The Rambus Opinions: CAFC rules on the “reasonable forseeability” of litigation

By Jason Rantanen

Micron Technology, Inc. v. Rambus Inc. (Fed. Cir. 2011) Download 09-1263
Hynix Semiconductor Inc. v. Rambus Inc. (Fed. Cir. 2011) Download 09-1299-1347
Panel: Newman (concurring and dissenting in part on Hynix), Lourie, Bryson, Gajarsa (concurring and dissenting in part on Micron and Hynix), and Linn (author of both Micron and Hynix)

Rambus is a company founded to commercialize inventions related to DRAM computer memory, including its RDRAM product.  It is currently engaged in several patent suits against manufacturers of SDRAM, an alternate type of memory technology.  Central to these cases is the issue of whether Rambus's destruction of documents – particularly a "shred day" during which thousands of pounds of documents were destroyed – prior to the commencement of the litigation constituted spoliation.  Given this case's discussion of document retention obligations, it has particular relevance for in-house counsel and outside counsel advising companies prior to the commencement of litigation.  Irrespective of the legal question discussed below, the specific factors discussed at pages 16-23 of Micron are worth reading.

Procedural Posture
Although addressing the same fundamental issue, Micron and Hynix arrived at the court in diametrically opposed postures.  In the district court proceedings on appeal in Hynix, Judge Whyte of the Northern District of California concluded that no spoliation occurred, while in the proceedings leading to Micron, Judge Robinson of the District of Delaware ruled that Rambus had engaged in spoliation and entered judgment in Micron's favor as a sanction.

Litigation Must Be Reasonably Forseeable
The resolution of the two appeals turned on the issue of when litigation became "reasonably forseeable," thus triggering Rambus's duty to preserve evidence.  Rambus argued that "to be reasonably forseeable, the litigation must be imminent, at least in the sense that it is probable and free of significant contingencies."  Micron and Hynix, on the other hand, aruged that reasonable forseeability incorporates no requirement of imminence. 

The five-judge panel unanimously concluded that the "reasonably forseeable" standard does not carry an imminence requirement.  While "[t]his standard does not trigger the duty to preserve documents from the mere existence of a potential claim or the distant possibility of litigation…it is not so inflexible as to require that litigation be imminent, or probable without significant contingencies,' as Rambus suggests."  Micron Slip Op. at 12-13. 

Applying this standard, the CAFC determined that the Micron district court had not clearly erred when it concluded that litigation was reasonably foreseeable prior to the critical 'shred day.'  In particular, the CAFC noted that (1) Rambus's document retention policy was not adopted for purposes of business management, but instead to further Rambus's litigation strategy; (2) Rambus was on notice of potentially infringing activities by particular manufacturers; (3) Rambus took several steps in furtherance of litigation prior to the key shredding party; (4) Rambus is the plaintiff-patentee [but see note below], and its decision whether to litigate or not was the determining factor in whether or not litigation would ensue; and (5) the relationship between Rambus and the DRAM manufacturers was not one of mutual benefit that subsequently turned sour, but rather was naturally adversarial. 

Arriving at a consistent outcome in Hynix, three of the five judges (Lourie, Byrson, and Linn) concluded that Judge Whyte erred in determining that litigation was not reasonably forseeable by inappropriately applying the stricter "imminent, or probable without significant contingencies" requirement for reasonable foreseeability.  "In Micron II, this court held that that standard does not carrry a gloss requiring that litigation be "imminent, or probable without significant contingencies….The district court here applied just such a standard. This is evident for three reasons." Hynix Slip Op. at 13 (internal citation omitted).

Yet although the reversal is clear, the majority's reasoning for its conclusion is not.  The majority first held that the district court erred not by considering the presence of contingencies at all, as one might expect given its preceding statement, but by failing to decide "that the resolution of these contingencies was reasonably foreseeable."  Id. at 14.  The majority's reason thus implicitly assumes that contingencies are, in fact, relevant to the foreseeability inquiry.  The second and third reasons – that the evidence of non-foreseeability was insufficient and the Hynix district court was the only court out of four to conclude that litigation was not reasonably foreseeable as of the critical date – similarly have little to do with whether the Hynix district court applied the correct legal standard.  The majority's actual analysis thus seems at odds with its statement that the foreseeability standard does not include a 'probable without significant contingencies' requirement given that its only relevant basis for reversal applies exactly such an analytical framework.  It is further puzzling because it involves an assessment of purely factual issues, not legal standards, to which it applies a de novo standard of review.

The Dissent
Writing in dissent in Hynix, Judge Gajarsa joined by Judge Newman takes the majority to task for its overriding drive to arrive at a consistent result in Hynix and Micron, a drive that in their view produced an erroneous result:

Seizing on the district court’s alleged grafting of an overly strict “gloss” on the reasonably foreseeable litigation standard, the majority claims that the district court erred as a matter of law by applying the wrong standard for spoliation. In so doing, the majority ignores the district court’s well-articulated understanding of the relevant Ninth Circuit law and its factual findings, which demonstrate that the district court applied the very standard that the majority now requires. The majority obtenebrates the facts presented in the district court’s opinion to resolve the conflict between the spoliation determinations in this case and in Micron.

Hynix dissent at 2.  In particular, the dissent criticizes what it perceives to be the majority's attempt to cloak its review of a factual determination as a legal issue in order to avoid applying a deferential standard of review, thus leading to an inconsistent result:

The majority created this standard in Micron with the intention of reconciling the district court’s spoliation determination in that case with the present one. The desires of an appellate tribunal, however, should not drive the outcome of decisions. Divergent district court opinions do not necessarily mean that the trial courts applied different “glosses” on a particular standard. Rather, this court must review district court opinions with the applicable standard of appellate review, which, in this case, is clear error or an abuse of discretion. In refusing to do so, the majority shirks its duties and performs its own factual analysis. The majority castigates the district court here for using a narrower “gloss” on the standard for spoliation just as the majority in Micron castigates the district court for holding the asserted patents unenforceable as a sanction for spoliation.

Id. at 3.

A Meaningless Remand?
Although affirming the Micron district court's conclusion that litigation was reasonably foreseeable, the majority nevertheless vacated the district court's sanction determination.  This move is unlikely to produce in any meaningful change.  The majority's basis for the remand was that the Micron district court opinion failed to provide sufficient detail or analysis in support of its conclusion that Rambus acted in bad faith, that Micron suffered prejudice, or that judgment in Micron's favor was the appropriate sanction to impose.  Given Judge Robinson's prior opinion, however, the inevitable result will be a more detailed opinion that reaches precisely the same conclusion on these issues. 

Judge Gajarsa dissented from this determination as well, again criticizing the majority for reviewing the facts and weighing the evidence "before it substitutes its judgment for that of the district court, deciding that based on the district court's thorough factual analysis, it would not have granted the dispositve sanctions."  Micron dissent at 2.

Note on consistency: In concluding that the Micron district court correctly concluded that litigation was reasonably foreseeable, the CAFC relied on the fact that Rambus was the plaintiff-patentee (and thus controlled the initiation of the litigation). Micron Slip Op. at 22.  Yet both the caption and the opinion's procedural background indicate that Rambus did not initate the lawsuits; rather they were declaratory judgment actions brought by Micron and Hynix. 

Disclosure: Although I had minimal exposure to the Rambus litigations while an attorney with Munger, Tolles, & Olson, note that my former employer did and continues to represent Rambus.

14 thoughts on “The Rambus Opinions: CAFC rules on the “reasonable forseeability” of litigation

  1. It didn’t work.

    They needed other art, using other elements combined in a different way, operating by different means achieving different results and not of their devising in order to complete their Rambus and make it functional. Some of that they filed patent on, but when they did, they didn’t disclose to the patent office that it formed new matter in the continuation of their Rambus inventions. They were already involved in the process of mining their earlier, defective, construction for claims to assert against the SDRAM standards they were participating in the development of.

    Now they have baseless claims and a history of spoliation, false and misleading statements under oath, litigation misconduct and forum shopping.

    Face it, as a patentee, when Pauline Newman goes all in on an opinion labeling you as a spoliator and another upholding the piercing of claims of privilege on the basis of good reason to believe there was intention to commit a crime and or fraud in the course of your litigation, it’s time to suddenly start acting like a good citizen for a change since the only remaining topic of discussion is which part of your anatomy would best benefit from corrective adjustment.

    So, far from “kicking dirt” I might actually be trying to be a bit helpful.

  2. Mud, I’m not following you. Did the original Rambus patent work for its disclosed purpose?

    You make a lot of arguments about what Rambus said about its technology post filing, most of it being untrue according to your account, but is all that really relevant to whether the original disclosure was operable for its disclosed purpose?

    If you were going to hold their patents invalid, what is the statutory reason? Merely kicking a lot of dirt towards Rambus in Rambus’s obvious efforts to have its own technology adopted or its rival pay royalties is unpersuasive. Rambus had every right to do what it apparently did so long as it did so within the law — but all that enforcement activity has nothing to do with validity, but with, perhaps, enforceability.

  3. “de novo” review of sanctions judgment based on underlying facts – are sanctions a legal, or an equitable, remedy? Equitable remedies, and the finding of associated facts, always are in the discretion of the court. Thus, de novo review (and a “clear error” standard) are appropriate, even regarding the court’s factual analysis. Here’s a shout-out to legal historians.

  4. Whether they withheld knowledge as of their initial filing date isn’t necessarily to the point when their prosecution of claims extended across more than fifteen years and a “family” of more than fifty patents. Their expressed purpose in those prosecutions was to prepare a “patent minefield” to attack a set of public standards they had participated in the establishment of. Since they clearly and convincingly knew of superior technologies and had abandoned their 1990 “invention” for any other purpose than to “salvage” claims that they could use to “shoot” competing interface specifications “in the head” how could anyone not consider that to be inequitable conduct in prosecution of claims while withholding information undoubtedly known to them and material to patentability.

    It’s not the “commercial viability” aspect of any combination of elements at issue here; they so quickly abandoned every aspect of their 1990 “invention” that it’s never been clear that even they attempted to reduce it to practice in anything other than the failed construction of their initial written description. Their willingness to mislead the public by implying that their original description was capable of providing benefit equivalent to what their commercial implementations had (and to use that implication to attack competing and often superior technologies in the marketplace) is the underlying problem with their patent practice and their litigation malpractice.

  5. Two cases, one set of facts?, how can the results be different? Allegedly, the Hynix court had the law wrong, but as Jason points out, it really didn’t.

    I was left with the definite impression that the law is as was actually applied by the Hynix majority, and the result should have been the result found by the Hynix DC.

    But the handwriting is on the wall. It reads, Rambus’s patents will be held unenforceable due to spoilation, based on bad faith, when the law itself on the issue was as clear as “mud” and at least one court and one Federal Circuit judge agreed they had followed the law.

    BTW, on the issue of bad faith, did Rambus rely on opinion of counsel and did that factor in?

  6. Dennis:

    Care to run a story on Disney tradmarking “Seal Team 6″?

    link to theatlanticwire.com

    It might be a good opportunity to teach and/or learn a little about trademarks, especially considering the shockingly lame part where they actually granted the trademark.

  7. Jason: “CAFC relied on the fact that Rambus was the plaintiff-patentee (and thus controlled the initiation of the litigation). Micron Slip Op. at 22. Yet both the caption and the opinion’s procedural background indicate that Rambus did not initate the lawsuits; rather they were declaratory judgment actions brought by Micron and Hynix.”

    It wasn’t well stated, but what they were referring to was not the actual Micron case before them but the litigation Rambus was getting ready for somewhere down the road or over the rainbow. Probably over the rainbow. The context of the CAFC statement was that the actions of those preparing to initiate a fight are scrutinized over a different time frame than those who get slammed with a subpoena out of nowhere.

    I’d be interested in others’ thoughts on the issue of outside IP counsel shredding docs.

    I am not going to shred clients’ docs until the SoL on malpractice runs — usually 5 or 6 years depending on jurisdiction. Period. I don’t care what instructions I get from the client. It is the client I’m protecting myself against.

  8. Without otherwise getting into any of the above, I am simply curious about what has always seemed to me to be very weak case law support for such an alleged defense of “no utility” [short of no operable specification enablement or a mere chemical intermediate]? Or, did you just mean to suggest that the only disclosed embodyment was clearly not the “best mode” known to the inventors as of their filing date because it was not commercially useful?

  9. Addressing the failed “consistency” that Rambus’ apologist (or advocate) wishes to advance:

    “The record is lengthy but uncomplicated. In 1990,
    Farmwald and Horowitz filed their first patent applica-
    tion directed to improving the speed with which computer
    memory can function. Rambus was founded the same
    year to commercialize this invention. Rambus developed
    its proprietary RDRAM technology, and licensed chip
    makers to manufacture memory chips incorporating this
    technology. Around this time, JEDEC was working to
    develop industry standard specifications for memory chips
    and the interfaces between memory chips and computer
    processor chips, eventually adopting its first SDRAM
    standard in 1993.
    In approximately 1992, Rambus
    learned of SDRAM and came to believe that the Farm-
    wald and Horowitz invention encompassed SDRAM.
    Rambus continued prosecuting multiple patent applica-
    tions in the Farmwald/Horowitz family, intending to
    obtain issued patent claims that covered SDRAM. Ram-
    bus thereafter pursued a two-prong business strategy: it
    licensed chip makers to manufacture chips that complied
    with Rambus’s proprietary RDRAM standards, and
    prepared to demand license fees and to potentially bring
    infringement suits against those manufacturers who
    insisted on adopting the competing SDRAM standard
    instead.”

    Rambus began its scheme to co-opt the JEDEC standards it was participating in the promulgation of concurrently with its joining of JEDEC. This is consistent with its prior behavior of surreptitiously monitoring the work of the IEEE Scalable Coherent Interface working group and attempting, in its own incompetent way, of cobbling together a local small system interface as a “knock-off” with (putatively, since the initial attempt was laughably non-functional) equivalent performance characteristics. It has never made showing that anyone is or was making, selling or using its “RDRAM” or “Rambus Access Controllers” without license (except perhaps when it terminated Samsung’s license “as a pretext” to bring an action based on claims it had previously lost due to litigation misconduct.

    Rambus’ complaints center around the fact that it was incapable of devising a successful data interface applicable to devices incorporating DRAM arrays, much less the “do it all: system interface they allegedly “envisioned” in their bongwater induced ravings from 1990. The fact that, even when all of their initial cock-ups were “fixed” they had a poorly implemented, overpriced set of moving target implementations that never provided the performance they touted in their “airwars” against more sensible engineering solutions should be no surprise considering how incompetent they showed themselves to be at the very start.

    All in all, the story is quite consistent. Instead of making a useful data interface adaptable to devices incorporating DRAM arrays, Rambus beavered away attempting to write claims that the SDRAM standards before them would “necessarily” infringe. Unfortunately, the only prospect for claims of priority was their useless 1990 “RamBus” specification, which they had already abandoned as useless by the time they began their scheme. In spite of nine years of effort, when they launched their “Lexington” litigation campaign in early 2000 with their preposterous attack on Sega and Hitachi, they still (unsurprisingly) had nothing with any basis in a disclosure of an actual utility to work with. Trying to play “defendent under attack” in light of the facts of their myriad abuses of the legal system and patent process is only consistent with their nauseatingly old M.O.

  10. Raspberries.

    Why did Rambus not disclose to the patent examiner during its prosecution of “clock” and “synchronous to an external clock” and “DLL/PLL” claims that there was no utility in the clocking scheme of its 1990 patent filing and that an alternative clocking scheme not devised by the claimants Farmwald and Horowitz was the “best mode” of practicing their “Rambus” invention subsequent to their collaboration with Toshiba in 1991?

    link to google.com

    link to google.com

    Why did they not disclose that the “clocked sense amp” samplers were not useful in either their Rambus system and had been abandoned around 1992 also?

    link to google.com

    Why didn’t they point out that the “midpoint clock” and “rising edge / falling edge” aspects of their “invention” of 1990 were already present in the work of one of Horowitz’ graduate students more than a year earlier than their initial patent filing?

    link to www-vlsi.stanford.edu

    “Latency registers” for controlling data output timing are there as well.

    Just the tip of the iceberg of what Rambus was thinking it could get away with. They can’t and won’t.

Comments are closed.