a P/BV of 1.4x or consistent with a more normalized mid-cycle range.
We are downgrading SanDisk to a Sell rating from Neutral with an unchanged price target of $18. Our new view on SanDisk is predicated on expectations for: 1) negative price elasticity that could limit the potential upside to earnings as a result of higher NAND flash prices leading to weaker unit or total bit capacity sales, 2) the approaching seasonal demand peak in October/November that could lead to weaker NAND flash ASPs starting in December, and thus serves as a negative catalyst for SanDisk shares given that it historically has tracked memory prices, 3) potential gross margin downside risk from the company’s royalty revenue stream which will see a pronounced decline starting in the December 2009 quarter and again in the March 2010 quarter due to new
Samsung royalty rates taking effect, and despite ongoing manufacturing cost reductions that should help to maintain or improve gross margin especially in a benign to improving price environment. Furthermore, SanDisk is also on track to move back to a more balanced fab-lite model where 25-30% of bits are sourced from external suppliers and thus could limit the upside to margins in coming quarters.
Negative Price Elasticity
We estimate that approx 85-90% of SanDisk’s revenues are from actual product sales and the majority of that comes from retail or bundled card sales vs. embedded NAND sales into mass storage applications. This large retail exposure leads SanDisk to be highly dependent on consumer trends and price elasticity. Though it is evident from company statistics over the past 7 years that an increase in ASP per bit will lead to a decrease in NAND content or bits per card sold, the R^2 is only 0.50. Since SanDisk is vertically integrated and currently purchases less than 5% of its bits sold from external suppliers, we believe the company has limited pressure to raise prices in order to maintain or improve gross margins given cost reductions enabled by the 32nm migration. However, depending on the financial health of SanDisk’s competitors and how much product margins are already compressed due to higher NAND contract prices, SanDisk may opt to follow competitors in raising prices or pursue less discounting in coming months. While it is unclear whether SanDisk would take advantage of the current environment to gain share, we do believe profitability is a key objective and maintaining market share.
We note that SanDisk does have a very modest amount of exposure to embedded NAND applications such as smartphones through its iNAND product family. Our own retail checks indicate that retail memory cards are still being discounted modestly whilst micro form factor cards are in some cases no longer being bundled with handsets/smartphones compared to the common practice of including at least a small capacity card. We believe much of the shelf space for flash products is devoted to USB flash drives and SD/Compact flash cards and generally less space for microSD or Memory Stick Pro Duo. We do note that SanDisk had the most shelf space of any other retail brand and fewer brands were competing for shelf space as a whole.
Seasonal Demand Peak Approaching
The manufacturing of products slated to be sold during the late November through year end holiday selling season typically concludes by late-October to mid-November and typically coincides with the peak in NAND flash price appreciation. Based on DRAMeXchange data, the NAND spot ASP M/M % change in October over the past 5 years was -5%, and compares to the up +9% seen in 2009 thus far. In the two month period ending in November, the 5-year average shows a -14% decline while the three month or quarterly Q4 average over the last 5 years is a -25% decline.
2005 was the last time ASP trends were up in October and the 2 month period ending in November. That period was marked by the strong growth of portable media players as a new demand catalyst, but seasonality still quickly came into play by December leading to a Q/Q ASP decline of -3%. We expect a similar trend to play out over the remainder of 2009 given that retailers are likely to not build significant amounts of inventory and the ramp of new smartphone products is likely timed to coincide with holiday launches and sales promotions.
New Lower Royalty Rate Could Affect Margins
From a gross margin perspective, we believe SanDisk’s core product manufacturing and sales operations should see stable to improving gross margin in Q3 and Q4 especially in a benign to improving price environment. From an overall gross margin perspective, the December 2009 quarter will be the first to take into account the new Samsung royalty rate (partial effect of 1 month) that is approximately half of the prior rate. Royalties remain a key or only source of profitability in SanDisk’s overall business given how competitive the retail NAND market remains.
Furthermore, SanDisk is also on track to move back to a more balanced fab-lite model where 25-30% of bits are outsourced and thus could limit the upside to margins in future quarters. This is not likely to have a significant impact until internal inventories are reduced to a more normalized level. If SanDisk’s procurement of external NAND supply coincides with further increases in NAND pricing due to limited supply growth, then we believe the incremental revenues could have lower margin associated with it.
From a valuation perspective, SanDisk is trading at a P/BV ratio of 1.4x and compares to trough-peak range of 0.3-8.0x with a more “normalized” range of 1.5-3.0x. The current 1.4x ratio is also in line with Micron’s current P/BV and historical range of 1.0-2.0x. On an EV/Sales basis, SanDisk is trading at 1.3x or at a discount to Micron’s 2.1x vs a “normalized” range of 1.5-3.0x which is similar to Micron. We believe that with a move to a more balance fab-lite manufacturing model that includes 25-30% of bits procured from outside suppliers and a royalty revenue rate from Samsung that is effectively half of the prior rate, SanDisk may not garner the same peak multiples as it has in prior cycles given the changing nature of its operating model.