Hello Jason and Aus.. :D
I couldn't help but respond to your posts regarding SNDK and SSTI.. every good argument deserves a good counterargument.. so maybe I could discuss the other side of things.. as to why SSTI is a great investment in itself.
<<...the biggest factor that separates the two from an investment perspective?>>
Jason, what do you think of when you think about flash? At least when I began my DD, I thought about flash cards.. flash for digital cameras. Quite certainly, that use of flash is the most well known and is the most consumer accessible b/c.. well, it's marketed as flash.
I don't mean to belittle anyone here by rerelating the basics, but they're important to tell SSTI's story and I want to make sure everyone is on the same page. There are two different primary types of flash memory that are in volume production now. NOR flash stores information in small chunks and is best suited for code applications. Much as you'd run an executable program on your computer, devices have a space in which it's "executable" is stored. NAND flash, on the other hand.. stores information in large segments, and as you can imagine.. this type of sequenced memory is best suited for data such as a graphic, or a sound file.. a large piece of binary information.
Understanding the needs of a digital consumer device is important in understanding what markets SSTI and SNDK serve. As you probably know from Ausdauer's excellent explanations in the past, SNDK focuses on data storage.. in fact, a large part of their business is in creating standards and hardware based on removable storage. Needless to say, data such as pictures, MP3s, archives, backups, etc.. all require large amounts of this NAND storage medium.. which is then packaged into an amazingly small little card. These cards go into things such as digital cameras, MP3 players, PDAs, etc.
SSTI on the other hand, focuses on almost the opposite end of the spectrum.. that NOR type memory used for storing code is this company's focus, at least at this time. You could certainly argue that SNDK is more "sexy" to the consumer in that it's a product that's more directly accessible to the public, etc.. which is certainly good for SNDK, b/c it is inherently more of a shareholder friendly concept. But to use that as an argument as to why it's a better investment than SSTI would be missing the point.
Suppose I were an engineer and I wanted to make an intelligent coffee maker. I wanted it to remember times in which my coffee supply was getting low so that it could call up Webvan and order some fresh beans. In this capacity, I would certainly need a piece of NOR flash to store the instructions on how the mechanics needed to work. Probably wouldn't need a flash card, however. This is exactly what places SSTI in a unique opportunity. I don't think flash producers were quite ready for this digital revolution, this internet/interconnectivity driven society in which all our devices are wired. All of these literally thousands of consumer devices will potentially all need NOR flash. Traditionally, NOR flash had quite limited uses: the BIOS for a computer's motherboard and other computer components. All that time, SSTI chugged away and had a nice niche business in that market.
In the meantime, they developed and patented a series of design and manufacturing patents regarding the creation of NOR memory which they dubbed "SuperFlash". These patents allow NOR to be made cheaper (because of less layers of metal required during the process) and smaller. I'll be happy to explain SuperFlash in more depth sometime, if necessary.. but to be brief, it quickly became well regarded as the best way to make NOR. Currently, companies such as TSMC, Sanyo, Motorola, Qualcomm, and IBM to name a few have licensed the technology.
When the internet and this "wired" revolution took place, with SuperFlash in hand, SSTI took the reigns of a market that exploded right before their eyes. Where other foundries and flash producers were focusing on higher densities and moving away from what was considered a niche market, SSTI was left as the acknowledged leader of low density NOR and the demand for this type of flash became very profitable for the company. Look through the financials yourself, they've gone from 125 million in revenues in 1999 to an expected *1 billion* in revenues next year.
Ausdauer makes some convincing, yet faulty arguments as to why SNDK is a better investment, let me quote some of them for you:
<<If demand for ultra-high density flash ramps up as steeply as some have predicted, it would appear that SanDisk, Toshiba and Hitachi will be leading this high growth flash sub-segment. The competitive advantage period for low or very-low density flash is limited.>>
We are entering into a new era of uses for flash today, and the methods for determining supply and demand, are needless to say, skewed in the favor of the "bigger is better" concept. For NAND memory, this is true.. you always need more room for data, and so demand for NAND is created on the insatiable need for more room. For NOR memory, however.. the demand is created through more and more diverse devices, of which all will potentially require NOR. Even the devices that use NAND require NOR as well. So to say that "the competitive window for low density flash is limited", is a guess at best. No one can point to the amount of demand that will be created by more and more consumer devices. It's limited only by the imagination. In the race to jump onto this NAND bus, most companies left the NOR workhorse behind, and their foundary upgrades reflect this fact:
From a recent Kaufman Bros. Analysis:
boards.fool.com
---
<<STRENGTH IN THE QUARTER IS BROAD BASED
Our channel checks and conversation with management indicated that low-density flash devices continue to be in tight supply while demand for flash is broad based. In the June quarter networking and wireless communications were the biggest YoY percentage gainers at 580% and 2668%, respectively (wireless off a small base). Additionally, the Intel (INTC $65 45/64) Firmware Hub Architecture (PC Bios)posted 500,000 units in its first shipping quarter. Our checks indicate that Internet computing and digital consumer products are similarly strong with communications devices continuing to step up.
TO INFINITY AND BEYOND (WELL AT LEAST BEYOND THE 3RD AND 4TH QUARTERS OF 2000)
We continue to acknowledge the supply of flash memory devices coming online over the next 12 to 24 months, however, much of this supply is of the NAND variety, which is typically for high-density mass storage applications. SSTI focuses on low-density NOR type flash focusing on code-storage applications. This delineation can be illustrated with SSTI's agreement with Intel to supply, as a licensee, compatible firmware hub architecture flash devices. We believe that Intel is willing to cede market share in this area because it currently has only a limited ability to supply 2 and 4 Mbit flash devices and the new capacity it is building is targeting high-density (16 Mbit and higher) flash devices. Additionally, the other significant flash capacity expansion announcements (from Fujitsu, Toshiba and Samsung) are by traditional NAND-type flash manufacturers and, as noted above, this type of flash is typically used in high-density mass-storage applications. The focus and excitement around the flash market has clearly been the high-density market. However, we would like topoint out that while every electronics device that requires high-density flash also requires low-density flash, every product that requires a piece of low-density flash for code storage does not require a piece of high-density flash.>>
---
If Intel, the largest producer of flash has contracted SSTI to create NOR for their 820 motherboards, doesn't that say something for the company's level of demand as well as their dominance of this segment? Many have argued that they have this stronghold primarily b/c other companies can't make a profit with their fabrication processes. SSTI's process however, is quite profitable for them as seen in the above figures.
<<Some of these producers will have ceilings which they will have difficulty penetrating due to competition from above.>>
Why would SSTI have any interest in "penetrating" this ceiling Aus refers to? It's like I explained to someone else the other day.. why climb to the top of the orange tree to eat when the branches are so weighed down that the fruit is within arm's reach? :D Developers aren't going to put more density than they need into these devices, it's not cost efficient. More importantly, however.. they can only fill 50% of the requests for their hardware.. you can only ramp up production so quickly! :)
<<And if pricing continues to increase, designers may go looking for alternatives such as EEPROM (mentioned in a recent article). Also, MRAM/FRAM will look to compete first with the low density code storage segment.>>
I frequent other boards, and this argument is the one used most often by short hypsters.. I'll allow Aus to speculate if he allows me to later.. :) First of all, MRAM and FRAM are older inferior technologies which both require special metal processes, have fragile structures and have had quite poor yields thus far.. they're definitely both nice concepts, but practical ramping of them is at best a couple years off.. you could certainly argue that in the meantime that SSTI will capture design wins and market share.. their product is out NOW.
If we were to speculate about flash cards (and SNDK for that matter), we could also make the argument that this same internet/interconnectivity revolution will make flash cards obsolete. This is the whole concept of bluetooth and wireless communication.. why would we need a flash card, if the photo or the song could be zapped from over the air? No denying that this market isn't on the minds of a lot of developers at this point..
I guess what one must acknowledge though, in summary is that both for the present time are excellent investments, and I have personally considered investing in SNDK.. but it'd certainly be easy enough to argue in favor of one or another based on what colors you choose to paint with..
all the best, docpaul |
