INTEL CORPORATION v. VIA TECHNOLOGIES, INC., (N.D.Cal. 2001)

WILLIAM ALSUP, UNITED STATES DISTRICT JUDGE.

INTRODUCTION

In this patent infringement case, the immediate question is whether the amended counterclaim should be dismissed for failure to state an antitrust claim and, if not, whether and to what extent the antitrust issues should be stayed and/or bifurcated. This order DENIES the motion to dismiss, DENIES the motion to stay discovery, and DENIES without prejudice the motion to bifurcate for trial purposes.

STATEMENT

The following well-pled facts taken from the counterclaim are presumed to be true for this motion. Intel Corporation is the dominant supplier of microprocessor chips used in desktop and notebook PCs in the United States, if not the world. The microprocessors relevant to this suit are referred to as "x86-compatible" because they use the x86 instruction software set. Microprocessors must communicate with other components such as memory or graphics-display devices and hard drives. To do so, they interconnect through another integrated circuit called a "chipset." To communicate with the microprocessor, each chipset must have bus specifications compatible with those of the microprocessor. The bus is the set of conductors connecting directly to the microprocessor. During the time period relevant to the counterclaim, there have been four different bus architectures used by the x86-compatible microprocessors made by Intel and another manufacturer, Advanced Micro Devices, Inc. ("AMD"), the leading competitor of Intel. A chipset designed to work with one bus architecture will not be compatible with any of the others.

A chipset must also be compatible with the buses that connect it to the PC components and peripheral devices. The so-called Peripheral Component Interconnect ("PCI") is an open standard for the bus that connects chipsets to hardware peripherals. This open standard allows vendors of hardware peripherals, such as modems and disk drives, to inter-operate with any x86-compatible chipset and microprocessor. Similarly, the Accelerated Graphics Port ("AGP") bus specification has been adopted as an open industry standard for the bus and interface between the chipset and the graphics chip. Graphics chip vendors expect that their graphic chips will work with any x86-compatible chipset as long as they use the AGP bus specification.

In sum, an x86-compatible chipset must be compliant with the PCI and AGP standards in order to be commercially viable. For the same reason, PCI-and AGP-compatible chipsets are necessary to enable x86-compatible microprocessors to compete.

* * *

In April 1996, Intel introduced AGP 1.0 as a new proposed bus standard for data transfers between all PC chipsets and all PC graphics chips. In May 1996, at an AGP developers conference attended by personnel of counterclaimant VIA Technologies, Inc., Intel encouraged industry use of the new standard, stated that it would be an open standard, and further stated that Intel would cross-license all of its patents necessary to implement the new standard on a royalty-free basis. At a subsequent AGP conference, about six months later, Intel again made the same representations and gave the same encouragement, making the same representation in writing. The AGP 1.0 specification released by Intel in July 1996 stated that any industry participant could execute a "reciprocal, royalty-free license" for any patents "necessary" to practice the AGP standard. In May 1998, Intel released the AGP 2.0 specification and likewise represented that a royalty-free cross-license was available to all industry participants for those patents "necessary" to implement the AGP standard.

In reliance on these and other Intel representations regarding the open AGP standard, VIA and the rest of the PC industry adopted AGP as the standard for the transfer of data between the chipset and the graphics chip on the PC. But for these representations, it is unlikely that they would have done so. Pursuant to the Intel representations, VIA then obtained a cross-license relating to the AGP standard. The license agreement did not call out any particular patents by name; rather, it licensed all patents "necessary" to implement the AGP standard, as the parties characterize it. The actual language is set forth in the footnote below.

According to a joint post-argument submission, the cross-license covers patent claims "that must be infringed in order to comply with the AGP Interfaces." The license states that it does not extend to features of a product which are not required to comply with the AGP Interfaces or to which there was a feasible alternative to infringing a given claim.

* * *

In June 1999, Intel commenced an infringement action against VIA in this district, accusing its chipsets of violating patents covering the AGP standard, or so it is alleged. In January 2000, Intel filed a similar proceeding against VIA in the International Trade Commission to block imports. In both, Intel asserted U.S. Patent No. 5,333,276 against VIA's "P5-compatible" chipset intended for use with AMD's K6 microprocessor and VIA's "EV6-compatible" chipset intended for use with AMD's Athion and Duron microprocessors. As stated, AMD is Intel's leading competitor in the x86-compatible microprocessor industry. Another patent asserted was U.S. Patent No. 5,740,385, discussed below. Intel also asserted both patents against VIA's "P6-compatible" chipset (intended for use with Intel microprocessors). VIA alleges these proceedings were maintained to monopolize or in an attempt to monopolize two markets in violation of Section 2 of the Sherman Act.

The two relevant markets alleged are: the market for PC microprocessors and the market for x8l6-compatible PC chipsets. As for the PC microprocessor market, Intel has long held a commanding lead, in excess of a seventy percent share. AMD, the number two in the market, lags behind at ten to twenty percent. As for x86-compatible PC chipsets, Intel commands a fifty percent or greater share. VIA is Intel's principal competitor in this market and is the principal third-party supplier of chipsets for AMD's Athion and Duron x86 microprocessors. Both markets are protected by high barriers to entry.

In order to dominate and control these markets and to suppress competition, it is alleged that Intel brought the two actions described above to prevent VIA from selling its P5-and P6-compatible chipsets in the United States (and elsewhere). With alleged predatory intent, Intel brought and maintained its prosecution of the '276 patent against VIA's products in order to place AMD and VIA at a competitive disadvantage. The counterclaim alleges that the assertion of the '276 patent was a sham because Intel knew at all times that its prior representations and its blanket license conferred on VIA the right to practice the art described in the '276 patent.

In July 2000, Intel modified its accusations. It dropped all the claims in its original complaint and asserted four different patents, including U.S. Patent No. 6,006,291 entitled "High-Throughput Interface Between a System Memory Controller and a Peripheral Device." This patent, allegedly, claims a monopoly on a "fast write" feature that is a key and necessary feature of the AGP 2.0 specification. In response, VIA again contends that it has been licensed to use whatever Intel patents are necessary to practice the AGP 2.0 specification, including the '291 patent. For Intel to assert otherwise, VIA alleges, is a sham, designed to block VIA from competing in the x86-compatible PC chipset market and to block AMD from competing in the PC microprocessor market by denying it a principal source of AGP 2.0-compliant chipsets.

* * *

In both the ITC proceeding and the action filed herein in June 1999, Intel originally asserted its U.S. Patent No. 5,740,385. Discovery, however, showed that a named inventor, George R. Hayek, an Intel employee, had willfully withheld from the Patent Trademark Office ("PTO") a key item of prior art, namely an Intel manual describing the Intel 82430 PC/SET Cache/Memory Subsystem. The manual had been published more than one year before the patent application. Although Inventor Hayek had sworn in his application that he knew he had a duty to disclose material information pursuant to 37 C.F.R. § 1.56 and had sworn that the claimed invention had not been described in a printed publication more than a year before the filing date, the detailed allegations, if true, show that Hayek's sworn statements were false. In addition, the manual contradicted arguments Intel's patent prosecution lawyers made to evade other prior art cited by the examiner provisionally to disallow the claims. In reliance on the material omissions and the arguments of Intel, the examiner eventually allowed the claims, and the '385 patent issued. When asked at his deposition if he had intended to deceive the PTO by withholding the manual, Inventor Hayek testified that he "could not recall." In July 2000, Intel deleted the '385 patent from the roster of patents asserted against VIA, but Intel has not disclaimed the patent and it remains a potential obstacle to the industry.

ANALYSIS

Intel's motion to dismiss is directed at important but narrow elements in VIA's antitrust claim for monopolization and attempted monopolization under Section 2 of the Sherman Act. For purposes of the instant motion, Intel must be taken to enjoy dominant positions in both alleged product markets and the two markets must be deemed properly framed, there being no challenge on these scores at this stage. The main issue now presented under Section 2 is whether the industrial tactics employed — namely the fraud on the PTO and the maintenance of alleged sham litigation — are immunized from liability by the Noerr-Pennington doctrine. See Eastern Railroad Presidents Conference v. Noerr Motor Freight, Inc., 365 U.S. 127 (1961); United Mine Workers v. Pennington, 381 U.S. 657 (1965).

Sham Litigation

In deference to the First Amendment and the right to petition the government for redress of grievance, a lawsuit ordinarily cannot be deemed predatory under the Sherman Act. One exception is where the lawsuit is a sham. To qualify as a sham, a lawsuit (i) must be objectively baseless and (ii) must be subjectively motivated by a desire to impose collateral, anti-competitive injury rather than to obtain a justifiable legal remedy. Professional Real Estate Investors v. Columbia Pictures Industries, Inc., 508 U.S. 49, 60-61 (1993); California Motor Transp. Co. v. Trucking Unlimited, 404 U.S. 508, 515 (1972); Nobelpharma AB v. Implant Innovations, Inc., 141 F.3d 1059, 1071 (Fed. Cir. 1998).

In the present case, if the blanket license in favor of VIA is eventually held to comprehend the '276 and '291 claims, and if it is determined that it was objectively baseless to think otherwise, and if it

Intel argues that the "objectively baseless" element could never be satisfied because VIA itself brought a counterclaim seeking declaratory judgment that the asserted patents were not infringed. Therefore, Intel contends, there must have been an objective basis under Rule 11 to have believed that the patents were not "necessary" to practice the AGP specification. This is counterfeit logic. Noninfringement may be proven by a license. Wang Labs., Inc., v. Mitsubishi Elec. America, Inc., 103 F.3d 1571, 1573 (Fed. Cir. 1997). In other words, the Court reads VIA's pleadings to contend that VIA practices the claims asserted but that no infringement ever existed because of the blanket license.

Intel asks the Court to take judicial notice of a report that VIA's expert (Dr. Edward S. Davidson) provided to the ITC. The report concluded that certain VIA products did not practice the '276 patent claims. On the surface, Intel seems correct that there is a troubling inconsistency. VIA, however, argues that the report was based on a pre-existing claim construction by Intel's expert that had already taken the claims outside the AGP specification. It is impossible, moreover, for the Court to correlate, one-to-one, the accused specific products listed in the report with the accused generic products listed in the counterclaim. The report is so evidentiary in nature that, at the pleading stage, the Court will not grant it conclusive effect. So too with the further evidence of inconsistencies catalogued in Intel's second supplemental post-hearing memorandum. is found that such claims were asserted with the subjective motive to suppress competition rather than to vindicate justifiable rights, then the assertion of those claims would, under the allegations made herein, be stripped of Noerr-Pennington immunity. That would not, by itself, prove a Sherman Act claim. The usual elements would still have to be proven. But it would qualify the sham litigation for use as a predatory act.

In this connection, Intel argues that VIA's counterclaim has not conceded that it even practices the claims of the '291 and/or '276 patents. Intel, however, affirmatively accused VIA of infringing those claims. That counts for something. At all events, VIA has come close enough in its pleading. VIA alleges that the '276 patent "purports to disclose and claim a scheme for implementing high-priority read and write commands that are necessary in order to comply with the AGP 1.0 and 2.0 bus interface standards" (¶ 95). VIA alleges that the title of the '291 patent "is a shorthand description of the AGP standard" (¶ 101), and alleges that "the '291 patent purports to cover a key feature of the AGP 2.0 specification" (¶ 104). These allegations are sufficient at the pleading stage to show that the patents in question disclose AGP inventions used by the accused products.

An unimmunized bad act by even a dominant firm does not translate automatically into a Section 2 violation. Intergraph Corp. v. Intel Corp., 195 F.3d 1346, 1354 (Fed. Cir. 1999). All other elements of a Section 2 violation must also be present. Aside from proper market definitions, there still must be injury to competition (in the case of monopolization) or a dangerous probability thereof (in the case of attempted monopolization). One or more bad trade acts, reprehensible though they may be, may not actually restrain competition or pose a dangerous probability of doing so. In the present case, therefore, it remains to be seen, on the facts, whether the alleged predatory counts truly restrained competition. It seems clear that the alleged sham litigation failed to chill VIA. It may have, as alleged, discouraged VIA's customers and potential customers from doing business with VIA by casting a cloud over the legality of its product line.

Two antitrust issues of law are raised that will not be finally decided on a pleading record. Intel argues that because its original complaint asserted non-sham patents in addition to the alleged shams (and/or because its amended complaint now asserts three patents that VIA does not allege are wrongfully asserted), Intel has not engaged in sham litigation as a matter of law. According to Intel, "the Noerr-Pennington doctrine immunizes an entire lawsuit if any one of its counts is not a sham" (Br. at 3). Intel is correct that one district court has so held. Dentsply Int'l, Inc., v. New Technology Co., No. 96-272 MMS, 1996 U.S. Dist. LEXIS 19846 (D.Del. Dec. 19, 1996). That ruling purported to follow a statement in Professional Real Estate Investors that "the lawsuit must be objectively baseless in the sense that no reasonable litigant could realistically expect success on the merits." 508 U.S. at 60 (emphasis added). In fairness, however, the Supreme Court was addressing a different issue, namely the objective (as opposed to subjective) component of the sham test and its use of the phrase "the lawsuit" was not directed at the instant issue. So too with Eden Hannon Co. v. Sumitomo Trust Banking Co., 914 F.2d 556, 565 (4th Cir. 1990). The Fourth Circuit rejected the sham litigation theory because the antitrust defendant's lawsuit was vindicated on appeal (in the same opinion). The Fourth Circuit did not address the more subtle issue of a single lawsuit with multiple claims, some of which are baseless, a "mixed" claims scenario.

Similarly, VIA's authorities do not really come to grips with the issue. VIA relies on the Handgards litigation from the Ninth Circuit. Handgards Incorporated v. Ethicon Incorporated, 601 F.2d 986 (9th Cir. 1979), applied the "sham litigation" rule in the context of the assertion of two patents, one of which was valid. In Handgards, Ethicon asserted two patents, the Orsini patent and the Gerard patent. Before trial, but after summary judgment, Ethicon dropped its claims regarding the Orsini patent. At trial, the court found that the Gerard patent was invalid under the "prior public use" rule. Handgards then brought an antitrust action against Ethicon for assertion of both the Orsini and Gerard patents and also sought declaratory relief that the Orsini patent was invalid. At the antitrust trial, the court found that the Orsini patent was valid, but the Gerard patent had been asserted in bad faith. On the next appeal, the Ninth Circuit held that Handgards had suffered antitrust injury as a result of the assertion of the Gerard patent, but had not suffered antitrust injury as a result of the assertion of the Orsini patent. Handgards, Inc., v. Ethicon, Inc., 743 F.2d 1282, 1296-97 (9th Cir. 1984). Although the fact pattern was on point, the decisions did not expressly address the problem of a "mixed" claims scenario.

The Court can imagine good arguments both ways. On the one hand, a plaintiff who adds a baseless claim to an otherwise bona fide lawsuit presumably does so for a reason, i.e., presumably to obtain an additional strategic advantage over and above those flowing from the assertion of non-sham counts. To immunize such bad-faith conduct because it is joined with good-faith conduct would eviscerate most of the sham litigation exception. Especially in patent litigation, an overreaching plaintiff will usually have a large portfolio of patents to press. It would be comparatively easy to find a colorable claim to join with baseless claims. On the other hand, evaluation and segregation of counts into "objectively bad" versus "objectively good" claims as well as "subjectively bad" versus "subjectively good" claims could require large amounts of work for a jury, not to mention claim-construction hearings (on patents no longer part of the case-in-chief). In a mixed-claim scenario, it would also be necessary for a jury to isolate and to trace the harms resulting solely by reason of the wrongful components of the litigation.

Cf. Paramount General Hospital Co. v. Jay, 213 Cal.App.3d 360, 368 (1989)("it would seem almost a mockery to hold that, by uniting groundless accusations with those for which probable cause might exist, the defendants could thereby escape liability" for malicious prosecution). Village of Lake Barrington v. Hogan, 272 Ill. App.3d 225, 237 (1995) (no authority for proposition that party can file sham litigation as long as one of several counts in the complaint had merit).

Given the uncertain state of the law on this issue, this order will not resolve it at the pleading stage. Rather, discovery will go forward. Discovery may reveal that the incremental effects of the supposed sham components were negligible or may show that they dominated the original complaint. Discovery may shed light on the actual factual pattern in need of a rule. While ultimately the Court may (or may not) agree with Intel's sweeping categorical rule of immunity, it should not do so at the pleading stage. See White Motor Co. v. United States, 372 U.S. 253, 261 (1963).

It is also unnecessary to reach the possible applicability of In re Dell Computer Corp., 121 FTC 616 (1966), concerning successful advocacy of an open standard by a dominant firm followed by its informing the industry that its use would infringe exclusive rights of the firm. VIA's counterclaim goes beyond Dell and presents an easier case. VIA alleges that Intel, after locking VIA and the industry into the AGP standard, brought sham suits against VIA for practicing the standard. There is, however, no allegation that Intel has threatened to sue others or has otherwise sought to prevent others from using the AGP standard. Since this order holds that the accused litigation is not immune at least this stage, it is unnecessary as yet to decide whether Dell-type conduct (without any sham lawsuits) would be actionable under the Sherman Act. For the same reason, it is unnecessary to decide whether a Dell-type theory is reconcilable with the statement of the Federal Circuit that "in the absence of any indication of illegal tying, fraud in the Patent and Trademark Office, or sham litigation, the patent holder may enforce the statutory rights to exclude others [under the patent] free from liability under the antitrust laws." In re Independent Service Org. Antitrust Litig., 203 F.3d 1322, 1327 Fed. Cir. 2000).

The Settlement/Release Defense

The alleged fraud practiced by Intel's inventor on the PTO lacks any immunity under Walker Process Equipment, Inc., v. Ford Machinery Chemical Corp., 782 U.S. 172, 177 (1965). Intel, however, argues that its settlement with VIA last summer, one year into the case, absolved it of liability for the Walker Process claim relating to the '385 patent. This is a matter of defense and will not be considered on a Rule 12 motion. Even Intel concedes that the settlement agreement preserved VIA's right to renew its Walker Process and sham litigation claims, relating to patents "at issue." Intel simply argues that the '385 patent was no longer "at issue" by the time of the settlement because it had already been withdrawn from the suit and the ITC proceedings. Such contract-interpretation questions should not be resolved on a motion to dismiss.

California Business Professions Code Section 17200

The only ground on which Intel moved to dismiss VIA's claim under California Business Professions Code § 17200 was that it is dependent on the survival of VIA's antitrust counterclaim. Because VIA's antitrust claim will not be dismissed at this stage, this argument fails.

The Bifurcation/Stay Motion

There will be no discovery stay. Discovery and motion practice may proceed on both the complaint and counterclaim. Contrary to Intel's characterization, the antitrust claims herein are prior in time, at least in this sense: Intel abandoned its initial battery of patents and refiled an amended complaint with a new battery of patents. The antitrust issues arose over the original set and were in play already, as evidenced by the reservation in the settlement agreement, before the amendment. The antitrust issues, therefore, have at least equal priority. The Court is concerned, moreover, over the allegation (not denied by Intel), that its inventor witness could "not recall" whether he had intended to defraud the PTO when he withheld the manual. The further passage of time may prejudice the ability of counsel to pin down the recollections of party witnesses. In addition, both sides are represented by legions of lawyers. There will be no genuine difficulty in their managing discovery for both suits.

Reasonable priority in discovery shall be accorded the settlement-scope and the license-scope issues so that a partial summary judgment motion thereon may be brought by August 2, 2001, to be heard on September 6, 2001. This does not mean that discovery on the issues in the complaint and answer should receive lower priority; it only means that a reasonable balance must be struck. The Court will decide later whether the trial will include both the patent and antitrust claims, whether (or to what extent) there will be any bifurcation for purposes of trial, and whether any adjustment in the trial date is warranted. Both sides are expected to work hard to meet the existing deadlines and should not anticipate any extensions.

Intel's post-argument request for further oral argument is DENIED. In its second post-argument submission, Intel's catalogs evidence that VIA has not been candid on this motion and, in fact, has been evasive or inconsistent in its previous litigation position concerning whether VIA practices the claims of the '291 and/or '276 patents. This evidence is outside the pleading and, as stated in footnote 2, will not be considered on the Rule 12 motion. For purposes of regulating discovery, however, the Court is troubled by Intel's showing. The Court directs VIA to file and serve sworn, direct, unequivocal and complete answers to the following questions by April 6, 2001:

(i) With respect to each patent claim in the '276 and '291 patents VIA contends was asserted in bad faith, identify the claim and identify precisely which, if any, accused VIA products ever read on said claim and

(ii) With respect to each such claim and any such accused VIA product reading thereon, state in full the basis for any contention by VIA that the AGP license covered said claim as to said product.

IT IS SO ORDERED.