Wednesday, September 19, 2018

Apple: Qualcomm's 13 German patent assertions part of "grand evil scheme" to force Intel out of wireless chip market

This is my second post on the Qualcomm v. Apple patent infringement trial held by the Mannheim Regional Court yesterday. In the previous post I reported on the alleged (non-)infringement and (in)validity of the patent-in-suit, EP2460270 on a "switch with improved biasing". While the case is too close to call, this patent assertion may fail on the merits just like the first one that went to trial in Mannheim. But the court might also, contrary to what the non-asserted independent claim 16 implies for claim construction purposes and despite a finding by the Swedish patent office that the patent lacks a sufficient inventive step over prior art presented by Apple, hold Apple liable for infringement and decline to stay the case pending a parallel nullity action. In that case, Apple's affirmative defenses--antitrust and licensing--will be outcome-determinative at least with respect to the availability of injunctive relief.

For a long time, it was hard to fend off even standard-essential patent injunctions in Germany on antitrust grounds (with or without a FRAND commitment, which German courts wouldn't deem enforceable by third-party beneficiaries anyway). It was arguably hardest in Presiding Judge Dr. Kircher's court. The situation improved after the Court of Justice of the EU ruling in Huawei v. ZTE; in a way, it already got a little bit better after the European Commission took action against Samsung and Motorola. But very regrettably, the thinking of German patent judges is still, by and large, that antitrust defenses are just part of a throw-in-the-kitchen-sink tactic of infringers.

The patents Qualcomm is asserting in Germany--at least the ones that have been discussed in hearings or trials--aren't standard-essential, which ups the ante for Apple's antitrust defense. However, the fact that Qualcomm's conduct has been deemed anticompetitive by competition enforcers in multiple jurisdictions ("Antitrust Grand Slam").

Judge Dr. Kircher stressed that yesterday's discussion of the antitrust issues in the case was just meant to be a round of opening statements, far from a final decision. However, the court's preliminary inclination is that

  • Qualcomm's conduct does not constitute a violation of Art. 102 of the EU Treaty (the abuse-of-dominant-position paragraph) in its own right,

  • nor does the court tend to believe that its German patent assertions against Apple constitute a violation of the conduct the European Commission fined it for (exclusivity arrangements with Apple depriving Intel of a fair chance to compete on wireless baseband chipsets; the court feels this is just a fundamentally different type of behavior).

The judge said Art. 102 TFEU violations are typically identified where a company abuses a dominant market position in order to cement its monopoly, with cases involving "divergent markets" (the one in which the dominant position is held and the one that is affected) representing a less common scenario. Therefore, he feels an antitrust defense of that kind should succeed only under exceptional circumstances. Then he is concerned about the "subjective" nature of Apple's theory that Qualcomm's (non-standard-essential) patent assertions are meant to force Apple to drop Intel as a baseband chipset supplier. He also pointed to the fact that Qualcomm denied any such intention. Also, he noted that Apple couldn't infringe Qualcomm's patents with impunity even if Qualcomm pursued anticompetitive goals. Apple's counsel--the antitrust part was mostly argued by Freshfields Bruckhaus Deringer's Dr. Frank-Erich Hufnagel--was quick to clarify that Apple wasn't seeking a "free ride," but prepared to take a license on reasonable terms. Apple is seeking the dismissal of the complaint, but on other grounds; its antitrust defense is all about injunctive relief.

First, I don't think "divergent markets" are nearly as much of an exception as the court indicated. Even the famous EU Microsoft cases involved the leveraging of Microsoft's desktop operating system monopoly and its effects on the markets for workgroup servers and media player software products. Second, the overarching objective of competition enforcement, be it merger control or antitrust in a narrow sense, is to preserve competitive restraints, meaning that a company can't just command any price it wants (without having to pay much attention to its competitors' price points) or impose all sorts of other terms. Competition is never perfect as evidenced by the prices Apple can command for its new iPhones, but the difference between those products and high-end Android smartphones is only gradual, while Qualcomm can charge, for the combination of SEP royalties and chipsets, a multiple of what its competitors demand. It's obvious that Intel's deal with Apple goes a long way toward some competitive restraints on Qualcomm's business decisions, so it's not far-fetched to see the intention.

In his opening statement on antitrust, Dr. Hufnagel explained to the court that--and why--the analysis should be holistic, taking into consideration the entirety of Qualcomm's conduct, which he described--using this exact English term--a "grand evil scheme" involving different actions and measures such as exclusivity rebates for customers/licensees, the refusal to license rival chipset makers, and a multiplicity of patent assertions in different jurisdictions, including (but not limited to) its ITC complaints targeting specifically Intel-powered iPhones (earlier this week, the ITC staff raised public-interest concerns over that litigation strategy) and the German cases, which in practical terms target only Intel-powered devices because all of the accused devices are sold with Intel baseband chips in Germany.

Dr. Hufnagel said Qualcomm was, by now, asserting a total of 13 patents against Intel-powered iPhones in Germany. Prior to that statement, I was aware of three patents asserted in Mannheim and seven (four of which are from the same family) in Munich. I'll try to find out about the three remaining assertions--and who knows how many other suits Qualcomm may file in the meantime.

Apple's counsel noted that Qualcomm could have simply sued the maker of the allegedly-infringing chipset (Avago/Broadcom), and said the "subjective"/"objective" distinction wasn't key under antitrust law: the focus should be on anticompetitive effects.

It was a powerful statement, but Judge Dr. Kircher took the (preliminary) position that the court had to focus on the patent-in-suit before it in the particular litigation. Dr. Hufnagel conceded that a single non-standard-essential patent assertion was not going to force Apple into submission and back into the Qualcomm fold all by itself. But the wider "grand evil scheme" would have that potential.

The discussion will continue on October 2 in a trial involving a different Qualcomm patent (EP3036768 on a "layout construction for addressing electromigration"). Without a holistic assessment of Qualcomm's overall conduct, Apple's antitrust defense is likely a lost cause except in a case where a patent couldn't be worked around within a reasonable time frame.

In the present case, it appears that Apple could just solve the problem by having its products for the German market made by Pegatron, one of its contract manufacturers. Pegatron was drawn into the litigation. Apple doesn't usually want to do so, but had to plead it into the case because of a lack of clarity regarding whether the patent-in-suit was covered by Pegatron's license deal with Qualcomm, which is subject to "capture periods" (presumably this means that patents applied for or granted during a certain period fall under a license agreement, a structure I've seen in the public redacted versions of settlement agreements). But Apple's primary contract manufacturer, Foxconn (Hon Hai), may not be licensed. Wistron was als mentioned. I believe Compal, for whatever reason, was not named.

The license-based affirmative defense was apparently raised at a relatively late stage. That discussion, too, may continue when other Qualcomm patents go to trial. But it's not as certain as the continuation of the "grand evil scheme" debate, which is actually raging around the globe.

Share with other professionals via LinkedIn:


Qualcomm switch patent asserted against Apple in Germany should be invalidated, Swedish patent office says

Yesterday's Qualcomm v. Apple trial took twice as long as the average Mannheim patent trial. In fact, the ventilation system was switched off in the late afternoon, so for the last hour, two doors had to be kept open. The courtrooms at the Mannheim Regional Court, Europe's leading venue for wireless patents, are famously windowless.

Presiding Judge Dr. Holger Kircher forthcomingly stated at the outset that this case was, in my words, too close to call (unlike the one that Qualcomm agreed to stay in June), thus the court had to elaborate on all our of Apple's defenses: non-infringement, invalidity (which German district courts don't determine, but they can and often do stay cases pending a parallel nullity or revocation proceeding in another forum), abusive conduct (antitrust), and licensing (through one or more contract manufacturers). I'll address the first two--the traditional defenses to patent infringement--in this post, and the affirmative defenses (the remaining two) in a subsequent post since there's an abundance of interesting things to report and comment on.

The patent-in-suit, EP2460270 on a "switch with improved biasing" ("biasing" in this context basically meaning that one voltage gets to control another), is not standard-essential. Essentiality hasn't been alleged by any party to the German Qualcomm v. Apple cases that have been heard so far. Nor is it related to wireless baseband processors: it's a general circuity patent covering a type of switch. It was mentioned during yesterday's trial that the chip allegedly infringing on the patent is supplied to Apple by Avago/Broadcom. But all of the accused devices come with an Intel baseband chip, a fact that will be relevant to the antitrust part of the next post.

In the combination of Judge Dr. Kircher's summary of the issues in the case and what Qualcomm's German counsel, Quinn Emanuel's Dr. Marcus Grosch, said yesterday, it turned out that Apple originally didn't even deny infringement but focused on invalidity. Then, after Qualcomm's reply in support of its complaint, Apple presented a non-infringement theory in a sur-reply. It wasn't said whether Qualcomm's reply brief or a change of mind triggered this, but normally a defense has to be presented in the defendant's first pleading, so I would attribute this to something Qualcomm said. This much is certain: the non-infringement argument is effectively a disputed construction of a claim limitation, not a factual question concerning the accused products.

As far as I understood, Apple argued that the patent-in-suit covered a certain chipset architecture involving a constant bulk voltage, with ground also being constant. But Qualcomm disagreed, arguing that "AC ground" stood for only the AC part, not DC. In any event, the court appeared less inclined to agree with Apple. But... and this was something very special:

Judge Dr. Kircher, who unlike German criminal judges does not have a duty to "investigate," felt the parties had not focused enough on an inconsistency between claim 1 (an apparatus claim; Qualcomm is asserting one or more apparatus claims) and independent claim 16, a method claim that refers--unlike the apparatus claim(s)--to only "the" (a single) control signal in connection with what turns on the switch, resulting in an input signal being passed through a plurality of transistors. The accused chip, however, has two control signals and would therefore, at least potentially, fall outside the claim scope.

I was profoundly impressed and still am. Apple v. Qualcomm is the commercially biggest patent-related dispute ever (at a minimum, the biggest one in the history of the mobile device industry). This patent is just a limited part of the wider dispute, but still this (unlike what Judge Alsup once mislabeled Oracle v. Google) is truly the World Series of IP. Yesterday one could see sky-high piles of case documents behind the bench. Still, with all that had been written in the build up to the trial, the presiding judge identified an issue that has the potential to be outcome-determinative and had not received much if any, attention.

It's not just because of one demonstration of extraordinary competence and dedication that I couldn't imagine a future Unified Patent Court (UPC) without this judge, despite fundamental disagreements in the past and the present with respect to the bearing that antitrust issues should have on patent infringement cases.

Unsurprisingly, Qualcomm argued that the distinction between one or two control signals was an irrelevant labeling question. I lack the expertise in electrical engineering to form an opinion, but I've watched numerous patent cases in (mostly) two major jurisdictions and this is the typical difference between an accused product and a properly-construed patent claim that routinely becomes a defendant's "get out of jail free" card. In fact, the world is littered with complaints that have been dismissed for lesser reasons.

However, Judge Dr. Kircher said that case law and literature were inconclusive as to whether or not a parallel claim could help a court construe another claim. Apparently, reasonable and reputable people disagree on this legal question because it hasn't been resolved. Better late than never. Here, it's a given that Qualcomm would appeal a stay by "complaining" (that's the term used here) to the Karlsruhe Higher Regional Court.

For the development of German (and, by extension, European) patent case law it would actually be good if the Mannheim court was prepared to take the risk of being overruled on appeal. I have no idea how the appellate panel under Presiding Judge Dr. Andreas Voss ("Voß" in German), who previously presided over another patent infringement chamber in Mannheim, or, in the ultimate consequence, the Federal Court of Justice (both in the city of Karlsruhe) would decide. My guess is that neither would pronounce a bright-line rule declaring it categorically impermissible to get mileage out of something that is found in the patent document itself. Why rely on descriptions, which aren't even updated as the claims are narrowed (which is also a huge issue in this case, with Apple alleging that Qualcomm pointed to a drawing that relates to something that the original application covered but the actually-issued patent does not), but not on claims, which the examiners focus on? Much less on anything found outside the patent document? It would probably the most likely outcome that some weight (within reason) would be given to what can be inferred or deduced from parallel claims, and in a case that is too close to call, it doesn't take too much to tip the scales.

It's fair to stay that no one in the room appeared to have a better idea of the issues in the case and the evidence in the record, across the board, than Judge Dr. Kircher. But in connection with invalidity, something apparently entered the record rather late, and while the court initially showed an inclination not to grant Apple's motion to stay since it's based on lack of an inventive step, while German courts normally grant stays only if there's a strong case for anticipation. The aforementioned Judge Voss ("Voß") always reminded counsel that a stay required a prior art reference constituting a "photographic image" of the patent-in-suit.

That's because German patent infringement courts don't like to take a position on whether a person of ordinary skill in the art would be likely to combine two or more prior art references: they're jurists, not engineers, unlike patent examiners.

Yet the hurdle for getting German patent infringement cases has, fortunately, become lower and lower over the years, as the regional and higher regional courts saw all too many cases in which they issued injunctions and later the patent died or was defanged as its claims got narrowed.

The court agreed with Apple that a new prior art reference (with an Asian name I didn't fully understand, but I'll find out whenever I find the time to request access to the nullity file), coupled with one named "Brindel" or "Brindle" that had been considered by the examiner in connection, represented a relatively stronger case for invalidation than one that the examiner found wasn't likely to become part of a combination. Dr. Grosch's efforts to deny or downplay this didn't persuade the court.

Freshfields Bruckhaus Deringer's Prince Wolrad of Waldeck and Pyrmont argued Apple's invalidity contentions. Toward the end I got the impression that the court was potentially having second thoughts. Just potentially, but that means a lot in a case where the original position was that a stay would not be warranted here.

The potential silver bullet here is that Apple had submitted the original patent and its proposed combination of prior art references (which are, according to Apple, also at the heart of its German nullity complaint pending with the Munich-based Federal Patent Court) to the Swedish Patent & Registration Office ("PRV"), and received an expert opinion from an examiner according to which the patent-in-suit does lack an inventive step in light of the prior art.

I haven't been able to find that opinion on the website of the Swedish patent office yet, but I'll keep trying.

Judge Dr. Kircher asked a key question: how many sucb opinions had Apple requested before obtaining this one? Apple's counsel, under the usual obligation to speak the truth, said it was just this one, and stressed that Apple wanted a neutral opinion based on just the prior art and no interaction with the examiner himself. That's the way it works in Sweden according to Apple. The Austrian and Swiss patent offices were also mentioned as providers of such opinions, but one of them no longer offers that service and the other one lets the petitioner communicate a lot with the examiner, which Apple's lawyers didn't want to do because they have a stronger case based on an uninfluenced opinion.

Apple didn't say so, but I might add that Sweden is a country with a lot of expertise in this field of technology, especially because of Ericsson.

It was again no surprise that Qualcomm's counsel would be dismissive of Apple's position. Dr. Grosch said that it would obviously simplify all patent cases if one could just fill out a web form somewhere. And it's not like the court indicated a change of mind. But I got the impression that Judge Dr. Kircher may indeed take the independent Swedish evaluation seriously. Maybe--just maybe--this does constitute a basis on which the court would make one of those rare exceptions in which a patent infringement case is stayed over inventiveness, not novelty, concerns.

There'll be a formal post-trial brief for which Apple was granted leave, and a legal brief (no new facts) which Qualcomm announced. The court will announce a decision on December 4, 2018. That one could be a final ruling, a stay, or whatever other procedural order.

Share with other professionals via LinkedIn:


Tuesday, September 18, 2018

Industry bodies support FTC's motion to require Qualcomm to license SEPs to rival chipset makers

It's a busy September on the FRAND front...

As I reported on the first of the month, the Federal Trade Commission brought a motion for partial summary judgment that may open up the wireless chipset market--by reminding Qualcomm of its self-imposed obligation to license rival chipset makers--even prior to the big antitrust trial in the Northern District of California.

It's odd that a mere reminder would be a potential game-changer, but that's the way it is because of Qualcomm's refusal to live up to the FRAND promise.

The strategic significance of the motion is underscored by an amicus curiae brief that two industry associations--ACT | The App Association and the Computer & Communications Industry Association (CCIA)--have asked Judge Lucy Koh to allow them to file (this post continues below the document):

18-09-17 ACT App Association & CCIA Amicus Brief by Florian Mueller on Scribd

The later the stage of proceeding, the more common it is for stakeholders to chime in. The summary judgment stage is definitely not where this typically happens. But the FTC's motion is special.

Also, I can't remember having seen a joint filing by those two organizations before.

There's a Mannheim trial today, so for logistical reasons I lack the time to comment on the brief in detail, but I did want to raise awareness. It's a good read, and if you do look at the PDF, please pay attention to what footnote 14 says about Europe (where a major wireless standard-setting organization, ETSI, is based) and Korea; and footnote 16, which quotes public statements by Qualcomm.

Share with other professionals via LinkedIn:


Monday, September 17, 2018

ITC staff raises public-interest concerns over ban of Intel-powered iPhones sought by Qualcomm

This morning, opening statements were delivered at the start of the evidentiary hearing in the investigation of Qualcomm's second ITC complaint (request for U.S. import ban). The complaint was filed last December. An earlier complaint by Qualcomm against Apple is at a more advanced procedural stage: a final initial determination (a preliminary ruling by an Administrative Law Judge, which is however subject to Commission review) was originally due last Friday, but after the ALJ originally in charge retired, Chief ALJ Bullock took over and extended the deadline by two weeks. In that earlier case, the Office of Unfair Import Investigations (OUII, commonly referred to as "the ITC staff") recommended an infringement finding with respect to one patent. Staff recommendations are not binding on ALJs, and even ALJs don't make the final decision: the Commission itself does. But what the staff says is often adopted.

This morning's hearing in the U.S. trade agency with quasi-judicial authority didn't start the way Qualcomm would have liked it to. The staff just presented the following recommendations:

  • None of the three remaining patents-in-suit ('336, '356, '674) is invalid (good for Qualcomm), but none is infringed (a disaster for Qualcomm). The latter would dispose of Qualcomm's case all by itself, so the other findings will be potentially outcome-determinative only if the final determination by the U.S. International Trade Commission's top decision-making body (which in the case I watched has sided with the staff to an even greater extent than the ALJs did) disagreed and identified an infringement. But in light of the other case pending and, even more so, the public interest in fair competition in the wireless chipset market, they're important at any rate:

  • With regard to the '336 patent, Qualcomm failed to satisfy the domestic industry requirement (rather unusual, and another hurdle it would have to overcome in the further proceedings).

  • A U.S. import ban would adversely affect competition in the baseband processor market and may adversely affect the production of alternative baseband processors. In case anyone forgot, Qualcomm's two ITC complaints against Apple don't target all iPhones--but only Intel-powered ones. Actually, Qualcomm sought to avoid issues with the ITC's public-interest considerations by limiting its request for an exclusion order to Intel-powered iPhones. But the alternative would just be Qualcomm-powered iPhones. Given that Intel is the only major competition Qualcomm is now--at long last--facing in the baseband chip market, I felt all along that this litigation tactic didn't alleviate the public-interest concern. Arguably, it exacerbated the situation, also from an antitrust point of view.

As Bloomberg's patent (and especially ITC) expert Sue Decker wrote in a pre-hearing report, "Apple is trying to underplay its dominance in the baseband chip market by focusing on how an exclusion order would supposedly harm Intel" and its "public interest argument would create a dangerous precedent." The ITC staff apparently disagrees with Qualcomm on the allegedly dangerous precedent, and either disagrees or just doesn't care about whether Apple is, as Qualcomm basically claims, opportunistically hiding behind Intel.

All of this is a surprise. Qualcomm has a huge patent portfolio and some of the best IP lawyers in the United States--firms that also have a tremendous track record at the ITC. I wouldn't have expected the ITC staff to clear Apple (and, by extension, Intel) of infringement of all three remaining patents-in-suit. But the ITC staff doesn't have a track record of being very sympathetic to public-interest arguments weighing against import bans. In this case, however, the anticompetitive effects of a hypothetical U.S. import ban have apparently been shown to be fairly serious.

Apple and Intel received far less support from the ITC staff in the other ITC investigation. It's impossible to know from the outside whether the ITC staff has become more open-minded about competition-centric considerations or whether this is attributable to a different body of evidence. But the importance of having Intel compete with Qualcomm in the baseband chipset market has been recognized, and there should be no difference between those cases in terms of the ultimate outcome. That's why today's staff recommendations pose a huge problem for Qualcomm with respect not only to the case heard this week but also to the earlier-filed one.

On the other coast, in the Northern District of California, Judge Lucy H. Koh can now realistically hope that she may never have to adjudicate a consumer motion (which was denied without prejudice, thus could be refiled later) to bar Qualcomm from the enforcement of a potential U.S. import ban against Intel-powered iPhones. The way things are going now validates her wait-and-see approach. And who knows, maybe the ITC staff read her decision and agreed with some of the concerns over Qualcomm's conduct that she raised in last month's order.

Jurisdictional differences exist, but the European Commission, which is preliminarily investigating an Apple complaint over Qualcomm's anti-Intel lawsuits targeting German iPhones, and the Mannheim and Munich courts may also take note of the findings in Washington, DC.

Share with other professionals via LinkedIn:


Coerced FRAND arbitration is Huawei's Trojan Horse in appeal of Samsung's anti-enforcement injunction

Huawei's lawyers have beaten my expectations with their reply brief in the Ninth Circuit antisuit (anti-enforcement) injunction appeal pending before the Federal Circuit. Reply briefs are often largely repetitive reinforcements of points previously stressed, and in this case it's practically already the second appeal as Huawei had (unsuccessfully) asked Judge William H. Orrick of the United States District Court for the Northern District of California to reconsider his decision to bar Huawei from enforcing a couple of Chinese patent injunctions prior to a San Francisco trial. But the Sidley team representing Huawei has filed something that will make the Federal Circuit hearing (date TBD) more interesting than I would have previously thought (this post continues below the document):

18-09-06 Huawei Reply Brief (Antisuit Appeal) by Florian Mueller on Scribd

I still doubt that the argument about whether the Ninth Circuit Gallo test replaces all of the Supreme Court's Winter preliminary-injunction criteria or only the irreparable-harm factor will be outcome-determinative: when upholding the injunction, Judge Orrick addressed the other Winter factors anyway.

With respect to international comity, Huawei is right that antisuit injunctions "virtually always" enjoin private parties, not courts, yet comity is always part of the consideration; in the international context one can even strike "virtually" because no U.S. court holds power over any foreign court. The truth is in the middle: there undoubtedly is a risk of alienating foreign judges if their rulings cannot be enforced, or if parties have to stay or withdraw suits in their courts, because of a U.S. injunction, but the degree of alienation is obviously nowhere near what it would be if a U.S. court based its decision on a principle such as the general primacy of U.S. law over the laws of all other countries.

With respect to China, let's face one thing: every U.S. judge knows that the rule of law, just like democracy, works differently in China. Huawei's lawyers portray the Chinese proceedings at 100% fair and comprehensive. Samsung's counsel obviously didn't suggest that the proceedings were unfair, nor did Judge Orrick say so in his order. The Federal Circuit will be diplomatic, too. But that doesn't mean that the appellate judges won't have their private and unspoken opinion anyway.

The strategic issue here is the one I mentioned in the headline: coerced FRAND rate-setting arbitration. In order to distinguish Huawei v. Samsung from Microsoft v. Motorola (with Samsung being the new Microsoft and Huawei being the new Motorola), Huawei points out that Microsoft said it would accept a court-determined FRAND rate without insisting on adjucation of all the defenses that Samsung is pursuing (though Judge Robart actually did hold some Motorola patents invalid anyway). But beyond differences between Huawei and Microsoft regarding the preconditions for rate-settings (which I simply attribute to the fact that the standards at issue in Microsoft v. Motorola were not nearly as critical to Microsoft's business as the ones in Huawei v. Samsung are to Samsung's core business, thus Microsoft was prepared to pay for invalid and non-infringed patents and exclusively concerned about injunctive relief), Huawei must deny that it is an unwilling licensor (not in the sense of unwillingness to extend a license, but to do so on FRAND terms) engaging in hold-up and instead argue that Samsung is an unwilling licensee engaging in hold-out. The one key issue in that context that comes up in different places in Huawei's appellate reply brief--one might also call it a red herring--is that Huawei seeks to leverage its proposal to have an arbitration panel set binding FRAND rates:

  • Huawei denies any accusations of hold-up by claiming that Samsung could avoid the enforcement of an injunction by simply agreeing to binding arbitration.

  • Huawei portrays Samsung as an unwilling licensee merely because of its reluctance, at this stage, to submit to arbitration.

Huawei points out that Samsung made a similar unwilling-licensee argument against Apple in that (in)famous ITC case that resulted in a presidential veto. That's historically accurate, and no one criticized Samsung (and the ITC majority at the time) more harshly than this blog did. What was wrong back then is still wrong today.

This blog has equally consistently opposed coerced arbitration over the years. There's nothing wrong with parties' consent to arbitration, except that it deprives courts of an opportunity to augment their body of FRAND case law (though they like to see their caseload reduced). What is highly problematic and often just hold-up by any other name is when a party that would rather have key determinations (on essentiality, infringement, validity, and on FRAND rates) made by courts of law faces enormous pressure to arbitrate. Coerced arbitration is a terrible thing and raises more issues than I can address in a single post, but let me highlight some that appear particularly relevant to the Huawei-Samsung situation:

While courts do hand down "sharp rulings" (as Apple called it in a filing five or six years ago), such as the FRAND determinations made by Judges Robart and Holderman that were way closer to zero than to the original royalty demands, arbitration tends to gravitate toward the middle. This means that a patentee can often achieve a supra-FRAND outcome by, which is an oversimplification but essentially correct, demanding twice that supra-FRAND rate. The willing licensee can't do more than argue a next-to-zero royalty (and even that is difficult without losing credibility with the panel). Simply put, the licensee can't counterbalance a multiple of a FRAND royalty by proposing a negative royalty.

In its reply brief, Huawei denies that it's adamant about receiving a 1.5% royalty, but it doesn't specify what else it is seeking, making it impossible from the outside to determine who's right. The Federal Circuit has access to the full record below and will form its opinion.

No less important is the following difference: while there is some room for forum-shopping, courts of law will simply decide a case based on its facts, statutes, and applicable case law, but arbitration is a much more free-form kind of dispute resolution and everything depends on the parameters. At the most, there are some arbitration courts that have certain procedural rules in place. But even on the procedural side the first fundamental disagreement may already arise when the licensee side would like to start with some validity, essentiality, and infringement determinations (such as of an exemplary set of asserted patents). Also, should one arbitration panel make such determinations, including rate-setting, for foreign jurisdictions or just for its local one? It's the parameters, stupid.

Coerced arbitration can disadvantage licensees in two ways. Either there's frightening uncertainty for a lack of parameters (which again means that a patent holder making unreasonably supra-FRAND demands has pretty good chances of getting a supra-FRAND outcome) or the parameters may outright disadvantage the would-be licensee.

If parties voluntarily submit to binding arbitration, after negotiating without undue pressure the procedural and substantive parameters, arbitration can be fair. But with the threat of enforcing an injunction that puts someone temporarily out of business, the parameters most probably won't be fair.

Hopefully the Federal Circuit won't adopt a proposal that would be tantamount to a de facto primacy of arbitration panels over courts of law when it comes to FRAND determinations.

Share with other professionals via LinkedIn:


Qualcomm keeps escalating patent dispute with Apple: new German lawsuit(s) filed

With a view to a Qualcomm v. Apple patent infringement trial in Munich on Thursday I contacted the Munich I Regional Court to check on the time. As for the biggest issue in that case, may I refer you to my recent post on how thin air can "practice" claim limitations unless the name of the game is the claim.

On the same occasion, I inquired about any Qualcomm v. Apple first hearings that might come up in the near term. Unlike other German courts, the Munich court holds a first hearing, not as formally focused on claim construction as a U.S. Markman hearing, prior to patent trials. A spokeswoman for the court kindly informed me that a first hearing in two parallel cases, targeting different Apple entities, has been scheduled for March 28, 2019.

The patent-in-suit in both cases is EP1988602 on a "mobile terminal with a monopole[-]like antenna."

Knowing the usual Munich timelines, it appears that the new complaints have only been filed recently. I guess just before the main summer vacation season here, or at least not long before.

I've also tried to find out from the Mannheim Regional Court (the other German Qualcomm v. Apple venue, where a trial will be held tomorrow in fact). The Mannheim court cannot announce new trial dates yet, and it doesn't hold first hearings. However, Qualcomm previously asserted the same number of patents in either venue, so there might also be a new Mannheim case, just that there's no official confirmation yet.

Qualcomm doesn't make phones, so it doesn't actually want the iPhone banned. Seeking injunctions against Apple is, however, a means to the end of forcing rival chipset makers, particularly Intel, out of the market, a strategic objective that has provoked an EU antitrust complaint by Apple, which the European Commission is investigating on a preliminary basis, and raises serious public-interest issues in the two ITC investigations of Qualcomm's requests for U.S. import bans. An evidentiary hearing (basically, a trial before an Administrative Law Judge) in the second one of those cases commences today, and I'll try hard to find out as soon as possible (last time I was slow) about the recommendations made by the Office of Unfair Import Investigations (OUII, commonly referred to as "ITC staff") and whatever the ALJ may indicate today or later this week.

Share with other professionals via LinkedIn:


Saturday, September 15, 2018

In Munich iPhone case, Qualcomm wants thin air to infringe software patent claim limitations: dangerous precedent possible

The name of the game is the claim. There's no German equivalent, at least none that rhymes, but I vividly remember Quinn Emanuel's Dr. Marcus Grosch stressing this point when he was defending Motorola Mobility against Microsoft, particularly in the Munich appeals court. Now, as counsel for Qualcomm in a German lawsuit against Apple that will be tried on Thursday and involves the iPhone's built-in Spotlight search technology, he's espousing the very opposite position. The former anti-software-patent campaigner in me dreads and hates the notion of software patents being given overbroad scope in litigation--way beyond any reaonable claim construction.

Whatever may come out of that Munich case (technically, a host of cases asserting different members of the same patent family and targeting different Apple entities) is rather unlikely to move the needle in the intercontinental Qualcomm v. Apple dispute. Based on how a first hearing went in May, even an infringement finding wouldn't result in an immediate injunction: the case(s) would most likely be stayed over validity concerns. But Munich is one of Europe's most important--and fastest--patent litigation venues, and Presiding Judge Dr. Matthias Zigann one of Europe's influential patent judges (among other things, he's a contributor to a leading reference on patent law). An infringement holding could set a dangerous precedent affecting many other cases before the Spotlight matter, if ever (since the patents-in-suit might all die anyway), would be appealed.

The issue here is related to the nature of patent-eligible inventions, to claim construction, and, ultimately, to the principles of infringement analysis.

Half a century ago, the Federal Court of Justice of Germany defined, in its Red Dove ("Rote Taube" in German) decision, what constitutes a patent-eligible invention: "Patent protection is available for a teaching of a planned course of action involving the use of controllable forces of nature in order to achieve a result with an identifiable causation." (emphases added)

If the forces-of-nature part was still applied, most software patents would be invalid anyway. But that's a different story. The issue in the Munich Spotlight case(s) is that Qualcomm's infringement allegation is based on the mere possibility of a screen output, in some edge cases, "only representing the usable communication channels [of a given user]" (emphasis added), without a planned course of action and an identifiable causation.

Qualcomm is now seeking to prevail on the basis of a total misreading of the 2005 Federal Court of Justice ruling in Shunting Trolley ("Rangierkatze" in German). As a result, even thin air could infringe key elements of a patent claim.

Here's my translation of the Shunting Trolley doctrine:

"A patent is definitely infringed if the limitations of the patent claim are practiced and the accused embodiment is objectively capable of delivering the features and effects covered by the patent. An infringement holding is not avoided because an apparatus is normally operated in different ways and customers therefore do not usually make use of the relevant teaching, even in cases in which a vendor explicitly recommends using its apparatus differently, provided that there is a possibility of the patented teaching being utilized."

I already vented outrage over Qualcomm's approach in my post-hearing commentary in May. One actually just has to take the very first sentence of the Shunting Trolley doctrine seriously in order to reject Qualcomm's infringement theory: "if the limitations of the patent claim are practiced". Not so here. You can find an exemplary set of those claims on Google Patents. The key passage here is in the penultimate paragraph of claim 1:

"presenting, on the display of the first communication terminal, a plurality of selectable items only representing the usable communication channels, responsive to sensing input of a title" (emphasis added)

The word "only" is key here. Qualcomm's counsel says: if a user only has usable communication channels (suzch as only one usable channel per user) in an address book, the search result will be limited to only usable communication channels. Unfortunately, there's no Markman proceeding in Germany, where claim construction would be determined first and where disputed claim limitations would have to be interpreted by the court in writing. Otherwise I have no doubt--no doubt whatsoever--that this claim element here would have been deemed a filter (a word that, therefore, appeared multiple times in my May post on this matter).

Therefore, under general principles of patent law and the way the first sentence of the Shunting Trolley doctrine restates one of them, Qualcomm would actually have to prove the existence of a filter. The fact that Spotlight normally displays information related to a user that does not constitute a "usable communication channel[]" strongly suggests that there is no such filter. (Theoretically, the display could aggregate information retrieved with multiple filters, but at least at the May hearing, Qualcomm didn't even make that allegation and Apple's counsel, Freshfields Bruckhaus Deringer's Wolrad Prince of Waldeck and Pyrmont, denied the presence of a filtering operation of the claimed kind.)

The discussion in May had me so concerned about how software patents might get overbroad scope in future German infringement cases that I read the entire Shunting Trolley opinion. As I had assumed, I found that it stops far short of the basis on which Qualcomm seeks to prevail here:

The mechanical patent asserted in Shunting Trolley covers a certain kind of braking system: a particular type of rail brake. Even the defendant couldn't deny that the accused product contained such a rail brake, nor could the defendant deny that under some rare circumstances, more or less accidentally, the rail brake could be activated, but stressed that a different brake, called "operational brake," always got activated when a user hit the brake, while the rail brake was not intended for normal use:

So the case turned on the relevance of either brake to the infringement analysis. The Shunting Trolley defendant naturally wanted the analysis to be limited to the non-infringing one of the two brakes:

The Federal Court of Justice, however, sided with plaintiffs and deemed the trigger irrelevant (even a mere coincidence is sufficient), and applied the usual rule that the additional presence of non-infringing components is irrelevant:

Qualcomm's position is correct on the left side of that chart, and at the top of the right side (deeming the trigger irrelevant), but throws the baby out with the bathwater by also deeming the claimed cause and causation (see the pink color in the image below) irrelevant, limiting the entire analysis to the effect:

The legal error in the pink part is that the Federal Court of Justice never said a mere coincidence could serve as a substitute for practicing a claim limitation (which takes us back to the identifiable causation it already stressed half a century ago). Nor did Shunting Trolley say that you only had to look at whether a braking occurred, ignoring what caused it and how. If we now apply the only reasonable interpretation of "only" in the given context--a filter--to the Qualcomm v. Apple case and correctly map it to the Shunting Trolley doctrine, we find that the user input is irrelevant to the analysis, but there must still be a filter, in the absence of which there simply is no infringement:

Thin air is not a filter. Otherwise, thin air could infringe a patent on a coffee filter if filtered coffee comes out at the bottom because filtered coffee is poured in at the top. It would be absurd, but it would be particularly bad for software innovation, given that patents on data processing often involve an output that can meet certain criteria with a particular set of input data, even if no patented technique is utilized. For another example of an analysis confined to a mere digital output, a screen displaying a different random number per second could be deemed to infringe a patent on a digital thermometer because, if you watch it long enough, it will sooner or later happen to display the actual temperature.

I still hope Qualcomm's infringement theory will fail. If not, then maybe some other defendant in some other case will make use of the above and get a total misreading of Shunting Trolley overturned...

Share with other professionals via LinkedIn:


Thursday, September 13, 2018

One year after telling judge its house was on fire, Qualcomm spends tens of billions on stock buybacks

Today Qualcomm proudly "announced that it has entered into accelerated share repurchase agreements (ASR Agreements) with each of Bank of America, N.A., Citibank, N.A. and Morgan Stanley & Co. LLC to repurchase an aggregate of $16 billion of Qualcomm's common stock, with an initial delivery to Qualcomm of approximately 178 million shares in the aggregate". Qualcomm CEO Steve Mollenkopf notes that the company is "continue to execute on [its] previously announced $30 billion stock repurchase program."

A few days ago, Qualcomm's stock hit $70 for the first time in three years.

It's just that all this good news is in stark contrast to what Qualcomm's lead counsel, Cravath's Evan Chesler, told Judge Gonzalo Curiel of the United States District Court for the Southern District of California about a year ago when Qualcomm was seeking a preliminary injunction requiring four Apple contract manufacturers to make royalty payments prior to a final ruling on the Apple v. Qualcomm dispute. Judge Curiel didn't buy Qualcomm's "irreparable harm" theory and denied the motion. The course of events since--showcased by this week's news on Qualcomm's stock price and the share repurchase programm--clearly has validated the judge's decision. It means Qualcomm has tens of billions to distribute, but claimed that Apple, through its contract manufacturers, withholding half a billion a quarter would put the whole company in jeopardy.

Just like in yesterday's post on a previously-unknown antitrust complaint lodged by Apple against Qualcomm in the EU, which has given rise to preliminary investigations, I'll point to MLex. Mike Swift reported on the San Diego hearing last year:

"The house is on fire when one the largest manufacturers in the world says I'm watching [what's happening in the Apple litigation] and I'm telling you that I'm stopping all payments as well," Qualcomm lawyer Evan Chesler said at a hearing in San Diego, California. "It ought not to be the case that the house must burn down before we can restore the status quo." (emphasis added)

The discrepancies between Qualcomm's courtroom arguments and shareholder communications (and financial facts) have previously cost Qualcomm credibility with financial analysts. Shortly after the said hearing, Barron's quoted Bernstein analyst Stacy Rasgon who found "the magnitude of the disparity troubling."

That report came a week after I had stated essentially the same thing on this blog.

If I were in Qualcomm's shoes, I'd really hope Judge Curiel isn't aware of Qualcomm's investor relations announcements, such as (especially) today's news of the acceleration of its huge share repurchase program. In case he is aware, Qualcomm will have an even bigger credibility problem with him than with some analysts or bloggers.

Share with other professionals via LinkedIn:


Wednesday, September 12, 2018

Qualcomm facing additional preliminary antitrust investigations in the EU over German iPhone lawsuits targeting Intel chips: MLex

MLex, the leading source of insight on regulatory risk (and by now a Bloomberg subsidiary), yesterday reported that the European Commission's Directorate-General for Competition (DG COMP) has asked Qualcomm questions about an antitrust complaint Apple had filed earlier this year and that I believe hasn't previously been reported on by anyone. According to Matthew Newman, one of the best-connected EU reporters (among other things, he was spokesman for a commissioner), the complaint relates to Qualcomm's patent enforcement against Apple in Germany (trials will take place in Mannheim and Munich over the next few months, with one in each city scheduled for next week) targeting iPhones that incorporate Intel baseband chips.

In January, the EU's competition authority slammed Qualcomm with a $1.2 billion fine for its exclusionary conduct in the years 2011-2016 when Apple was contractually precluded from sourcing baseband chipsets from Qualcomm's competitors such as Intel. As Commissioner Vestager explained at the time, Qualcomm had "taken measures to avoid competition on the merits" in its quest for total market dominance.

The way I understand the new MLex report, the Apple complaint that the EU is now investigating on a preliminary basis alleges that Qualcomm is pursuing the same monopolization objective just by different means: in this case, through the enforcement of patents against Intel-based iPhones in Germany. In the EU's largest economy, injunctive relief is available as a legal remedy, i.e., without the kind of equitable analysis U.S. courts perform under eBay, though even German courts can deny injunctive relief, or stay entire cases, over antitrust considerations, under the ancient Roman dolo agit doctrine, according to which no one can claim something he has an obligation to return immediately. For standard-essential patents, German courts must and do follow the Huawei v. ZTE ruling by the Court of Justice of the EU. However, the German Qualcomm v. Apple cases, unlike Huawei v. ZTE, do not involve FRAND-pledged standard-essential patents.

I would put it this way: in order to originally keep, now force Intel out of the wireless baseband chipset market, Qualcomm has first used a carrot--giving Apple a better deal in exchange for exclusive arrangements--and then, after Apple refused the carrot and started buying chips from Intel anyway, a stick: patent infringement lawsuits.

Over the carrot, Qualcomm has already been fined (a decision that it is appealing in court). In July, the Commission sent Qualcomm a supplementary Statement of Objections (SO) on predatory pricing. And now the question is whether the ongoing preliminary investigations relating to the stick will result in formal investigations, which in turn could lead to yet another SO and, potentially, another fine.

Such escalation would not be unprecedented. Google knows what it's like when you start being investigated over one issue (vertical search engines), then over the next one (Android, with a record fine), and when there are stakeholders who are prepared continue to bring new complaints unless certain kinds of behavior come to an end.

It has worked out for me in logistical terms to be able to attend next week's Qualcomm v. Apple trials in Germany. I still have some things to do in Germany and will therefore be able to report on what's going on there. A patent family Qualcomm is asserting in Munich raises a broader concern over the scope of software patents in Germany that sends a shiver down the spine of a former anti-software-patent campaigner, but the antitrust implications of cases specifically targeting Intel-powered iPhones (given that Intel is the only major competitor, thanks to its iPhone deal with Apple, that Qualcomm is facing at the moment) are even more important in the global scheme of things.

Share with other professionals via LinkedIn:


Sunday, September 9, 2018

Texas jury verdict exposes Huawei's conflict between inbound and outbound licensing of FRAND-pledged patents

This is a follow-up to my retweet (of three days ago) of a TechRights article:

Huawei is indeed torn between two roles: that of a licensor of standard-essential patents (SEPs), and that of a licensee.

Just at a time when Huawei is trying to prove Samsung's infringement of some of its SEPs in court (and seeking a whopping 1.5% royalty), a group of three affiliated non-practicing entities (Optis Wireless Technology, Optis Cellular Technology, and Panoptis Patent Management) has prevailed over Huawei in the United States District Court for the Eastern District of Texas on all five patents-in-suit: four FRAND-pledged wireless SEPS and one non-FRAND-pledged (but potentially still essential) H.264 codec patent. And the jury awarded a total of $10.6 million, with each of the four FRAND-committed patents, however, accounting for only an average of approximately $0.7 million. In other words, the non-FRAND patent was deemed roughly ten times more valuable. While it's theoretically possible that the non-FRAND H.264 patent was considered more valuable for technical reasons than the FRAND-pledged wireless SEPs, Judge Rodney Gilstrap, the nation's busiest patent judge in recent history, had instructed the jury that the FRAND promise needed to be kept in mind when determining damages.

This is the damages section of the verdict form (click to enlarge; this post continues below the image):

Most of the Optis/Panoptis patents-in-suit were transferred to that entity by Ericsson--a practice that is called "privateering" and that this blog has frequently criticized over the years. A former Ericsson executive runs Optis/Panoptis, and McKool Smith, the rising-star litigation firm representing Optis/Panoptis, frequently litigates patents on Ericsson's behalf.

Ericsson and Huawei had a dispute years ago. Apparently Ericsson came under serious pressure in China and more or less backed out (a face-saving exit at best). But Ericsson's privateers are still going after Huawei.

Huawei has yet to prevail on the merits in its dispute with Samsung. Same with Qualcomm in its dispute with Apple. Those two companies (who are at loggerheads with each other, yet have somewhat similar patent enforcement strategies) are demanding many billions of dollars before establishing the infringement of a single valid patent.

For all owners of wireless SEPs seeking to enforce them (also including Huawei when the shoe is on the other foot, and Qualcomm), every decision (here, a jury verdict) that awards only a modest amount of damages/royalties is bad news. Huawei has a significant U.S. market share, and if its infringement of four wireless FRAND-pledgedd SEPs just entitles a patent holder to about 700 grand per patent, it's hard to see how Huawei could demand many billions from Samsung.

After this jury trial, there'll be a bench trial at which Judge Gilstrap will have to determine whether Optis/Panoptis had previously made Huawei a licensing offer on FRAND terms, which Huawei denies. This, as well as the usual post-trial motion practice seeking a reduction of damages, will force Huawei to take positions contrary to the ones it's espousing in its dispute with Samsung.

Huawei makes very good devices. I've bought some for my app development company, and it's one of two Android brands that I've repeatedly recommended to others. It would be great if Huawei's priority was its product business, in which case it should be in the FRAND camp. Its dispute with ZTE earlier this decade was more of a statement. To me it looked like Huawei wanted to demonstrate that it takes pride in its contributions to wireless innovation. But the longer its dispute with Samsung takes, the more I fear that Huawei is actually aspiring to become the next Qualcomm...

It will be interesting to watch the further proceedings in the Eastern District of Texas.

Share with other professionals via LinkedIn:


Apple asks court to find that Qualcomm cannot claim billions of dollars for breach of gag order

There's a third summary judgment motion related to Qualcomm's business practices that is worth talking about, though it is admittedly a distant third given the enormous potential of the FTC's motion regarding the licensong of rival chipset makers and Apple's motion targeting Qualcomm's "double-dipping" (chipset sales + patent royalties). In a lower-profile motion that nevertheless highlights a major issue, Apple has asked Judge Gonzalo Curiel of the United States District Court for the Southern District of California to throw out Qualcomm's counterclaims according to which Apple has to repay rebates amounting to billions of dollars because it breached a "Business Cooperation and Patent Agreement" (BCPA) through its efforts to instigate and broaden antitrust investigations into Qualcomm's conduct (this post continues below the document):

18-09-01 Apple Motion Re. R... by on Scribd

In order to understand the context of the motion, one needs to go back to last year's (amended) complaint, in which Apple stated the issue as follows ("Nature of the Action", paragraph 1):

"Qualcomm pursues its illegal practices through a secret web of agreements designed to conceal and obfuscate its conduct. In at least one such agreement, Qualcomm inserted a gag order that prevented an aggrieved party from seeking relief that could curb Qualcomm's illegal conduct, in an effort to keep courts and regulators in the dark and its coerced customers quiet." (emphasis added)

A "gag order" relating to information that stakeholders give to regulators is highly problematic. Allowing this kind of scheme would enable companies with huge leverage to impede the work of competition authorities to the point that antitrust action would be both less likely to happen and far less likely to succeed.

Most of Apple's motion focuses on compliance ("we didn't breach the agreement"), not on (un)enforceability, but at least Section A.5 makes a public-policy argument:

"Any interpretation of the BCPA that would prevent Apple from responding to the [regulatory] agencies' requests (whether based on subject matter or the alleged truthfulness thereof) or would allow Qualcomm—the target of agency investigations—to receive repayment of billions of dollars based on Qualcomm's own (incorrect) view of their contents would contravene established public policy. The public policy of California favors full disclosure of concerns about unlawful conduct to governmental investigators." (emphasis in original)

"Apple had a legal duty to comply with all subpoenas and CIDs [Civil Investigate Demands] from the FTC. Moreover, FTC rules of practice state that the FTC 'expects all parties to engage in meaningful discussions with staff.' [...] Inherent in this expectation is freedom from retaliation if the investigation's target dislikes a response. For this reason, responses to CIDs are generally confidential 'to safeguard the rights of individuals under investigation and to protect witnesses from retaliation.'"

"The same is true in Korea. The KFTC has stated that it 'relies heavily on third parties to gain information” relevant to ongoing investigations and to detect anticompetitive activity in Korea. [...] And Korean antitrust law expressly states that an investigative target cannot retaliate against, or give 'any disadvantage' to, another entrepreneur for 'cooperating with investigations' of unfair trade practices."

Apple's legal argument in the summary judgment motion is mostly about evidentiary failure on Qualcomm's part (for an example, Apple says there's no evidence Tim Cook discussed the KFTC investigation with his counterpart, whom Qualcomm didn't even seek to depose), about timing (Apple's first written correspondence with certain regulators occurred in eah case after the regulator had launched an investigation; and the BCPA expired at some point, so Qualcomm can't claim a breach based on what occurred post-expiration), and so forth. If Qualcomm convinces Judge Curiel that there is room for factual dispute, the motion may be denied (though Apple could still prevail on any of this at trial). If Apple's strategy works, the court will hold that nothing Apple did constituted a violation of the BCPA (reasonable interpretation provided). And if my wish was granted, the court would sua sponte place a lot more emphasis on the enforceability part than Apple did.

A few more quotes from Apple's complaint show the importance of the underlying issue:

"Qualcomm's recent effort to cover its tracks—by punishing Apple for providing truthful testimony at the request of government regulators—underscores the lengths to which Qualcomm will go to protect its extortion scheme."

"The BCPA carved out, as it must, an acknowledgment that Apple has a responsibility to respond to enforcement agencies' requests for information. But in restraining Apple from initiating action or bringing concerns to law enforcement, Qualcomm conditioned billions of dollars on Apple's silence before courts and regulators about Qualcomm's business practices. And Qualcomm is now interpreting that agreement to retaliate against Apple for responding to requests for information about Qualcomm's practices from competition agencies, inhibiting law-enforcement review of Qualcomm's anticompetitive practices."

"232. Specifically, Qualcomm offered to pay Apple the nearly $1 billion it owed if Apple would, in Qualcomm's words:

(i) publicly and specifically retract and correct each of Apple's misstatements about Qualcomm to regulatory agencies [...]; (ii) inform the relevant agencies that such statements were and are untrue; (iii) disclose Apple’s correspondence with any agencies relating to any investigation of Qualcomm;10 (iv) provide any and all additional facts to regulators and Qualcomm relating to Apple's dealings with Intel concerning any possible or actual consideration from Intel to Apple relating to Apple's implementation of WiMax or the use of Intel chips; and (v) provide Qualcomm with the requested information about any communications between Apple’s senior executives and Samsung.

233. Thus, in an extraordinary and transparent effort to manipulate regulatory iinvestigations into its anticompetitive behavior, Qualcomm offered to repay Apple nearly $1 billion in withheld BCP Payments if Apple recanted its true and, in many cases, sworn testimony before government agencies and instead gave false testimony favorable to Qualcomm."

In light of the above, the term "gag order" is actually an understatement. Qualcomm wanted to be exonerated by using Apple, in exchange for a billion-dollar payment, as its mouthpiece. Demanding a retraction of a statement is reminiscent of what certain communist dictatorships used to do during the Cold War. They, too, obligated people to publicly speak against their previous statements.

While it's understandable that Qualcomm will try anything in a multi-billion-dollar dispute to save money and, therefore, claims that Apple breached the agreement, the most important question is why Qualcomm put a "gag order" in place and made a billion-dollar offer for retracting statements Apple had made to regulators. Why would Qualcomm have done all of that in the first place if its business model was as legal as it keeps telling courts and the general public?

Share with other professionals via LinkedIn:


Thursday, September 6, 2018

Apple, Foxconn et al. want to end Qualcomm's double-dipping practice (chips and patents) once and for all

The previous post was about a magistrate judge's annoyance at Qualcomm's gamesmanship. But what's sure to cause Qualcomm and its lawyers a lot more headache is the pressure from two important summary judgment motions: the FTC's motion in the Northern District of California seeking to clarify Qualcomm's self-imposed obligation to license rival chipset makers and--which is what this present post is about--a motion by Apple and four contract manufacturers (Foxconn, Compal, Wistron, and Pegatron) in the Southern District of California, leveraging the Supreme Court's 2017 Lexmark ruling against Qualcomm's double-dipping practice (this post continues below the document):

18-08-31 Apple Et Al. Motion for Partial Summary Judgment by Florian Mueller on Scribd

For a recap of something discussed in greater detail last year, Qualcomm's business model is basically like this:

  • As Judge Koh recalled in a recent order, "an analysis conducted by Qualcomm in 2015 showed that revenues from Qualcomm's licensing program were 'equivalent in size to the sum of ~12 companies with a form of technology licensing,' including leading cellular SEP licensors such as Ericsson, Nokia, and Interdigital."

  • Customers have to accept an out-of-this-world royalty rate in order to be able to buy Qualcomm's chips.

  • Those who don't want to buy Qualcomm's chips must still pay patent license fees and will sooner or later be "persuaded" to enter into exclusivity arrangements under which their total cost (the sum of chip supplies and patent royalties) is reduced (but apparently still very high).

The above sounds like a "great" (for Qualcomm, not for the economy at large or for society) deal. However, Apple and its four contract manufacturers say that "Lexmark forbids this double-dipping."

The doctrine of patent exhaustion, which the Supreme Court clarified, upheld and reinforced in Lexmark, is fundamentally irreconcilable with the notion of a company selling chipsets and collecting patent license fees from the customers buying its chips. But Qualcomm appears to have believed for a long time that it would get away with it. One, because no one would dare challenge it. Two, by virtue of three structural elements that the summary judgment motion addresses:

  • "Qualcomm divides the chipset sales and software license across two agreements"--one for the purchase of the chipsets and another one, named Master Software Agreement, for the software license.

  • Foreign sales (such as to Asia-based contract manufactures) don't help as Lexmark clear stated that "[a]n authorized sale outside the United States, just as one within the United States, exhausts all rights under the Patent Act."

  • The motion labels as "corporate shell games" a structure under which different Qualcomm entities sell certain products and technically own certain patents. More than six years ago I already reported and commented on a related restructuring.

The motion invites Judge Gonzalo Curiel to take a Lexmark attitude and focus on a simple commercial question: are we talking about authorized chipset sales or not? If authorized, there's exhaustion; if Qualcomm wants to argue that exhaustion doesn't kick in, it has to show that something was unauthorized about the sale.

The most technical question here ("technical" in a literal sense, not legally technical) is whether, as the motion says, "Qualcomm’s chipsets and software indisputably are a matched pair by design, each intended to be used with, and only with, the other." The moving parties describe this as an undisputed fact. It will be interesting to see whether (and if so, how) Qualcomm will seek to raise doubts about it, but absent a single reasonable use case for that software outside of Qualcomm's chipsets and for Qualcomm's chipsets without the related software license, I guess it just can't deny what matters here.

Exhaustion can only be determined on a patent-by-patent basis (even if certain key questions, particularly anything related to Lexmark, will overlap). It's not a portfolio question.

That's because it must be shown for each patent that it is practiced by Qualcomm's chipsets. It wouldn't be practical to have one or more courts make this determination for thousands of patents. And a motion for partial summary judgment is subject to rather strict page limits--the whole idea of summary judgment is that you raise an issue that the court can easily decide. Apple and the likes of Foxconn therefore picked three exemplary patents (from the ones Apple and the contract manufacturers tackled before). Footnote 1 says that "[a] favorable ruling on this motion would inform the parties as to the exhaustion issues for the other patents." In other words, the parties need some guidance, but then they should be able to figure out exhaustion for numerous other patents.

Footnote 6 explains, in other words, that the movants are simply in their right to present alternative claims: should they prevail on exhaustion, infringement won't be reached. While the motion doesn't say so, this would even apply in a case where the theories aren't just alternatives but where they would be contradictory. Here, however, it's technically still possible that Qualcomm's chips practice a patent (thereby triggering exhaustion, provided that a sale is authorized and a court is underwhelmed by an attempted end-run based on a separate software license and a nested corporate structure) while, for an example, Intel's chips don't. However, to the extent that a patent is truly standard-essential, it may be practiced by Qualcomm and infringed by everyone else--but who cares if it's exhausted?

The FTC's motion for partial summary judgment in San Jose has the potential to force a settlement of that entire antitrust matter. If the San Diego motion on exhaustion succeeds, there will still be a host of financially and strategically important issues to resolve, but the case would be streamlined in a highly important way and, in a best-case scenario, it's possible that no (alternatively, fewer) patent infringement questions would have to be taken to trial. And above all, either motion would have major impact on the market.

Share with other professionals via LinkedIn:


Magistrate judge rebukes Qualcomm for iPhone patent infringement allegations it originally chose not to bring

It's pretrial season in the Huawei-Samsung and Apple-Qualcomm FRAND disputes, thus a flurry of motion practice. Before I'll get to discuss the strategically extremely important exhaustion issue (in light of the Supreme Court's 2017 Lexmark) in Apple v. Qualcomm, there's also a short but interesting ruling that United States Magistrate Judge Mitchell D. Dembin handed down in the Southern District of California on Tuesday (this post continues below the document):

18-09-04 Order on Motion to Strike Qualcomm Allegations by Florian Mueller on Scribd

The procedural context, which the order summarizes at the start, is that Apple's original complaint involved, among other things, requests for declaratory judgment on invalidity and non-infringement of nine Qualcomm patents. With the first amended complaint, Apple tackled another nine Qualcomm patents. As an Apple filing noted (see the update paragraph at the end of this June 2017 post), Qualcomm--quite surprisingly!--elected not to bring compulsory infringement counterclaims, forever precluding Qualcomm from bringing such charges.

With the trial approaching, Qualcomm may have had second thoughts regarding its choice. Its lawyers apparently preferred to have their experts opine on infringement with a view to everything the jury will have to decide.

Apple didn't accept that change of mind and brought a motion to strike. The motion succeeded with respect to certain passages of multiple expert reports, which Qualcomm (unless it successfully appeals this order to Judge Curiel) won't be allowed to leverage at trial. The motion failed with respect to some others (the judge concluded that those expert opinions focused on standard-essentiality of certain patents and valuation; in a way, essentiality is often tantamount to infringement).

The following harsh words show that Magistrate Judge Dembin was annoyed by the way in which multiple Qualcomm expert reports were inconsistent with Qualcomm's decision not to bring infringement counterclaims with respect to the original set of nine patents:

"Qualcomm's counsel know that in a declaratory judgment action by a licensee against a patentee seeking an order of non-infringement, the patentee, Qualcomm, bears the burden of persuasion of infringement. [...] Qualcomm made the tactical decision not to assert infringement and thus avoid certain discovery obligations as mentioned above. In its expert designations, Qualcomm chose not to disclose that certain experts expressly would opine on infringement and assert that Plaintiffs are infringing patents-in-suit. Qualcomm will be held accountable for the consequences of its tactical decisions." (emphasis added)

"To the extent that Qualcomm claims they have disclosed in discovery their views regarding infringement and, consequently, there is no surprise and no prejudice, is unavailing. Rules are rules and tactical decisions have consequences[.]"

In other words (not in the order): you make your bed and you lie in it.

Share with other professionals via LinkedIn: