This last week was interesting.
On Tuesday, Lazard Capital Markets (LCM) essentially confirmed that SanDisk is indeed inside Apple’s iPhone and upcoming iPad- for 32GB capacities at least. I have posted LCM details here.
Lots of positives there and interesting angles, but those will have to wait for another day. Suffice it to say that small form-factor SSDs for mobile devices are already here.
On Wednesday February 17, Morgan Stanley (MS) released a 14 page analyst report on SanDisk- upgrading the company to “Overweight” with a $37 price target. Excerpts from that report follow this post.
While MS points to a $37 price target as its “base case,” two other scenarios for 2010 are included on page 2- a “bear case” with an $18 price target; and a “bull case” with a $48 price target.
For the bull scenario, MS posits $4 EPS for 2010 with a multiple of 12x. MS gets kudos for slipping in that $4- while maybe not so probable, not a number out of reach.
Elsewhere in the report, MS notes that SanDisk is currently trading at 12x consensus and 9x MS estimates (NTM) – at 5 year trough levels. MS also notes that SanDisk has historically traded within a P/E range of 12x to 25x.
In other words, in MS’s view, if the stars were to align- even $48 could rather be low if SanDisk were to be awarded a higher multiple. Nobody’s saying that yet, but its there for the whispering.
Investment Thesis, Key Value Drivers, & Key Debates
MS’s Investment Thesis:
“We think the risk-reward is attractive in next 12 months as 2H10 undersupply in NAND flash memory and demand upside from accelerating smartphone and solid state drive (SSD) adoption sets the stage for moderate price decline and margin upside for SNDK. As such, we believe management guidance and consensus estimates for 2010 appear low and will likely result in positive estimate revisions over the next several quarters.”
MS’s Key Value Drivers:
“Faster reduction of production costs compared to competition. Rollout of new flash-based products, especially Gen 3/4 SSDs.”
MS’s Key Debates:
“Is NAND in oversupply or undersupply this year? We think moderate undersupply.
Are +30% product gross margins sustainable? We think yes.
Is valuation still attractive despite strong run-up last year? We think yes.”
Give MS credit. From the looks of this report, they seem to understand the SanDisk story and its potential for 2010.
To my mind, the SanDisk story circa 2010 seems to come down to market NAND supply-demand. MS is ahead of the curve in acknowledging the prospects of undersupply- not balance.
If this is the way it turns out, undersupply could drive strong NAND pricing, which in turn could drive strong ASPs which in turn should drive product margin strength. The tighter the supply, all the better for SanDisk for 2010.
Its worth noting that MS is expecting new and improved “Gen 3/4 SSDs” from SanDisk. Also MS thinks SanDisk’s guidance appears low. I’m inclined to agree.
So here are the three MS scenarios, from negative to most positive:
MS doesn’t spend much time on this scenario, as they clearly think it unlikely- unless the market crashes.
“Bear Case $18 [price target]
1.0x CY10 book value per share
Further deterioration in consumer spending spoils 2010 NAND supply-demand recovery.
• Lower than expected smartphone growth ~10% on consumer spending concerns
• Oversupply of 10% on aggressive tier 1 NAND capacity additions in 2H10 and/or below 70% bit demand growth
• Severe price decline of ~50% – 60%”
This is the scenario that MS thinks most likely and the scenario on which they base their “fair value $37” price target.”
“Base Case $37 [price target]
12x times our base case CY10 EPS of $3.10
NAND undersupply in 2H10, demand upside from smartphone adoption and moderate price decline.
• Smartphone growth of 37%
• Undersupply of 3% on disciplined 15-20% tier 1 NAND capacity additions in 2H10
• Moderate price decline of ~30%Y/Y
• Cost reduction of 40% from planned transition to 24nm technology
• SNDK bit growth of 70%
• Bit demand growth of 80% – 90%
Product gross margins of 34% – 35%”
MS believes that smartphone growth is now the biggest driver of NAND demand and that smartphone adoption will accelerate in 2010 with units increasing 37% Y/Y and continuing to grow at this pace through 2013.
MS feels NAND memory content in smartphones will nearly double to 13GB in 2010. Overall, MS estimates that NAND in smartphones will grow 156% Y/Y and represent over 25% of overall NAND demand in 2010. See graphic above.
MS thinks tablet PC’s, including the Apple iPad, could drive meaningful NAND consumption in 2010 and 2011, up to an additional 2-4% of total demand, given the high level of NAND content likely to ship in such devices.
Interestingly MS believes that SSD growth from the enterprise market presents an upside for SanDisk in 2010. This might simply be wishful thinking or something else. We shall see. I expect we’ll get an SSD update at next week’s SanDisk investor’s day.
Looking at MS’s key numbers shown above, undersupply feels right, as does SanDisk’s cost reduction of 40% and bit growth of 70%.
The key will be NAND pricing. If MS is right and there is a ~30%Y/Y price decline, product gross margins of 34% – 35% should be in the bag.
If the NAND prices decline by less than 30%, the numbers shift towards MS’s bull case.
The bottom line of MS’s base case for 2010 is an EPS of $3.10. This is predicated on a number of factors, but the most important is a NAND pricing decline of 30%±. If NAND prices decline less, SanDisk’s GMs will increase accordingly and EPS for the year could approach or exceed $4.
MS’s bull case:
“Bull Case $48 [price target]
12x times our bull case CY10 EPS of $4
NAND undersupply throughout 2010, demand upside from smartphone adoption and stable prices.
• Better than expected smartphone growth above 45%
• Undersupply of 10% on limited capacity additions in 2010
• Stable prices of down 10-20%
• Cost reduction of 40% – 50% in 2010 from faster transition to 24nm technology
• SNDK bit growth of 80%-90%
• Bit demand growth of +100%
Product gross margins of +35%”
MS has thrown all possible upsides into their bull case.
While it would certainly help to get smartphone growth above 45%, NAND undersupply of 10%, a faster transition by SanDisk to 24nm, SanDisk bit growth at or above 80% and/or bit demand growth of +100%, the most likely bull scenario boils down to NAND pricing.
If NAND pricing in 2010 declines 20% or less and ASPs only decline 20%, $4 EPS for SanDisk looks like a distinct possibility- even with a SanDisk cost reduction of 40% and bit growth of 70%.
It’s worth noting that if we are in for another period of undersupply in 2010, pricing could swing more that anticipated
Eli’s Oppenheimer comments on how much NAND pricing can swing triggered by only a 5% imbalance in demand/supply:
“This [demand/ supply balance] is very sensitive. I mean the things that drove pricing up by a factor of 100% was a 5% imbalance in favor of demand. The demand was 5% higher than supply. The prices doubled.”
So there we have MS’s three scenarios. Unless the world economy collapses MS feels SanDisk will be very profitable in 2010. How profitable is very profitable? That answer seems to hinge on NAND pricing.
**** excerpts from Morgan Stanley 02.21.10 report below****
February 17, 2010
Upgrade to Overweight on
Rating Equal-weight to Overweight
Price Target NA to $37.00
Margins to drive significant earnings upside. We anticipate 400-500bps of gross margin upside on stable NAND pricing driven by undersupply in 2H10. As such, we view management’s full year guidance as conservative. Our revised CY10 estimate of $3.10 is 34% above consensus.
What’s New: Recent industry checks suggest NAND pricing to decline modestly in 1Q or down 10% Q/Q. Our prior EW rating was based on a much steeper pricing decline of down 20-30%. In addition, our updated proprietary capex model suggests memory capex is skewed more towards DRAM this year ~65%, which boosts our confidence in disciplined NAND supply.
Where we differ: While we remain cautious on the group, we prefer semis with exposure to strong product cycle fundamentals. We view SNDK as a key beneficiary of solid-state drives for enterprise use and tablet PCs (iPad) in 2H10. We believe these two products can generate 10-15% of incremental demand in addition to smartphones, which consensus sees as the primary NAND demand driver.
Valuation looks attractive on revised estimates and recent pull back. SNDK stock has declined ~17% YTD on expectations reset of seasonal demand weakness.
Risk-reward looks attractive at ~1:2 at current levels, with $48 as our bull case and $18 as bear case. SNDK trades at 9x our revised CY10 EPS estimates or significantly below its 5-year historical range of 12 – 25x.
Where we could be wrong: Faster than expected NAND supply by Tier I NAND makers.
What’s Next: SNDK analyst day February 26
• We think the risk-reward is attractive in next 12 months as 2H10 undersupply in NAND flash memory and demand upside from accelerating smartphone and solid state drive (SSD) adoption sets the stage for moderate price decline and margin upside for SNDK. As such, we believe management guidance and consensus estimates for 2010 appear low and will likely result in positive estimate revisions over the next several quarters.
Key Value Drivers
• Faster reduction of production costs compared to competition.
• Rollout of new flash-based products, especially Gen 3/4 SSDs.
• Is NAND in oversupply or undersupply this year? We think moderate undersupply.
• Are +30% product gross margins sustainable? We think yes.
• Is valuation still attractive despite strong run-up last year? We think yes.
• Stabilization of NAND ASPs.
• Faster adoption of smartphones and solid state drives.
• Supply constraints at competition, caused by yield issues.
• Oversupply led by aggressive capacity addition by NAND makers
• Risks: SanDisk operates its business by anticipating a 40-50% annual decline in the price per Megabyte sold, and the company continues to be a cost leader,
with solid execution on its technology roadmap. However, it has little control over price decline if competition decides to a) add more supply or b) becomes
aggressive on pricing in 2010…
Summary & Conclusions
We are upgrading SNDK from Equal-weight to Overweight with a new price target of $37, representing 35% potential upside. While the stock has tripled from trough levels last year, we believe consensus estimates do not fully reflect potential upside to NAND flash product margins – amid a backdrop of stable pricing and favorable supply-demand fundamentals – and will be revised higher through the year. In addition, we think many investors view memory stocks purely as short-term “cyclical plays” and could be overlooking the longer-term secular drivers for NAND flash memory usage in smartphones, tablet-style portable devices, and solid-state drives (SSDs).
Why upgrade to OW now?
We turn constructive on SNDK based on recent checks that suggest NAND flash pricing will track above our prior expectations for a seasonal decline of +20-30% in Q1. In addition, our updated proprietary capex model suggests memory capex is skewed more towards DRAM this year ~60-65%, which boosts our confidence in disciplined NAND supply. Our bullish outlook on NAND flash demand is also consistent with MS Wireless Equipment team’s recent forecast of above consensus 37% Y/Y smartphone growth over the next 2-3 years. (See Ehud Gelbum’s report “Smartphones: Top of the Second, Not Bottom of the Ninth” dated 1/4/2010 for
As a pure play on NAND flash supply-demand, we think above positive market trends uniquely benefit SNDK over the next 12 to 18 months by supporting a) moderate price declines for SNDK’s retail card and OEM component products, b) 35% or better product gross margins on cost reduction from transition to 24nm technology, and c) modest investment in capex and JV’s to strengthen the company’s balance sheet. We also think SNDK’s actions from last year to reduce captive wafer supply and develop a broader OEM channel should continue to move the company towards a less capital intensive and sustainably profitable model over future cycles.
Guidance Appears Conservative, Raising FY10 Estimates
Given our bullish outlook on NAND fundamentals this year, we think management guidance for 2010 appears conservative, and we adjust our estimates above full year revenue and margin guidance. This lifts our FY10 EPS estimate to $3.10 or 34% above consensus (see Exhibit 1). We arrive at our new PT of $37 by applying a 12x multiple, the low end of SNDK’s historical trading range and the average for our global memory
MS FY10/11 Estimates 30-50% Above Consensus
Rev EPS GM* Rev EPS GM*
MS (Old) $3,888 $2.00 36.7% NA NA 34.4%
MS (New) $4,706 $3.10 39.7% $6,038 $3.90 38.7%
Consensus $4,350 $2.32 37.3% $4,789 $2.26 35.9%
* Combined royalty and product gross margin
Source: FactSet, Morgan Stanley Research Estimates
Debate 1: Is NAND memory in oversupply or undersupply this year?
Market’s view: Supply and demand will be balanced.
Our view: We see moderate undersupply due to lack of new NAND capacity and demand upside from smartphones, tablets, and SSDs.
We believe NAND will be in undersupply for the next 12-18 months. After three years of oversupply and +60% annual price declines, we believe SNDK and the flash industry overall are in a “sweet spot” of balanced supply and demand. SNDK and peers have been reluctant to add new capacity after returning to profitability and appear focused on repairing their balance sheets and pursuing sustainable growth.
The Morgan Stanley global memory team estimates a 3% undersupply for NAND flash memory in 2010 based on 70% bit supply growth and 80% bit demand growth (see Exhibit 2). We think memory makers are proceeding cautiously with plans to add new wafer start capacity and instead direct capex investments to node shrink and multi-level cell (MLC) technology that effectively increase “bit” production without requiring construction of costly new fabs. We believe investors underestimate the degree to which supply for NAND is constrained in the near-term given 1) low level of NAND equipment shipments and 6 -12 month lead-time to ramp new capacity after orders are placed, and 2) reluctance by SNDK and peers to make significant capex investments given their desire to repair balance sheets on the back of the prior cycle.
ROIC for NAND makers just approaching attractive levels.
The prior investment cycle for NAND capacity in 2007-08 was preceded by sustained profitability of 10-15% ROIC for SanDisk and other Tier I NAND makers (see Exhibit 3). After experiencing losses in 2008 and historically low levels of profitability for most of 2009, SanDisk’s 4Q09 product gross margin reached 36% and ROIC broke through the 10% threshold at which we believe the company would consider adding new capacity. SNDK and peers Samsung and Toshiba have reinforced this view of limited supply growth in 2010 in recent public announcements and analyst calls.
In SanDisk’s 4Q earnings call, the company announced plans to spend $700M to $900M on capex and JV investments in 2010, compared to $1,857 and $1,800M invested in 2007 and 2008, respectively. We believe the bulk of SanDisk’s outlay will be put towards technology spending and 10% incremental wafer capacity utilizing empty clean room space in Yokkaichi Fab 4. Until industry ROIC reaches 15% historical peak levels sometime in 2H10, we believe SanDisk and peers will remain cautious on aggressively adding new capacity. Moreover, given current equipment shipment levels, which are 50% below peak, and 6-8 months delivery lead times, we think NAND supply will be short of demand through end of 2010 and likely to mid-2011 (see Exhibit 4).
In addition, our proprietary semiconductor capex model shows memory related spending in 2010 will be skewed 65%/35% towards DRAM over NAND, which makes us more constructive on NAND supply-demand this year (see Exhibit 5). We do not believe DRAM capacity is fungible for NAND use given wide process technology gap.
Accelerated smartphone growth remains biggest driver of NAND demand. MS Wireless Equipment team believes that smartphone adoption will accelerate in 2010 with units increasing 37% Y/Y and continue to grow at this pace through 2013 as popular examples like Blackberry, iPhones, the Motorola Droid and Palm Pre are joined by the introduction of newer midrange models in coming years.
Additionally, we believe the average NAND memory content in smartphones nearly doubles to 13GB in 2010 as consumers increasingly rely on these devices as a replacement for their digital music player, camera and video recorder for everyday use (see Exhibit 6). iSuppli estimates that an average of 35.2GB will be used in each iPhone sold in 2010, possibly representing the largest market for NAND flash. Overall, we estimate that NAND in smartphones will grow 156% Y/Y and represent over 25% of overall NAND demand in 2010 (see Exhibit 7).
We believe the new category of tablet PC’s, including the recently announced Apple iPad, could drive meaningful NAND consumption in 2010 and 2011, up to an additional 2-4% of total demand by our estimates. Given the high level of NAND content likely to ship in such devices, successful launches of similar new tablet products could make this category a significant driver of NAND demand in the next 2-3 years.
SSD growth from enterprise market presents upside. Currently, SSD’s are most commonly deployed in niche enterprise storage and mobile applications where performance benefits outweigh their high relative cost. Our checks indicate that IT departments are increasingly adopting these solutions as they recognize the productivity gains afforded by faster read times, power savings and durability from SSD’s. And while SSD’s are still not mainstream, we estimate that NAND used in these products will grow by 280% Y/Y and represent ~10% of overall NAND demand in 2010.
Low expectations for consumer SSD adoption until 2011. We do not expect SSD’s to meaningfully replace hard disk drives (HDD) in consumer PC’s and notebooks in the near term given a price per GB that is currently ~10 times that of traditional HDD’s. As such, we believe that NAND component prices will have to fall significantly from ~$2 per GB to below $1 per GB for SSD’s to be competitive in the marketplace with traditional HDD’s. With NAND in shortage over the next 12-18 months, we would not expect prices to reach these levels until 2011 at the earliest.
Where we could be wrong. A significant deterioration of consumer demand, particularly in US and Europe would negatively impact our smartphone and SSD demand assumptions.
Leading indicators. NAND equipment shipment rates, smartphone unit data, and SSD penetration rates.
Debate 2: Are +30% product gross margins sustainable?
Market’s view: No. Lower royalty revenue stream and use of non-captive supply will cap margin upside.
Our view: Yes. Undersupply and transition to 24nm technology gets us to 35% product gross margins in 2H10.
Stable NAND pricing. NAND flash memory prices have declined 50-60% annually in the past five years from rapid supply growth and 50% annual reduction in manufacturing costs achieved by NAND makers through process technology transitions and increasing scale. We believe contract prices bottomed in December 2008, doubled in 2009 on absolute basis as NAND makers retired 200mm old capacity and cut utilization rates to unprecedented ~70% levels (see Exhibit 8). On a per bit basis, NAND pricing was down 20%Y/Y in 2009.
Given our expectations for undersupply this year, we expect NAND flash pricing per bit to fall 20-30% Y/Y in 2010, or still half the historical rate of annual decline. While seasonal consumer buying patterns could result in moderate 5-10% near-term decline in absolute pricing, introduction of new smartphone and mobile internet devices with high NAND content such as the Apple iPad or 64GB iPhone by mid year should drive absolute pricing for mainstream 32Gb NAND components higher starting April, in our view.
Transition to 24nm further bends SNDK’s cost curve. SNDK lowered NAND manufacturing cost per bit by 52% in 2009 (similar to prior years) through transition to 3-bits per cell and 32nm process technology. In 2010, we expect the company to lower costs by another 40% by completing the transition to 32nm (was 10% of output in 4Q) and beginning transition to 24nm process technology in 2H10. We expect cost reductions to exceed price declines by ~10% in 2010 (see Exhibit 9).
Expanded OEM sales should benefit margins and accelerate growth. We believe investors underestimate the positive impact of SNDK’s expanding OEM channels and sales to private label card resellers in Asia to margins. SNDK’s share of revenue from OEM customers increased to 56% in 4Q09 from 40% in the year ago period through increased penetration with mobile handset manufacturers and private label card resellers (see Exhibit 10 and 11). We believe the expanded OEM presence represents an attractive opportunity for SNDK due to 1) “stickier” base of tier I customers such as Apple and Nokia that prefer qualified, reliable suppliers like SNDK, 2) increased exposure to high growth emerging markets in Asia through “white label” channel, and 3) higher margins on component sales to OEMs compared to packaged products due to lower associated marketing and sales expense.
Restructured captive supply lowers inventory risk. SanDisk procures NAND flash wafers through multiple joint ventures (JVs) with Toshiba, whereby the two companies share investments in manufacturing capacity and technology development. In January 2009, SNDK and Toshiba transferred 20% of the JV’s wafer capacity to Toshiba, effectively reducing SNDK’s captive supply by 10%. Although the restructuring reduces SNDK’s captive wafer capacity, we think potentially lower margins related to purchasing wafers from non-captive sources is more than offset by lower inventory risk and reduced fixed costs of the company’s “fab-light” model.
Royalty revenue set for upside surprise. SNDK licenses various NAND flash IP to Samsung and other memory makers. With the renegotiation of the Samsung royalty agreement in May 2009 to half the prior rate, we estimate overall royalty revenues decline roughly 10% Y/Y in 2010 with possible upside if NAND price and demand exceed current expectations, as we anticipate. Royalties contribute positively to the overall margins.
Where we could be wrong. Process issues delay transition to 24nm node and delay cost reduction roadmap, weaker than expected consumer demand causing oversupply in NAND
Leading indicators. Progress updates on node transitions. Industry checks on NAND demand,
Debate 3: Is valuation on SNDK still attractive despite strong run-up last year?
Market’s view: No, market is already pricing in balanced supply/demand and management guidance.
Our view: Guidance appears conservative and our estimates are above street. On a P/E and EV/S basis, valuation looks attractive.
Price target of $37 represent 35% upside. SNDK shares have roughly tripled from 2009 trough levels on a 100% rebound in NAND flash memory prices, return to +30% gross margins, and recovering demand primarily driven by accelerating adoption of smartphones with increasing NAND content. Although the best entry point was clearly 9-12 months ago, we believe SNDK shares remain attractive on conservative management guidance and consensus estimates not recognizing likely upside to margins and NAND undersupply from multiple demand drivers. On our price target of $37, we see 35% upside from current levels.
Company guidance and consensus estimates appear low. In SNDK’s 4Q09 earnings call, management guided FY10 overall revenues to $4.0 – $4.4B, royalty revenue to $320M – $360M, and product gross margins to 31% +/- 3%. Full year operating expense was guided to $725M – $750M and other income to $40M. Given our constructive view on NAND flash market and company fundamentals, we model FY10/FY11 revenues estimates of $4.7B/$6.0B above consensus of $4.35B/$4.8B and EPS estimates of $3.10/$3.90 above consensus of $2.32/$2.26 (see Exhibit 12).
Our 12-month price target of $37 is based on a P/E multiple of 12x applied to our FY10 EPS estimate of $3.10 per share. Although the stock has historically traded within a P/E range of 12x – 25x, we think lower royalty revenues from new Samsung agreement and persistent concerns on oversupply implies the low end of this range will be applicable going forward.
Attractive on P/E, P/B and EV/S. Based on historical valuation data, SNDK currently trades below the 5-year average on P/E, P/B, and EV/S metrics (see Exhibits 13-15).