SanDisk (NASDAQ: SNDK) The Flash Memory Pure Play|
SanDisk, a leading edge semiconductor manufacturer
and inventor of CompactFlash and the MultiMediaCard,
is well-positioned to benefit from the upcoming demand
for removable flash memory cards in the Post-PC era.
company website ==> sandisk.com
Part II of II
III. A Three-Pronged Business Model
SanDisk's balance and stability is based on three separate revenue streams.
see synopsis ==> Message 13371770
How does SanDisk make money? What are the principle sources of revenues? The answer to these two questions is much more complex than one might imagine. Several years ago SanDisk could have been described as a fabless, flash memory design house. Frequently smaller companies that are just starting out are unable to pony up sufficient cash for an investment in a fabrication plant that larger competitors can afford. As such they are "fabless" and generally thrive by coming up with innovative designs and offering them to larger companies or competitors in return for royalties or manufacturing considerations. With such an arrangement profitability is reduced because the company does not directly manufacture all components of the finished product. The flip side is that the company is less exposed to investment risk by avoiding substantial manufacturing overhead.
As a fabless company SanDisk has survived on three basic revenue streams:
i) direct sales (retail and OEM) of finished product,
example: a finished CompactFlash card relabelled by Kodak
or purchased for sale to consumers by BestBuy or Ingram Micro
ii) royalty/license fees based on "naked" flash memory chip manufactured by flash chip manufacturing competitors for card assembly competitors,
[example: flash chips manufactured by Hitachi under license and sold to Lexar Media for card assembly]
iii) royalty/license derived from CompactFlash assembly patents.
[***example: completely assembled CompactFlash cards manufactured and sold by SanDisk's competitor, SSTI]
[***note: As will be discussed below, the CompactFlash card assembly patents are contested and do not contribute significantly to current revenue streams.]
SanDisk is unique as a flash card manufacturer in that they control the production of their own flash wafer, design their own card controllers, and assemble their own finished product. This manufacturing control gives SanDisk the potential to become the lowest cost producer of finished flash memory cards. Because of the overwhelming demand for flash chips and steadily increasing competition for flash memory fabrication space the cost of running purely fabless operation has increased. In order to circumvent these issues in the long run SanDisk has teamed with one of its fiercest competitors, Toshiba Corporation, in a joint venture known as FlashVision LLC. This cooperative effort will create a leading edge, state-of-the-art fab in Virginia. The details of this joint venture have been known for several months since the original Memorandum of Understanding early last Fall. The importance of this venture cannot be understated.
The SanDisk/Toshiba Joint venture is arguably the most important cooperative agreement in the history of the company. Dr. Harari categorized the decision as "a critically important milestone".
Please visit the links below for further details:
1) Message 11474396
2) Message 11471706
The FlashVision JV melds SanDisk's extensive flash IP, NOR cell design expertise and multilevel cell technology with Toshiba's advances in NAND technology and superior lithographic skills. It sets the stage for great manufacturing efficiencies for both parties while minimizing overlap and competition in the market place (see Shareholders' Meeting notes for details below).
Dr. Harari emphasized the importance of controlling production in a recent interview...
It is clear that the decision to create FlashVision LLC was a move to break out of the restrictive nature of a fabless arrangement, especially as vacant space for custom fabrication of flash wafers is essentially spoken for. This decision was spurred by SanDisk's rapid growth and maturation and the enormous size of the target markets. SanDisk simply outgrew the fabless mold that was cast early on in the life history of the company. Additionally, the joint venture between SanDisk and Toshiba also distributes SanDisk's manufacturing risk over a larger geographic area with the bulk of production in 2001 and beyond likely coming off production lines in Manassas, VA, the site of the FlashVision fab (see investment risks below).
IV. The SanDisk Intellectual Properties
The collective power of over 100 individual patents remains an important intangible asset for SanDisk.
SanDisk has made more than 100 individual contributions to the field of flash memory. This collection of novel ideas has a substantial, yet relatively intangible value. These patents include not only core technology of flash cell architecture and design, but also original ideas pertaining to the implementation and successful integration of flash memory within central processors and host devices. I divide the various patents into three distinct categories: core technology, component technology and card assembly. The value of this body of work may not reside in any individual patent, but in the broad scope that this prolific work encompasses. This ability to build on a series of related advances with the general discipline of flash memory is a hallmark of SanDisk.
The result of this collective effort has just recently started to bear fruit. SanDisk had an undeniably successful year defending its intellectual properties in 1997. In fact, 1997 was a pivotal year for the company. During that Spring and Summer a rapid succession of legal victories established SanDisk as a major player in the high density flash memory market. Three industry giants, Toshiba Corporation, Hitachi and Samsung, all entered into cross-licensing agreements with SanDisk. These agreements came on the tail of previous successful licensing arrangements with Intel, Sharp and Matsushita.
This series of legal victories started in March of 1997 when Samsung was forbidden by the ITC to import their own variety of flash memory chips which were determined to infringe on SanDisk intellectual property.
Samsung decision ==> techweb.com
There was initially some reluctance on Samsung's part to comply with the enforcement of the SanDisk patents in question which lead to the need for ITC intervention. Samsung, at the time a licensee of Toshiba's NAND technology, was placed in a disadvantageous position in August of 1997 when SanDisk brokered a separate co-licensing agreement with Toshiba...
"The agreement may also strengthen a SanDisk suit filed in federal court, charging Samsung and its subsidiary, Samsung Semiconductor Inc., San Jose, with violating two key NAND-based patents, according to industry observers. Samsung, which has refused to pay royalties to SanDisk, is itself a NAND-flash licensee of Toshiba and could have been hoping its partner would serve as a legal ally, said Bruce Bonner, an analyst with Dataquest Inc., San Jose. But now that Toshiba has struck a deal with SanDisk and has agreed to a royalty schedule, it is unclear if Toshiba will be able to plead Samsung's case, he said. 'This [cross-license] makes it much, much more difficult to come up with what [Samsung] would consider a good deal,' Bonner said. "[Samsung] got one of their legs shot out from under them, partly because their chief technology partner just defected to the enemy, and also because they are relatively new to the flash market and don't have an extensive flash portfolio to fight back with."
see ==> techweb.com
Dr. Harari stated at that time that "Samsung ha(d) said that they will design around the '338 patent to avoid being subject to the ruling, but I doubt they can."
A short time later Hitachi succumbed to pressure to recognize the strength of the SanDisk patent portfolio "in an effort to steer clear of the legal quagmire that has halted shipments of Samsung's" flash memory chips. Thus, the third large domino fell in rapid succession.
see ==> techweb.com
It is important to recall that these agreements antedated the wide-spread acceptance and distribution of CompactFlash and MMC in the US. Instead, the patents in dispute related to core flash memory technology at the chip level and were not limited strictly to NOR cell designs that are SanDisk's specialty. The crucial portions of these patents likely relate to unique functions at the chip level such as addressing and erasing of cell blocks, detection and elimination of individual cell failure, "wear leveling" and other maintenance functions necessary to prevent data corruption, the most essential feature of any data storage device regardless of type. The tangible value of these cross-licensing arrangements is difficult to estimate as the terms of each agreement is highly confidential. Some of the agreements undoubtedly call for a "swapping" of technology, while others command payments in the form of royalties which are presumably linked to production. This is suggested by the gradual, yet steady rise in royalty payments which now total approximately $12,000,000 to $12,500,000 on a quarterly basis.
Perhaps the greatest event of 1997 was the awarding of an unrelated patent that has become the most fiercely contested portion of SanDisk's IP. This is the so-called '987 patent that deals with disk drive emulation. It is this functionality that allows products of such as the CompactFlash cards to be universally accepted ("plug-n-play") in a variety of host devices. As is explained at the SanDisk Internet site...
"SanDisk's long history of technological innovation and industry leadership has led to the development of a flash family of products unparalleled in its scope and breadth. The company has five flash data storage product lines: the PC Card ATA FlashDisk, CompactFlash, MultiMediaCard, IDE FlashDrive and the IDE Flash Chipset. In addition to high density, patented SanDisk flash chips, all of the products include an intelligent controller which effectively places the burden of card management on the card itself, not on the host system in which it is used."
Many have doubted that the '987 could be enforced successfully. In order to prove the validity of the '987 disk drive emulation patent SanDisk engaged in licensing discussions with a competitor, Lexar Media, a company specializing in flash memory card assembly such as CompactFlash. These discussions were not concluded successfully and a lawsuit ensued. After two years of legal maneuvers Lexar Media was found recently found guilty of contributory infringement. SanDisk was awarded a partial summary judgement against Lexar Media who has subsequently decided to pursue a "patent nullification" trial in US Patent Court. The Lexar claims have been aggressively pursued by SanDisk. It would seem highly probable that the decision to go after Lexar Media has, in part, to do with the fact that many of the Lexar Media executives are prior SanDisk employees. In fact, the Lexar Media CEO, Mr. John Reimer, worked in the marketing department at SanDisk prior to breaking away to start the new company. In court documents to date it has been purported that Mr. Reimer admitted that Lexar infringed on SanDisk patents potentially exposing the company to payment of further damages if "willful infringement" could be demonstrated. The case has many complex features that are beyond the scope of discussion here. Perhaps the most peculiar facets is the fact that several SanDisk competitors utilized Lexar Media technology in their own flash products at about the same time the licensing discussions with Lexar Media were proceeding. Indeed, several of the founding directors of Lexar Media have since divested their interests in the company to start yet another flash card assembly company that recognizes SanDisk's IP. The details for those who are interested are included in the link that follows...
Whether SanDisk will prevail in the upcoming '987 patent nullification trial is unknown. Some industry analysts who were skeptical that SanDisk would prevail early on in the proceedings seem to be softening their stance...
[see ==> techweb.com
"Alan Niebel, director of non-volatile memory research at Semico Research Corp. (Phoenix), said the SanDisk patent is broad because it covers what was a novel concept when it was issued in the late 1980s. While inclusion of a controller on a flash card is fairly common today, many of the vendors using the technology are SanDisk licensees. That list includes Intel, Samsung Electronics, Hitachi, Sharp and Silicon Storage Technology. "Anything that emulates a hard drive using flash that is removable and has a controller is covered by that patent," said Niebel. "I think it's too broad, and that's what Lexar is arguing, but so far SanDisk has been able to prevail. The patent has been unchallenged to date."
The fact that SanDisk competitors such as Hitachi, Samsung and Toshiba are content to sell unassembled flash components or controllers to much smaller CompactFlash card assemblers and not assemble or sell CompactFlash cards themselves here in the US suggests that the perceived legal consequences for infringement of the '987 patent is prohibitive and a strong deterrent to entry into the US markets. It is unclear whether the patent will be able to be enforced in the Japan where a separate lawsuit against Mitsubishi Electric is currently contemplated after licensing discussions deteriorated. Those interested in further conjecture regarding this particular matter are directed to the following links...
1) Message 13517245
2) Message 13792202
Perhaps the second most promising patent holdings in the SanDisk patent vault is the technology (and engineering talent) recently acquired from INVOX. This privately-held company is a technology leader in the area of multi-level cell (MLC) construction. The current MLC technology used by SanDisk called "D2" is currently used to double wafer production capacity by encoding two bits of data in a single memory cell. This requires the three separate electrical potentials in a single cell rather than one (and a ground state). The future development of a "D3" MLC (seven electrical potentials plus ground) or a "D4" (15 electrical potentials plus ground) may be facilitated by technolgy from INVOX. Gigabit flash memory density will likely require the engineering of higher degrees of MLC. Invox's technology using analog cell technology, NOR flash cells and digital-to-analog conversion to assign 256 unique energy potentials in a single cell.
see ==> eetimes.com
It is understandable, therefore, that a former Invox lead engineer will head SanDisk's trek to Gigabit flash chips.
"SanDisk will hire an undisclosed number of Invox employees under the agreement, including Geoff Gongwer, vice president of engineering at Invox. Gongwer, who has worked at Atmel, Intel, and other chip makers, has been appointed program director of SanDisk's 1-Gbit flash-memory program."
see ==> semibiznews.com
V. Competing Technologies
Are there alternative solutions which may dethrone flash memory?
SanDisk's competition is encountered on several levels. First, chip manufacturers such as Hitachi and Samsung will compete with FlashVision LLC to produce the lowest cost, highest density flash memory. The overall market for flash memory could be oversupplied if capacity is built to rapidly. This eventuality is difficult to predict. The concept of "market elasticity" suggests that as higher and higher density flash chips are produced, new applications will be created to gobble them up. It may be some time before world-wide flash output saturates all of the potential applications that exist. In fact, it is difficult to predict what new megamarkets may arise purely on the basis of the availability of extremely high capacity flash memory cards. Thus, SanDisk has placed most of the pressure on itself to meet skyrocketing demand for its products. It tends to see other flash manufacturers as promoters of flash technology rather than true competitors. This sentiment was clearly stated by Dr. Harari in a recent interview:
Other levels of competition include card manufacturers who are producing flash memory cards such as CompactFlash from alternative sources. These manufacturers are needed to help meet demand and serve as "second sources" for OEM's to keep pricing competitive. As members of the CompactFlash Association they help to establish CompactFlash as the de facto standard for consumer electronics applications. If, however, these companies design around SanDisk's card assembly patents the company could enter into a commodity environment with diminishing margins. The importance of the Lexar Media suit cannot be understated here.
The last level of competition comes from non-flash technology or alternative flash form factors. The mechanical drives such as the IBM microdrive and the Iomega Clik! have proven to be less competitive that previously expected. The microdrive has several disadvantages that have been discussed at length on the Silicon Investor SanDisk Thread. These disadvantages include size, entry cost, power consumption and ruggedness. The microdrive will likely be relegated to a niche market catering to higher end professional photographers and certain hobbyists. The Iomega Clik! has not penetrated any of the mentioned Post-PC markets and is incompatible with the storage slots selected by OEM's. Power consumption, switching costs to OEM's and ruggedness issues typical of rotating disc drives will likely limit the popularity of Clik!
The final viable competitors are non-CompactFlash/non-MMC flash form factors. Included in this category is SmartMedia and the Sony Memory Stick. Because the SmartMedia card will be a FlashVision product, SanDisk is likely to benefit from the continued sale of SmartMedia. The general trend has been for increasing CompactFlash and MMC design-in wins a the expense of SmartMedia. As such, SmartMedia poses a diminishing threat to SanDisk. Likewise, The Sony Memory Stick, essentially a duplicative flash form factor which has no enhancements or advantages when compared to CompactFlash, is struggling for design-ins. In the final analysis it also remains unclear how Sony will manufacture Memory Stick given their relative lack of experience in flash memory design and lack of flash chip manufacturing capacity. It is not beyond the realm of possibility that Sony may become a FlashVision customer in the future. Currently they have expressed an interest in partnering with Lexar Media for the production of Memory Stick. Because Lexar also has no flash memory fabrication capacity this relationship may offer limited production capacity to Sony.
VI. Investment Risks
Forces which could derail SanDisk in the near-term and beyond.
SanDisk's greatest risks at present, in addition to the competition described above, include political instability in Taiwan and the risk of natural disaster on the island. There has been a recent change in the ruling party in Taiwan after elections this March. The relationship with the mainland Chinese remains unpredictable and potentially unstable. Political skirmishes between these two countries could have a material effect on SanDisk's production capacity.
Additionally, Taiwan suffered a devastating earthquake in September of 1999 that damaged necessary infrastructure and halted production of SanDisk's fabrication line at UMC. The structural damage to UMC has been minimal, however disasters of similar magnitude could materially impact SanDisk's production capacity that is currently concentrated on the island.
The following post is extremely helpful in understanding the situation on the island:
The 2000 SanDisk Shareholders' Meeting
PLEASE REFER TO THE ANNUAL REPORT FOR A COMPLETE LIST OF INVESTMENT RISKS, COMPETING TECHNOLOGIES AND POSSIBLE POTHOLES IN THE ROAD TO SUCCESS FOR SANDISK
Best of Luck to All SanDisk Longs!
SanDisk...See the BIG picture!
[formerly: What are their prospects?]
started by Bhag Karamchandani on Jan 31, 1996
Would any one care to enlighten the rest of us as to how SanDisk is positioned in the Flash memory market in relation to their competition and earnings potential?