To: jim kelley who wrote (54288)
From: Estephen Wednesday, Sep 20, 2000 5:58 PM ET
Reply # of 54321
Micron the Ethical (Pt. 1/3)
Thanks, Jim and ptnewell of yahoo for the great post!!! ..Est
by: ptnewell 9/20/00 12:52 pm
Msg: 162402 of 162556
While practicing the ethically-challenged politics of Depression era Texas, Lyndon Baines Johnson proclaimed, "If you can take their money, drink their liquor, sleep with their women, and still vote against ‘em, you’re an honest politician." The only U. S. President to enter the Oval Office a pauper and leave a multi-millionaire, Johnson always insisted that he had firmly developed set of principles.
Contagious principles, it seems. Latitudinally opposed Idaho might appear to have little in common with Texas of the thirties. Known mostly as a haven for neo-Nazis, separatists militia groups, and the Klan, Idaho is also home to the last American Dramurai, Micron. Having taken $500 million of Intel’s money, offered to ensure adequate supplies of Rambus’s RDRAM, and after having made billions more off Rambus’s 1990 invention of synchronous DRAM (SDRAM), Micron has decided that its own peculiar brand of principles require its uproots exertion to quash the upstart company. CEO Appleton’s motto appears to be, "You can take their money, and live off their IP, so long as you work your butt off to undermine them."
Micron, along with Hyundai, announced plans to ramp up RDRAM production in the early spring of 2000. These announcements had the result of freezing Samsung and NEC RDRAM production at levels inadequate to supply the workstation market. The resulting shortages led to extraordinarily high prices for RDRAM in the "spot" market. That artificial shortage and the resulting high prices were then used as powerful ammunition by the anti-Rambus (pro-DDR) camp. It was mid-May before Samsung realized what Micron and Hyundai were doing, and Samsung then doubled its own RDRAM production. Samsung increased the production of 128-megabit RDRAM from 2 million units/month to about 4 after June, contrary to its earlier plans to hold production steady until the arrival of the Pentium 4. Micron and allies (such as Semico) then energetically pushed for Intel not to launch the Pentium 4 with an RDRAM only chipset because of those same shortages.
However Intel has held firm to an RDRAM-only launch of the Pentium 4. (Because RDRAM cannot constitute the majority of DRAM before 2003, Intel of course must eventually introduce an SDRAM version of the Pentium 4.) Very recently, NEC has joined Samsung in increased RDRAM production, leading to a rapid drop in the spot prices – which however remain well above actual production costs. Incidentally, Dell was the big winner of the entire episode. Having pre-ordered adequate RDRAM supplies from Samsung, Dell dominated the workstation market in spring 2000 as never before.
Micron still worked the press energetically, however. In July, a CNET report stated that Micron was cutting RDRAM production back from 7% in the first half to 0% in the second. In fact, Micron not only shipped no RDRAM in the first half, they were not even qualified to do so by Rambus and Intel. (Micron has quite recently qualified a 128-megabit RDRAM unit, but appears not to have shipped a single stick). Thus Micron managed the impressive trick of making it appear RDRAM production was falling at a time when it was rapidly rising.
9/20/00 12:52 pm
Msg: 162404 of 162556
The most overt step action taken by Micron, of course, is the lawsuit alleging anti-trust violations against Rambus. The lawsuit smears Rambus with bracing boldness, yet does so with such exquisite mastery that plausible deniability is retained. Much is made in the complaint of patent continuations and extensions granted Rambus. Much is also made of Rambus having participated in the committee meetings of a trade organization, JEDEC. Micron is clearly hoping that readers will infer that something untoward is involved. Indeed, many Rambashers, drawn largely from the ranks of Intel haters, have gleefully drawn precisely this unstated smear. A few media reports have actually stated that Micron is making such allegations.
They are not. (The text of Micron’s complaint can be found here: (http://www.swiftsolution.com/sites/rambusite/getpage.asp?pag... MU_Suit). Modifying a patent application to incorporate JEDEC material into a Rambus patent dated 1990 would be a federal crime, and if Micron had plausible evidence of such, they would have supplied it to the Department of Justice. In fact neither Micron nor anyone else has cited anything in the 1990 patents which is misdated. By now it is safe to assume lawyers for Micron, Hyundai, and Infineon have carefully checked the original 1990 application (as for that matter, must have Toshiba, NEC, Oki, and Hitachi, not to mention the U.S. Patent office’s own safeguards). Anything amiss would have been loudly trumpeted long since.
There is nothing sinister or even unusual about the subdivision of the original 1990 patent application. The invention of synchronous DRAM involved, to the U.S. Patent Office, not one invention but several. For example, the CAS latency register, that allows the CPU to know when to expect requested data is one individual invention (ultimately U.S. Patent 5953263) contained in the original application date April 18, 1990. It was reasonable for a tiny startup, still many years from going public, to attempt to fit multiple related inventions into one patent application fee. It was reasonable of the patent office to require subdivision (and multiple application fees).
Thus to believe that Rambus is not the inventor of synchronous DRAM requires either willful ignorance or great credulousness. Micron, of course, is not such a babe in the woods. Although encouraging the smear, Micron is basing its actual legal position on other grounds. Micron argues that by joining JEDEC in 1992, Rambus knowingly entered into an implied contract requiring them to surrender their crown jewels, the extraordinarily valuable 1990 patents. In exchange, Rambus received the dubious benefits of joining that debating society. Note that unlike the Dell case Micron cites, Rambus signed no documents explicitly requiring patent disclosures or sharing (at least not with JEDEC).
Rambus argues that JEDEC disclosure guidelines circa 1992 were not mandatory, were not widely adhered to by others, and that in any event they did comply with those guidelines. For example, Rambus states that they did formally raise the 1990 patents with JEDEC in 1993. Some of these issues of fact may have to be resolved in court. Note however that both Intel and NEC, the world’s largest and second largest chipmakers, signed non-disclosure agreements with Rambus by 1991 to gain access to Rambus technology. Certainly these, and probably other key members of JEDEC were fully aware of Rambus’s patents. (Incidentally, Rambus also owns European patents on the same technology. European patent applications are public shortly after filing. I have not been able to determine when the Rambus European patents became publicly available
Ignore this User | Report Abuse
Micron the Ethical (Pt. 3/3)
by: ptnewell 9/20/00 12:53 pm
Msg: 162405 of 162558
Indeed, it is questionable whether SDRAM is really a JEDEC standard, or whether it is the Intel standard that really matters. Intel has for many years laid down SDRAM specifications for DRAM manufacturers to match. Given Intel’s enormous clout, it seems clear that the DRAM manufacturers are at least as concerned about Intel’s specifications as JEDEC’s. There is no good reason to believe that Intel would be reluctant to chose a memory standard because it would bring Rambus royalties.
Micron doubtless thought it had solid strategic reasons for bringing such an implausible case. For example, they appear incensed that Rambus is charging higher royalties on DDR than on RDRAM. They may hope the lawsuit allows them to negotiate from a position of greater strength. Micron may also hope that the discovery process will turn up something damaging to Rambus. By launching the suit now they likely hoped to freeze out any further Dramurai from signing with Rambus. Micron and Hyundai certainly coordinated strategies (based on reports in the Micron-friendly Cahner’s media publications); they likely hoped to intimidate Rambus with two more lawsuits (with utterly different approaches) while Rambus was already occupied suing Infineon. Finally it often pays just to look tough. Other patent holders with weaker claims (e.g., with a single very general patent) are after the Dramurai; Micron could reasonably hope that a strong show of opposition to even the holder of iron-clad multiple patents would send a discouraging message to those others. No doubt, Micron feels it could offer Rambus a settlement at any time it chooses and have it speedily accepted.
Micron has miscalculated. The huge chipmaker NEC signed with Rambus despite the litigation (surely testimony as to the strength of Rambus’s patents). With RDRAM shipments rising strongly, and the new NEC payments on SDRAM to begin shortly, Rambus is very well funded for legal battle. Far from intimidated by facing three simultaneous lawsuits, Rambus has upped the ante to four simultaneous lawsuits – apiece. Rambus is suing Hyundai, Micron and Infineon individually in France, Britain, and Germany, in addition to the U.S. action. Rambus has also filed with the U.S. International Trade Commission to block imports of DRAM from Hyundai. Rambus appears to be sending a deliberate message that Micron and Hyundai will never win through intimidating Rambus by litigation costs. Moreover, the German courts, in particular, offer much quicker trials. As the party with the strongest case, a quick trial date is naturally to Rambus’s liking. Indeed, it was after Rambus filed a case in Germany against Hitachi, forcing the latter to face a quick trial, that Hitachi folded.
Rambus has consistently maintained that it will not license adjudicated infringers. Which means: if Micron goes the distance and loses, they lose the right to manufacture SDRAM or DDR. Of course that still leaves Micron with EDO. Micron is proud to be the leading manufacturer of this dying DRAM technology. Meanwhile the major Dramurai with a tradition of innovation and forward thinking – Number one Samsung, giant NEC, and PlayStation2 designer Toshiba – are all betting on RDRAM as the future. All are on excellent terms with Intel and Rambus.
The pro-DDR, anti-Rambus alliance of Micron, Hyundai, and Infineon is not likely to survive long under the triple threat of early trial dates in Germany, soaring RDRAM market share, and a DDR movement that has remained stalled at "a few months away" for years. Some expect Infineon to break ranks first, since it has the earliest trial date (December, 2000). However my money is on Micron to cut a deal, leaving its partners alone in the lurch. It would be in accordance with their principles. |
