LINN, Circuit Judge.

Rambus Inc. ("Rambus") appeals the decision of the United States District Court for the District of Delaware holding that the twelve Rambus patents asserted against Micron Technology, Inc., Micron Electronics, Inc., and Micron Semiconductor Products, Inc. (collectively, "Micron") are unenforceable due to Rambus's spoliation of documents. Micron Tech., Inc. v. Rambus Inc., 255 F.R.D. 135 (D.Del.2009) ("Decision"). Rambus also appeals the district court's order piercing Rambus's attorney-client privilege on the basis of the crime-fraud exception, Micron Tech, Inc. v. Rambus Inc., No. 00-792 (D.Del. Feb. 10, 2006) ("Privilege"), and denial of Rambus's motion to transfer to the Northern District of California, Micron Tech, Inc. v. Rambus Inc., No. 00-792 (D.Del. June 14, 2007) ("Transfer"). For the reasons discussed below, this court affirms-in-part, vacates-in-part, and remands.

I. BACKGROUND

This case and the companion case of Hynix Semiconductor Inc. v. Rambus Inc., 645 F.3d 1336 (Fed.Cir.2011) ("Hynix II") (decided contemporaneously herewith), concern a group of U.S. patents issued to Rambus covering various aspects of dynamic random access memory ("DRAM"). Although semiconductor memory chips have been used in computers for decades, advances in other aspects of computer technology by the early 1990s created a bottleneck in the ability of computers to process growing amounts of data through the memory. At least two related methods were discovered of building memory chips (and the interfaces between memory chips and computer processors) in a way that eliminated or minimized this bottleneck. The founders of Rambus, Mike Farmwald and Mark Horowitz, developed one of these methods, which Rambus later commercialized as Rambus DRAM, or RDRAM. The original Rambus applications claim the inventions included in RDRAM. Rambus believed that Farmwald's and Horowitz's invention was broad enough to encompass synchronous dynamic random access memory, or SDRAM, the other type of new memory technology.

Farmwald and Horowitz did not initially file patent applications with claims explicitly directed at SDRAM. However, after Rambus's tenure and resignation as a member of the standard setting Joint Electron Devices Engineering Council ("JEDEC"), Rambus amended its claims to cover the SDRAM technology adopted as the standard by JEDEC. See generally Rambus Inc. v. Infineon Techs. AG, 318 F.3d 1081 (Fed.Cir.2003) ("Infineon") (discussing Rambus's participation in JEDEC). The patents at issue here and their enforceability against SDRAM products have been the subject of numerous suits in district courts, the Federal Trade Commission, the International Trade Commission, and this court. However, this court has never finally and definitively resolved the question of whether Rambus engaged in spoliation in connection with this litigation.

The present appeal began when Micron filed a declaratory judgment action against Rambus, asserting that Micron's production of SDRAM products do not infringe Rambus's patents and that Rambus's patents are invalid, unenforceable, and violate antitrust laws. The district court separated the case into three proceedings: (1) unenforceability due to spoliation, (2) invalidity, and (3) infringement. The court held a bench trial on the spoliation issue, and concluded that the patents in suit were unenforceable against Micron because Rambus had engaged in spoliation by intentionally destroying relevant, discoverable documents in derogation of a duty to preserve them. The district court thus did not reach the validity or infringement issues. On appeal, Rambus argues that the trial court clearly erred in determining that Rambus spoliated documents, acted in bad faith, and prejudiced Micron. Rambus also argues that the district court abused its discretion by dismissing the case as a sanction for the spoliation. Rambus also puts forth two procedural arguments: (1) that the district court erred by requiring the production of documents allegedly subject to attorney-client privilege; and (2) that the district court erred by denying Rambus's motion to transfer the litigation to the Northern District of California.

The record is lengthy but uncomplicated. In 1990, Farmwald and Horowitz filed their first patent application 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 Farmwald and Horowitz invention encompassed SDRAM. Rambus continued prosecuting multiple patent applications in the Farmwald/Horowitz family, intending to obtain issued patent claims that covered SDRAM. Rambus 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.

The first prong of Rambus's strategy went smoothly for some time. In 1996, Intel licensed the RDRAM technology and adopted it as the memory interface technology for its next generation microprocessors. Rambus negotiated licenses with eleven DRAM manufacturers to produce RDRAM-compliant chips for Intel's use. By the fall of 1999, though, these manufacturers had failed to deliver the promised manufacturing capacity, and Intel was therefore beginning to rethink its adoption of RDRAM. Rambus contends that only after RDRAM failed to become a market leader in late 1999 did it to put into action the second prong of its business strategy, to seek licensing revenue (and litigation damages) from those manufacturers adopting SDRAM.

Micron disagrees, arguing that Rambus was planning litigation against SDRAM manufacturers at the same time it was seeking to license RDRAM manufacturers.

In 1997, Rambus hired Joel Karp as its vice-president in charge of intellectual property, and on January 7, 1998, Karp was directed by Rambus's CEO Tate to develop a strategy for licensing and litigation. Karp then met with several transactional attorneys at Cooley Godward. Because they were not litigation specialists, they referred Karp to Dan Johnson, a litigation partner at Cooley Godward. Karp met Johnson on February 12, 1998. At the meeting with Johnson, Karp discussed licensing accused infringers, mentioning royalty rates that were so high that Johnson said "you're not going to have a licensing program, you're going to have a lawsuit on your hands." Karp said Rambus needed to get "battle-ready," by which he meant that Rambus needed to be ready for litigation. Johnson also advised putting into place a document-retention policy. In March 1998, Karp presented his proposal for a licensing and litigation strategy to the Rambus board of directors. He proposed a 5% royalty on SDRAM, a rate within the range that had prompted Johnson to say that litigation would inevitably follow. In the course of presenting the litigation strategy, Karp recommended implementing a document-retention policy.

In August or September 1998, Rambus hired outside counsel to perform licensing and patent prosecution work as well as to begin preparing for litigation against SDRAM manufacturers. In October 1998, Karp advised Rambus executives that he was planning to assert Rambus's patents against SDRAM manufacturers in the first quarter of 2000, explaining that there were good business reasons for the delay in bringing suit, particularly Rambus's interest in getting licensing revenues from RDRAM manufacturers, who would be the same parties it would seek to license for the production of SDRAM. In November 1998, Rambus executives held an offsite strategy meeting. The meeting notes show that Rambus planned to eventually assert its patents against SDRAM, even if the RDRAM adoption strategy succeeded. In approximately December 1998, Karp drafted a memo describing a possible "nuclear winter" scenario under which Intel moved away from RDRAM. The memo outlined plans for suing Intel and SDRAM manufacturers, saying that "by the time we do this, the proper litigants will be obvious." The memo also noted that infringement claim charts for Micron devices had already been completed by December 1998. On April 15, 1999, Karp met with Rambus's outside counsel at Fenwick & West to "discuss [Rambus's] patent portfolio and potential litigation."

Thereafter, in 1998, Rambus also began implementing the portion of Karp's litigation strategy that required the institution of a document-retention policy. In the second quarter of 1998, Rambus established "Top Level Goals" for "IP Litigation Activity." These goals included "[p]ropos[ing] [a] policy for document retention." In the third quarter of 1998, Rambus established "Key Goals" for "IP Litigation Activity." These goals included "[i]mplement[ing] [a] document retention action plan." On July 22, 1998, Karp presented the finished document retention policy to Rambus employees. The slides used for this presentation were titled "BEFORE LITIGATION: A Document Retention/Destruction Policy." The policy explicitly stated that destruction of relevant and discoverable evidence did not need to stop until the commencement of litigation. Despite the policy's stated goal of destroying all documents once they were old enough, Karp instructed employees to look for helpful documents to keep, including documents that would "help establish conception and prove that [Rambus had] IP."

The document destruction policy extended to the destruction of backups of Rambus's internal email. On March 16, 1998, an internal Rambus email discussed the "growing worry" that email backup tapes were "discoverable information," and discussions began regarding how long to keep these backup tapes. On May 14, 1998, Rambus implemented a new policy of keeping email backup tapes for only 3 months. Karp said that keeping tapes for any longer period of time was shot down by "Rambus'[s] litigation counsel." Consistent with this policy, in July 1998, Rambus magnetically erased all but 1 of the 1,269 tapes storing its email backups from the previous several years. The one exempted was a document that helped Rambus establish a priority date, and, as discussed below, Rambus went through great lengths to restore that document from the backup tapes.

In addition to destroying the email backup tapes, Rambus began destroying paper documents in accordance with its newly-adopted document-retention policy. On September 3-4, 1998, Rambus held its first "shred day" to implement the policy. In April 1999, Karp instructed Lester Vincent, Rambus's outside patent prosecution counsel at Blakeley Sokoloff, to implement the Rambus document-retention policy with respect to Rambus documents in Vincent's possession. Vincent complied, discarding material from his patent prosecution files. Vincent continued discarding material through at least July 1999. He discarded draft patent applications, draft patent claims, draft patent amendments, attorney notes, and correspondence with Rambus.

In June 1999, the first patent in suit issued. On June 24, 1999, Karp was instructed by the Rambus CEO to "hammer out ... our strategy for the battle with the first target that we will launch in October [1999]." In June 27, 1999, Rambus established its "IP 3Q '99 Goals," including goals for "Licensing/Litigation Readiness." These goals included "[p]repar[ing] litigation strategy against 1 of the 3 manufacturers," being "[r]eady for litigation with 30 days notice," and "[o]rganiz[ing] [the] 1999 shredding party at Rambus." Planning for litigation continued when, on July 8, 1999, Fenwick & West prepared a timeline for the proposed patent infringement suits showing that Rambus planned to file a patent infringement complaint on October 1, 1999.

On August 26, 1999, Rambus held the "shredding party" it had planned as part of its third-quarter intellectual property litigation readiness goals. Rambus destroyed between 9,000 and 18,000 pounds of documents in 300 boxes.

Litigation did not ultimately start as planned on October 1, 1999. Still, conditions eventually deteriorated to the point that Rambus felt it could no longer delay the litigation it had started planning in early 1998. As noted above, in the fall of 1999, several RDRAM manufacturers failed to deliver on their promised production of RDRAM chips, causing Intel to rethink its commitment to RDRAM. On September 24, 1999, Karp spoke to Rambus executives, telling them that the industry did not respect Rambus's intellectual property and that Rambus would "have to ultimately pursue remedies in court." Karp asked the board to approve his licensing and litigation strategy, and the board did so. In October 1999, Rambus approached Hitachi, seeking license payments for Hitachi's manufacture of SDRAM. In November 1999, negotiations with Hitachi broke down. Rambus instituted a litigation hold in December 1999, and Rambus sued Hitachi on January 18, 2000. The suit against Hitachi was settled on June 22, 2000. In the meantime, Rambus negotiated SDRAM licenses with Toshiba, Oki, and NEC. Rambus continued to litigate against the members of the chip-making industry by bringing suit against Infineon on August 8, 2000. Rambus, Inc. v. Infineon, 155 F.Supp.2d 668, 671 (E.D.Va.2001). Before that litigation began, Rambus's in-house counsel reminded Rambus executives on July 17, 2000, to continue destroying drafts and other materials related to license negotiations.

On August 18, 2000, Rambus approached Micron about the possibility of Micron taking a license for its SDRAM production. Micron filed a declaratory judgment action against Rambus in the District of Delaware on August 28, 2000, asserting invalidity, non-infringement, and unenforceability. The following day, Hynix Semiconductor filed a similar declaratory judgment suit against Rambus in the Northern District of California. Hynix Semiconductor, Inc. v. Rambus, Inc., 591 F.Supp.2d 1038 (N.D.Cal.2006) ("Hynix I"). The issue of whether Rambus had destroyed relevant documents after it had a duty to begin preserving documents was litigated in both suits. The Northern District of California reached the issue first. Following a bench trial, that court ruled in January 2006 that "Rambus did not actively contemplate litigation or believe litigation against any particular DRAM manufacturer to be necessary or wise before its negotiation with Hitachi failed, namely in [November] 1999." Id. at 1064. The Northern District of California ruled that this made Rambus's adoption of its document-retention policy in mid-1998 a permissible business decision, and the destruction of documents pursuant to that policy did not constitute spoliation. Id. The appeal of that decision is the subject of the companion Hynix case decided herewith. Hynix II, 645 F.3d 1336.

Meanwhile, in the Micron litigation in the District of Delaware, Micron sought access to communications between Rambus and its attorneys relating to the adoption of Rambus's document-retention policy. Courts in Hynix and Infineon had previously required production of these documents, and in February 2006, the District of Delaware agreed after finding that the adoption of the policy on the advice of counsel raised the likelihood that Delaware and California criminal statutes prohibiting destruction of evidence had been violated. The court held that the attorney-client privilege could be breached under the crime-fraud exception because Rambus and its counsel had possibly committed a crime. Following this decision and the favorable ruling of the Northern District of California on the spoliation issue, Rambus sought on February 14, 2006, to have the Micron case transferred to the Northern District of California. The District of Delaware denied the motion to transfer.

In November 2007, the District of Delaware held a bench trial on the unclean-hands issue asserted by Micron. Stopping short of reaching the unclean-hands claim, the district court found that Rambus had engaged in spoliation; the court accordingly entered judgment in Micron's favor as a spoliation sanction. The court found that litigation was reasonably foreseeable to Rambus "no later than December 1998, when Karp had articulated a time frame and a motive for implementation of the Rambus litigation strategy." The district court thus ruled that documents destroyed after December 1998 were intentionally destroyed in bad faith. The district court concluded that the only reasonable sanction for the intentional destruction of documents was to hold Rambus's patents in suit unenforceable against Micron. Rambus timely appealed.

II. DISCUSSION

A. Spoliation

As the Supreme Court has noted, "[d]ocument retention policies, which are created in part to keep certain information from getting into the hands of others, including the Government, are common in business. It is, of course, not wrongful for a manager to instruct his employees to comply with a valid document retention policy under ordinary circumstances." Arthur Andersen LLP v. United States, 544 U.S. 696, 704, 125 S.Ct. 2129, 161 L.Ed.2d 1008 (2005) (internal citation and quotation marks omitted). Thus, "a party can only be sanctioned for destroying evidence if it had a duty to preserve it." Zubulake v. UBS Warburg LLC, 220 F.R.D. 212, 216 (S.D.N.Y.2003). The duty to preserve evidence begins when litigation is "pending or reasonably foreseeable." Silvestri v. General Motors Corp., 271 F.3d 583, 590 (4th Cir.2001). See also West v. Goodyear Tire & Rubber Co., 167 F.3d 776, 779 (2d Cir.1999) (applying the same standard). Thus, "[s]poliation refers to the destruction or material alteration of evidence or to the failure to preserve property for another's use as evidence in pending or reasonably foreseeable litigation." Silvestri, 271 F.3d at 590. This is an objective standard, asking not whether the party in fact reasonably foresaw litigation, but whether a reasonable party in the same factual circumstances would have reasonably foreseen litigation.

When litigation is "reasonably foreseeable" is a flexible fact-specific standard that allows a district court to exercise the discretion necessary to confront the myriad factual situations inherent in the spoliation inquiry. Fujitsu Ltd. v. Fed. Express Corp., 247 F.3d 423, 436 (2d Cir. 2001). This standard does not trigger the duty to preserve documents from the mere existence of a potential claim or the distant possibility of litigation. See, e.g., Trask-Morton v. Motel 6 Operating L.P., 534 F.3d 672, 681-82 (7th Cir.2008). However, it is not so inflexible as to require that litigation be "imminent, or probable without significant contingencies," as Rambus suggests. Reply Br. of Rambus at 4. Rambus's proposed gloss on the "reasonably foreseeable" standard comes from an overly generous reading of several cases. See Burlington N. & Santa Fe Ry. Co. v. Grant, 505 F.3d 1013, 1032 (10th Cir.2007) (noting that "[a] spoliation sanction is proper where (1) a party has a duty to preserve evidence because it knew, or should have known, that litigation was imminent, and (2) the adverse party was prejudiced by the destruction of the evidence" (emphasis added); Trask-Morton, 534 F.3d at 681 (citing Burlington for the proposition that "courts have found a spoliation sanction to be proper only where a party has a duty to preserve evidence because it knew, or should known, that litigation was imminent," but holding that "Motel 6 had no reason to suspect litigation until—at the earliest—Morton's attorney sent Motel 6 a demand letter" after the alleged spoliation (emphases added)). Burlington merely noted that imminent litigation was sufficient, not that it was necessary for spoliation, and on the easy facts of Trask-Morton, it was decided that the alleged spoliator did not even "suspect" litigation. This court declines to sully the flexible reasonably foreseeable standard with the restrictive gloss proposed by Rambus in light of the weight of contrary authority and the unnecessary generosity that such a gloss would extend to alleged spoliators. See Silvestri, 271 F.3d at 591; West, 167 F.3d at 779 ("Spoliation is the destruction or significant alteration of evidence, or the failure to preserve property for another's use as evidence in pending or reasonably foreseeable litigation."); Kronisch v. United States, 150 F.3d 112, 126 (2d Cir.1998) ("This obligation to preserve evidence arises when the party has notice that the evidence is relevant to litigation... as for example when a party should have known that the evidence may be relevant to future litigation."); MOSAID Techs. Inc. v. Samsung Elecs. Co., 348 F.Supp.2d 332, 336 (D.N.J.2004) (noting that a litigant "is under a duty to preserve what it knows, or reasonably should know, will likely be requested in reasonably foreseeable litigation"); Scott v. IBM Corp., 196 F.R.D. 233, 249 (D.N.J.2000) (same). See also United States v. Rockwell Int'l, 897 F.2d 1255, 1266 (3rd Cir.1990) (holding that for attorney work product to be shielded by the work product privilege, "[l]itigation need not be imminent ... as long as the primary motivating purpose behind the creation of the document was to aid in possible future litigation.") (internal citations omitted)). Moreover, it would make little sense to enjoin document destruction only when the party clears all the hurdles on the litigation track, but endorse it when the party begins the race under the reasonable expectation of clearing those same hurdles. Thus, the proper standard for determining when the duty to preserve documents attaches is the flexible one of reasonably foreseeable litigation, without any additional gloss.

After carefully reviewing the record, the district court determined that "litigation was reasonably foreseeable no later than December 1998, when Karp had articulated a time frame and a motive for implementation of the Rambus litigation strategy." Decision, 255 F.R.D. at 150. In coming to this conclusion, the district court applied the correct standard, noting that "[a] duty to preserve evidence arises when.... litigation is pending or imminent, or when there is a reasonable belief that litigation is foreseeable." Id. at 148.

This court reviews the district court's factual findings, such as the date at which litigation was reasonably foreseeable, for clear error. Citizens Fed. Bank v. United States, 474 F.3d 1314, 1321 (Fed. Cir.2007); Port Auth. of N.Y. & N.J. v. Arcadian Corp., 189 F.3d 305, 315 (3d Cir.1999); Brewer v. Quaker State Oil Ref. Corp., 72 F.3d 326, 334 (3d Cir.1995). Rambus argues that when litigation was reasonably foreseeable is a "mixed question of law and fact reviewed de novo." However, the cases it cites do not support such a standard in this context. For example, Travelers Indemnity v. Ewing, Cole, Erdman & Eubank, 711 F.2d 14 (3d Cir.1983), addressed the "issue of whether the level of care exercised by the defendant measured up to the standard expected of reasonably prudent architects" as a mixed question of law and fact. Id. at 17. However, that question is more about whether a duty is breached than when the duty commenced. Similarly inapposite, Pell v. E.I. DuPont de Nemours, 539 F.3d 292 (3d Cir.2008), concluded that "the District Court's determination that 1972 is the appropriate adjusted service date is a mixed conclusion of law and fact," and that this question is broken down into its "components and [the appeals court applies] the appropriate standard of review to each component." Id. at 305. Pell does not specify whether the date at which the duty arises is a law component or a fact component, and thus does not persuade this court to review the issue de novo. In a variety of contexts, foreseeability of an event is a traditional issue of fact, and is reviewed with deference to the district court. Cates v. Dillard Dep't Stores, Inc., 624 F.3d 695, 697 (5th Cir.2010) (noting that whether a risk of harm was reasonably foreseeable is a question of fact); Citizens Fed. Bank v. United States, 474 F.3d 1314, 1321 (Fed.Cir.2007) (noting that "[f]oreseeability is a question of fact reviewed for clear error" in the damages context); United States v. Cover, 199 F.3d 1270, 1274 (11th Cir.2000) (reasonable foreseeability of a co-conspirators actions is a question of fact). This court likewise applies a clear error standard of review.

The district court found that Rambus destroyed relevant, discoverable documents beginning in July 1998, with the first major shred day occurring in September 1998. The court found that the destruction continued at least through November 1999, with another major shred day occurring in August 1999. In addition, the district court found that Rambus ordered its outside patent prosecution counsel to purge his files relating to the prosecution of the prospective patents in suit in April 1999. There is ample evidence to support all these findings, and they are not seriously disputed even by Rambus. The exact date at which litigation was reasonably foreseeable is not critical to this decision; the real question is binary: was litigation reasonably foreseeable before the second shred day or after? Therefore, the question this court must answer is whether the district court clearly erred when it determined that, at some time before the second shred day in August of 1999, litigation was reasonably foreseeable. This court cannot conclude that the district court clearly erred for at least the following five reasons.

First, it is certainly true that most document retention policies are adopted with benign business purposes, reflecting the fact that "litigation is an ever-present possibility in American life." Nat'l Union Fire Ins. v. Murray Sheet Metal Co., 967 F.2d 980, 984 (4th Cir.1992). In addition, there is the innocent purpose of simply limiting the volume of a party's files and retaining only that which is of continuing value. One might call it the "good housekeeping" purpose. Thus, where a party has a long-standing policy of destruction of documents on a regular schedule, with that policy motivated by general business needs, which may include a general concern for the possibility of litigation, destruction that occurs in line with the policy is relatively unlikely to be seen as spoliation. Here, however, it was not clear error for the district court to conclude that the raison d'etre for Rambus's document retention policy was to further Rambus's litigation strategy by frustrating the fact-finding efforts of parties adverse to Rambus. This is a natural reading of getting "[b]attle ready." The preparation of the document retention policy was one of Rambus's "IP Litigation Activity" goals in the second and third quarters of 1998. When the finished document retention policy was presented to Rambus employees, the presentation slides used were titled "BEFORE LITIGATION: A Document Retention/Destruction Policy." The policy explicitly stated that destruction of relevant and discoverable evidence did not need to stop until the actual commencement of litigation. Despite the policy's stated goal of destroying all documents once they were old enough, employees were instructed to look for helpful documents to keep, including documents that would "help establish conception and prove that [Rambus had] IP," and they did keep these documents. Moreover, on March 16, 1998, an internal Rambus e-mail noted a "growing worry" that email backup tapes were "discoverable information," and discussions began regarding how long to keep these backup tapes. On May 14, 1998, Rambus implemented a new policy of keeping email backup tapes for only 3 months. Karp said that keeping tapes for any longer period of time was shot down by "Rambus'[s] litigation counsel." Karp also noted that if anyone had questions about the document retention policy, they could contact him, but that he "would prefer to discuss [the] issue face to face," and that if they did send e-mails, to "keep them brief, and keep the distribution narrow." Shortly after the email backup destruction policy was instituted, all of Rambus's old backup tapes were destroyed. Taken together, the implementation of a document retention policy as an important component of a litigation strategy makes it more likely that litigation was reasonably foreseeable Cf. United States v. Adlman, 134 F.3d 1194, 1203 (2d Cir.1998) (adopting a test for work product immunity where a document is prepared in anticipation of litigation where the document "can fairly be said to have been prepared or obtained because of the prospect of litigation") (emphasis added) (citing 8 Charles Alan Wright, Arthur R. Miller, and Richard L. Marcus, Federal Practice and Procedure § 2024, at 343 (1994)).

Second, Rambus was on notice of potentially infringing activities by particular manufacturers. Once the patent issued, the gun was loaded; when the targets were acquired, it was cocked; all that was left was to pull the trigger by filing a complaint. While it may not be enough to have a target in sight that the patentee believes may infringe, the knowledge of likely infringing activity by particular parties makes litigation more objectively likely to occur because the patentee is then more likely to bring suit. Here, numerous internal documents manifest Rambus's plan "to play [its] IP card with the DRAM companies" against SDRAM products, either through a patent infringement or a breach of contract suit. See Decision, 255 F.R.D. at 138-48 (noting that even in the early 1990s, Rambus was already "concerned that DRAM manufacturers were using Rambus'[s] technology to develop their own competing DRAMs," and detailing Rambus's campaign to capitalize on non-compliant products' infringement); id. at 144 ("The [Nuclear Winter Memorandum] indicated specifically that Rambus already had claim charts showing that Micron infringed one of the Rambus patents."). See also Br. of Rambus's at 34 ("Rambus therefore feared that demanding licenses on non-compatible products (let alone initiating litigation) would risk undermining its relationships with the very DRAM manufacturers its business strategy depended upon."). In addition, the bulk of the discussions between CEO Tate, Karp, and Rambus's attorneys related to SDRAM and Rambus's licensing (as Rambus argues) or litigation (as Micron argues) plans. Either way, Rambus was on notice of activities it believed were infringing. Cf. Schmid v. Milwaukee Elec. Tool Corp., 13 F.3d 76, 81 (3d Cir.1994) (overturning a district court spoliation sanction in part because the plaintiff's expert's destruction of evidence occurred when "no suit had been filed and Schmid did not know whether he had a basis for instituting suit."). Indeed, Rambus was more than on notice because, by its own admission, it actively broadened its claims to cover JEDEC standard-compliant products, and, according to the testimony of CEO Tate, it knew that those products would infringe its claims.

Third, Rambus took several steps in furtherance of litigation prior to its second shredding party on August 26, 1999. Karp had already concluded that Rambus would "need to litigate against someone to establish [a] royalty rate and have [the] court declare [the Rambus] patent[s] valid," had prioritized defendants and forums, had created claim charts and determined an expected timeline for litigation that it would "launch in October [1999]," and had as its goal to "be ready for litigation with 30 days notice" "against 1 of the 3 manufacturers" by the third quarter of 1999. On June 24, 1999, Karp was instructed by CEO Tate to "hammer out ... our strategy for the battle with the first target that we will launch in October [1999]." The first steps toward this litigation were spelled out on June 27, 1999, when Rambus established "IP 3Q '99 Goals," including goals for "Licensing/Litigation Readiness." These goals included "[p]repar[ing] litigation strategy against 1 of the 3 manufacturers," being "[r]eady for litigation with 30 days notice," and "[o]rganiz[ing] [the] 1999 shredding party at Rambus." Planning for litigation continued when, on July 8, 1999, Rambus's outside litigation counsel, Fenwick & West, prepared a timeline for the proposed patent infringement suits showing that Rambus planned to file complaints on October 1, 1999. Indeed, the second shredding party was itself part of Rambus's third-quarter intellectual property litigation readiness goals.

Rambus strongly argues that the steps it did not yet take in furtherance of litigation, i.e. the contingencies, compel a finding that litigation was not reasonably foreseeable. Rambus cites the contingencies accepted by Judge Whyte in the companion Hynix case as precluding Rambus from reasonably foreseeing litigation:

Hynix, 591 F.Supp.2d at 1062. It is of course true that had these contingencies been cleared, litigation would have been more foreseeable. However, it was not clear error to conclude that overcoming the contingencies was reasonably foreseeable. For example, Rambus makes much of the inadvisability of jeopardizing its relationship with the manufacturers through litigation over SDRAM, because those same manufacturers were producing RDRAM, which Rambus hoped would become the market leader. However, as was made clear in the Nuclear Winter Memorandum, if RDRAM did not become a market leader, Rambus would go after the manufacturers of SDRAM and if RDRAM did become a market leader, and the RDRAM ramp "reache[d] a point of no return," then Rambus could come out from "stealth mode," and could then "ROCK THE DIRECT BOAT" because the manufacturers would be locked in to the RDRAM standard. Hence the use of definitive language of future intention, such as asking "WHAT'S THE RUSH [to assert patents against RDRAM partners]?" and noting that it should "not asserts patents against Direct [RDRAM] partners until ramp reaches a point of no return (TBD)." (emphasis added). Similarly, obtaining product samples would certainly be a reasonably foreseeable event, particularly because Rambus had explicitly broadened its claim coverage in prosecution to cover standard-compliant products, which, by the terms of the standard, all the manufacturers would meet. It was also reasonably foreseeable that the manufacturers would reject Rambus's licensing terms, because Karp proposed a five percent royalty rate to the board in March 1998 that attorney Johnson had called "ridiculous," and that the Cooley attorneys informed him would result in a lawsuit. In December 1998 or January 1999, Karp opined that in situations where Rambus was "not interested in settling," they should propose a royalty rate between five and ten percent, and noted that "we should not be too concerned with settlement at this point and should push for very high rates." It is thus not clear error to conclude that Rambus reasonably foresaw that the manufacturers would reject its licensing offer. The same is true for the other listed contingencies. Thus, Rambus's preparations for litigation prior to the critical date, including choosing and prioritizing manufacturers to sue, selecting forums in which to bring suit within a planned time-frame, creating claim charts, and including litigation as an essential component of its business model, support the district court's decision that Rambus reasonably foresaw litigation before the second shredding party on August 26, 1999.

Fourth, when Rambus sued Hitachi on January 18, 2000, it was the plaintiff-patentee, and its decision whether to litigate or not was the determining factor in whether or not litigation would in fact ensue. In other words, whether litigation was reasonably foreseeable was largely dependent on whether Rambus chose to litigate. It is thus more reasonable for a party in Rambus's position as a patentee to foresee litigation that does in fact commence, than it is for a party in the manufacturers' position as the accused.1

Fifth, as discussed above, the relationship between Rambus and the manufacturers involving RDRAM did not make litigation significantly less likely, it only delayed the initiation of litigation until the manufacturers were either too invested in RDRAM for the SDRAM litigation to negatively impact Rambus's sales, or until Rambus had no choice but to sue because RDRAM was rejected. In general, when parties have a business relationship that is mutually beneficial and that ultimately turns sour, sparking litigation, the litigation will generally be less foreseeable than would litigation resulting from a relationship that is not mutually beneficial or is naturally adversarial. Thus, for example, document destruction occurring during the course of a long-standing and untroubled licensing relationship relating to the patents and the accused products that ultimately become the subject of litigation is relatively unlikely to constitute spoliation, while destruction of evidence following repeated failures of a licensee to properly mark products or remit royalties, is more likely to constitute spoliation. Because the relationship regarding RDRAM did nothing to make litigation significantly less likely, and because Rambus and the manufacturers did not have a longstanding and mutually beneficial relationship regarding SDRAM, Rambus cannot use its delay tactics regarding RDRAM to undermine the other considerations herein discussed.

Rambus argues that the district court clearly erred in setting December 1998 as the date at which litigation was reasonably foreseeable, because the only happening on that date was the issuance of the Nuclear Winter Memorandum, which addressed Rambus's potential response to the "very unlikely" scenario that Intel would drop its support for RDRAM. Rambus argues that a document addressing such a contingency cannot form the basis for reasonably foreseeable litigation. The district court found that litigation was reasonably foreseeable "no later than December 1998, when Karp had articulated a time frame and a motive for implementation of the Rambus litigation strategy." Decision, 255 F.R.D. at 150. The important inquiry is not whether a particular document made litigation reasonably foreseeable, but whether the totality of the circumstances as of the date of document destruction made litigation reasonably foreseeable. As discussed above, there was no clear error in the district court's holding that they did.

This court thus affirms the district court's determination that Rambus destroyed documents during its second shred day in contravention of a duty to preserve them and, thus, engaged in spoliation.

B. The District Court's Choice of Sanction

District courts have the "inherent power to control litigation," West, 167 F.3d at 779, by imposing sanctions appropriate to rectify improper conduct by litigants. Schmid, 13 F.3d at 78. Such sanctions may include dismissal. Leon v. IDX Sys. Corp., 464 F.3d 951, 958 (9th Cir. 2006). The particular sanction imposed is within the sound discretion of the district court in exercising its inherent authority and in assuring the fairness of the proceedings before it. See Silvestri, 271 F.3d at 590 (quoting Chambers v. NASCO, Inc., 501 U.S. 32, 45-46, 111 S.Ct. 2123, 115 L.Ed.2d 27 (1991)) ("The right to impose sanctions for spoliation arises from a court's inherent power to control the judicial process and litigation, but the power is limited to that necessary to redress conduct `which abuses the judicial process.'"). As such, the district court's choice of sanction is reviewed for an abuse of discretion. Mindek v. Rigatti, 964 F.2d 1369, 1373-74 (3d Cir.1992).

Rambus challenges the district court's imposition of the dispositive sanction of dismissal, arguing that Micron failed to prove bad faith or prejudice, and that the district court was limited to applying some lesser sanction than dismissal. This court addresses Rambus's arguments in turn.

i. Bad Faith

To make a determination of bad faith, the district court must find that the spoliating party "intended to impair the ability of the potential defendant to defend itself." Schmid, 13 F.3d at 80. See also Faas v. Sears, Roebuck & Co., 532 F.3d 633, 644 (7th Cir.2008) ("A document is destroyed in bad faith if it is destroyed `for the purpose of hiding adverse information.'") (citation omitted); In re Hechinger Inv. Co. of Del., Inc., 489 F.3d 568, 579 (3d Cir.2007) (noting that bad faith requires a showing that the litigant "intentionally destroyed documents that it knew would be important or useful to [its opponent] in defending against [the] action"); Anderson v. Cryovac, Inc., 862 F.2d 910, 925 (1st Cir.1988) (finding bad faith "where concealment was knowing and purposeful," or where a party "intentionally shred[s] documents in order to stymie the opposition"); Gumbs v. Int'l Harvester, Inc., 718 F.2d 88, 96 (3d Cir.1983) (noting that an adverse inference from destruction of documents is permitted only when the destruction was "intentional, and indicates fraud and a desire to suppress the truth") (citation omitted). The fundamental element of bad faith spoliation is advantage-seeking behavior by the party with superior access to information necessary for the proper administration of justice.

Here, the district court's analysis of bad faith follows its conclusion on spoliation and does not fully explain the factual underpinnings of its bad faith determination:

Decision, 255 F.R.D. at 150. A determination of bad faith is normally a prerequisite to the imposition of dispositive sanctions for spoliation under the district court's inherent power, and must be made with caution. In determining that a spoliator acted in bad faith, a district court must do more than state the conclusion of spoliation and note that the document destruction was intentional. See Mathis v. John Morden Buick, Inc., 136 F.3d 1153, 1155 (7th Cir.1998) ("That the documents were destroyed intentionally no one can doubt, but `bad faith' means destruction for the purpose of hiding adverse information.") (emphasis added). From the district court's sparse analysis, this court is unable to determine whether the district court applied the applicable exacting standard in making its factual determination that Rambus acted in bad faith.

The district court's opinion alludes to several key items, including: (1) facts tending to show that Rambus's document retention policy was adopted within the auspices of a firm litigation plan rather than merely carried out despite the reasonable foreseeability of such litigation, e.g., Decision ¶¶ 17, 53, 55 and n. 29; (2) facts tending to show the selective execution of the document retention policy, e.g., Decision ¶ 13 and n. 23, 27; (3) facts tending to show Rambus's acknowledgement of the impropriety of the document retention policy, e.g., Decision ¶¶ 6, 38 and n. 24, 47; and (4) Rambus's litigation misconduct, Decision ¶¶ 37-39. While these items may lead to a determination of bad faith, the district court did not make clear the basis on which it reached that conclusion.

"It is not our task to make factual findings," Golden Hour Data Sys., Inc. v. emsCharts, Inc., 614 F.3d 1367, 1380 (Fed. Cir.2010), and we will leave it to the district court's sound discretion on remand to analyze these, and any other, relevant facts as they apply to the determination of bad faith, see Thomas v. Capital Sec. Servs., Inc., 836 F.2d 866, 873 (5th Cir. 1988) (en banc) ("[T]he district court will have a better grasp of what is acceptable trial-level practice among litigating members of the bar than will appellate judges.").

We note that the district court applied a "knew or should have known" standard in its bad faith determination. On remand, the district court should limit its bad faith analysis to the proper inquiry: whether Rambus "intended to impair the ability of the potential defendant to defend itself," Schmid, 13 F.3d at 80, without regard to whether Rambus "should have known" of the propriety of its document destruction.

Litigations are fought and won with information. If the district court finds facts to conclude that Rambus's goal in implementing its document retention policy was to obtain an advantage in litigation through the control of information and evidence, it would be justified in making a finding of bad faith. If, on the other hand, the district court determines that Rambus implemented its document retention policy for legitimate business reasons such as general house-keeping, a finding of bad faith would be unwarranted. Without a finding either way, however, "the opinion explaining the decision lacks adequate fact findings, [and] meaningful review is not possible." Dennison Mfg. Co. v. Panduit Corp., 475 U.S. 809, 811, 106 S.Ct. 1578, 89 L.Ed.2d 817 (1986). This court therefore remands for the district court to further assess the factual record in reaching a determination on bad faith.

ii. Prejudice

Prejudice to the opposing party requires a showing that the spoliation "materially affect[s] the substantial rights of the adverse party and is prejudicial to the presentation of his case." Wilson v. Volkswagen of Am., Inc., 561 F.2d 494, 504 (4th Cir.1977) (internal quotation marks omitted). In satisfying that burden, a party must only "come forward with plausible, concrete suggestions as to what [the destroyed] evidence might have been." Schmid, 13 F.3d at 80 (emphases added). See also Leon, 464 F.3d at 960 ("[B]ecause any number of the 2,200 files could have been relevant to IDX's claims or defenses, although it is impossible to identify which files and how they might have been used.... the district court did not clearly err in its finding of prejudice."). If it is shown that the spoliator acted in bad faith, the spoliator bears the "heavy burden" to show a lack of prejudice to the opposing party because "[a] party who is guilty of... intentionally shredding documents ... should not easily be able to excuse the misconduct by claiming that the vanished documents were of minimal import." Anderson v. Cryovac, Inc., 862 F.2d 910, 925 (1st Cir.1988). See also Coates v. Johnson & Johnson, 756 F.2d 524, 551 (7th Cir.1985) ("The prevailing rule is that bad faith destruction of a document relevant to proof of an issue at trial gives rise to a strong inference that production of the document would have been unfavorable to the party responsible for its destruction.").

It is undisputed that Rambus destroyed between 9,000 and 18,000 pounds of documents in 300 boxes. The district court concluded that the destroyed documents were relevant to at least the following defenses, which would have been "illuminated by evidence of a non-public nature, e.g. by internal Rambus documents": "unenforceability due to patent misuse and violation of the antitrust and unfair competition laws (based in part on Rambus's conduct at JEDEC), as well as inequitable conduct." Decision, 255 F.R.D. at 150-51. Documents relating to Rambus's conduct at JEDEC, together with documents reflecting Rambus's instructions to its patent prosecution counsel concerning its conduct at JEDEC, could have helped resolve Micron's claims relating to patent misuse, antitrust violations, and unfair competition. Documents reflecting Rambus's knowledge of relevant prior art references could have helped resolve Micron's inequitable conduct claims. On the other hand, because it is not clear what documents were destroyed, it may be, as Rambus argues, that all the documents destroyed were either redundant or irrelevant to the trial.

The proper resolution of this issue turns largely on whether Rambus has the burden to show lack of prejudice or Micron has the burden to show prejudice. As discussed above, this turns on whether the district court, on remand, concludes that Rambus was a bad faith spoliator. The question of prejudice is therefore also remanded.

iii. Dispositive Sanction

In addition to reassessing on remand its determination of bad faith and prejudice, the district court should also explain the reasons for the propriety of the sanction chosen (if any) based on the degree of bad faith and prejudice and the efficacy of other lesser sanctions.

Dismissal is a "harsh sanction," to be imposed only in particularly egregious situations where "a party has engaged deliberately in deceptive practices that undermine the integrity of judicial proceedings." Leon, 464 F.3d at 958 (internal citations omitted). This court agrees that such sanctions should not be imposed unless there is clear and convincing evidence of both bad-faith spoliation and prejudice to the opposing party. Shepherd v. ABC, 62 F.3d 1469, 1472, 1477 (D.C.Cir.1995) (noting that dismissal requires proof by clear and convincing evidence); Gates Rubber Co. v. Bando Chem. Indus., Ltd., 167 F.R.D. 90, 108 (D.Colo. 1996) (requiring clear and convincing evidence because "[t]o do otherwise would be to contravene the strong public policy which favors adjudication of cases on their merits"). Moreover, the presence of bad faith and prejudice, without more, do not justify the imposition of dispositive sanctions. In gauging the propriety of the sanction, the district court must take into account "(1) the degree of fault of the party who altered or destroyed the evidence; (2) the degree of prejudice suffered by the opposing party; and (3) whether there is a lesser sanction that will avoid substantial unfairness to the opposing party and, where the offending party is seriously at fault, will serve to deter such conduct by others in the future." Schmid, 13 F.3d at 79 (emphases added). See also Leon, 464 F.3d at 958 (noting that the district court must consider "(1) the public's interest in expeditious resolution of litigation; (2) the court's need to manage its dockets; (3) the risk of prejudice to the party seeking sanctions; (4) the public policy favoring disposition of cases on their merits; and (5) the availability of less drastic sanction"). The sanction ultimately imposed must be commensurate with the analysis of these factors.

The district court must "select the least onerous sanction corresponding to the willfulness of the destructive act and the prejudice suffered by the victim." Schmid, 13 F.3d at 79 (citing Jamie S. Gorelick, Steven Marzen and Lawrence Solum, Destruction of Evidence, § 3.16, p. 117 (1989)). While the district court noted that "[s]anctions such as adverse jury instructions and preclusion of evidence are impractical, bordering on meaningless, under these circumstances and in the context of a typical jury trial," and that "the simple imposition of fees and costs is wholly inadequate under the facts of this case," Decision, 255 F.R.D. at 151, it did not explain why only dismissal would "vindicate the trifold aims of: (1) deterring future spoliation of evidence; (2) protecting the defendants' interests; and (3) remedying the prejudice defendants suffered as a result of [Rambus's] actions." See West, 167 F.3d at 780.

If the district court again concludes on remand that there was bad faith and prejudice, the record evidence may indeed justify a dispositive sanction, but the seriousness of such a sanction warrants an analysis of all of the factors discussed above. Cf. Roadway Express v. Piper, 447 U.S. 752, 764, 100 S.Ct. 2455, 65 L.Ed.2d 488 (1980) (noting that because "inherent powers are shielded from direct democratic controls," they "must be exercised with restraint and discretion").

C. Piercing of the Attorney-Client Privilege

The district court's spoliation rulings depended in part on evidence from communications between Rambus and its attorneys; these communications were in the record only because they had been ordered produced by Rambus after the district court pierced the attorney-client privilege that otherwise would protect the communications from disclosure. Rambus appeals the privilege-piercing ruling, arguing that the district court erred by finding that Micron had made the required prima facie showing that Rambus had committed or intended to commit a fraud or crime and that the attorney-client communications in question were in furtherance of that crime or fraud.

Rambus is correct that the crime-fraud exception to the attorney-client privilege requires such a prima facie showing. In re Grand Jury Subpoena, 223 F.3d 213, 217 (3d Cir.2000). But making a prima facie showing is "not a particularly heavy" burden, In re Grand Jury Investigation, 445 F.3d 266, 274-75 (3d Cir.2006), and this court agrees with the district court that Micron did indeed carry that burden here. Specifically, there was enough evidence to find a likely violation of § 135 of the California Penal Code, which provides that

As discussed above, there is ample evidence that Rambus destroyed documents in its possession knowing that they would likely be forced to be produced in litigation and intending to prevent that production. There is also ample evidence that Rambus devised this strategy partly on the basis of the advice it received from its outside counsel. The only question therefore is whether the documents Rambus destroyed were "about to be produced in evidence," or whether the delay of some months between Rambus's destruction of the evidence and Rambus's final decision to file suit against Hitachi eliminates any possibility that this element of the California statute could be satisfied.

Rambus argues that the "about to be produced in evidence" element of § 135 has been interpreted to "connote[ ] an immediacy or temporal closeness" between the destruction of the evidence and the time when it was to be produced and that such "temporal closeness" is missing here. People v. Prysock, 127 Cal.App.3d 972, 1000-01, 180 Cal.Rptr. 15 (Cal.App. 5th Dist.1982). However, Prysock is distinguishable. There, the criminal defendant committed murder and, shortly thereafter, burned the clothes he wore while committing the crime. Id. at 981, 180 Cal.Rptr. 15. Because no "law enforcement investigation... had started" at the time the defendant destroyed his clothes, and because "law enforcement was [not] ... looking for the" clothes at the time they were destroyed, the California court found no violation of § 135. Id. at 1001, 180 Cal.Rptr. 15. In Prysock, then, the defendant controlled the timing of the destruction of relevant evidence, while law enforcement, acting independently, controlled the timing of the initiation of the investigation that would trigger the application of § 135. Here, by contrast, Rambus controlled the timing of both events. It would make no sense to allow Rambus to escape liability, criminal or otherwise, by intentionally destroying evidence and then waiting for an arbitrary period of time before choosing to file suit. Moreover, it is not clear that California courts would interpret the statute in such a way as to make the initiation of the "investigation" the time when Rambus filed its complaint. Instead, the statute could just as reasonably be interpreted such that it would be violated by the intentional destruction of evidence at any time after Rambus began preparing to file this suit. Interpreted this way, Rambus's preparations to file this suit in early 1998 began the civil equivalent of the criminal law-enforcement investigation in Prysock, and Rambus's destruction of documents in late 1998 and 1999 violated § 135 if, as noted above, Rambus intentionally destroyed those documents with the intention of keeping them from being produced in the ensuing litigation. This is clearly a different situation from that encountered in Prysock, and this court sees no error in the district court's distinguishing that case. Because the district court properly found that Micron made a prima facie showing that (1) Rambus willfully destroyed documents it knew would have to be produced in the litigation it intended to initiate against Hitachi, (2) Rambus destroyed those documents in order to keep them from being produced, and (3) Rambus began destroying those documents based on communications from its litigation counsel advising it to begin destroying discoverable information, this court agrees with the district court's use of the crime-fraud exception to pierce the attorney-client privilege.

D. Denial of Transfer

The final issue on appeal concerns the district court's denial of Rambus's motion to transfer this case to the Northern District of California. As discussed above, Rambus filed its first suit against SDRAM manufacturers in January 2000. Micron filed the present declaratory judgment action against Rambus in August 2000. The following day, Hynix filed suit against Rambus in the Northern District of California. The present case was stayed at Rambus's request on June 28, 2002, with the stay apparently lifted in late 2004. Meanwhile, proceedings continued in the Hynix matter in the Northern District of California. In that case, in January 2006, the U.S. District Court for the Northern District of California decided that there had been no spoliation by Rambus. Following that favorable ruling, Rambus moved to transfer the present case to the Northern District of California on February 14, 2006. The Delaware district court denied the motion to transfer on March 29, 2006. At the time of Rambus's motion to transfer, the present case had been pending for five and a half years, and there was a trial schedule already in place.

In an oral ruling from the bench, the district court referred to Rambus's motion to transfer "as clear and obvious a case of forum shopping as has probably ever existed in the federal judiciary," thanks to the motion's coming hard on the heels of the favorable spoliation ruling in the Northern District of California. The district court weighed several factors, including the forum-shopping allegation (weighing against transfer), the plaintiff's choice of forum (weighing against transfer), the convenience of the witnesses (not weighing in favor of either transferring or not transferring), and the fact that trial was scheduled and imminent (weighing against transfer). Based on these factors, the district court denied the transfer motion.

This court reviews this issue under the law of the relevant regional circuit. The Third Circuit reviews denials of motions to transfer for abuse of discretion. Jumara v. State Farm Ins., 55 F.3d 873, 878 (3d Cir.1995). Rambus argues that the district court abused its discretion here because there were other cases regarding the same patents pending in the Northern District of California, and avoiding having the same issues litigated in multiple forums is the "paramount consideration" when deciding whether to transfer a case. But, as Micron points out, treating consolidation of related cases as "paramount" generally involves transferring all the cases to the same forum where the earliest-filed action is pending. Here, the earliest-filed case is the present one, pending in Delaware; the cases in the Northern District of California were all filed later. Thus, the "paramount consideration" might help in an argument to transfer the Hynix case to the District of Delaware, but it does nothing to support Rambus's argument in favor of transferring this case to the Northern District of California.

Moreover, every other factor either is neutral or supports the district court's decision not to transfer the case: (1) Rambus waited over five years to ask for this case to be transferred, (2) the motion to transfer came just months before a scheduled trial, and (3) Rambus filed the transfer motion only a month after receiving a favorable ruling by the proposed transferee court, strongly suggesting forum shopping. This forum-shopping should be discouraged, arguing strongly in favor of denying the transfer motion. The district court also pointed out that the convenience of the witnesses did not favor either forum, because most of the witnesses were employees of or consultants to the parties and could therefore be encouraged to testify in either forum, even if they could not be compelled to testify in Delaware. This was correct, not least because Rambus had earlier tried to move a related case out of the Northern District of California, arguing that it was inconvenient for its witnesses to testify there. Given that both parties were incorporated in Delaware, they had both willingly submitted to suit there, which weighs in favor of keeping the litigation in Delaware. Finally, Rambus had previously filed other litigation regarding the same patents (against Hitachi) in the District of Delaware, suggesting that it had no difficulty litigating a patent infringement suit in that court. Given that all the relevant factors either favor denying the transfer motion or are neutral, this court holds that the district court did not abuse its discretion by refusing to transfer this case to the Northern District of California, and therefore affirms the district court's denial of Rambus's motion to transfer.

III. CONCLUSION

For the reasons stated above, this court affirms the district court's determination that Rambus spoliated documents, but vacates the district court's dismissal sanction, and remands for further consideration consistent with this opinion.

AFFIRMED-IN-PART, VACATED-IN-PART, AND REMANDED.

COSTS

Each party shall bear its own costs.

GAJARSA, Circuit Judge, concurring-in-part and dissenting-in-part.

While I agree with the majority that there was spoliation of evidence by Rambus, I dissent from that part of the majority's opinion that remands for a reexamination of the evidence for bad faith and vacates the district court's sanction award. Even though the majority applauds with one hand the district court's "inherent power to control litigation," West v. Goodyear Tire & Rubber Co., 167 F.3d 776, 779 (2d Cir.1999), with the other hand it strangles this power by vacating the district court's sanction award. Indeed, the majority does not review the district court's sanction award for an abuse of discretion, instead it reviews the facts and weighs 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 dispositive sanctions. Because we should not "disarm the [district] court of its important power to police its proceedings to ensure transparency and predictability and to discourage mischievous conduct by litigants," I dissent.1 Transclean Corp. v. Bridgewood Servs., Inc., 290 F.3d 1364, 1374 (Fed.Cir.2002).

The district court found that Rambus' conduct "impugned" the very integrity of the judicial system. Micron Tech., Inc. v. Rambus Inc., 255 F.R.D. 135, 151 (D.Del. 2009) ("District Court Op."). In so doing, Rambus also abused the privilege of owning a patent monopoly. "As recognized by the Constitution, [a patent] is a special privilege designed to serve the public purpose of promoting the `Progress of Science and useful Arts'" and "is an exception to the general rule against monopolies and to the right to access to a free and open market." Precision Instrument Mfg. Co. v. Automotive Maint. Mach. Co., 324 U.S. 806, 816, 65 S.Ct. 993, 89 L.Ed. 1381 (1945). Thus, the public has a "paramount interest in seeing that patent monopolies spring from backgrounds free from fraud or other types of inequitable conduct...." Id. Here, Rambus abused its privilege by intentionally—as found by the district court—destroying evidence in bad faith to protect its exclusive monopoly.

Instead of recognizing this abuse by Rambus, the majority searches to find a needle in the haystack because, in its collective superior judgment, Rambus' conduct does not require taking away that privilege. In fact, the majority fails to consider the "high hurdle" that Rambus must overcome in showing that the district court abused its discretion. Tex. Digital Sys., Inc. v. Telegenix, Inc., 308 F.3d 1193, 1218 (Fed.Cir.2002). In so doing, the majority reweighs the evidence and decides that "several key items" on which the district court relied "may lead to a determination of bad faith," but the basis on which the district court "reached that conclusion" was not "clear." Majority Op. at 1327.

As an appellate court, we should not decide whether the facts before us "may" lead to a conclusion that we agree with, but whether by so concluding the district court abused its discretion. Indeed, "[t]he question, of course, is not whether ... the Court of Appeals, would as an original matter have [resolved the case in the same way as the District Court]; it is whether the District Court abused its discretion in so doing." Nat'l Hockey League v. Metro. Hockey Club, Inc., 427 U.S. 639, 642, 96 S.Ct. 2778, 49 L.Ed.2d 747 (1976) (citations omitted) (rejecting appellate court's reweighing of evidence and upholding district court's imposition of terminating sanctions for discovery violations as this did not amount to an abuse of discretion).

Here, the district court followed the appropriate Third Circuit standard and provided ample basis in fact for its decision to award dispositive sanctions. District Court Op. at 148-51. In the Third Circuit, a spoliating party acts in bad faith when it "intended to impair the ability of the potential defendant to defend itself." Schmid v. Milwaukee Elec. Tool Corp., 13 F.3d 76, 80 (3d Cir.1994). Under this standard, the district court did not abuse its discretion in finding that Rambus did, in fact, act in bad faith.

First, Rambus used its document retention plan to disguise and hide its destruction of relevant documents. Rambus "instructed patent counsel to purge [its] patent files," which would have at least been relevant to inequitable conduct. District Court Op. at 150. Further, Micron's defenses of patent misuse and violations of unfair trade and antitrust laws could all be "illuminated by evidence of a non-public nature, e.g., by internal Rambus documents," id. at 151, which almost certainly could have been included in the 300 boxes of documents destroyed in the second shred day in August 1999, id. at 145, or the 480 boxes destroyed on December 28, 2000, id. at 147. The district court, however, did not make a blanket determination that Rambus' document destruction impeded all of Micron's defenses. In fact, the district court found Micron's ability to assert anticipation and obviousness would not have been impaired by Rambus' spoliation, as the prior art used to assert such defenses is publicly available. Id.

Second, Rambus' document retention policy informed employees they should "LOOK FOR THINGS TO KEEP," including documents that would help establish conception but "expunge" "documents questioning the patentability of Rambus inventions." Id. at 142 n. 26. This policy remained in effect even after December 1998, the date after which destruction of documents was deemed to be spoliation. Id. at 150.

Third, Rambus' own documents (or, more accurately, those that did not make it to the shredding bin) demonstrate that it was aware that its document retention policy resulted in destruction of documents relevant to litigation. Outside counsel Neil Steinberg e-mailed Rambus executives on July 12, 2000 explaining his desire for a new document retention policy that "is similar to the previous policy—however, this time the IP group will attempt to execute the policy more effectively." Id. at 147 n. 57. In addition, Rambus' numerous misrepresentations about its document retention policy during the litigation are evidence, as found by the district court, of a guilty conscience. Id. at 147-48, 151.

In criticizing the district court's sanctions award, the majority claims that the district court must explain the propriety of the sanction "based on the degree of bad faith and prejudice and the efficacy of other sanctions." Majority Op. at 1328. This misstates the analysis a district court must undertake to award sanctions for spoliation in the Third Circuit. Schmid requires that a district court determine:

13 F.3d at 79 (citations omitted). With regard to the first factor, the district court must determine the "degree of fault" of the spoliating party, not the degree of bad faith. It is incongruous to establish a "degree" of bad faith—a party either did or did not act in bad faith. Indeed, Schmid recognized this by defining bad faith as whether the spoliator "intended to impair the ability of the potential defendant to defend itself...." Id. at 80 (emphasis added). Requiring degrees of bad faith is the equivalent of finding whether or not a party is "just a little bit pregnant." The majority's desire for the district court to define how "bad is bad" is contrary to Third Circuit law, which this court must apply.

The majority further states that the district court failed to satisfy the third Schmid factor by not explaining how holding Rambus' patents unenforceable would deter future spoliation of evidence, protect Micron's interests, and remedy the prejudice to Micron. Majority Op. at 1329-30. Not only is this contrary to the record, but the majority is now creating requirements for the imposition of dispositive sanctions that do not exist in the controlling regional circuit law. As explained above, the district court did not abuse its discretion in finding that Rambus acted in bad faith or that the destruction of these documents prevented Micron from mounting an appropriate defense. Further, the district court specifically found that any sanction other than a dispositive one would be "impractical [and] border[ ] on meaningless" due to the egregiousness of Rambus' conduct. District Court Op. at 151. Indeed, Rambus' conduct "impugned" "the very integrity of the litigation process." Id. Obviously, a dispositive sanction will serve to deter others from the egregious conduct seen here. There is no better way for the district court to have complied with the third Schmid factor.

In vacating the sanctions award, the majority has called the firing squad to the ready, the squad cocking their guns and taking aim, but instead of shooting the appropriate and culpable party, the squad aimed at the district court's proper determinations of fact. The majority selectively chooses those facts that support its desired outcome, while ignoring those that do not. Weighing evidence as a fact finder is not our function as an appellate court. If the evidence that was considered and weighed by the district court is objectively analyzed by this court under the abuse of discretion standard, it would lead all reasonable people to affirm. See Gen. Elec. Co. v. Joiner, 522 U.S. 136, 143, 118 S.Ct. 512, 139 L.Ed.2d 508 (1997) (holding that the appellate court erred in excluding expert testimony by "applying an overly `stringent' review to that ruling [and thereby] fail[ing] to give the trial court the deference that is the hallmark of abuse-of-discretion review"). On remand, the district court would not, in my judgment, need to review any additional evidence; it may only be required to parse the facts more specifically and again determine that the only appropriate sanction for Rambus' egregious conduct is dismissal of this suit. Moreover, I agree with the district court that under these facts, such a sanction would be appropriate.

In substituting its own views for those of the district court, the majority directly interferes with the sound discretion of the trial courts in managing their cases and prevents them from protecting the litigation process, which they are inherently bound to do. Cooter & Gell v. Hartmarx Corp., 496 U.S. 384, 404, 110 S.Ct. 2447, 110 L.Ed.2d 359 (1990) ("Deference to the determination of courts on the front lines of litigation will enhance these court's ability to control litigants before them."). Because the majority ignores this essential and inherent power of the district court, I dissent from its vacateur of the sanction imposed by the district court.