Monday, April 19, 2010

Sandisk (NASDAQ:SNDK): Into the Strait of Messina; Downgrading to Underperform - Pacific Crest

Pacific Crest is downgrading Sandisk (NASDAQ:SNDK) to Underperform from Sector Perform saying they see fair valuation in the mid-to-high $20's.

Model and market changing; risks increasing. Firm says they think the NAND market has moved from a growth business to a cyclical business, perpetually in search of an application or market. SanDisk apparently believes this to be true, because instead of investing in extending their brand, the company has chosen to sell wafers and finished product to the Asian module houses that are largely considered its competitors.

It appears that the market views the opportunity in SNDK as if the days of heady growth are back, or as if the company’s admirable job of restructuring is operations will drive long-term and structurally higher profitability. Pacs sees it differently. They see a company in transition, and one in search of a next act. Unfortunately, a next act does not appear to be on the horizon any time soon. In short, they have become more cautious for the following reasons:

- A notable increase in “Alternative Channel” activity
- The negative implications of selling into alternative channels: negative ROI of increasing market share in traditional card market and the burden of captive manufacturing
- This strategy’s impact on SNDK’s multiple
- Realization that the appropriate multiple on even the most optimistic EPS assumption yields full value (peak) at current prices.

This alone would be sufficient to drive a rerating of most stocks, yet there are many things we can add to the above list. These include:

- The prospect of additional wafer capacity coming on line in 2H2010
- Marginal free cash flow generation even under optimistic scenarios for profitability
- A royalty stream running at 50% of prior year levels from its major licensee, Samsung
- The need to take on the major NAND players Samsung, Toshiba, Hynix and Micron, all of which have superior scale, in the embedded market. Three of those companies have a product portfolio that allows them to bundle/integrate other products
- No new volume “killer app” for NAND in the near term; the adoption of notebook SSD looks to be several years away

Notable Alternative Channel Activity Increase in a Supposedly Healthy Market
This is the issue that has recently gotten the firm's attention. After discussions with numerous supply chain contacts in Asia, it appears that SanDisk has notably increased its efforts in this area. SanDisk began this effort in Q3 of last year; the company revealed that it would sell wafers, unmarked cards/drives, and downgrade products into the memory module industry. Last year, this strategy was interpreted as an effort to work off excess inventory that SanDisk had been carrying because of the down-cycle. Conventional wisdom holds that the current NAND environment is healthy—strong pricing, no incremental capacity, iPad launch, and so on. Yet what Pacs found was a meaningful increase in availability of SanDisk wafers and unbranded cards in the Asian module market during 1Q2010. Several module houses said they can get meaningful quantities of unbranded cards and 3-bit per cell wafers from SanDisk, and they heard from one contact that even traditional 2 bit/cell MLC wafers were for sale. Pacs believes the pricing is often aggressive, and they think that Samsung now views SNDK as its primary competition for business with the Asian module houses.

Hard to Spin This Positively, Except Perhaps in the Short Term
There are a number of ways to view this development. First, it could imply that overall NAND industry demand is less healthy than it appears, as it begs the question: why would SanDisk be willing to enable its competitors in a market that is thought of as marginally undersupplied? Second, it could reveal that the market for X3 (3-bit/cell) devices is not as robust as hoped, as this is the primary product that seems to be available to Asia module houses. The commentary about aggressive pricing also seems to be at odds with how SNDK characterized X3 pricing at its analyst day in February.

“Pricing competitive to X2: Expands Gross Margin and Improves Return on Invested Capital”

- SanDisk Investor Day Presentation, p. 51

Third, it could imply that SanDisk does not think that further share gains of its own brand yield a positive ROI.

Implies Different, and Lower, Multiples for SNDK
Pacs argues, that the morphing of the business model should lead to a lower multiple more in line with those associated with either the NAND semi manufacturers SanDisk is beginning to resemble, the competition to which it is selling, or a combination of both. It suggests a P/E ratio closer to 8x to 9x or a P/S ratio closer 1.2x to 1.3x.

In recent conversations regarding SanDisk, the firm notes they have heard the case made that the company could earn as much as $4.00 in 2010, if it can post revenue upside and expand margins through the year. While they think numbers in this range are highly unlikely, they assume for the purposes of this exercise that the company reaches those numbers. Given what they think are the appropriate multiples for the changing business, its current price implies full valuation (paying the median multiple for peak earnings is typically a stretch). If EPS comes closer to Pacs' number, or consensus, the stock should have downside from here (based on their multiple). From otheirperspective, the valuation has more chance of deflating from here.

At $2.50 in earnings, downside goes to $25. The risk/reward tilts heavily toward risk—it’s time to exit the shares.

Notablecalls: I think this is a fairly significant call from Pacific Crest's Semiconductor team, headed by Kevin Vassily. The comments of meaningful quantities of unbranded cards in the alternative channel spell trouble. Not to mention the more L-T oriented comments regarding the morphing of Sandisk's business model.

SNDK will report in two days, which makes the call even more important. Note that most other firms (have two tier-1's out this morning) are calling for a strong quarter. But do note that while J.P. Morgan is calling for strong numbers, even they are mindful of current valuation and are reiterating their Neutral rating.

I think these comments from Pacs will hit SNDK stock today. I'm guessing the stock will be down 1pt+ today, putting $36.50 level in play.


Sam said...

There is so much that is simply wrong in that "analysis" that I have to put the word in quotes. (1) Sandisk and their partner Toshiba are the only companies to perfect x3 (3 bits per cell NAND, in case the analyst isn't up to the jargon), which gives them at least a 10-20% cost advantage over competitors. This enables them sell in alternative markets while maintaining good margins. And by selling fewer of their own products (but keeping their premium brand), they keep the end markets from getting too glutted, and actually reduce their own selling risk. The analyst completely misunderstands Sandisk's strategy and the technological strengths that enable to do this profitably. (2) Their royalty stream isn't 50% of what it used to be; the guy is math and analysis challenged. This misunderstanding of their royalty stream is common, and has been rebutted many times, both logically and in Sandisk's results, so I won't bother to do again here. (3) He apparently thinks that the booming growth in cell phones and the advent of the iPad and other tablets aren't NAND killer apps--this is just nonsense. Cell phones are in inning 2 of their growth, and the national anthem is still being sung in the tablet market. (4) He is just plain wrong when he says that Micron and Hynix have "superior scale" in NAND. While Samsung and Toshiba do produce more, Toshiba is their partner, Samsung pays them royalties, and Sandisk has superior controller technology to the latter. It is that superior controller technology which enables them to be the low cost producer along with their partner Toshiba, and enables them to make better than average profits while being a merchant NAND producer/seller. The analyst's assignment of an 8x or 9x PE is just nonsense for a company that is on the cutting edge of technology and, contrary to what the analyst says, that is supplying huge growth markets for at least the next 5 or 6 years (cell phones and tablets now, SSDs will join the party by sometime in 2012). That said, I hope he and others short the heck out of the stock. He will be covering in the 40s or 50s, shaking his head about how "crazy" the market is.

Sam said...

To clarify one comment I made: it is true that Micron and Hynix have more overall sales than Sandisk, but the majority of their sales are in DRAM, not NAND. Sandisk not only sells more NAND than either of them, but they have a better balance sheet and more advanced technology as well. It is their controller technology--which is protected both by patents and by its very complexity--that is their greatest advantage over their competitors. Analysts continually fail to appreciate the importance of that advantage in the NAND sector, and thus keep confusing NAND with DRAM. That is their primary error, and that is why Sandisk and Toshiba will be the dominant players in this dynamic sector for years to come, a sector that is just beginning to come into its own.

market operator said...

That being said, $36.50 trades and then some. Good call.

Sam said...

Is there a reason why you chose to give Pacific Crest's bearish view of Sandisk prominence while ignoring Morgan Stanley's bullish report, just issued today?

bl said...

SNDK was a gap down that kept going down who cares what the analyst(2) said. Forget the mumbo jumbo fundies and talk, and speculation,, and read the chart...short term tradng here is what NC is all about I think. Tuesday SNDK might me up 3 pts. So what unless you are a long term holder.

notablecalls said...

Sam, love the comments! Love'em!

Pls keep them coming. There is nothing more I like than a second (& well informed) opinion.

The reason why I didn't highlight the positive MSCO call is bc it was issued intraday (sometime after noon, I think). The MSCO defense was ill-timed as the stock plunged another full pt before recovering with the rest of the mkt.

Sam said...

Well, Vassily succeeded in getting his name mentioned on CNBC as Maria asked Eli Harari about his call. Eli just smiled as he repeated his (and my) belief that the market opportunity for NAND is bigger than most analysts think. In light of Sandisk's Q1 .95 of earnings, it is fair to say that Vassily's $2.41 estimate is now rendered completely absurd. And frankly, his belief that Sandisk should be given a PE of 8 or 9 is also ridiculous, given the way the company is executing and their gorilla status in a market that, driven by mobile phones now (with a powerful assist from tablets) and with SSDs to come in the next 18-24 months, is entering the tornado stage.