|
Tech Law Journal |
 |
| News, records, and analysis of legislation, litigation, and regulation affecting the computer, internet, communications and information technology sectors | |
|
|
|
| |||||
|
This document is split into four parts:
This document was created by Tech Law Journal by converting a MS Word 8.0 version into HTML. Several features were eliminated in the conversion, including double spacing, and paragraph indentations. Footnotes were converted into sidenotes. Copyright 1998 Tech Law Journal. All rights reserved. [begin page 1] 1. Whether the district court erred in concluding that Intergraph is likely to prove monopolization or attempted monopolization by Intel in violation of the Sherman Act, 15 U.S.C. § 2 (1994), where the district court (a) did not have reliable data showing Intel has a monopoly share of any relevant primary (upstream) market, (b) did not properly define a relevant secondary (downstream) market that Intel is attempting to monopolize, (c) did not find a specific intent to monopolize a secondary market based on Intel’s response to Intergraph's assertion of its patents against Intel and its customers, and (d) did not establish any of the elements necessary to show that it was dangerously probable that Intel would monopolize any market in which Intergraph competes. 2. Whether the district court erred in granting a mandatory preliminary injunction requiring Intel to treat Intergraph as a preferred customer and supply it with samples of Intel’s patented microprocessors and copyrighted materials and trade secrets, in advance of their release to the public, on the ground that such advance products and materials are, under the Sherman Act, an "essential facility," which must be provided to Intergraph without compensation, while Intergraph remains free to pursue its patent infringement lawsuit seeking billions of dollars in damages and an injunction against Intel’s manufacture and sale of the very microprocessors sought by Intergraph. 3. Whether the district court erred in holding that Intel had no legitimate business reason for ceasing to treat Intergraph as a preferred customer, without adequately assessing whether Intel’s conduct furthered [begin page 2] competition and whether a mandatory injunction would serve to protect competition rather than just protecting Intergraph. 4. Whether the district court erred in identifying joint sales presentations by Intel and some of Intergraph’s competitors as being conspiracies in restraint of trade where such presentations would not result in an injury to competition, and in finding coercive reciprocity on the basis of Intel’s purported tying of Intergraph’s advanced receipt of Intel’s technology to Intel’s obtaining a license under the Intergraph patents, when the conduct involved would not foreclose competition in any way. 5. Whether the district court erred in construing a one-page letter to Intergraph from an Intel field sales engineer to be an enforceable contract of specific duration that superceded the terms of the parties’ written Non-Disclosure Agreement ("NDA"), even though the letter lacked an offer, acceptance, consideration, or specific terms and expressly referred to the allegedly superceded NDA. 6. Whether the district court erred in (a) ruling that a bilateral termination provision in a written NDA executed by two sophisticated commercial enterprises to protect their confidential information was "unconscionable" both as written and as applied, (b) alternatively finding that "reasonable notification" of termination of the NDA required at least 18 months advance notice, and (c) issuing a mandatory preliminary injunction that contains directives wholly unrelated to remedying the alleged wrong and which is wholly inconsistent with the specificity requirements of Fed.R.Civ.P. 65. [begin page 3]
Intel is a Delaware corporation, with corporate headquarters in Santa Clara, California, and is a leading manufacturer and supplier of microprocessors (also known as central processing units or "CPUs") for use in a wide variety of computer applications (A4). Intergraph has its principal place of business in Huntsville, Alabama and manufactures high-end computer workstations, used principally in the generation of computer-aided graphics (A11-12). The record does not suggest that Intergraph makes CPUs or that Intel makes workstations. In 1987, Intergraph purchased a microprocessor known as the "Clipper," including the related patent rights. Between 1986 and 1993, Intergraph used the Clipper technology in its workstations (A12-14). By the end of 1993, "Intergraph had discontinued further development of its own Clipper microprocessors" and began to incorporate Intel microprocessors into its workstations (A14). Intergraph concedes that it switched from the Clipper technology without any commitment from Intel "to provide [to Intergraph] a perpetual supply of chips, pre-released chips, or confidential information" (A14 n.24). In 1994, Intergraph sought access to Intel’s confidential information regarding microprocessor products in development to aid in the early design of Intergraph workstation products based on such new products. Intel agreed, but only upon execution of a written NDA requiring each party to maintain the confidentiality of any proprietary information exchanged between them. Since then, Intel and Intergraph have shared a business relationship under written NDAs that enabled Intergraph to [begin page 4] receive Intel’s proprietary materials, such as advance technical information and engineering samples of unannounced products, which contain Intel’s proprietary patented, copyrighted, and trade secret technology. Consistent with Intergraph’s initial understanding that Intel had never promised a continual supply of unannounced CPUs or confidential technical information, the NDA recites that "[n]either party has any obligation to disclose Confidential Information to the other" (A17 n.29). In addition, the NDA confirms that neither party has an obligation to buy or sell products, and that there is no implied or express license grant from either party to the other (A17 n.30). Either party had the right to terminate the NDA at-will and request the return or destruction of confidential information disclosed under the agreement (A15-17). In 1997, three years after the first NDA was signed, Intergraph began threatening some of Intel’s customers (who are also Intergraph’s competitors), alleging that their use of Intel’s microprocessors infringed Intergraph’s Clipper patents (A19). In response to indemnification requests from these threatened customers, Intel sought to negotiate a cross-license with Intergraph to solve the rapidly escalating patent dispute. Id. However, the parties were unable to reach a mutually agreeable solution. When it became apparent that a cross-license would not be satisfactory to Intergraph, Intel began treating Intergraph as a "regular" rather than a "preferred" customer. Intel ceased providing Intergraph with proprietary Intel materials, such as advance technical information, proprietary and patented engineering samples of unannounced products [begin page 5] and pre-released microprocessors. Pursuant to the NDAs, Intel also asked Intergraph to return Intel’s trade secrets (which Intergraph refused to do). Notwithstanding the unresolved patent dispute, Intel continued to provide Intergraph with released products and related non-confidential information. On November 17, 1997, Intergraph filed a complaint asserting three counts of patent infringement under its Clipper patents and various state law claims, but no antitrust claims (see A211-64). Although Intergraph’s complaint sought to enjoin Intel from making, using or selling its microprocessors, four days after filing the complaint, Intergraph filed for a mandatory preliminary injunction that would require Intel to make and sell those very same microprocessors and provide Intergraph an unlimited and continuous supply of pre-released microprocessors and technical data (A265-66). Two weeks later -- after Intel had responded to Intergraph’s motion -- Intergraph amended its complaint, adding new antitrust counts as the purported basis for the previously-requested injunction (A101-02, A211-12). In pertinent part, Intergraph’s amended complaint alleges that Intel’s decision not to supply Intergraph with sample and pre-released CPUs or to predisclose technical information constitutes "unlawful refusals to deal with Intergraph, denial of access to essential products and technical information that Intergraph needs to compete, and improper leverage and use of Intel’s monopoly in microprocessors and chips to monopolize or attempt to monopolize downstream markets such as the [begin page 6] workstation market in which Intergraph competes" (A102). However, the amended complaint does not allege that Intel competes with Intergraph in "the workstation market." On December 8, 1997 -- days after the antitrust claims were first raised -- the district court held a preliminary injunction hearing (A2). Without prior notice to Intel, the court permitted Intergraph to supplement the written record for its antitrust allegations by presenting live testimony (A114-15). Four months later, the court issued an order setting forth the mandatory preliminary injunction against Intel from which this appeal is taken (A1-80). That mandatory injunction requires Intel to restore Intergraph to a "preferred" customer status, and, among other things, to provide Intergraph with advanced chip samples and technical documentation, a supply of production chips, and involvement in Intel marketing activities (A75-80). As the basis for its injunction order, the Alabama district court described a series of flawed antitrust liability theories, and concluded that Intergraph was likely to succeed against Intel on every one of them. In addition, the district court construed a one-page letter by an Intel field sales engineer to be an enforceable contract of specific duration that superceded the parties’ NDA and warranted specific performance. Many of the district court’s factual findings supporting these erroneous legal theories were taken from the uncorroborated hearsay testimony of Mr. Wade Patterson, the President of Intergraph’s Computer Hardware Division in Huntsville, who testified in person at the hearing [begin page 7] (A114). Based on such evidence, which the district court itself conceded may ultimately prove to be "illusory" (A4 n.4), the court concluded that Intergraph was likely to prevail on its antitrust claims. However, the district court’s findings (based on self-serving statements by Intergraph, not supported by objective evidence) highlight that the only grounds on which the district court could premise the requisite injury to competition was on the legally impermissible basis that injury would be suffered allegedly by a single competitor -- Intergraph:
Other than these and similar findings regarding injury or harm to Intergraph, the district court made no findings with respect to showing an injury to competition generally as required by the antitrust laws. [begin page 8] The district court’s misplaced desire to protect Intergraph rather than competition generally is further revealed by the finding that "[i]f . . . consumers cannot get the most advanced product from Intergraph, they will get it from Intergraph’s competitors who have not been denied access to the advance information, sample chips, and production chips that Intel has denied to Intergraph" (A27). The court did not explain how there can be any antitrust violation if consumers can get the products they want from others in a competitive market. Notably, the district court received no evidence and thus made no findings as to whether or how any perceived injury to Intergraph’s ability to compete would adversely affect competition generally in the relevant market in which Intergraph competes, or would reduce the level of innovation, or would deny consumers access to improved technology (other than perhaps Intergraph’s), or would reduce competition on the basis of price and quality. Instead, the district court merely offered a conclusory declaration that harm to one competitor equates to harm to competition (A15, A26-37). In keeping with the theme of its findings, the district court’s legal conclusions suggest a quest to find any legal basis to alleviate the perceived short-term harm to Intergraph’s employees resulting from its management’s decision that Intergraph is likely better off in the long run by suing Intel for billions of dollars. Among the many alternatives offered by the district court as to why Intergraph might be likely to prevail on all of its antitrust claims are: (a) the supply of advanced Intel CPUs and technical information is an "essential facility" because "competitors cannot effectively compete in the relevant markets without access to them" (A44); [begin page 9] (b) Intel may be "leveraging" a monopoly in a primary (upstream) market for high-performance CPUs or Intel-branded CPUs to thwart "competition by Intergraph" in a secondary (downstream) "graphics subsystems relevant market" (A45-46); (c) Intel’s course of conduct amounts to "a form of coercive reciprocity" (A46-47); (d) "Intel has no legitimate intellectual property basis with which it can refuse to supply Intel microprocessors and technical information to Intergraph" (A48); and (e) "Intel has entered into one or more agreements and contracts in restraint of trade" (A50). As will be shown herein, the district court overlooked the necessary elements on every one of these theories. Despite glaring legal and evidentiary gaps in the record, the court nonetheless granted the extraordinary mandatory preliminary injunction requested by Intergraph.
As an alternative justification for its injunction, the district court concluded that it is likely that Intergraph would establish that a one-page letter written to Intergraph by an Intel field sales engineer "constitutes an enforceable agreement of definite duration" (A54 & n.45). Notwithstanding this conclusion, the district court was unable to pin down the allegedly definite duration, holding Intergraph was entitled to specific performance of the letter until "Intel develops the Deschutes and Merced [microprocessor] programs, and perhaps longer" (A55) or "at least through 1999 . . . and perhaps through the year 2000" (A57). According to the court, that letter also supercedes the parties’ pre-existing written NDAs, even though (a) the letter expressly references the continued applicability of the NDAs, and (b) the NDAs contain terms [begin page 10] expressly contrary to the ordered relief. To reach this result, the district court found it necessary to disregard the express reference in the letter to the executed NDAs so as to avoid having enforcement of their termination provisions render the letter contract "illusory" (A54). The district court also held that the termination provisions in the NDAs were unconscionable, either at the time they were made or if invoked by Intel to prevent Intergraph from receiving advance access to Intel’s products and technical information (A57-65). While the court characterized its injunction as being "specific performance" of the letter, the terms of the order do not resemble the terms of the letter. As written, the district court’s injunction requires Intel (a) to supply Intergraph with "all Intel product information," in the manner Intel supplies "Intergraph’s similarly situated competitors" (A75); (b) to "set aside a supply" of CPUs for Intergraph on "an advance basis," and in "such quantities as forecasted by Intergraph" (A76); and (c) to "supply Intergraph with Production Chips not yet available from Intel’s authorized distributors," in "such quantities as forecasted by Intergraph" or in the amount supplied to "Intergraph’s similarly situated competitors" (A78). None of these terms can be found anywhere in the letter itself. Moreover, the order contains no guidance on how Intel is to treat Intergraph like allegedly similarly situated customers, each of whom Intel treats differently based on its own attributes, needs and contributions. Finally, the district court required Intergraph to post a bond in the amount of only $25,000 to cover the entirety of any harm to Intel if it is determined that Intel has been wrongfully required to turn over its intellectual property (A79). [begin page 11] After several years of harmonious dealings, Intergraph began to assert a scope for its Clipper patents that seeded apprehension among Intel’s customers (Intergraph’s competitors) about Intel’s continued right to manufacture and sell its flagship microprocessors. Intel tried to resolve this patent dispute by negotiating a cross license of intellectual property. Intergraph rebuffed this proposal, as it had a right to do, presumably because it viewed a cross license as less desirable than trying to parlay its Clipper patents into billions of dollars in royalties or damages from Intel and its customers. However, Intel had a corresponding right to cease predisclosing its core intellectual property to a company that was challenging Intel’s right to make, use, and sell products incorporating that very same intellectual property. The court erred in overriding this fundamental legal right. For responding as it did to Intergraph’s assertion of its intellectual property against Intel’s core microprocessor business, Intel has been branded a monopolist and a likely antitrust violator, denied its copyright and patent rights, forced to divulge its trade secrets, penalized for using its business judgment as to how best to protect its confidential information and intellectual property, compelled to do business with a litigation opponent, and threatened with contempt if it does not follow an order too vague to satisfy the specificity requirements of Fed. R. Civ. P. 65(d). Neither antitrust law, contract law, tort law, nor the law of injunctions supports this result. Intel was treated in this manner, although it did no more than lower Intergraph’s customer status from "preferred" to "regular." Despite the [begin page 12] ongoing dispute, and although it had no obligation to do so, Intel continued to provide Intergraph with released microprocessors and technical information. Intel only ceased providing its proprietary, pre-release CPU products and confidential information to Intergraph, thereby putting Intergraph on the same footing as any of Intel’s many other "regular" customers. Thus, this dispute concerns Intergraph’s ability to use the antitrust laws to compel continued receipt of confidential Intel materials as early as Intergraph competitors that are not suing to enjoin Intel from supplying its microprocessors to anyone. Contrary to the district court’s notion of fairness, the Eleventh Circuit and several other courts have made it clear that a defendant’s refusal to deal with a litigation opponent is entirely permissible. Courts also have held that parties have no duty to predisclose proprietary information and may elect to maintain a competitive advantage by retaining their confidential information. In addition, it is well established that a patent owner has no obligation to license or sell its patented products to anyone. Here, Intel has been compelled by injunction to do all of the above, and more, under circumstances which Intergraph can exploit to obtain extrajudicial discovery, to create a commercial advantage over its rivals, or to try to develop additional infringement claims. The antitrust laws do not support such a result. As the primary basis for its order, the court theorized that Intel has a microprocessor monopoly (the purported primary or upstream market) and is "leveraging" that monopoly to gain an illegal "advantage" in the undefined graphics subsystems market (the purported secondary or downstream market) by refusing to supply advance-sample and pre-released chips and [begin page 13] information to Intergraph. But there is simply no duty to deal in intellectual property. Even a purported monopolist controlling an essential facility has no obligation to sell its patented goods, divulge its copyrighted materials, or disclose its trade secrets. Indeed, even where intellectual property is not involved, a purported monopolist has no duty to deal unless it is using its monopoly power in one market to gain a monopoly in another market. There is no evidentiary record or legal analysis to support such a "leveraging" claim here. First, the record provides legally-insufficient grounds to support a finding that Intel has monopoly power in either of the alleged relevant primary markets (high-end CPUs or Intel-branded CPUs) identified by the district court. For the first market, there are no market shares showing that Intel has monopoly power. The second is simply not a cognizable antitrust market under Eleventh Circuit precedent. So there is no primary market monopoly for Intel to leverage. Second, the court’s findings do not suggest that Intel could gain a monopoly in the identified secondary market ("graphics subsystems") by refusing to deal with Intergraph. Before the district court could legitimately conclude that Intel could gain a second, "downstream" monopoly, a great deal of additional evidence, findings, and legal analysis would have been needed. Specifically, the court needed to define carefully the downstream market that was in danger of being monopolized, so that it could determine the probable effect of Intel's challenged conduct on competition in that market rather than on Intergraph alone. Here, there is neither evidence nor findings supporting such a market definition, merely alternating [begin page 14] references to "workstations" or "graphics subsystems" without any indication of what products actually compete against each other. The district court also should have determined whether entry barriers protect the properly-defined secondary market from new competition, so that it could assess Intel's potential to monopolize that market. Once again, there is no record evidence or findings on that subject. After defining a market and assessing entry barriers, the court needed to determine whether Intel had a dangerous probability of gaining a monopoly (instead of just whether it might gain an unspecified advantage) in the secondary market as a result of the alleged conduct. Courts typically require an upstream monopolist to have at least a 50% share of a downstream market before such a dangerous probability can be found. Not only did the district court fail to make a finding of Intel’s market share in the graphics subsystems "market," but the evidence of record shows that Intel is not even a competitor in that market. Next, the district court should have evaluated, but failed to evaluate, whether Intel had a specific intent to monopolize the alleged downstream market. Had the court done so, it would have found that all the record evidence on this subject shows no such intent. Instead, Intel’s purpose was to resolve an intellectual property dispute that threatened Intel’s core business. Merely downgrading Intergraph's customer status in the face of such a threat is no evidence at all of a specific intent to monopolize. The court did discuss whether Intel controlled an essential facility. But the evidence did not support the court’s finding that Intel had a monopoly share of a primary market, so the essential facility finding also [begin page 15] was without foundation. Moreover, as a matter of law, a firm controlling an essential facility has no obligation to deal unless it is attempting to use that essential facility to monopolize a downstream market. This is simply not the case here. Finally, even if a firm were attempting to leverage an essential facility, the firm would still have a right to refuse to deal if the essential facility was protected by intellectual property, which is the case for Intel’s microprocessors and confidential materials. Simply put, the district court’s injunction only serves to protect Intergraph, despite the Supreme Court’s repeated admonitions that the antitrust laws are only to protect competition, not competitors. In protecting just Intergraph, the court displaced the antitrust laws and gave effect to its unique view that a defendant must promote the business of a litigation adversary suing it for patent infringement. The law is clearly to the contrary. Intergraph may have had a right to assert and sue on a good faith belief of infringement, but Intel had a corresponding right to propose a cross license. Intergraph could refuse to negotiate a cross license so as to continue its quest in court for billions in damages and an injunction, but Intel had a corresponding right to cease providing its intellectual property to Intergraph. By ignoring Intel’s rights, the district court allowed Intergraph to "have its cake and eat it too." The court enabled Intergraph to sue Intel for infringement in hopes of extracting damages or royalties under the threat of an injunction, without suffering any inconvenience in its business activities from Intel’s invocation of its intellectual property rights. That result simply cannot be upheld. Equally inappropriate is the district court’s attempt to justify its injunction as a form of "specific performance" enforcing a one-page letter [begin page 16] from an Intel field sales engineer that the district court improperly construed to be a valid contract of specific duration. While the court tried to reason its way around the absence of offer and consideration, it did not even try to show how there had been an "acceptance." That fundamental flaw should have ended the analysis right there. Nevertheless, the court continued its analysis, stating it could overcome any open terms, such as the lack of a specific duration, through "gap-filling" techniques, but it never did. In holding that the letter "contract" superceded the parties’ written NDAs, the court declined to enforce the NDAs on grounds that it would render the newly-formed letter contract "illusory." Such circular logic must be rejected. Next, in what surely is a first between two Fortune 1000 companies, the court held that allowing Intel to invoke the termination clause in the parties’ executed NDAs would be unconscionable. A provision designed to protect confidential trade secrets cannot be unconscionable. And even if it could be, such a finding does not support the ordered relief. Furthermore, the court’s "reasonable" termination requirement of "perhaps through the year 2000" is clearly unreasonable. Moreover, even striking the NDA termination provision would not oblige Intel to supply additional trade secrets or microprocessors. In issuing its injunction, the court also inappropriately required Intel’s treatment of Intergraph to be the same as Intel’s treatment of "similarly situated" workstation competitors. But if the alleged harm to competition is in the "graphics subsystems" market, there is no nexus between the remedy (targeted at workstations) and the supposed wrong (relating to graphics subsystems competition). In addition, the remedy [begin page 17] does not even protect competition in workstations. The court’s mandatory preliminary injunction merely protects a single competitor. Such mandatory relief is contrary to law because it will not protect competition, will not preserve the district court’s jurisdiction over the merits, and will not benefit the general public. The order itself also offers such inadequate guidance to Intel that it violates the specificity requirements of Fed. R. Civ. P. 65. Finally, the order is flatly at odds with Intergraph’s core patent complaint.
Although that complaint seeks a permanent injunction preventing Intel from selling its
microprocessors, Intergraph has now obtained a preliminary injunction compelling Intel to
continue to sell those very same microprocessors. The irony is that Intergraph portrays
itself as the defender of competition. However, it is merely defending its own interests.
And, Intergraph’s interests are clearly antithetical to competition. If Intergraph
were to prevail on its patent claims and secure an injunction, Intel could not continue to
supply any of Intergraph’s competitors. The resulting chaos would dwarf the effects
Intergraph has alleged in its antitrust claims. The law of the Eleventh Circuit generally governs the propriety of preliminary injunctive relief and the likelihood of success on the merits of Intergraph’s purported antitrust and contract claims. Panduit Corp. v. All States Plastic Mfg. Co., 744 F.2d 1564, 1574-75 (Fed. Cir. 1984); see also Bonner v. City of Prichard, 661 F.2d 1206, 1207 (11th Cir. 1981) (adopting [begin page 18] prior Fifth Circuit holdings as binding precedent). Under this law, the grant of a preliminary injunction, while reviewed under an "abuse of discretion" standard, remains subject to rigorous scrutiny on appeal. Mistakes of law are entitled to "no deference," Cuban Am. Bar Ass’n v. Christopher, 43 F.3d 1412, 1424 (11th Cir. 1995), and mixed questions of law and fact are similarly reviewed de novo. Duke v. Smith, 13 F.3d 388, 392 (11th Cir. 1994). A preliminary injunction is a "drastic remedy," Café 207, Inc. v. St. Johns County, 989 F.2d 1136, 1137 (11th Cir. 1993), "never to be indulged in except in a case clearly demanding it." Buffalo Courier-Express v. Buffalo Evening News, 601 F.2d 48, 59 (2d Cir. 1979). The Eleventh Circuit permits this remedy only if the plaintiff proves all of the following: (1) a substantial likelihood of success on the merits; (2) irreparable injury to plaintiff in the absence of relief; (3) a balance of harms that tips in plaintiff's favor; and (4) a furtherance of the public interest. Warren Publ’g v. Microdos Data Corp., 115 F.3d 1509, 1516 (11th Cir.), cert. denied, 118 S. Ct. 397 (1997). A plaintiff's "failure to sustain [its] burden with regard to any one of the elements will cause the motion to be denied." Devereaux v. Colven, 844 F. Supp. 1508, 1509 (M.D. Fla. 1994) (emphasis added).
The Eleventh Circuit has flatly rejected the "sliding scale" recited by the district court that the "severity of hardship may lessen the showing needed on the merits" (A69). See Snook v. Trust Co. of Georgia Bank, N.A., 909 F.2d 480, 483 n.3 (11th Cir. 1990). Moreover, the district court failed to recognize that Intergraph’s burden is even higher when seeking a mandatory preliminary injunction, particularly in a complicated antitrust [begin page 19] case.1 Such injunctions, which require rather than prohibit affirmative action by the enjoined party, are permitted only "in the rare instances in which the facts and law are clearly in favor of the moving party." Exhibitors Poster Exch. v. National Screen Corp., 441 F.2d 560, 561 (5th Cir. 1971); see Justin Indus. v. Choctaw Sec., L.P., 920 F.2d 262, 268 n.7 (5th Cir. 1990) (movant must show "clear entitlement to the relief under the facts and the law"). In actuality, the district court’s findings and legal conclusions in this case suffer from all four types of error identified in Fraige v. American-National Watermattress Corp., 996 F.2d 295, 297 (Fed. Cir. 1993), any one of which requires reversal: "An abuse of discretion exists when (1) the lower court’s decision is based on an erroneous conclusion of law, or (2) on a clearly erroneous finding of fact; (3) the record contains no evidence on which the lower court rationally could have based its decision; or (4) the court’s decision is clearly unreasonable, arbitrary or fanciful." When the law and record are properly considered, Intergraph has no likelihood of prevailing on the merits of any of its claims. Thus, under any standard of review, the district court’s injunction can only be reversed. [begin page 20]
Although the district court premises its antitrust analysis on an unsupported assumption that Intel is a monopolist in a primary microprocessor market (see pages 42-45, infra) and discusses whether that alleged monopoly was unlawfully acquired or maintained (A40-42), the court’s resulting injunction cannot be supported by that theory. An injunction requiring Intel to supply Intergraph with Intel’s proprietary CPU materials cannot affect Intel’s supposed microprocessor monopoly because Intergraph does not compete in the sale of microprocessors (A14). To justify the ordered relief, the district court needed to find a market in which Intel and Intergraph are alleged rivals. The court named (but did not define) a "graphics subsystems" market (which is not set forth in the amended complaint) and then wrongly concluded that Intel has an affirmative duty to deal with one of its supposed competitors in this undefined "market" (A42).
Contrary to the district court’s apparent assumption, the "antitrust laws do not compel a company to do business with anyone." American Key Corp. v. Cole Nat’l Corp., 762 F.2d 1569, 1578 (11th Cir. 1985). Even a monopolist may deal with whomever it chooses and has no general duty to help its competitors. United States v. Colgate & Co., 250 U.S. 300, 307 (1919); see Aspen Highland Skiing Corp. v. Aspen Ski Co., 472 U.S. 585, 600-01 (1985). A "monopolist can (like the rest of us) be expected to act selfishly, and in some cases consumers will be hurt[;] in the long run they [begin page 21] will be hurt more if juries are allowed to burden a monopolist with a positive duty of assisting competitors."2 A refusal to deal is only actionable if it is undertaken with a purpose to create or maintain a monopoly and there is a dangerous probability that the purpose will be achieved. Lorain Journal Co. v. United States, 342 U.S. 143, 147 (1951). Platitudes about affirmative duties are no substitute for a rigorous analysis of each of the elements necessary to determine whether there is a dangerous probability that Intel will obtain a monopoly. It was clear error for the district court to conclude that Intel had likely violated the Sherman Act without undertaking the proper antitrust analysis (A40-42). Preliminary injunctions must rest on "established theories of antitrust liability" (A42) and require adequate evidence supporting them. [begin page 22] Because there must be a nexus between the wrongful conduct and the ordered relief, Stanley v. University of Southern California, 13 F.3d 1313, 1324 (9th Cir. 1994), the only conduct that is arguably germane to the injunction on appeal is Intel’s alleged refusal to supply Intergraph nonpublic technical information and not-yet released products. And, because Intergraph is in the graphics subsystems business, not the microprocessor business (A14), Intel’s conduct could only be unlawful if undertaken to extend or leverage its alleged power into graphics subsystems. (See A44).
A "leveraging" claim cannot stand unless a defendant uses its "monopoly power in one market to acquire a monopoly in another market." Davis v. Southern Bell Tel. & Tel. Co., 1994-1 Trade Cas. (CCH) ¶ 70,510, at 71,770-71 (S.D. Fla. 1994) (emphasis added); see Ad-Vantage Tel. Dir. Consults. v. GTE Dir. Corp., 849 F.2d 1336, 1347 (11th Cir. 1987); cf. Mr. Furniture Warehouse, Inc. v. Barclays Am./Comm’l Inc., 919 F.2d 1517, 1522 (11th Cir. 1990) ("In the absence of any purpose to create or maintain a monopoly, the [Sherman] act does not restrict the long recognized right of trader or manufacturer engaged in an entirely private business freely to exercise his own independent discretion as to parties with whom he will deal") (quoting United States v. Colgate & Co., 250 U.S. 300, 307 (1919)). Merely describing the alleged upstream monopoly as being an "essential facility" does not change the analysis. Even accepting the district court’s declaration that advanced access to Intel’s own nonpublic, confidential information would constitute an [begin page 23] "essential facility," the theory still requires that the monopolist deny access to the allegedly "essential facility" for the purpose of acquiring another monopoly.3 Here, the record simply does not support such a leap.
Attempted monopolization requires proof "(1) that the defendant has engaged in predatory or anticompetitive conduct with (2) a specific intent to monopolize, and (3) a dangerous probability of achieving monopoly power." Spectrum Sports, Inc. v. McQuillan, 506 U.S. 447, 459 (1993); see Levine v. Central Fla. Med. Affiliates, Inc., 73 F.3d 1538, 1555 (11th Cir. 1996). [begin page 24] A proper analysis cannot begin and end with the district court’s naked observation that Intel "may" be attempting to obtain a monopoly in graphics subsystems (A27). To do the requisite analysis of an attempted monopolization claim, the court needed to define the market, determine each firm’s market share, decide whether there are high entry barriers into that market, and determine whether potential harm to Intergraph alone would allow Intel to monopolize that market. Spectrum Sports, 506 U.S. at 456. The court also needed to find that Intel had a specific intent to monopolize the market. McGahee v. Northern Propane Gas Co., 858 F.2d 1487, 1500 (11th Cir. 1988). The district court did none of this. Given the last-minute pleading of Intergraph's antitrust claim and its concomitant failure to supply sufficient evidence to support every element of that claim, this shortcoming may be understandable, but it is nonetheless fatal to the court’s mandatory injunction.
Proof of the relevant markets is essential to a Section 2 claim.4 Moreover, the relevant markets cannot simply be "named." Instead, they [begin page 25] must be defined by assessing the interchangeability of product use or the cross-elasticity of demand. United States v. E. I. du Pont de Nemours & Co., 351 U.S. 377, 404 (1956). "Stated differently, a ‘market’ is the group of sellers or producers who have the ‘actual or potential ability to deprive each other of significant amounts of business’." Rebel Oil Co. v. Atlantic Richfield Co., 51 F.3d 1421, 1434 (9th Cir. 1995) (citation omitted); see Levine, 72 F.3d at 1552. While the district court declared that Intel and Intergraph were direct competitors (A26), it never addressed or analyzed whether there was meaningful competition between them in an antitrust sense. That is, could Intel and Intergraph deprive each other of significant amounts of graphics subsystems business? If they could not, they are not competitors for antitrust purposes and are not in the same antitrust market. Cf. Ad-Vantage Tel., 849 F.2d at 1342 (other forms of advertising are not acceptable substitutes for Yellow Pages® advertising).
If the district court had properly attempted to define the graphics subsystems market, it would not have found any record evidence on interchangeability to support a finding that Intel’s and Intergraph’s products could deprive each other of significant amounts of business (A110, 138-40). Nothing indicates that Intel has ever taken a graphics subsystems customer from Intergraph or vice versa (A120-34). Notably, Intergraph’s 10K report, filed on March 30, 1998, does not even mention [begin page 26] Intel in a listing of its competitors.5 See A151-54. (Although Intergraph’s 10K report is not part of the record below (it was published after the hearing), it is properly subject to judicial notice. See Fed. R. Evid. 201(b). Hence, neither the record nor the district court’s analysis would support a conclusion that Intel and Intergraph are in the same downstream market. See Allen-Myland, Inc. v. IBM Corp., 33 F.3d 194, 201 (3d Cir. 1994) ("To the extent that the district court’s alleged errors were in formulating or applying legal principles [in defining the market], our review is, of course, plenary."). Conspicuously, the court’s lumping of Intel’s and Intergraph’s graphics subsystems products into one "market," without any consideration of interchangeability in performance, price, or use, is not only erroneous, but it stands in direct contrast to that court’s acknowledgment of the importance of such factors when defining the microprocessor markets (A33). Because Intel cannot monopolize a market in which it has not been shown even to compete, the injunction must be reversed.
Even if Intel is assumed to have a monopoly in a relevant primary market (which, as explained below, it does not), and even if Intel’s refusal to grant preferences to Intergraph were deemed predatory (which it is [begin page 27] not), the district court’s antitrust analysis fails unless there is a supportable finding that Intel dangerously threatens to achieve a monopoly in the so-called "graphics subsystems" secondary market. Once again, there is not. To have "a dangerous probability of monopolizing a market, the defendant must be close to achieving monopoly power." U.S. Anchor Mfg., 7 F.3d at 994. Under the Sherman Act, a market share between 30 and 50 percent is rarely sufficient to threaten a monopoly, and a share below 30% is virtually never sufficient. See Norton Tire Co. v. Tire Kingdom Co., 858 F.2d 1533, 1535 (11th Cir. 1988) (20% share insufficient); ABA Section of Antitrust Law, Antitrust Law Developments 298-99 nn.398-99 (4th ed. 1997) ("ALD 4th"). Here, there are no findings and no record below that Intel has even a one-percent share of a graphics subsystems "market" (A120-34). Moreover, even if Intel itself had a sufficiently large share of the relevant secondary market, any conduct by Intel foreclosing Intergraph from that market still could not materially advance a plan to monopolize. That is because Intergraph is too small for the alleged exclusion to "appear reasonably capable of contributing significantly to creating or maintaining monopoly power." Instructional Sys. Dev. Corp. v. Aetna Cas. & Sur. Co., 817 F.2d 639, 649 (10th Cir. 1987). The analysis is the same for an essential facility. As Professor Areeda observed, "[a] single firm’s facility . . . is ‘essential’ only when it is both critical to the plaintiff’s competitive vitality and the plaintiff is essential to competition in the marketplace." Phillip E. Areeda, Essential Facilities: An Epithet in Need of Limiting Principles, 58 Antitrust L.J. 841, 852 (1990)(emphasis added). Indeed, there is no evidence in this record that Intergraph has as much as [begin page 28] a one percent market share. (Cf. A121 ("There are a lot of other people that also sell graphics subsystems.")). Intergraph is therefore certainly not "essential to competition in the marketplace."
Finally, even if, contrary to the record, both Intel and Intergraph had large shares of the purported graphics subsystems market, Intel could not dangerously threaten to monopolize that market unless entry into that market is difficult.6 As noted, the district court was quite willing to recognize the importance of entry barriers when analyzing primary microprocessor markets (A10, 40), but did not even pay lip service to this prerequisite when analyzing graphics subsystems competition. If the district court’s analysis had been consistent, any consideration of the latter market would have revealed a complete absence of such entry barriers. See, e.g., (A150) ("[T]he market for 3D graphics accelerators is extremely competitive and subject to rapid change. The Company expects competition to increase in the future from existing competitors and from new market entrants."). [begin page 29]
A required element for Intel’s purported antitrust violation in this case is a "specific intent . . . to achieve a monopoly" in the secondary market for graphics subsystems. McGahee, 858 F.2d at 1500. For purposes of that analysis, both objective and subjective intent must be considered; the defendant’s subjective intent cannot simply be ignored. Id. Here, the record provides no basis for finding either an objective or subjective intent to monopolize by Intel. Instead, the evidence shows that Intel’s actions were taken solely in response to a patent dispute initiated by Intergraph (A104 n.3). See also page 29, infra. Indeed, Intel did not engage in the challenged practices until after Intergraph threatened Intel and its customers (Intergraph’s competitors) with patent infringement. There is nothing in the record that suggests that the parties’ patent dispute over the Clipper patents had anything to do with graphics subsystems. Even if the dispute had some relationship to that business, as noted above, Intel does not have a noticeable share (if any at all) of the graphics subsystems business; so it would be improper to assume that Intel intended to embark on a futile scheme to monopolize that market by refusing to deal with Intergraph. Intergraph’s own share of that market is so small that any such efforts would have no effect, whether or not Intergraph stayed in the market. Similarly, the lack of evidence of entry barriers would doom any attempt to monopolize as a matter of law. Moreover, Intergraph took the position, and the district court accepted, that Intel was actually trying to help Intergraph’s competitors in that [begin page 30] market, which is hardly the tactic of a firm attempting to develop its own monopoly (A29-30). According to the district court, Intel had no legitimate business reason for ceasing to provide Intergraph with pre-release CPUs and technical information. In the court’s view, "[t]he patent dispute that arose between the parties could and should be resolved without linking it to the supply of products and information that are essential to Intergraph’s business survival" (A25-26). While that might be preferable in the district court’s view, it is not the law. Indeed, Intel’s reasons are very legitimate, and should have been dispositive as a matter of law. The district court implicitly conceded as much when it stated that it could think of no rational or legitimate reason why Intel would terminate its NDAs with Intergraph "[o]ther than Intel’s position that it does not wish to do business with those who sue it" (A29).
Having failed to make the findings necessary to support a conventional attempted monopoly claim, the district court found instead that Intel’s actions were unlawful because it would thereby gain an unidentified "advantage" in the undefined graphics subsystems market (A45-46). That finding was legal error because it is simply not unlawful for an alleged monopolist in one market to refuse to deal if the alleged [begin page 31] monopolist would gain nothing more than an "advantage" in another market. The advantage theory relied upon by the district court stems from a Second Circuit opinion, Berkey Photo, Inc. v. Eastman Kodak Co., 603 F.2d 263 (2d Cir. 1979). Subsequent to Berkey, however, the Supreme Court admonished that § 2 of the Sherman Act "makes the conduct of a single firm unlawful only when it actually monopolizes or dangerously threatens to do so." Spectrum Sports, 506 U.S. at 459. Because gaining an advantage falls short of threatening a monopoly, the "clear weight of authority has rejected claims under § 2 based upon leveraging to gain competitive advantage." Wojcieszek v. New England Tel. & Tel. Co., 977 F. Supp. 527, 536 (D. Mass. 1997). This infirmity of the "advantage" doctrine was widely established even before the Supreme Court sounded its death knell in Spectrum Sports. In Alaska Airlines v. United Airlines, 948 F.2d 536, 548 (9th Cir. 1991), the court explained that the "anticompetitive dangers that implicate the Sherman Act are not present when a monopolist has a lawful monopoly in one market and uses its power to gain a competitive advantage in a second market." Gaining an advantage "fails to meet the second element necessary to establish a violation of Section 2." Id. There is simply no violation unless "the monopolist uses its power in the first market to acquire and maintain a monopoly in the second market." Id. In Fineman v. Armstrong World Industries, 980 F.2d 171, 205 (3d Cir. 1992), the Third Circuit held that condemning a mere advantage would disregard the statutory scheme of the Sherman Act, which gives [begin page 32] effect to the prevailing view that conspiracies typically present a far greater threat to competition than the acts of a single firm. For that reason, § 1 of the Sherman Act, which governs concerted activity, is violated when a conspiracy threatens an unreasonable "restraint of trade," while § 2, governing unilateral conduct, requires monopolization or an attempt to monopolize. Id. at 205 (citing Copperweld Corp. v. Independence Tube Corp., 467 U.S. 752, 774 (1984)); see Twin Labs., Inc. v. Weider Health & Fitness, 900 F.2d 566, 570 (2d Cir. 1990) (observing that Berkey’s "advantage" theory is mere dictum).
With the exception of the district court in this case, all district court cases within the Eleventh Circuit have also rejected the advantage theory in reliance on the Supreme Court’s ruling in Spectrum Sports. See Aquatherm Indus. v. Florida Power & Light Co., 971 F. Supp. 1419, 1432 (M.D. Fla. 1997) (citing Key Enterprises7 for proposition that a "leveraging" claim requires proof of an attempt to monopolize); Davis, 1994-1 Trade Cas. ¶ 70,510, at 71,770-71. Thus, the district court’s implicit rejection of Spectrum Sports in favor of the discredited view in Berkey is legal error, and should not be accepted.
The Berkey court required that "a ‘substantial’ amount of competition [must be] foreclosed" to constitute an "advantage," 603 F.2d [begin page 33] at 276, suggesting that the court there was using a § 1 rather than § 2 standard to determine whether the monopolist had gained an unlawful advantage. Even under Section 1, conduct that harms competitors is not unlawful unless it also harms competition. Brunswick Corp. v. Pueblo Bowl-O-Matic, Inc., 429 U.S. 477, 488 (1977) (the antitrust laws "were enacted for the protection of competition, not competitors") (quoting Brown Shoe Co. v. United States, 370 U.S. 294, 320 (1962) (emphasis in original); Cargill, Inc. v. Monfort of Colo., Inc., 479 U.S. 104, 116-17 (1986) (antitrust laws protect "competition, not competitors," and "[i]t is in the interest of competition to permit dominant firms to engage in vigorous competition); Spectrum Sports, Inc., 506 U.S. at 458 (antitrust laws prohibit "conduct which unfairly tends to destroy competition itself"); Mr. Furniture Warehouse, Inc., 919 F.2d at 1522 ("to constitute a violation the monopolist's activities must tend to cause harm to competition; unrelated harm to an individual competitor or consumer is not sufficient"); Products Liab. Ins. Agency, Inc. v. Crum & Forster Ins. Cos., 682 F.2d 660, 663-64 (7th Cir.1982) ("[T]here is a sense in which eliminating even a single competitor reduces competition. But it is not the sense that is relevant in deciding whether the antitrust laws have been violated. . . The consumer does not care how many sellers of a particular good or service there are; he cares only that there be enough to assure him a competitive price and quality."). Using a position in one market to gain an advantage in another market will not harm competition unless "a significant fraction of buyers or sellers are frozen out of a market." Jefferson Parish Hosp. Dist. No. 2 v. Hyde, 466 U.S. 2, 45 (1984) (foreclosure of 30% of the market is clearly [begin page 34] not sufficient) (Brennan, J., concurring); accord Amey, Inc. v. Gulf Abstract & Title, Inc., 758 F.2d 1486, 1503-04 (11th Cir. 1985). As explained by Herbert Hovencamp, Federal Antitrust Policy: The Law of Competition and Its Practice 266 (1994): "If a monopoly manufacturer sells to 50 retailers, and then arbitrarily cuts one of them off, the retail market will remain competitive. There is no plausible way that such a refusal can result in lower output or higher prices." The court did not, and could not, explain how Intel, who may have at most a de minimis share of the graphics subsystems "market," could harm competition or gain an advantage by refusing to deal with a single de minimis competitor in a crowded market. One need only consider the impact to competition of Intergraph exiting entirely. For example, if Intel and Intergraph each had a one-percent share of the business when Intergraph left the market, and all remaining competitors proportionally shared in Intergraph’s business, Intel would gain only 1/100th of a percent of the market. As a matter of law, such a small gain in market share could not harm competition and could not therefore gain a firm the requisite advantage required by the Berkey court. Yet, even that small increase may overstate Intel’s potential gain. There is no finding or evidence of interchangeability between Intel’s and Intergraph’s graphics subsystems products that would suggest that they even compete in the same "market." A purported monopolist, even one that controls an allegedly essential facility, cannot violate the antitrust laws by injuring competitors, or even competition, in a market in which it does not compete. See, e.g., Official Airlines Guides, Inc. v. FTC, 630 [begin page 35] F.2d 920 (2d Cir. 1980); Ad-Vantage Tel., 849 F.2d at 1348; Ferguson, 848 F.2d at 983; Interface Group, 816 F.2d at 12. The lower court’s passing remark that Intel leveraged monopoly power into the market for "workstations" (A44) further highlights the court’s failure to apply the proper standard in evaluating the impact of Intel’s conduct. There is no evidence that Intel competes in any "workstation market." As the above cases make clear, even a monopolist cannot gain an advantage in, let alone monopolize, a second market in which it does not compete. The court clearly erred when it failed to apply the appropriate standard to graphics subsystems, as well as workstations.
Even if it is assumed, contrary to everything in the record, that Intel’s decision to lower Intergraph’s "preferred" customer status injured competition in a relevant market, such conduct would still not violate the Sherman Act. Intellectual property holders, including Intel, have an absolute statutory right to exclude all others from making, using, or selling patented inventions (35 U.S.C. §§ 154, 271(a) (1994)); the exclusive right to reproduce copyrighted works, make derivative works, distribute copies of copyrighted works (17 U.S.C. §§ 106(1)-(3) (1994)); and the right to control trade secrets (cf. 18 U.S.C. § 1832 (1994)). The right to exclude others is "the essence" of the patent grant, Dawson Chemical Co. v. Rohm & Haas Co., 448 U.S. 176, 215 (1980), and gives a patent owner a legal "right to refuse to sell . . . [its] patented [begin page 36] products." Ethyl Gasoline Corp. v. United States, 309 U.S. 436, 457 (1940). Likewise, a copyright owner "may refrain from vending or licensing and content himself with simply exercising the right to exclude others from using his property." Fox Film Corp. v. Doyal, 286 U.S. 123, 127 (1932); Stewart v. Abend, 495 U.S. 207, 229 (1990).
Virtually every court of appeals has concluded that an owner "cannot be held liable under Section 2 . . . by refusing to license the patent to others." Miller Insituform, Inc. v. Insituform of N. Am., 830 F.2d 606, 609 (6th Cir. 1987).8 Given the "historic kinship between patent law and copyright law," Sony Corp. of Am. v. Universal City Studios, 464 U.S. 417, 439 (1984), a unilateral refusal to license a copyright similarly cannot violate the antitrust laws.9 [begin page 37] In 1988, Congress further protected this right of exclusion by amending the federal patent statute to provide that "[n]o patent owner . . . shall be . . . deemed guilty of misuse or illegal extension of the patent right by reason of his having . . . refused to license or use any rights to the patent." 35 U.S.C. § 271(d)(4) (1994) (emphasis added). This amendment "illustrates that a patent holder's unilateral refusal to deal cannot constitute unlawful leveraging of monopoly power." In re Independent Serv. Orgs. Antitrust Litig., 989 F. Supp. 1131, 1136 (D. Kan. 1997). According to Data General Corp. v. Grumman Systems Support Corp., 36 F.3d 1147, 1187 (1st Cir. 1994), this amendment may "herald the prohibition of all antitrust claims . . . premised on a refusal to license a patent." To be sure, an intellectual property owner remains subject to the antitrust laws for conduct that is outside the scope of the patent and copyright laws, such as for tying a patented product to an nonpatented one, or for a price-fixing conspiracy involving copyrighted works. See United States v. Line Mat’l Co., 333 U.S. 287, 308 (1948). However, the existence of Intel’s intellectual property ends the "refusal to deal" analysis with respect to those protected products and thus should have ended the district court's analysis here. [begin page 38] Overlooking the long line of cases absolutely protecting an intellectual property owners’ unilateral right to deny access or rights to that intellectual property, the district court adopted a distorted view of Image Technical Services v. Eastman Kodak Co., 125 F.3d 1195, 1215-18 (9th Cir. 1997), cert. denied, 66 U.S.L.W. 3704 (1998), a lone contrary case. In doing so, the district court ignored Image Technical's acknowledgment that there is "no reported case in which a court has imposed antitrust liability for a unilateral refusal to sell or license a patent or copyright," 125 F.3d at 1216, and its holding that a monopolist’s "desire to exclude others from its [protected] [patented] work" is a "presumptively valid business justification for any immediate harm to consumers." Id. at 1218. According to the Image Technical court, that presumption can be rebutted only by showing that the asserted justification was a pretext hiding a subjective attempt to monopolize. Id. It was rebutted in that case based on the defendant's testimony that "patents did not cross [his] mind" and the defendant’s refusal to deal in both goods covered by intellectual property rights and those not covered by intellectual property rights. Id. at 1219. The Image Technical presumption is based on a footnote in Eastman Kodak Co. v. Image Technical Services, Inc., 504 U.S. 451, 480 n.29 (1992), which stated that "power gained through some natural and legal advantage such as a patent, copyright, or business acumen can give rise to liability if a seller exploits his dominant position in one market to expand his empire into the next" (internal quotation omitted). However, [begin page 39] the Court in Kodak was merely rejecting a broad immunity from tying claims. See id. As later noted in Data General, 36 F.3d at 1185 n.63, the Kodak footnote on which Image Technical relied only means that "concerted and contractual behavior that threatens competition is not immune from antitrust inquiry simply because it involves the exercise of copyright privileges." The Ninth Circuit in Image Technical also purported to rely on the First Circuit’s decision in Data General for the proposition that any right under the patent and copyright laws to withhold protected materials is "rebuttable" with evidence showing the patent to be a pretext hiding a subjective intent to monopolize. However, this conclusion squarely conflicts with Data General’s holding that unilateral refusals to license patented inventions never violate the antitrust laws. 36 F.3d at 1186. Not surprisingly, the only later federal case to consider Image Technical flatly rejected it. In re Independent Serv. Orgs. Antitrust Litig., 989 F. Supp. at 1134 ("We decline to follow the Ninth Circuit’s holding" and "find that where a patent or copyright has been lawfully acquired, subsequent conduct permissible under the patent or copyright laws cannot give rise to any liability under the antitrust laws.").
In short, no court before or since Image Technical has concluded that the lawfulness of an intellectual property owner’s exercise of its statutory rights turns on its subjective motivation. In an analogous context, the Supreme Court has held that intent is entirely irrelevant to an [begin page 40] exercise of intellectual property rights.10 There is good reason for the Court's position: it is unlikely fact finders could distinguish an unlawful intent to exclude in order to monopolize a market from a lawful intent to exclude in order to reap the rewards of a lawfully-issued patent. Even under Image Technical’s isolated view, the district court erred in failing to afford Intel the benefit of the presumption that even a monopolist may refuse to provide its intellectual property unless such refusal is not genuinely based on intellectual property considerations and is simply a pretext for creating a monopoly. 125 F.2d at 1219. Instead, the court turned an allegedly rebuttable presumption on its head into a virtual per se rule of liability by declaring that "Intel has no legitimate intellectual property basis with which it can refuse to supply . . . Intergraph" (A48). In this case, Intergraph’s complaint, its testimony and its brief all concede that Intel’s refusal "aris[es] solely out of the patent dispute" (A104 at n.3). Nor does the evidence suggest any intent to monopolize Intergraph’s graphics subsystems market. And, if Intergraph’s allegations are any guide, Intel has been helping, not trying to exclude, Intergraph’s competitors -- not the sort of conduct one would expect from a would-be monopolist, see page __, supra and __, infra. Clearly such conduct is not a pretext for a scheme to monopolize. [begin page 41]
Even a "monopolist is permitted, and indeed encouraged, by Section 2 to compete aggressively on the merits" and "any success that it may achieve through the process of invention and innovation is clearly tolerated by the antitrust laws." Berkey, 603 F.2d at 281. Thus, it is also well established that Intel may "keep its innovations secret from its rivals as long as it wishes" and has no obligation to pre-disclose them in the tangible form of a pre-release microprocessor or in written form. Id.; see also California Computer Prods. v. IBM, 613 F.2d 727, 744 (9th Cir. 1979) (IBM "was under no duty to help [plaintiff] or other peripheral equipment manufacturers survive or expand."). The Supreme Court has particularly emphasized the importance of trade secret protection "at the developmental stage, before a product has been marketed and the threat of reverse engineering becomes real." Bonito Boats, Inc. v. Thunder Craft Boats, Inc., 489 U.S. 141, 161 (1989) ("During this period, patentability will often be . . . uncertain" and "the protection offered by trade secret law may ‘dovetail’ with the incentives created by the federal patent monopoly"). By not prereleasing samples and confidential information, Intel did no more than withhold trade secrets from Intergraph during the development stage of its microprocessors before these products were available for purchase by imitators that reverse engineer Intel’s products. Under the law, Intel was well within its rights to do so. [begin page 42]
The district court also did not adequately support a conclusion that Intel even had a primary monopoly to leverage.11 In that regard, the court "found" that Intel has a monopoly in two purported "markets": (1) "high-performance microprocessors," seemingly defined as the "CPUs or microprocessors that are supplied to high-end workstations," and (2) "the separate relevant market for Intel CPUs" (A38). As shown below, both determinations constitute reversible error because: (a) the record does not support the court’s finding as to Intel’s market share in a "high-performance microprocessor" market; (b) the market share found for the "high-performance microprocessor" market does not satisfy the threshold for monopolization; and (c) the court did not adequately support its asserted "Intel CPU market." The court noted in passing that Intel has about a 60% share of microprocessors used in workstations (A10; see also A15 n.26 ("ratio of about two to one")). If intended to be a market share finding (A38), it was clearly unsupported. The court’s sole citation was to the vague testimony of Intergraph’s Wade Patterson, who at first thought he recalled reading a [begin page 43] publication (not in the record) reporting that Intel was "going to be in the one and a half million kind of range this year," out of a total 2.5 million chips (which would be a 60% share) (A118-19), but then acknowledged having ignored 200,000 chips made by Digital (which would reduce Intel’s share to 56%). However, the district court may have incorrectly disregarded Digital’s share, as its order appears to have inappropriately taken judicial notice that Intel was allegedly acquiring Digital’s Alpha microprocessor (A8 n.16). In actuality, that simply is not true. Intel has absolutely no ownership interest in the Alpha (A210). A few minutes later, Mr. Patterson had second thoughts, testifying "I’m pretty sure that there are more Intel workstations" than workstations without Intel CPUs, and finally conceding, "I don't know the relative breakdown between the two" (A135). Such vague and shifting speculation cannot serve as a basis for any legitimate market share finding. Even a properly based finding showing a 50%-60% share would not, without more, support a finding of monopoly power. Colorado Interstate Gas Co. v. Natural Gas Pipeline Co., 885 F.2d 683, 594 n.18 (10th Cir. 1989) ("lower courts generally require a minimum market share of between 70% and 80%."); Holleb & Co. v. Produce Term. Cold Storage Co., 532 F.2d 29, 33 (7th Cir. 1976) (60% insufficient).
Equally flawed is the district court’s alternative finding that Intel has a monopoly of "Intel CPUs." Absent exceptional conditions, "one brand in a market of competing brands cannot constitute a relevant product market." Domed Stadium Hotel v. Holiday Inns, 732 F.2d 480, 488 (5th Cir. [begin page 44] 1984).12 Nevertheless, the court found an "Intel-only" market principally because, once Intergraph elected to use Intel’s microprocessors, Intergraph allegedly became "’locked-in’ to Intel's microprocessor technology and cannot feasibly switch to other microprocessors" (A39). However, this finding (even if technologically correct, which it is not) cannot support the existence of an Intel-only market because it fails to analyze the "pre-selection" competition between Intel and its competitors.Although the Supreme Court in Kodak suggested the possibility of a market of "locked-in" customers, it held that such a market is possible only in the absence of adequate pre-selection competition. 504 U.S. at 475-76. Under the clear law of the Eleventh Circuit, being "locked-in" is not enough by itself to define a market; instead, the choices available to customers that are not "locked-in" must also be evaluated. In rejecting a government contention that the lock-in of some customers supported a narrow product market definition, the Eleventh Circuit explained: "The [begin page 45] Government failed to consider competition in the pre-formulation . . . industrial market -- competition before [defendants’ product] has been selected as an ingredient." United States v. Engelhard Corp., 126 F.3d 1302, 1306 (11th Cir. 1997).13 The bottom line is this: the district court’s failure to assess the pre-selection competition between Intel and others requires rejection of its "locked-in" finding because it does not satisfy the Eleventh Circuit’s Engelhard test. Consequently, the court could not properly conclude that Intel had a "monopoly" in any primary market that it could leverage. Nor could the court properly conclude that Intel CPUs were an essential facility because a prerequisite to such a claim is the "control of the essential facility by a monopolist." MCI, 708 F.2d at 1133. Furthermore, the essential facilities doctrine only condemns the extension of a monopoly from one market to another. See id. at 1132; Interface Group, 816 F.2d at 12. Thus, the district court’s entire leveraging analysis is infected with error: there is no upstream monopoly and no threat of obtaining one downstream. [begin page 46] Even if there was any legal basis for a likelihood of success on the merits of the refusal-to-deal claim, it could not support the broad injunction issued. There must be a clear nexus between the relief granted and the anticompetitive effect arising from the unlawful conduct. See page __, infra; Stenberg v. Cheker Oil Co., 573 F.2d 921, 924 (6th Cir. 1978) ("[W]e are required to examine the terms of the injunction carefully to determine that the remedy fits the wrong which the district court found."). There is no such nexus here. The wrong is allegedly Intel’s attempt to monopolize a graphics subsystems market. Yet, the court’s discussion of harm centers on Intergraph’s workstation business, not graphics subsystems, and its order is impermissibly structured to remedy the perceived harm to Intergraph’s workstation business, not to competition in the graphics subsystems market (A23, 35-36). Finally, the court’s order prohibits Intel from treating Intergraph differently from "similarly situated competitors" (A75-79), but made no findings that Intel provides advanced chip samples or prediscloses trade secrets to any of Intergraph’s graphics subsystems competitors. That is reversible error.
Because the district court’s injunction directs Intel to deal with Intergraph, the antitrust claims that do not involve a refusal to deal can [begin page 47] offer no basis for the relief granted.14 For example, Intergraph's claim of conspiracy involves no more than Intel allegedly conspiring with Intergraph’s competitors to take away Intergraph’s customers for graphics subsystems by making joint sales presentations (A29, A50). The court did not find, and Intergraph did not allege, that Intel’s alleged refusal to deal is part of that alleged conspiracy and the order does not prohibit the conspiracy alleged. In any event, the agreements described in the district court’s memorandum are simply not restraints of trade. Sales presentations bolstering customers’ products enhance competition; they do not restrain it. Courts have "never doubted that the restraint of which the Sherman Act speaks is a restraint on competition." Phillip E. Areeda, Antitrust Law ¶ 1502, at 370 (1986). Nowhere does the court explain how such dealings injure competition generally as opposed to injuring Intergraph alone, a fatal deficiency for a Section One claim. See, e.g., American Key Corp., 762 F.2d at 1579; Burdett Sound, Inc. v. Altec Corp., 515 F.2d 1245, 1249 (5th Cir. 1975) ("[I]t is simply not an antitrust violation for a manufacturer to contract with a new distributor, and . . . terminate his relationship with a former distributor."). The alleged conspiracies are, in any event, completely inconsistent with Intergraph’s claim that Intel is trying to monopolize the graphics subsystems market in which Intergraph competes. If that were the case, [begin page 48] Intel would not have attempted to convince customers "to buy from Intergraph’s competitors" (A29). See TV Communications Network, 964 F.2d at 1026-27 (Conspiracy allegation failed to state a claim where defendants "would have no rational motive" to conspire.). If anything, Intergraph’s assertions reveal that Intel does not compete in the secondary market. Similarly, the coercive reciprocity theory identified by the district court (A46), but found nowhere in the complaint, cannot justify the relief because the order does not purport to prohibit that conduct either. Nor does the conduct amount to coercive reciprocity. Such a claim requires that a supplier and customer exclude others from a particular market by virtue of a reciprocal agreement to purchase from each other. Betaseed, Inc. v. U & I Inc., 681 F.2d 1203, 1216 (9th Cir. 1982). Here, there is simply no allegation that an Intel-Intergraph agreement is preventing others from supplying Intergraph. See American Key Corp., 762 F.2d at 1579 n.8 ("conspiracy is an essential element of all Sherman Act Section 1 violations.").
There is no written or oral agreement that requires Intel to provide Intergraph with its advance-sample and pre-released microprocessors or other proprietary information. Indeed, the only relevant agreements, the NDAs, expressly provide that neither party is obligated to disclose any confidential information to the other (A14 n.24; A16-17 & n.29). Each NDA disavows any "obligation to buy or sell products" and allows both [begin page 49] parties to "terminate this Agreement at any time without cause upon notice to the other party" and to "cease giving Confidential Information to the other party without liability or request in writing the return of Confidential Information previously disclosed" (A17 n.30, A16-17 n.29). Circumventing the express language of the parties’ NDAs, the district court construed a one-page letter (A20) from an Intel field sales engineer as an enforceable contract, which bound the parties on a non-terminable basis for an allegedly definite period and entitled Intergraph to obtain, inter alia, advance-sample and pre-released microprocessors and other proprietary information from Intel. That conclusion, transforming a non-obligatory, terminable-at-will business relationship into a non-terminable, multi-year contract worth millions to Intergraph and very little to Intel (A54), is reversible legal error. Under Alabama law, the elements that create a contract include offer and acceptance, consideration, and mutual assent. Fant v. Champion Aviation, Inc., 689 So. 2d 32, 37 (Ala. 1997). Where, as here, extrinsic evidence is lacking, the court must "look to the terms of the agreement to determine the intent of the parties." Sharer v. Creative Leasing, 612 So. 2d 1191, 1194 (Ala. 1993). "It is not the province of the court to make . . . a contract for the parties." Muscle Shoals Aviation, Inc. v. Muscle Shoals Airport Auth., 508 So. 2d 225, 228 (Ala. 1987). Even a cursory review of the one-page letter reveals that the parties never manifested the requisite intent to create a binding [begin page 50] agreement superceding the express terms of the NDAs. Intergraph, for example, never manifested any intent to be bound by the letter, and the court’s opinion is barren of any finding that Intergraph accepted Intel’s purported "offer." That omission by itself is fatal to the court’s legal conclusion that the letter can be construed as a binding contract. The letter also lacks valid consideration – i.e., "an act, a forbearance, a detriment, or a destruction of a legal right, or a return promise, bargained for and given in exchange for the promise." Smoyer v. Birmingham Area Chamber of Commerce, 517 So. 2d 585, 587 (Ala. 1987). The district court points to "‘continued support’ (as well as payment for Intel Products)" (A54) as adequate consideration from Intergraph, but the letter merely indicates an "appreciat[ion for] Intergraph’s continued support" and does not ask for any reciprocal payment (A20 n.35).
The letter is also missing sufficiently definite terms to be given effect as a contract, failing to include such basic terms as price, quantity, timing of sale or shipment, or duration. See Alabama Nat’l Life Ins. Co. v. National Union Life Ins. Co., 151 So. 2d 762, 766 (Ala. 1963). Recognizing such defects, the district court turned to the Alabama UCC’s "gap-filler" provisions (A54 n.45), but did so improperly because there still must be an "inten[t] to make a contract." Alabama Code § 7-2-204(3) (1997); see Port City Constr. Co. v. Henderson, 266 So. 2d 896, 899 (Ala. Civ. App. 1972). Here, the many gaps in the one-page letter show that [begin page 51] the parties never intended for the letter to operate as an enforceable agreement.15 The district court ignored the only express "term" in the letter, i.e., "multiple programs efforts," that "are currently being managed under Non-Disclosure [Agreements]." (A20 n.35 (emphasis added)). Disregarding this clear language, the court held that the reference to multiple programs somehow created a contract of "determinable duration, one that is not terminable at will" (A53). Recognizing that its interpretation was completely contrary to the letter's express reference to the NDA, which has an at-will termination provision and disavows any supply obligation, the district court refused to enforce this part of the letter because doing so would render "illusory" the contract already construed to exist (A54). Under the court’s own logic, if the court had taken the only express provision in the letter into account before deciding whether it constituted a contract, the court would have been forced to conclude that the parties never intended the letter to be a contract because the parties' express intent would render the letter "illusory" and "meaningless." Only by ignoring the parties’ expressed intent until after concluding that the letter [begin page 52] was a contract could the district court then turn and throw out the "illusory" and "meaningless" reference to the NDAs (A54). Simply put, creating a contract by such tortured logic and circular reasoning is legal error.
On its face, the letter expressly incorporates the unambiguous terms of the NDAs, which therefore "must be enforced as written." P & S Bus. v. South Cent. Bell Tel. Co., 466 So. 2d 928, 931 (Ala. 1985).16 In addition to contravening the plain language of the letter, the court also ignored that "[w]here general propositions in a contract are qualified by the specific provisions, the rule of construction is that the specific provisions in the agreement control." Paladino, 134 F.3d at 1058 (citation omitted). The letter’s general reference to multiple programs cannot, as a matter of law, override the letter’s specific provision that Intergraph’s involvement with those programs would be "managed under" and thus subject to the NDAs. [begin page 53] Finally, nothing in the letter evidences the district court’s suggestion that there may have been fraudulent conduct by Intel (A66-67). There is no finding or evidence that, at the time the letter was sent, Intel did not have an intention to honor it or that the letter contained any misrepresentations of fact on which Intergraph could reasonably rely.17 The district court’s injunction is based on neither the language of the letter nor any of the court’s own "gap-filling" findings. Instead, the court orders specific performance "in the same manner and the same terms as is done by Intergraph’s similarly situated Competitors" (A76). Even where the court does allude to the "contract," it improperly imposes obligations upon Intel based upon its relationships with others, not Intergraph. (See, e.g., A79 ("Intel shall include Intergraph as an ‘active member of the Intel Inside program,’ and provide it all rights, privileges and opportunities made available to all other members")). Without any findings as to how the letter itself offers and grants Intergraph the same "right, privileges and opportunities" afforded Intergraph’s similarly situated competitors, the district court clearly abused its discretion in ordering preliminary specific performance. The district court’s failure to find specific contractual provisions supporting the order, through gap-fillers or otherwise, highlights another [begin page 54] substantial failure of the mandatory relief ordered. Because a mandatory preliminary injunction cannot issue unless the rights of the parties are entirely clear (see cases cited at page14, supra), the court's failure to identify each party’s contractual rights should have prevented granting this extraordinary, mandatory relief.
A contract provision is not unconscionable at the time that the contract was executed unless (1) meaningful choices were unavailable; (2) the term unreasonably favors one party; (3) there existed unequal bargaining power; and (4) the term is patently unfair or oppressive. See Ala. Code § 7-2-302 (1997); Layne v. Garner, 612 So.2d 404, 408 (Ala. 1992); Wilson v. World Omni Leasing, 540 So. 2d 713, 717 (Ala. 1989). "[R]escission of a contract for unconscionability is an extraordinary remedy usually reserved for the protection of the unsophisticated and the uneducated," Wilson, 540 So. 2d at 716, and was certainly not intended for Fortune 1000 companies like Intel and Intergraph.18 [begin page 55] Intergraph was not denied any "meaningful choice" in its decision to use Intel microprocessors. Until 1993, "Intergraph used the Clipper microprocessor in all its workstations," thereby "lessening its own reliance on outside sources for its supply of microprocessors" (A13). Not only was Intergraph not "locked-in" to using Intel CPUs at that time, but Intergraph switched with full knowledge that Intel (1) had not committed "to provide [to Intergraph] a perpetual supply of chips, pre-released chips, or confidential information" and (2) had not committed to any "continued or ‘perpetual business relationship’ with Intergraph" (A14 n.24). See Corenswet, Inc., 594 F.2d at 134 (clause permitting a unilateral termination of distributorship without cause was not unconscionable where party had meaningful choice). When Intel and Intergraph executed their first NDA in 1994, they were memorializing the arrangement both parties understood to exist in 1993. That cannot be unconscionable. The NDAs did not require Intergraph to make any further commitment or otherwise deprive Intergraph of any "meaningful choice"; it simply gave Intergraph means by which to receive Intel’s trade secrets to which Intergraph was not otherwise entitled at the time Intergraph began to use Intel CPUs. See In re Pennsylvania Tire Co., 26 B.R. 663, 671 (Bankr. N.D. Ohio 1982) (termination without cause or prior notice is commercially reasonable where terminated party would have undertaken the same conduct in the absence of any agreement). [begin page 56] Where an at-will termination provision is available to both parties, such as in the NDAs, it is neither unreasonably favorable nor patently unfair. See Jones Distrib. Co. v. White Consol. Indus., Inc., 943 F. Supp. 1445, 1462 n.9 (N.D. Iowa 1996); Highway Equip. Co. v. Caterpillar, Inc., 908 F.2d 60, 65 (6th Cir. 1990). A termination provision covering the disclosure of proprietary technology and information is enforceable as a matter of law because it bears a reasonable relation to the business risks associated with the sharing of such information. See Blalock Mach. Equip. Co. v. Iowa Mfg. Co., 576 F. Supp. 774, 778-79 (N.D. Ga. 1983); Roberson v. Money Tree, 954 F. Supp. 1519, 1521 (M.D. Ala. 1997). Here, the district court’s "unconscionability" holding impermissibly alters the parties’ risk allocation in contravention of Alabama law. Ala. Code. § 7-2-302 Comment (1) (1997) ("The principle [of unconscionability] is not [one] of disturbance of allocation of risks."); Southland Farms, 575 So. 2d at 1079 (rejecting unconscionability because "[t]his Court has confirmed the validity of risk-shifting provisions in the commercial context"). Striking termination provisions as unconscionable "can only destabilize the institution of contract, increase risk, and make the parties worse off." Industrial Reps. v. CP Clare Corp., 74 F.3d 128, 131-32 (7th Cir. 1996). In recognition of these principles, "courts have uniformly rejected claims of unconscionability of at-will termination clauses." See Jones, 943 F. Supp. at 1462 (citing many cases supporting the majority position upholding at-will termination clauses). The district court erred in trying to force this case within two cases dealing specifically with exculpatory clauses in Yellow Pages® advertising [begin page 57] contracts (A58-59). In Morgan v. South Cent. Bell Tel. Co., 466 So.2d 107, 117 (Ala. 1985), the court applied six "public interest" criteria that are inapplicable here and were not even addressed by the district court. The other, Allen v. Michigan Bell Tel. Co., 171 N.W.2d 689, 692 (Mich. Ct. App. 1969), is based on Michigan law, not the UCC, and has been rejected and severely criticized as contrary to the fundamental principle that "business parties should be able to allocate liability risks – a principle with which ‘[n]early all other cases agree’." J. White and R. Summers, 1 Uniform Commercial Code 241-42 (4th ed. 1995).
For the same reasons, the district court erred in concluding that the NDAs’ bilateral termination provision is unconscionable as applied against Intergraph. See Ala. Code § 7-2-309(3) (1997). For this result, the court relied on Shell Oil Co. v. Marinello, 307 A.2d 598 (N.J. 1973), a "franchise-termination" case never addressed in Alabama. Other jurisdictions with unconscionability doctrines similar to Alabama’s have expressly rejected Marinello or limited its application to franchise relationships.19 Moreover, courts have held that an express bilateral at-will termination provision is not unconscionable, either at the time of its execution or based upon the results of its operation. See Advanced [begin page 58] Plastics Corp. v. White Consol. Indus., 828 F. Supp. 484, 491 (E.D. Mich 1993), aff’d mem., 47 F.3d 1167 (6th Cir. 1995). Thus, even if the UCC stretches from goods to covering an exchange of trade secrets, the district court’s "unconscionability" holding is contrary to Alabama law and the great weight of other authority, and is clear reversible error. Finally, the court's flawed unconscionability findings do not support the affirmative relief ordered. Because the NDAs expressly eschewed any obligation to provide trade secrets, a finding that the NDAs’ termination provisions are unconscionable cannot possibly oblige Intel to supply trade secrets to Intergraph. See Pratt v. Colonial-Sales-Lease Rental, Inc., 799 F. Supp. 1132, 1133-34 (N.D. Ala. 1992) ("[R]elief available [for unconscionable term] is that of judicial intervention disallowing the enforcement of the contract or offending portion." (emphasis added)).
The harm to Intel, Intel’s customers and personal computer users if the injunction is allowed to stand will far outweigh any cognizable harm to Intergraph, if it is reversed. First, the injunction deprives Intel of the control of its intellectual property and threatens similar innovators. Second, Intel, its customers, and users will be harmed by the constraints placed on Intel’s ability to allocate scarce resources to their highest valued use. Third, these parties are also harmed if Intel can no longer pursue a cross-licensing policy, which enhances competition. Fourth, as the courts have recognized in similar circumstances, Intel is threatened with substantial harm if forced to deal with a litigation opponent. Fifth, [begin page 59] Intergraph suffered no legally cognizable harm that would tip the balance the other way. Thus, the balance that must be struck is on the side of preventing harm to Intel, its customers, and the public generally. See United States v. Jefferson County, 720 F.2d 1511, 1519 (11th Cir. 1983). The order deprives Intel of its right as a patent holder to selectively license and sell. In re Independent Serv. Orgs. Antitrust Litig., 989 F. Supp. at 1140-41; see e.g., Genentech, Inc. v. Eli Lilly & Co., 998 F.2d 931, 949 (Fed. Cir. 1993); and cases cited on page __-__, supra. It also deprives Intel of its ability to protect its trade secrets from pre-disclosure. Courts protect these rights to encourage the process of invention and innovation. Depriving Intel of these rights will have the opposite effect. Second, the injunction deprives Intel of its right to use its best business judgment to allocate its intellectual property. See, e.g., In re Independent Serv. Orgs. Antitrust Litig., 989 F. Supp. at 1140-41. In exercising that judgment, Intel considers the needs of its customers and computer users. The order now trumps Intel’s business judgment. So if, for example, Intel must allocate scarce resources, it cannot allocate them to the customers with the greatest needs. Nor is it clear that Intel can exclude Intergraph from marketing events even if Intergraph’s system shows the Intel CPU in a poor light and customers would be led to believe that Intel endorsed the Intergraph system. Before the court issued its order, Intel used its best business judgment to strategically promote its product and the PC platform that has become so important to the American economy. The court order harms Intel, its customers, and the PC users by displacing that strategic judgment with a burdensome prorationing scheme. Previously, Intel [begin page 60] allocated limited supplies of microprocessors according to its specific strategic needs (A107 at ¶ 16). Under the mandatory preliminary injunction, Intel cannot focus its efforts towards customers best positioned to promote Intel-based systems. This irreparably interferes with Intel’s policy of sharing proprietary information and technology in accordance with each customer’s ability to work cooperatively with Intel so that both companies’ strategic goals are met (A142 at n.8). For example, some companies, such as Compaq, Dell and Hewlett Packard have impressive knowledge bases and significant resources, which they apply to the development of Intel-based products. These "cross investments" (A148) justify an enhanced relationship between such companies and Intel. Third, the preliminary injunction order encourages anticompetitive conduct and discourages procompetitive conduct. Intergraph apparently wishes to use its Clipper patents to enjoin Intel from selling microprocessors and to pursue recovery of maximum royalties or damages from Intel and others. In contrast, Intel advocated a cross-license of patents. Intel’s proposal to Intergraph would have furthered competition on the merits because it increases the number of competitors and avoids patent owners charging exorbitant royalties on top of each other. See Town of Concord v. Boston Edison Co., 915 F.2d 17, 24 (1st Cir. 1990) ("Prices that squeeze a ‘second-level’ firm will benefit consumers whenever the ‘second-level firm’ is itself a monopolist.") (Breyer, J.); Fishman v. Estate of Wirtz, 807 F.2d 520, 563 (7th Cir. 1986) (Easterbrook, J., dissenting) ("That successive monopolies injure consumers is a proposition on which there is unanimous agreement.") (collecting authorities). Far from being exclusionary, Intel’s conduct is [begin page 61] inclusionary, helping to further competition among firms and encouraging firms to produce products with the best possible technology and the lowest possible prices, which is the central goal of the antitrust laws. See Northern Pac. Ry. v. United States, 356 U.S. 1, 4 (1958). Fourth, the order’s requirement that Intel entrust its trade secrets to a litigation opponent creates further potential for serious and irreparable harm. See, e.g., Thayer Plymouth Ctr. v. Chrysler Motors Corp., 63 Cal. Rptr. 148, 151 (Ct. App. 1967); Humboldt Oil Co. v. Exxon Co., U.S.A., 695 F.2d 386, 389 (9th Cir. 1982). A customer that sues Intel is simply not "similarly situated" to one that maintains amicable relations. "[T]he relationship between a manufacturer and his customer should be reasonably harmonious; and the bringing of a lawsuit by the customer may provide a sound business reason for the manufacturer to terminate their relations." House of Materials, Inc. v. Simplicity Patterns Co., 298 F.2d 867, 871 (2d Cir. 1962); see also Zoslaw v. MCA Distrib. Corp., 693 F.2d 870, 889 (9th Cir. 1982); Corenswet, Inc., 594 F.2d at 134 n.3. Once Intergraph filed suit, every communication and action between the parties could end up as part of this or other litigation. As the party enjoined, Intel must guard against Intergraph misusing the injunction for extrajudicial discovery. Even with such precautions, how is Intel to know whether Intergraph is asking for information to satisfy a business interest or to obtain a litigation advantage? In light of the high stakes in the ongoing litigation, the injunction will surely require the court to provide close oversight to assure that neither party treats the other unfairly. Reflecting this difficulty, the court warned it might appoint a special master to [begin page 62] supervise the parties’ relationship if they fail to adhere to the order (A75). Courts do not hesitate to overturn preliminary injunctions that require such "constant supervision." Natural Resources Defense Council, Inc. v. United States EPA, 966 F.2d 1292, 1300 (9th Cir. 1992); see Original Great Am. Chocolate Chip Cookie Co. v. River Valley Cookies, Ltd., 970 F.2d 273, 277 (7th Cir. 1992) ("courts should be "reluctant to issue ‘regulatory’ injunctions . . . that constitute the issuing court an ad hoc regulatory agency to supervise the activities of the parties."). Nor should a court "be called upon to weld together two business entities which have shown a propensity for disagreement, friction, and even adverse litigation." Corenswet, Inc., 594 F.2d at 134 n.3. Fifth, Intergraph suffered no legally cognizable harm. There was no showing that Intel’s treatment of Intergraph as a regular, rather than preferred customer, is any different from Intel’s treatment of any other graphics subsystems supplier. Since that is the market where competitive harm is alleged, that is also the market in which Intergraph must show it was injured. It did not do so. Finally, the order is flatly at odds with Intergraph’s core patent complaint and the relief it seeks. Although Intergraph’s complaint seeks a permanent injunction preventing Intel from selling its microprocessors, Intergraph sought and received preliminary relief compelling Intel to do exactly the opposite – to sell microprocessors to Intergraph. Similarly strange circumstances prompted one court to vacate an injunction, holding that when a movant seeks an injunction that "is not only unrelated, but directly contradictory to, the injury for which it seeks redress in the underlying complaint, then a preliminary injunction [begin page 63] simply should not issue." Omega World Travel v. TWA, 111 F.3d 14, 16 (4th Cir. 1997). Accordingly, even if Intel had violated the antitrust laws, it would afford no basis for a mandatory injunction that strips Intel of its copyright and patent rights, forces it to pre-disclose trade secrets (without compensation), compels it to do business with a litigation opponent, puts it under a pervasive regulatory scheme, prevents it from using its business judgment to best promote its product, and threatens it with contempt if it does not do what it cannot understand. In depriving Intel of these rights the district court took one party’s side in a garden-variety patent infringement suit, albeit one with large stakes. Intergraph remains free to withhold its purported intellectual property from Intel without any downside business risk while Intel is compelled to provide its intellectual property without compensation.
Every order granting a preliminary injunction must "be specific in its terms" and "describe in reasonable detail . . . the act or acts sought to be restrained." Fed. R. Civ. P. 65(d). Injunctions must "protect those who are enjoined ‘by informing them of what they are called upon to do or to refrain from doing. . . .’" Hughey v. JMS Dev. Corp., 78 F.3d 1523, 1531 (11th Cir.) (citation omitted), cert. denied, 117 S. Ct. 482 (1996). "[B]asic fairness requires that those enjoined receive explicit notice of precisely what conduct is outlawed." Schmidt v. Lessard, 414 U.S. 473, 476 (1974). The injunction requires Intel to treat Intergraph like "similarly situated competitors" (A75-6), but the court makes no attempt to define [begin page 64] what that means. In fact, the court’s list of allegedly "similarly situated competitors" just muddies the waters. Many of the Intel customers listed as being "similarly situated competitors" (such as Compaq, Dell, Hewlett Packard, and IBM) are large manufacturers that -- unlike Intergraph -- purchase substantial quantities of products from Intel and are "Validation Partners" for new Intel products (A76, A135B). The court orders Intel to treat Intergraph just like these large customers, yet also indicates that Intergraph need not be conferred "Validation Partner" status. Compounding the confusion, it is impossible to define "similarly situated competitors" because Intel has unique relationships with each of its customers. In these discretionary relationships, each customer may get treatment that will be better than other customers in some respects, and worse in others. While the district court stated that Intel need not give Intergraph most-favored-nation status (A74), Intel has no way of discerning whether Intergraph is being treated "no better and no worse" than similarly situated customers (Id.). So how should Intel allocate scarce products? The order says proportionally, but Intel does not use a proportional system to allocate to its other customers (A107 at ¶ 16). How is Intel to interpret the command to provide Intergraph "marketing involvement" and include it in "new product introduction events"? (A79) Must Intel include Intergraph in all events, even if the event was intended for only a single customer? Must Intel include Intergraph if Intergraph’s system shows the Intel CPU in a poor light? The order does not say. According to the order, Intel also must supply Intergraph with trade secrets supplied to "similarly situated" customers (A75-79). More [begin page 65] problems of ambiguity abound. Must Intel divulge trade secrets that Intergraph has no objective need for? Must Intel divulge trade secrets used exclusively by a competitor of Intergraph? Must Intergraph be given requested trade secrets that no other company has asked for? Must Intel communicate in person or in writing? May Intel decline to provide information if, in Intel’s judgment, it is unduly burdensome or expensive to respond? These are just a few of the unresolved questions raised by the court's vague mandatory preliminary injunction. Surely such an order is not consistent with Fed.R.Civ.P. 65(d). [begin page 66] For the foregoing reasons, the mandatory preliminary injunction entered by the district court in favor of Intergraph should be reversed.
|
|
