Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : SILICON STORAGE SSTI Flash Mem -- Ignore unavailable to you. Want to Upgrade?

To: Mr. Miller who wrote (615)	7/3/2000 3:52:22 PM
From: Mr. Miller	Respond to of 1881

Flash Memory: White-Hot, But Why? Fayad Abbasi Jun 23 2000 streetadvisor.com In the past several months, flash memory has become a pretty hot topic in the world of semiconductors. Device makers will do almost anything to get their hands on flash memory at the moment. As we’ve written before, an executive even demanded recently that all his employees turn over their PDAs so the flash could be taken out and used on the company’s production line. If that’s not desperation, we don’t know what is… So the recent shortage has given flash quite a bit of press, but most reports don’t explain how it works or why it’s so important. We’re going to change all that. Let’s start with the basic functions. Flash is a non-volatile memory device, which means that, unlike most memory (such as DRAM), it can hold its contents once its power has been tuned off. Flash memory is similar to Electronically Eraseable Programmable Read Only Memory, or EEPROM. EEPROM is used to store basic instructions for simple devices or startup instructions for more complex devices such as PCs. The information is stored on EEPROMs by “burning” or writing it byte by byte at a higher-than-normal voltage. Burning cannot be done quickly and is usually done on a special machine. Flash has two advantages over EEPROM. First, writing to the memory doesn’t require a higher voltage and can be done on the fly. Second, flash memory stores items in blocks, so writes can be done quickly. All types of memory are composed of a network of transistors called “logic gates.” These gates consist of inputs and outputs that operate on Boolean logic. All digital devices use electronically transmitted binary values (0s and 1s ) to operate, and these values are combined in the gates to produce an output value. Here are some basic Boolean operations: AND: if all inputs are 1, the output is 1; otherwise the output is 0. OR: if all inputs are 0, the output is 0; otherwise it is 1. NOT: the output is the opposite of the input. 0 à 1 and 1 à 0. NAND: if all inputs are 1, the output is 0; otherwise the output is 1. NOR: if all inputs are 0, the output is 1; otherwise the output is 0. Memory cells are composed of cross-coupled NAND or NOR gates called “flip flops.” Each flip flop holds 1 bit of data. There are 23 bits in a byte, 210 bits in a kilobyte, and 220 bits in a megabyte. So, millions of these flip-flops are required for each megabyte of memory. Aside from all the manufacturer-specific “bells and whistles,” there are two basic types of flash: NAND and NOR. These names have to do with the basic structure of the logic gates and the Boolean logic of the chip. Most standard types of flash are NOR, while newer ones use NAND logic. Flash memory gets its name from the way the chips are erased. The memory cells are organized in such a way that an entire section can be erased in one operation known as a flash. When a block of memory is erased, a process known as Fowler-Nordheim tunneling takes place. This is where electric current crosses over a dielectric material to disrupt the electrostatic field in each floating gate. Flash memory has several applications, but most of them revolve around the same premise. Many different electronic devices require instructions when they start or “boot up”. A non-volatile memory is needed to store these instructions. In the past, EEPROM was used to store this data, but the information could not be updated without removing the chip. Firmware updates were usually non-existent because the chip or device had to be mailed to the manufacturer. Nowadays, flash memory is used to store this data because it can be updated quickly by special software. The best example of this evolution of technology is PC BIOS. Flash memory allows for easy updates and bug fixes, and most peripheral devices such as video cards contain their own BIOS, which can be easily updated. The other common use of flash is what we will call “medium-term storage.” Data is not temporary so it must be stored, but at the same time, it’s not permanent so it must be changeable. Digital cameras store pictures, and MP3 players store songs using flash memory. Cell phones and PDAs also use flash to store information. SanDisk [SNDK] even has a flash memory device for PCs that emulates a hard drive. Intel [INTC], AMD [AMD], and other large memory producers all produce flash memory. However, there are very few pure players in the flash industry—Silicon Storage Technology [SSTI] and SanDisk are the two that specialize in devices containing flash memory. Until production capacity can meet the growing demand, flash producers will continue to hold the winning hand.

To: Mr. Miller who wrote (615)	7/3/2000 4:29:18 PM
From: Allegoria	Read Replies (1) \| Respond to of 1881

There has been a good on-going discussion on the SNDK and G&K threads concerning SSTI. I posted this message on the SNDK thread, but thought in case people here didn't get to see it that it might be nice to repost it here: Ausduaer said: ********************************IMPRECISION***** There is a lot of imprecision in the way the flash memory market is perceived. If I were making a low density product with little barrier to entry I would worry about a larger competitor coming in and whittling down margins. I can't see how low density commodity flash can remain a high margin item forever. On the other hand, high density flash and, in particular, proprietary miniaturized flash cards sporting the latest flash technology and surrounded by a deep moat (the IP) would seem to be a harder market to crack and disrupt. I also believe the demand for flash cards (in multiple forms) will be difficult to flood in the near term because of the mass appeal, the rapid anticipated growth of the target markets, and the protracted state of undersupply we are battling. ***************************************************** Let me start by saying I am long both SNDK and SSTI, so I am not intending to 'bash' either company nor the products they produce nor the markets in which they participate. "If I were making a low density product with little barrier to entry I would worry about a larger competitor coming in and whittling down margins." I think you have this completely backwards. Let's first look at these larger competitors you have mentioned: Intel and AMD are No. 1 and No. 2 in worldwide flash memory market share respectively. Flash currently represents about 30% of revenues for AMD and is in the high single digit percentages for Intel. And never loose sight of the "real over-suppliers": Sony, Matsushita, Fujitsu, Atmel Corp. and especially Hitachi. And now the talk is "IBM's 1GB Microdrive, the latest in IBM's range of after-dinner mint-sized disk drives. IBM thinks the new drives will "turn today's digital still cameras into genuine movie cameras" by almost trebling the storage capacity accessible from a standard Compact Flash Type II interface. (Is this competition for SSTI or SNDK?) AMD just unveiled two new flash memory chips for next generation cell phones that rank as some of the fastest on the market. The 32-megabit and 64-megabit chips will begin showing up in 2001 and 2002 as cell phones increase their functionality. Chase H&Q says: "there is not another company in our mind who is executing better in flash than AMD." Roughly 75% of those sales are from high-density flash memory, which carry better margins. smartmoney.com Fujitsu Ltd. & AMD said Tuesday (May 23) that they will build a new flash memory production facility and by the end of fiscal year 2001, the new fab is expected to churn out 13 million flash chips per month. Output will then increase to 52 million pieces per month by the end of fiscal year 2002. By that time, the Fujitsu & AMD joint venture's total production of flash memory devices will increase to 200 million chips per month, up from the current level of 36 million per month. Toshiba Corp. and SanDisk Corp. also recently announced plans to increase production of NAND flash memory more than sixfold at Toshiba's joint venture facility, Dominion Semiconductor (Manassas, Va.). And even a Samsung executive said: "We are converting some DRAM production lines [over] to flash memories." So it is fairly obvious that the 'giants' are moving into flash, but what kind of flash are they targeting, w.r.t. SNDK & SSTI? You know SNDK concentrates in the data storage segment area whereas SSTI concentrates in the code storage segment, so they are not direct competitors. A data storage device is used to store large amounts of data such as music and pictures whereas a code storage device (read: embedded memory that stores code). is primarily used to store the program code that runs the device. When comparing SSTI & SNDK remember this: currently,applications for flash code storage devices far outnumber applications for flash data storage devices, as supported by the 3/30/00 CSFB report. (No link avail.). SSTI is positioning itself to dominate the code storage market. Silicon Storage has staked out its turf in the "lower density" end of the market -- meaning chips that have anywhere from 512K to 8MB of storage capacity. This is the space being abandoned by bigger flash-memory players like Intel and Atmel, which favor higher-capacity code-storage chips more commonly used in cell phones. But that exodus is a good thing for Silicon Storage. Intel recently announced that they are LEAVING the market FOR LOW-DENSITY FLASH in order to concentrate on the HIGH density market. SSTI is the first (and only?) Intel licensee to supply flash memory products compatible with Intel's 800 Series Hub Architecture chipsets. "The announcement that Intel selected SST as the first licensee of it FWH architecture indicates that Intel believes that SST has the quality, reliability, and foundry relationship necessary to support this enormous market opportunity. While we believe Intel will license others to participate in this market, it is our belief that SST will have a significant technological lead and cost advantage over its future competitors.."---CSFB 3/30/00 report. And AMD and others have reported new fab construction concentrating on the HIGH density flash market. I have not seen (not to say I haven't missed it!) one announcement about anyone targeting the low density market. Has anyone seen an announcement concerning low-density fabs coming on-line? So for me it is fair to conclude that there is more evidence that the competition is moving into the high density arena then the low density arena. You also mentioned "...a low density product with little barrier to entry" Aus, this is EXACTLY SSTI's greatest strength: The company's proprietary technology and focus on the low-density end of the market should protect it from competitors. SSTI is positioning itself to dominate the code storage market. No one can compete on price (both product AND manufacturing) & quality. "Currently, availability is tightest at 2 and 4 Mbits (density). Getting low-density flash is a major challenge. The ADSL market took off. And cable modems, in the United States and North America, and Internet appliances are selling like wildfire," Eugene Feng (SSTI guy) said. "These devices all require 2- and 4-Mbit flash, and many vendors are walking away from that market since it isn't very cost-efficient to make." From: June 26, 2000, Issue: 1217 techweb.com "Most people thought that low density equals low margins," explains chief executive Bing Yeh (CEO SSTI). "That may be true for our competitors, but not for us." Why? The reason is that Silicon Storage uses a patented process known as SuperFlash. "SSTI's SuperFlash technology is a NOR type, split-gate cell architecture which uses a reliable thick-oxide process with fewer manufacturing steps resulting in a low-cost, nonvolatile memory solution with excellent data retention and higher reliability". SST has over 20 patents on its SuperFlash technology. Importantly, these patents cover not only the flash product itself, but the PRODUCTION PROCESS which is also licensed out to other manufactures, creating a royalty stream. Current licensees include Analog Devices, IBM, Motorola, National Semiconductor, Samsung Electronics Co. Ltd., SANYO Electric Co., Ltd., Seiko Epson Corp., TSMC-Acer Semiconductor Manufacturing Co. (TASMC) and Taiwan Semiconductor Manufacturing Co. Ltd. (TSMC). Therefore any increase by these companies only increases the royalty stream for SSTI…a win-win situation! What is the advantage of this process? The SST SuperFlash technology typically uses a simpler process with fewer masking layers, compared to other flash EEPROM approaches. The fewer masking steps significantly reduce the cost of manufacturing a wafer. Reliability is improved by reducing the latent defect density, i.e., fewer layers are exposed to possible defect causing mechanisms. For a comparison of superiority of SST SuperFlash versus the other major approaches to production see: ssti.com In lay terms, one poster put it simply: "for each technology change, i.e., .50um,.35um, .25um, .15um, etc., SST's products will always have smaller die and peripheral circuitry than its competitors, such as AMD, Intel, Atmel, etc. You can always shrink the die size for each technology change, but you cannot easily shrink the peripheral circuitry,and smaller die size per unit always means lower cost!" This process brings down costs in part by allowing the company to use smaller die sizes. It also makes for more reliable products, because thick oxide layers on the chips don't break down as much as competitors. From that same poster: (http://www.siliconinvestor.com/readmsg.aspx?msgid=13968487) If a price war does come, it would be with "high-density" Flash, not SST's segment market, again, if you understand the value of "SuperFlash", then you would see SST's value proposition. So called "high-density" really is a moving target, today's high-density is tomorrow's low-density, and due to SST's lower cost structure due to "SuperFlash", it can always dominate its chosen market and others will move away because they are not price competitive. You are seeing Intel and the others are doing just this now… In the end what we are interested in is the almighty Share Price. From the last report, SSTI's margins are rising to 41.5%, an 11.2% increase over the previous quarter. With such a tight supply, you can expect margins to increase further. And remember that SSTI is growing revenues from a smaller base than SNDK. SSTI has considerably less # of outstanding shares and therefore has potentially better revenues / share metrics than SNDK (IF the nominator grows!). With similar gross margins and revenue % growth rate, SSTI will translate to better EPS. The market pays for growth and you are seeing this reflected in the share price. Look at the growth numbers, but don't forget to factor in the recent announcement from SSTI (300% growth…) Jun-99 Sep-99 Dec-99 Mar-00 SNDK Revenue 52.5 67.5 82.7 109.4 Sequential Growth 19% 29% 22% 32% YoY Growth 68% 111% 63% 148% SSTI Revenue 23.0 35.1 48.3 62.3 Sequential Growth 25% 53% 38% 29% YoY Growth 37% 94% 167% 240% And finally, remember that "The postage stamp-sized SD Card, developed by Matsushita, Toshiba Corp and U.S.-based SanDisk Corp" means that SNDK is in a three way split for profits… Please everybody, do not get me wrong: I love SNDK as evidenced by my investment in the company. But this "low density - low profit - greater competition" theory is, IMO dead wrong. Hope I have helped & good luck to all, Eric