SanDisk: Risk Down $3, Reward Up $30

Memory chip pricing has shown no sign of stabilization for either DRAM or NAND flash drives over the last six months, and expectations for the next six months aren't more encouraging as we head into the seasonally weaker time of year (but we know all this already). Actually, this week we could see spot pricing and activity pick up a bit as buyers return from the Consumer Electronics Show, though any rebound will be short-lived. That said, we think investors should start to look at SanDisk Corp. (SNDK) on the long side.

No, pricing will not get unexpectedly better and more capacity is still slated to come online. However, NAND demand is on the verge of another exponential expansion and SanDisk is just too cheap at these levels to ignore. Less importantly, we are also seeing some reductions in DRAM build plans by Taiwanese firms pushing out tool orders as they are about to experience EPS losses greater than Q1 revenues.

Avian Securities has correctly determined the direction of SNDK stock through the last four years both on the upside and the downside. We were positive from 26 to 75 in '04; we were negative from 80 to 50 in '05; positive again from 40 to 52 in '06 and negative again from 52 to 40 in '07. We know what seasonality should be. We know what Street expectations are. We know how much capacity is coming on line. We have a reasonably good assessment of new product releases, and most importantly we closely monitor both spot and contract pricing with regard to normal seasonal trends.

In early September, we saw a bunch of capacity come online, a delay in solid state drive adoption, and price declines in what was supposed to be a seasonally strong period. So, we quickly reversed our opinion to a negative one. Today we are turning positive on SNDK again.

NAND, as you likely know, is non-volatile memory used to store information on portable devices like cameras, music players, cell phones, and thumb drives. SanDisk, through a joint venture with Toshiba, is neck-and-neck with Samsung as the largest and lowest-cost producer of NAND. It also holds a large chunk of NAND IP for which many of its competitors, including Hynix and Samsung, pay SNDK a royalty.

In this note we want to focus on three major drivers for NAND which, although they have all been discussed over the last nine months, are just starting to emerge. (As is typical of these markets, just as these drivers begin to develop, investors - having eagerly anticipated their arrival - are throwing in the towel.) The three drivers are:

1. Solid State Drives (SSDs) for more energy efficient/reliable notebooks and servers.
2. NAND on PCs to help boost boot-up times and overall performance.
3. The proliferation of smart phones which will require significantly more NAND.

All three opportunities are in their infancy but should emerge in the next six months and will combine to expand NAND demand by several orders of magnitude.

The MacWorld conference on Wednesday will highlight the opportunity of SSD with the possible launch of an ultra mobile notebook. We should see a rapid pace of new SSD announcements and launches over the next three months. NAND on or next to the motherboard will get a lift from a new Vista update in mid '08 and as Intel (INTC) refreshes its product lineup. Finally, smart phones are just starting to explode as data plans and integrated WiFi reduce the cost of getting connected.

The recent debacle in memory pricing was a result of these drivers failing to materialize in '07, and while we are convinced they will occur in '08, manufacturers hurt by recent expansion will end up underestimating their impact and therefore extend the next cycle.

Let's look at each of these opportunities individually.

• 260 million PCs will be sold worldwide in '08 growing between about 10% and 15% annually. Eliminating 30% of the lower end PCs, we are left with about 200 million PCs in '09 that will ship with between 2 and 5 GB of NAND versus zero GB of NAND in PCs today. This means an incremental 700 million GB annually will be sold into PCs growing at about 15% annually (50% loading growth and 10% PC unit growth).
• 120 million notebook PCs will be sold this year with incremental demand growing at 25% a year. Primarily due to NAND's durability advantage (but also potentially due to speed and energy considerations), we believe that roughly 10% of notebooks in '09 will ship with an average of 128 GBs of solid state storage adding an additional 2.4B GB of NAND demand versus zero today – growing at a rate of 50% annually over the next five years. Moreover, with Robson, hybrid HDDs, and Vaulter or some similar product potentially penetrating much of the rest of the premium notebook market, we could easily see another 400M GB of flash in solutions not involving SSDs. Servers could also contribute to this demand at the edge (though the significantly smaller market for enterprise-class products obviously will limit the impact of these products on overall market economics). This segment will actually develop most quickly as demonstrated by EMC Corp.'s (EMC) announcement on Monday that it will include flash drives to improve the performance of its Symmetrix systems.
• 1.1 billion cell phones were sold worldwide on '07. Fifty percent of them have slots for flash cards. If 20% of handsets sell with 5 GB of NAND in '09, we are talking about 1.1 Billion GB of NAND or 900 million GB of incremental demand. Samsung estimates 731 MB of average NAND loading per phone in '09. So, if you use Samsung's numbers we are in the same ballpark at 900 million GB of NAND – a 267% increase from '07.

Our back-of-the-envelope math gets us roughly 4.5B GB of incremental NAND demand in '09 versus 1 billion GB shipped in '06 and about 2 billion GB shipped in '07. Don't forget this opportunity is on top of incremental demand growth from current sources such as MP3 players, digital cameras, USB storage devices, and so on. A digital camera upgrade cycle, the growth of NAND devices in every car, refrigerator, game console, and the continued growth of music and now portable video players will further drive demand. All this math would be less interesting if SanDisk wasn't the industry leader. I would argue that during these periods of famine, SanDisk will increase its manufacturing lead over competitors like Micron Technology (MU), Hynix, even Samsung, and especially the smaller players. Tightening credit availability worldwide advances this notion as the higher-cost producers will have trouble funding additional improvements. Also larger losses on the DRAM side limit diversified memory manufacturers incrementally versus SanDisk which has no direct exposure to DRAM.

We can see all this evidenced in SNDK's Q3 report, as SNDK was able to maintain EPS even in a very difficult market environment for NAND. SNDK has about $2.3B in cash, or about $10 a share; after netting out debt they have just under $5 a share in cash. Our '08 SNDK estimates are in-line with the Street at $4.8B and $2.50. The stock is currently trading at 12x '08 earnings, or 10x '08 earnings net cash. More importantly, in our view in light of uncertainty over the next six months and its impact on earnings, SNDK is best positioned to survive the current downturn and to exit any malaise in a more dominating competitive position. As the above stated demand ramps up and soaks up the current excess production, we expect '09 pricing to firm and SNDK numbers to grow significantly.

We are convinced that we are at or within three months of the bottom of a two-year NAND cycle and that SNDK will significantly appreciate from current levels over the next two years.

Disclosure: Author has no position in SNDK.

Comments

[rd4sndk]

I'm sorry but you are late to the party as are most analyst and MM. Dispite all of your partificating around nand spot pricing, Sandisk has never been and is currently not dependent to much extent on spot pricing of nand. One need only look at 2006 numbers where you analyst were expecting, for most of that year, for sandisk to report disappointing earnings. Stock price fail based on your flawed understanding of Sandisk business model a value proposition. However, earnings increased 25% while Revenue increase over 75%. Need more proof that you're full of hot air? Look no further than 1Q07 when nand pricing fail 24% during the quarter but Sandisk, unlike other semi providers that report huge losses in their Nand operations, reported inline earnings (excluding impact of mystems acquisition).

Sandisk global market share expansion and improved cost efficientcies from its latest card assembly plant, increasing average card density is what's currently fueling its top/bottom line growth. Additionally, inventories are low going into the 1Q08 at a time when YoY demand is accellarating.

Although I agree with some of your reasoning, your inability to evaluate Sandisk as card producer and the only one of its kind leaves me to reject your analysis for investment advise. However, these monkeys that listen to you control the price of the stock and for that reason I appauld your positive although ill imformed opinion.

[User 140904]

I think you are wrong regarding cash/share. It's true they have $10 cash per share but they also have $5/share in liquid long term investment, which cancels out with debt, so they still have about $10/share net cash/share.

[tessant]

Sandisk has taken up many hours of research and ticker watching for me. I cant believe its gotten to this price level. Right here its a no brainer for a long term investment; just dont buy short term calls..

Scott W
growthportfolio.ning.c...
investment social network

[reinharden]

> expand NAND demand by several orders of magnitude.

> Our back-of-the-envelope math gets us roughly 4.5B GB
> of incremental NAND demand in '09 versus 1 billion GB
> shipped in '06 and about 2 billion GB shipped in '07.

Um, going from 2 billion GB in 07 to even 6 billion GB in 09 would represent increasing demand by a factor of 3.

An order of magnitude requires increasing demand by a factor of 10. Two orders of magnitude would be at least a factor of 100. Several orders of magnitude requires at least a factor of 1000...

reinharden

[All Tech]

Regarding Net Cash per share, user 140904 is correct. You missed a big chunk of cash in the LT Investments line item. As of Sept 30, 2007 Sandisk had $970mm in cash, $1.334Bn in ST Investments, and $1.824Bn in LT Investments for a total of $4.128Bn. Subtracting the $1.25Bn in the convertible (which is not due til 2013 and only pays 1% per annum) you have $2.878Bn in Net Cash or $12.51 in Net Cash / share. At todays stock price of $27 - $12.51 = $14.49 of value for company operations which should produce $2.50-$3.00 in '08 and $3.50-$4.00 in '09. So that gets you a multiple of 5.27x '08 and 3.9x '09 estimates.

But the best part is that Sandisk will put up another $200mm+ of cash in the just finished Q4 so the actual numbers are even lower and the company as of today has $13.50 in net cash, leaving $13.50 for compay operations or 4.9x '08 and 3.6x '09. Which my friends is pretty damn cheap for the best run company in the fastest growing part of the semiconductor sector and arguable the entire technology sector.

[GJELOSH197]

avi what to make NOW of your 'DOWN 3, UP 30' SNDK thesis!!!??