Regarding the ML downgrade, I downloaded the entire report and have the following comments:
SanDisk reported to its shareholders at the last conference call following the release of earnings that it was experiencing increased demand for the low end compact
flash units destined for OEM use. This means that several camera manufacturers are shipping cameras with 8 mb or 16 mb
flash cards, the bare minimum that would be practical, particularly in the higher end 2.1 and 3.3 megapixel resolution cameras. These cards go for as low as about $2 a
unit and help the camera manufacturers price their products more competitively. The cards themselves don't generate much profit for SanDisk, as noted earlier to shareholders. And the prices of those cards are down to the point where they become loss leaders. Maybe Merrill Lynch sees this
pricing strategy as a sign of weakness. I see it as a sign of strength. Here's why:
If you buy a 2.1 megapixel camera, such as the Nikon 950 that I bought last year, the first accessory you buy is a larger capacity flash card, most of which are made by SanDisk. More than 2 million digital cameras have been sold so far this year. Most of the cameras contained 8 mb flash cards, and the camera owners subsequently went out and bought typically 32 or 64 mb cards. I bought a 96 mb card, and some serious users have also bought 128 mb cards, which sell for over $200 a piece. Thus, by pricing its 8 mb cards at a very low level, SNDK enticed camera manufacturers to include that card as original equipment, and at the same time assured itself a retail market for higher capacity cards.
All one has to do is check on a retailer such as amazon.com, as noted on this thread, to see that virtually every SanDisk flash card with 32 mb or higher capacity is out of stock and back ordered, with a waiting period of two weeks or more. I find it difficult to understand how ML could assume that there was a weakening of demand in the face of what we know about retail sales.
Another weakness I found in the ML analysis concerns the MMC cards for the MP3 music players. These cards are now selling very well, but their introduction was delayed by
uncertainty over the effect of copyright litigation. Turns out that the secure digital format developed by SanDisk is very appealing to the music publishers and therefore should
end up as the format of choice. The increasing sales of MP3 players using the SD MMC does not seem to have received any consideration whatsoever in the ML analysis.
So I take issue with the ML assumption that demand is dropping, based on lower pricing. In fact, part of the lower pricing, if not most of it, comes from better yield on silicon wafer production. Yield improves with experience,
and SanDisk has been buying product from several fabricators who recently completed new factories in Taiwan. Furthermore, SNDK increased its production of double density cards (which effectively store twice the capacity on the same size chip), lowering the manufacturing cost of a given capacity chip. I don't see any of this analysis in the ML report, and that is why, in my view, ML reached a conclusion that is simply not
supported by the facts. In short, lower prices reflect manufacturing and design efficiencies, not slack demand. In fact, as unit prices drop, demand goes up, just as it did with microprocessors during the last ten years.
Another indicator of increasing demand for digital cameras, which, as Merrill notes, comprise roughly half of SNDK sales, is the weakening demand for conventional film, as noted in recent Kodak reports to shareholders and investment analysts. Kodak tried to make the best of it by attributing lower film sales to dealers maintaining lower inventories. In fact, film sales are down (except for throwaway cameras) precisely because the really serious photographers are switching to digital. I bought my Nikon 950 in July, 1999. The last time I used conventional film was August, 1999. I used to buy about 30 to 40 35mm cartridges a year. I used to send out over 400 xmas cards a year, each with a conventional color print. Multiply my change in buying habits by about a half million, and you can see what is
happening to Kodak, and also what is happening to SanDisk.
Finally, there seems to be very little recognition of other new applications for nonvolatile flash memory. For instance, hand held computers, which represent the greatest growth area for those who otherwise might have purchased more expensive notebook computers, are using flash memory for storage, and most of the flash memory units (except for equipment made by Sony) are using either SNDK flash cards or cards for which a royalty is paid to SNDK. The MemoryStick,
furthermore, appears to use components made by Lexar, which just lost a patent suit to SNDK. As best as I can tell from
the settlement agreement, it may be necessary for royalties to be paid to SNDK on MemorySticks sold after March 2001. That's because the MemoryStick has features that Lexar now acknowledges may use SanDisk intellectual property. The ML analysis does not take into consideration the additional royalties SNDK will likely receive as consumers buy more and
more appliances using removable flash cards.
That leaves open the issue of embedded memory using flash cards built into circuitry in cell phones, some hand held computers, set top TV boxes, and other appliances that need a cheap form of data storage. This is a market new to SNDK, and there is no indication that SNDK will not be able to compete with companies like Intel and Advanced Micro Devices. SNDK has entered this market through joint
arrangements with companies like Toshiba, whose technology in chip production and design in this area is as advanced as that of INTC or AMD.
There is not one word, finally, in the ML analysis of management quality. Maybe they don't know how to measure it, but they could begin by looking at the cash on hand and comparing that figure, in dollars per share, with similar
data from other new high tech companies. SNDK has the best, most conservative cash position of any of the group--further
evidence that these guys know how to run a company in the best interest of its shareholders.
We'll just have to keep our fingers crossed. I expect that other investment firms will eventually come out with analyses that dispute the ML conclusions, and further that the Fed will be forced to lower interest rates, thereby setting in motion some recovery. But what I don't know is how long it will take for the larger investment firms to realize that there is a difference in the fortunes of companies now lumped together in the semiconductor sector, regardless of the types of semiconductors they make.
As one of my clients, a portfolio manager for a Manhattan based firm, has noted, these analysts from big investment firms are not God.
Art Bechhoefer |
