Friday, July 19, 2019

Vestager serves Qualcomm a double whammy for dessert: second EU antitrust fine in as many years

The term of the Juncker Commission is nearing its end, and while I'm far from enthusiastic about his successor, I'm relieved that neither that conservative-in-name-only Weber nor "Poor Man's Bernie" (Timmermans, who's less reasonable than the real Senator Sanders) got the top job. However, my preference among reasonably likely candidates would have been Margrethe Vestager, the EU's competition commissioner, as I made clear on social media, despite disagreeing with some parts of her regulatory activism, such as the "state aid" case against Ireland and certain aspects of the Android case.

Last year I wrote that Qualcomm "won" the "Antitrust Grand Slam" when the European Commission joined the U.S. Federal Trade Commission and a couple of Asian regulators in fining Qualcomm. The Commission imposed a fine of €997 million ($1.2 billion) over exclusionary conduct in the years 2011-2016 when Apple was precluded from sourcing baseband chipsets from Qualcomm's competitors such as Intel. That exclusive dealing is one of the four counts on which the U.S. FTC defeated Qualcomm in court this year (my previous post discussed some support Qualcomm got for its motion for an enforcement stay).

Yesterday's fine, based on a supplemental Statement of Objections that came down in July 2018, amounts to "only" 242 million euros (272 million U.S. dollars), so it now got a "double whammy" from the EU. The latest one is about predatory pricing. At first sight, that's counterintuitive. We all know that the allegations usually brought against Qualcomm, besides exclusive dealing, are all about maximizing revenues even in the very short term, not just for the long haul. However, we have to keep in mind that Qualcomm is not just "a monopoly" (in the sense of U.S. antitrust law; over in the EU, this is called "market dominance"), but a dual monopoly: its SEP portfolio bestows monopolistic rights (not only on Qualcomm but also on any other patent holder, provided at least one patent in a portfolio is truly standard-essential), as does its position in certain segments of the chip market, and those monopolies are mutually-reinforcing as a result of Qualcomm's practices.

So even if Qualcomm (now quoting an EU Commission statement) "sold certain quantities of three of its UMTS chipsets below cost to Huawei and ZTE, two strategically important customers, with the intention of eliminating Icera, its main rival at the time in the market segment offering advanced data rate performance," it doesn't mean that phones became cheaper or that Qualcomm lost money on those devices in the short term. Much to the contrary, we can assume that Qualcomm was able to easily afford price dumping on those low-end chipsets because it would impose its patent tax on those devices anyway (even if those device makers had used non-Qualcomm chips).

Having said that, the second EU antitrust fine imposed on Qualcomm during Mrs. Vestager's first term is perfectly consistent with findings by various antitrust regulators around the globe, and with Judge Koh's landmark ruling.

Share with other professionals via LinkedIn:

Patent enforcement against delivery of non-executable data: Federal Circuit more balanced than Germany's Federal Court of Justice

The Munich-based CIPLITEC (Center for Intellectual Property Law, Information and Technology) is a recently-created forum for an exchange of views between the leading Bavarian law schools and practitioners. Yesterday, Dr. Ralf Uhrich, in-house patent litigation counsel at Google and in charge of all European patent infringement cases to which Google is a party (always as a defendant unless I missed something), gave a talk at Ludwig Maximilian University, as per CIPLITEC's invitation, on the subject of "patent protection for data." I had previously seen Dr. Uhrich at various German patent trials while he was a Quinn Emanuel associate, and at a couple of conferences in the area.

As a result of CIPLITEC's inclusive approach, there were 200 or 300 people in the audience as per my estimate; most of them were students, but I also spotted a number of law professors, patent attorneys, and patent litigators. In response to strong demand, this event had to be moved to a larger room than the one they had originally reserved. I attribute this to the topic, to many people's interest in Google's views (though one of the organizers stressed that Dr. Uhrich was expressing his personal views as opposed to engaging in advocacy on his company's behalf), and to Dr. Uhrich's great personal reputation in the German legal community. By the way, if you didn't know who he is, you might be led to believe he's the younger and smarter brother of Liverpool FC manager (and this year's UEFA Champions League-winning coach) Juergen Klopp.

The starting point for the presentation was the European legal framework for patent-eligible subject matter: Article 52 of the European Patent Convention (EPC). As Dr. Uhrich explained, the exclusions set forth there appear broader than they are applied when it comes to entire patents: a token reference to a computer or even just a storage medium salvages a claim, but non-technical claim limitations (though a more efficient use of computing resources is considered technical) are irrelevant to a novelty or inventiveness analysis. Having campaigned against software patents in the EU, my personal view is that the European Patent Office as well as national courts act as if Art. 52 EPC didn't exist as far as entire patent claims are concerned, and while the situation is indeed substantially better with respect to what counts in a novelty or inventiveness context, the statutory exclusions have simply been gutted. A 1969 decision by the Federal Court of Justice of Germany, Rote Taube ("Red Dove"), defined technicity as the planned use of controllable forces of nature for the purpose of achieving a specific result, and on that basis it would be possible to distinguish a software-controlled braking mechanism or other advanced in an applied natural science from pure advances in software such as more efficient memory (or bandwidth) usage through data compression.

The good news in the early part of Dr. Uhrich's presentation was that even the EPO doesn't grant patents that claim a data structure per se. So the issue here is not one of patentable subject matter in the strictest sense, but of the scope given to patent claims at the enforcement stage. To share the bad news upfront, the effect of an overreaching infringement theory can be just as bad as straightforward patent claims on data formats. But, at least for now, the related case law in the United States is fundamentally better than in Germany, though this may be attributable in no small part to the historic happenstance of what cases were put before the courts in what sequence--and what questions for review the parties raised.

Just so there is no misunderstanding: Dr. Uhrich's academic talk was nonjudgmental, so when you find words like "good news" and "bad news" here, rest assured they're just my opinion. He may or may not agree depending on context.

The enforcement-related main part of Dr. Uhrich's talk started with a 19th-century holding by the German Reichsgericht (Imperial Court), Methylenblau, involving a patent covering a chemical manufacturing process that was employed outside of Germany, but the resulting product entered the German market. The key doctrine there was that the scope of protection of a manufacturing patent potentially extends to the output if the substance so produced is an integral part of the patented process. On that basis, the Reichsgericht remanded the matter to the trial court.

The legal tradition that started with Methylenblau wouldn't have had to inevitably lead to a high-court decision, more than a century later, that data sequences generated by a patented data processing operation are afforded the same degree of protection (potentially, as it's always subject to the specific facts of a case). Not only is there a fundamental difference between physical goods and non-physical data but what makes this doubly unreasonable is the blatant inconsistency of such an outcome with the statutory exclusion of patents on "computer programs as such." Unfortunately, it nevertheless happened.

In 2012, the Bundesgerichtshof (Federal Court of Justice of Germany) handed down a decision on whether data storage media manufactured outside of, but imported into, Germany might infringe a video encoding patent, EP0630157 on "systems and methods for coding alternate fields of interlaced video sequences," a patent declared essential to the MPEG 2 video standard. While the patent holder lost the case due to a combination of other reasons, particularly patent exhaustion (the video data was generated with a licensed tool), the decision held that the case could not be dismissed on the grounds of the accused products containing data sequences as opposed to an encoder (be it a physical device or a piece of software).

I have read the MPEG-2-Videosignalcodierung (MPEG 2 video signal encoding) decision, and there is no reference in it to the statutory exclusion of patentable subject matter under the EPC, suggesting very strongly that counsel for defendant didn't raise this point with the top court. German patent litigators rarely recur to the Article 52 exclusions as they have been rendered as meaningless as I explained above when it comes to entire patent claims. Many of my allies in the movement opposing software patents suspect that patent litigators wouldn't even want Article 52 to be given meaning since it would discourage the filing of many patent applications and infringement cases in the first place. Whatever the reason may be, I had a conversation in 2014 with a Federal Court of Justice patent judge (who was on the panel of five judges who made the MPEG 2 decision, though one never knows where a member of a German panel actually stood since there is no such thing in Germany as a dissenting opinion), and I mentioned my opposition to software patents. He noted that their decisions are often misunderstood as being exceedingly permissive in this regard just because the questions put before them--which may just be about novelty or inventiveness, for instance--don't even enable them to address subject matter. I'm still grateful for a very interesting conversation on a train ride from Bayreuth to Nuremberg after a conference, so I don't want to criticize the Federal Court of Justice for a decision that I regard as inconsistent with the letter and the spirit of Article 52 EPC. However, I would strongly encourage other litigants (knowing how many litigators read this blog) to raise the Article 52 question when the opportunity arises in connection with patent enforcement against pure byte sequences.

One of the reasons for which I believe it would be worth trying is that, as Google's Dr. Uhrich explained yesterday, the Federal Court of Justice seized the opportunity afforded to it by a non-digital patent case, Rezeptortyrosinkinase II, to clarify (well, there's a floating border in case law between clarifications and corrections) that the MPEG 2 doctrine doesn't cover just any information resulting from a patented process. In that case decided in 2016, the Federal Court of Justice did not see a patent infringement in the mere communication of the result of a medical analysis from the country in which the analysis is performed to the country in which the process is protected by a patent. In that case, the result was basically just a binary yes-or-no piece of information on whether a certain medical condition was present.

The relevant line-drawing is centered around the question of whether the transmitted data bears actual and technical characteristics engraved by the patented process. In the case of MPEG 2 video data, that was deemed to be potentially the case as the encoding process results in a particular data structure. Apart from the general concerns I raised above, I do have an MPEG-2-specific concern, which is that the key characteristic of MPEG 2 is "lossy" compression, so at least to the extent that data is removed (so as to allow for more efficient compression), it's hard to see how it can simultaneously be "engraved." However, despite disagreeing with the MPEG 2 doctrine, I certainly see the difference between a specific data format on the one hand and some abstract information on the other hand.

However, it will take more decisions to get a better idea of where exactly the line is drawn (and ideally one of the cases would eliminate the problem on the basis of Article 52 EPC). In the Q&A part, I brought up a case I had watched in Mannheim earlier this decade. A cellular standard-essential patent covered an algorithm for calculating a number sequence used as a key (like a cryptographic key) for mathematical operations that transformed data (the data to be sent over the wireless network) in a way that minimized undesirable physical effects at the radio frequency level. The key was clearly standard-essential, just like a cryptographic key prescribed by a security standard would be; but the asserted patent claim, in the preliminary opinion of the panel of judges, probably wasn't because the accused baseband chipsets simply had the number sequence hardwired instead of performing the patented process to derive them. That case was somewhere in the middle between MPEG 2 and Rezeptortyrosinkinase II, but never reached the higher courts due to a settlement.

Thankfully, Dr. Uhrich also drew a comparison between German and U.S. case law on patent enforcement against data sequences. In Bayer v. Housey Pharmaceuticals (2003), the United States Court of Appeals for the Federal Circuit affirmed the dismissal of an infringement claim because "infringement under 35 U.S.C. § 271(g) is limited to physical goods that were manufactured and does not include information generated by a patented process, and because the physical goods here (drug products) were not 'manufactured' by a process claimed in the asserted patents." The opinion was authored by Circuit Judge Timothy Dyk, joined by then-Chief Judge Mayer and now-Chief Judge Prost.

The term "manufacture" plays a key role in U.S. patent law. As some of you may remember, it was key to the Samsung v. Apple Supreme Court appeal related to the "article of manufacture" based on which a design patent holder would be entitled to an unapportioned disgorgement of an infringer's profits. The term "manufacture" alone, coupled with an almost-originalist interpretative standard that takes into account what lawmakers really meant way back when, enabled the Federal Circuit to decide against what would have been a similarly expansive school of thought as the one of the Federal Court of Justice of Germany.

Here comes Judge Sharon Prost again, who in most contexts (with exceptions like design patent damages proving the rule) takes very balanced positions. Meanwhile she had become Chief Judge, and she authored the Federal Circuit opinion in ClearCorrect v. ITC, a decision that Google's Dr. Uhrich also explained yesterday. In that case, the ITC had ordered an import ban on data generated outside the U.S. but sent to the U.S. for the purpose of 3D printing. It's not unheard of for the ITC to have an expansive view of its jurisdiction, even including digital data transfers, but the appeals court made clear that it disagreed with what the ITC had already held prior to ClearCorrect, which was that the statutory term "articles" "should be construed to include electronic transmission of digital data [...]."

The way things work, there's no doubt that some patent-asserting plaintiffs are still going to try to push the envelope of data format patentability in the United States. But at least for now, they'll be facing an uphill battle whenever they try.

What is clearly needed is a pushback against overreaching patent enforcement in Germany. Yesterday's academic presentation was neither a campaign speech nor particularly alarmist. Expressing a personal--not corporate--view, Dr. Uhrich responded to a question from the audience with a reference to other forms of intellectual property protection for data, such as database rights (a big thing in the EU, by the way) and copyright law.

I mentioned my Art. 52 EPC concern, but I also think that regardless of software patent-eligibility, the MPEG 2 decision is bad as a matter of policy because if an encoder is patentable, then so is, typically, the related decoder. And that's what the enforcement of codec patents should be limited to.

Share with other professionals via LinkedIn:

Wednesday, July 17, 2019

DOJ, DOE, Pentagon, Ericsson support Qualcomm's Ninth Circuit motion for stay of FTC's antitrust remedies

As the longest-standing and staunchest Donald Trump supporter among IP bloggers, I must admit I'm more than a little bit disappointed at three very recent events, two of which are related to the mobile industry. I keep my fingers crossed for Four More Years no matter what, but

  • the "go back" suggestion had a factual basis with respect to only one "Squad" member (Ilhan Omar);

  • it wasn't a good idea to conflate security and trade issues with respect to Huawei, so I honestly don't know what to think of that new bipartisan Lex Huawei proposal (was there a security issue in the first place, or was it more of a pretext? either way, the Administration's approach to Huawei doesn't look principled at this stage); and

  • yesterday the Department of Justice tripled down on its pro-Qualcomm advocacy I already criticized on previous occasions (as recently as last month), with declarations by the Department of Defense (Scribd link) and Department of Energy (Scribd link) attached (this post continues below the document):

19-07-16 DOJ Statement of I... by Florian Mueller on Scribd

Antitrust Assistant Attorney General Makan Delrahim's subordinates made a bizarre filing in early May when they asked Judge Lucy H. Koh of the United States District Court for the Northern District of California to hold a special remedies hearing. I said "bizarre" because of the substance of the brief, the timing (more than three months after the San Jose bench trial), and the way the DOJ antagonized the FTC. That was the first time they were in the tank for Qualcomm (not counting public comments by Mr. Delrahim, a former Qualcomm outside counsel). The second time, in connection with Qualcomm's appeal of Judge Koh's certification of a consumer class, their intervention was infinitely more reasonable. But yesterday's Statement of Interest (of the United States, as the DOJ is authorized to speak on behalf of the federal government regardless of whether an independent government agency like the FTC agrees) is closer in (un)reasonableness to the DOJ's first pro-Qualcomm filing than to the second.

The district court's well-reasoned ruling is the FTC's biggest success in a long time. The DOJ should have more respect for the independent Federal Trade Commission and for the independent judiciary. Instead, the brief, filed yesterday with the United States Court of Appeals for the Ninth Circuit, arrogantly asserts that the FTC and Judge Koh failed to figure out the law.

The DOJ attacks Judge Koh's decision from three angles: merits (liability), remedies, and the public interest. As for the public-interest part, the DOJ mostly relies on the aforementioned declarations by two other departments, and Reuters' Stephen Nellis accurately described the gist of those statements as follows:

While the public-interest part is the most appropriate aspect for a Statement of Interest, the accuracy and relevance of a public-interest argument is not a function of how many sister departments say essentially the same thing. This case is about patent licensing. It's not about Qualcomm's ability to develop and sell great 5G chips being hamstrung. By contrast, there have been antitrust cases that impacted product design, such as the Media Player part of the first EU case against Microsoft (though Microsoft ultimately managed to comply with the decision in a way that minimized its effects). In the Qualcomm case, we're talking about one company raking in roughly 25% of the industry's total patent licensing revenues, and collecting 80% or 90% of the totality of what some device makers such as Huawei spend on patent royalties. And this is a company that has just spent tens of billions of dollars on stock buybacks in recent years. It's not like Qualcomm would have to, or would want to, cut back on R&D as a result of this antitrust case.

If Qualcomm's supporters in the Administration really wanted to make a compelling case, then we'd find some financial analysis, even if based on only rough estimates of inherently limited reliability, in the DOJ's filing and/or the supporting declarations. They'd be talking about how many billions they believe Qualcomm would not make as a result of the immediate enforcement of the FTC's remedies, and what impact this would have on Qualcomm's R&D strength and its influence on the evolution of the 5G standard. Instead, they just posit that anything of substantial impact would automatically pose a threat to national security.

Qualcomm's dilemma is that numbers, whether provided by Qualcomm itself or by its governmental allies, would potentially impact future negotiations, in which they'll try to defend their royalty demands to the greatest extent possible. And in order to have a chance to persuade the Ninth Circuit, the numbers (right or wrong) would have to be dramatic, which would also scare investors. But without numbers, the public-interest argument here is a non sequitur. Even a non-starter. With numbers, one could at least perform a plausibility check.

On the merits (liability) side, the DOJ's filing is perfectly consistent with Mr. Delrahim's philosophy that patents are sacrosanct rights that antitrust law shouldn't interfere with. That philosophy, which one might even call a religion, is not the law, however. Nor should it be.

The most important aspect of the DOJ's merits-based argument is that a duty to deal must be a rare exception. In this regard, the DOJ's brief focuses very much on Judge Koh's citation to the Supreme Court's Aspen Skiing decision. I have re-read the related citations in the district court's findings of fact and conclusions of law, and nowhere does Judge Koh say or suggest that Qualcomm's obligation to extend FRAND licenses to rival chipset makers is like a photographic image of the fact pattern in Aspen. The Aspen case was about competing ski resorts in Aspen, CO. Originally, there were three resorts owned by three companies, but then one of them acquired one of the others, and created a fourth. When that consolidator owned three of the four resorts, it was no longer willing to continue a previously very popular and profitable multiarea-ticket cooperation with the sole remaining competitor on the previous--or any other reasonable--terms. The district court put the case before a jury, which identified an antitrust violation in the form of a refusal to deal, and both the Tenth Circuit and the Supreme Court affirmed.

I'm aware and, while I disagree, respect that many lawyers are very skeptical of the notion of an antitrust duty to deal with competitors. For an example, I know at least one lawyer who absolutely wants the likes of Qualcomm to be required to grant SEP licenses to rival chipset makers, but, as a matter of principle, purely on the basis of contract--not antitrust--law. Granted, Aspen is not the least controversial Supreme Court antitrust ruling in history. But its key holding--that a refusal to deal may constitute an antitrust violation under special circumstances--is still the law of the land and probably won't ever be overruled. There have been decisions like Trinko that stress the importance of not imposing a duty to deal in too many cases. But none of that makes Aspen an exclusive and extremely narrow path to the conclusion Judge Koh reached in FTC v. Qualcomm with respect to chipset licensing.

The DOJ argues that Judge Koh misapplied Aspen in two ways. First, the DOJ argues Qualcomm never voluntarily licensed other chip makers, but just "erroneously relied on its interpretation of Qualcomm’s FRAND obligations to standard-setting organizations, as required by their IP policies, as contractually compelling Qualcomm to license rival chip makers." This is at odds with the way Qualcomm portrayed (in its disputes with the FTC as well as Apple) its decision to make CDMA an industry standard, and to participate in standardization in general, as an act of generosity, or at least as a sound business decision. Second, the DOJ claims Judge Koh incorrectly held that Qualcomm discontinued (for anticompetitive reasons) something profitable by stopping at some point to license other cellular baseband chipset makers: "Qualcomm realizes greater profits by licensing at the end-device rather than the chip level." That's what Qualcomm had told the IRS as we know. But there's a wealth of evidence in the record, such as testimony by Intel and MediaTek, as to the anticompetitive effects of Qualcomm's rivals being unable to sell a licensed product to customers (and unable to control the effective bottom-line cost their customers incur). And on page 43 of the findings and conclusions, Judge Koh wrote that "the Court's focus is 'upon the effect of [the defendant's] conduct, not upon the intent behind it.'" (though anticompetitive malice is relevant, too)

Qualcomm's motion cited to some Ninth Circuit decisions, trying to argue that only a refusal to deal that is a bad business decision in the short run but a good one in the long run because of competitors being driven out of the market can be an antitrust violation. But that's not even what Aspen says. The dominant ski resort operator in Aspen presumably also had some very short-term benefits because a three-resort ticket (for all three resorts owned by that operator) was an attractive package in its own right, and nothing in Aspen says that only a long-term investment in the total annihilation of competitors constitutes an antitrust violation. What appears more important is that Qualcomm's refusal to deal had and still has anticompetitive effects on the chipset market. Qualcomm is/was not just one monopoly like the Aspen defendant, but the case is about two mutually-reinforcing monopolies (chipsets and patents, which are monopoly rights by definition, especially standard-essential patents). Qualcomm killed two birds with one stone by maximizing its patent royalties and greatly reducing the competitiveness of other cellular baseband chipset makers.

While it's unsurprising that Ericsson also supports Qualcomm's motion for a stay (Scribd link), the angle is interesting. Ericsson focuses on its use of Qualcomm chips in its base stations. However, Ericsson's motivation as a patent licensor to refuse to license chipset makers is well-documented, as is the fact that Ericsson itself once complained about such refusal to deal.

Another brief in support of Qualcomm has been filed by former Federal Circuit chief judge and pro-patent extremist Paul Michel.

Without a doubt, the Ninth Circuit will give some weight to the DOJ (plus DOD plus DOE) statement. But the FTC and its amici will now get to respond, and while the DOJ brief does contain some food for thought, I believe it can be effectively countered. However, at this stage it's about a motion to stay enforcement, so there won't be a full-blown analysis of the relevant questions. Qualcomm will be granted a stay if an analysis of limited depth results in a conclusion that there would be irreparable harm from immediate enforcement, that Qualcomm is reasonably likely to prevail on the merits, and that a stay is in the public interest. This hurdle is a lower one for Qualcomm than the actual appeal.

Share with other professionals via LinkedIn:

Friday, July 12, 2019

Ninth Circuit grant's Qualcomm motion to expedite appeal of FTC's antitrust victory

On Wednesday, the United States Court of Appeals for the Ninth Circuit granted Qualcomm's motion to expedite its appeal of the FTC v. Qualcomm antitrust ruling (this post continues below the document):

19-07-10 Order Granting Qua... by on Scribd

These are the two key paragraphs:

"The opening brief and excerpts of record are due August 9, 2019; the answering brief is due October 4, 2019; and the optional reply brief is due October 25, 2019."

"This case shall be placed on the first calendar available upon completion of briefing."

That's a tight schedule for such a complex case, but the FTC's litigation team has been very efficient and will probably be able to craft a great response to Qualcomm's opening brief in the eight weeks they will have.

This schedule slightly increases Qualcomm's chances of obtaining a partial stay of enforcement while the appellate proceedings are ongoing. The length of such a stay is part of the consideration. However, even with this schedule it may still be spring or so before the Ninth Circuit actually hands down a decision, given the scope and scale of this case.

By the way, Samsung filed a motion to intervene, but only with respect to the sealing of its highly confidential effective royalty rates under its 2018 deal with Qualcomm.

Share with other professionals via LinkedIn:

Thursday, July 11, 2019

United States Trade Representative launches investigation into France's digital tax for good reasons--and should look into copyright law, too

Yesterday the United States Trade Representative (USTR), Robert Lighthizer, announced, officially at the behest of President Donald Trump, the initiation of a Section 301 investigation (i.e., an analysis of whether a foreign government violates a trade agreement or acts unreasonably or discriminatorily against U.S. commercial interests) of France's digital services tax. The next formal step will be a Federal Register notice. This is what Mr. Lighthizer said:

"The United States is very concerned that the digital services tax which is expected to pass the French Senate tomorrow unfairly targets American companies. [...] The President has directed that we investigate the effects of this legislation and determine whether it is discriminatory or unreasonable and burdens or restricts United States commerce."

The President and the USTR are rightly concerned. France, with its almost plan-based highly-centralized economy and statist approach (= the opposite of Reagan's take that one has to get government out of the way) is an abysmal digital-industry failure, and TIMSS, the leading international math skills analysis, shows that French students are pretty much at a level with Third World countries, which means things are only going to get worse in the "Hexagon." Geographically, France is part of Europe; in educational terms, it's the worst country in the entire EU. By comparison, Singapore takes 25 times as many students (relative to the total number of students) to the top-performing level; Russia, ten times as many; and even the U.S., with its oft-criticized educational systems, seven times the number of France (again, this is a relative measure, so large countries don't have any advantage). While Germany performs 2.5 times as well as France (still far behind not only East Asia but also the U.S. and even a country like Kazakhstan), its politicans are incompetent and/or ideological enough to believe that France is their best partner in innovation policy, instead of trying to stay away from the failed French approach as much as possible.

There are two French computer game makers I think highly of: Ubisoft and (in the segment of "snackable" minigames) Other than that, I can't even think of a French technology product that would matter.

The French digital-tax initiative is ill-conceived because they don't want to tackle the real issue. The real issue is simply that the EU as a whole is a fundamentally-flawed supranational structure that does more harm than good. I sometimes recommend a great Wall Street Journal article (which doesn't even mention all of the problems), "Incredible Shrinking Europe."

Just like it was stupid and irresponsible in the first place to put a common currency in place without a common economic policy (as a result, the European Central Bank hasn't increased its interest rates even once in more than a decade, while the U.S. had half a dozen hikes during the same period), it was also incompetent and irresponsible to create a "Single Market" without some minimum tax standard or, in the alternative, an easy way to exclude members taking unfair advantage of this by positioning themselves as a low-tax access point to a market of 500 million consumers. As a conservative I'm all for tax competition, but fair tax competition and not just leeching.

The EU and the large member states of the eurozone are so poorly run that they didn't even seize the historic opportunity they had when Ireland needed a bailout. They could have conditioned the bailout on Ireland agreeing to some minimum tax standard. Obviously, leeches like Luxembourg (Juncker's country) wouldn't have liked this anyway, but Ireland is the #1 problem in this regard.

The Apple "state aid" case is the only crazy thing Mrs. Vestager did during her first term (other than that, I disagree only gradually, not fundamentally, such as with respect to some aspects of the Android case; I'm now looking forward to her second antitrust hammer--the final one for this term--coming down on Qualcomm soon; and on Twitter I repeatedly voiced the view that she was the best potential candidate for the presidency of the European Commission). That Apple-Ireland case has nothing to do with "state aid" and everything to do with "buyer's remorse" in the sense of the EU now seeing the problems that a Single Market without a common fiscal policy (at least a minimum tax standard) creates. Apple is not responsible for the EU's structural issues.

What France is doing with its digital tax is really odd. It's not a sales tax because there is no physical sale occurring in France when Facebook, for instance, displays an advertisement. Nor is it a tax on profits. Instead, France argues that because major digital platform companies are very profitable, foreign entities owe France a percentage of revenues attributable to the French market.

France wanted to make this happen at the EU level, but never got real traction as Germany was reluctant to support this with a view to potential backlash affecting its automotive industry. So France decided to implement something at the national level, and I'm glad the U.S., under its best president in decades, will seriously consider some retaliation in order to dissuade France from this idiocy. Maybe the U.S. International Trade Commission, with its investigative resources, will also be of help in the process. The top-listed candidate of Macron's party in this year's EU Parliament elections made it very clear that they view Google, Amazon, Facebook and Apple as enemies of the state, or collectively as the equivalent of a rival world power, all of which is downright insane.

While I'm not going to do any more copyright reform-related posts on this blog (maybe a new blog further down the road), I would like to just say that Articles 15 and 17 (previously Articles 11 and 13) of the EU Copyright Directive adopted this year are also the equivalent of a digital tax discriminating against U.S. Internet platform makers. France was the driving force behind Article 17 (upload filters), while Germany was more interested in Article 15 (link tax). In fact, France politically blackmailed Germany by threatening to block the Nord Stream 2 pipeline deal with Russia at the EU level if Germany hadn't supported Article 17 of the copyright bill. That is not a conspiracy theory. It was confirmed by reliable sources and reported by Frankfurter Allgemeine Zeitung, and it was obviously no coincidence that both the gas pipeline issue and copyright reform were the only two "A items" on the EU Council's April 15, 2019 agenda (so as to make it clear to Germany, which was having second thoughts, that derailing copyright reform would trigger some energy-related blowback).

The "upload filter" paragraph of the copyright bill is practically a digital tax because it creates a liability regime that is so strict that even the term "strict liability" the way it is reasonably understood in the U.S. would be an understatement. It gives enormous leverage to copyright holders (many of whom are European collecting societies and publishers) against major Internet platform companies in licensing negotiations. There are, as we all know, many very significant U.S. copyright holders (Hollywood, music industry etc.), and they'll benefit from this, too, but the European share of copyrighted works consumed in Europe is hugely greater than the European share of digital platforms used by Europeans. Therefore, this is another means of unreasonably and discriminatorily sucking money out of major U.S. Internet platform companies. It's like the digital tax, but indirect: collecting societies and publishers will initially receive the money, but obviously this will result in incremental tax payments in Europe, including France.

EU member states have some limited flexibility--a modicum of wiggle room--regarding the transposition of Article 17 into their national laws. France is already pressing ahead with the most draconian and unbalanced implementation imaginable, and other European countries may follow, though there's clearly less enthusiasm for this elsewhere.

I will follow the Sec. 301 investigation of the digital tax issue, and I strongly recommend to major Internet platform companies and the industry bodies representing them to raise the EU Copyright Directive, or at least its transposition into French law, in this context. It really is an indirect digital tax. Doing so now might dissuade other EU member states from adopting France's copyright extremism. While that is not what President Trump was elected for, it would happen to have significant benefits for European Internet users, too.

Share with other professionals via LinkedIn:

Tuesday, July 9, 2019

Qualcomm files motion with Ninth Circuit for partial stay of FTC antitrust remedies

Within only a few days of Judge Lucy H. Koh's order unsurprisingly denying Qualcomm's motion for a stay of the Federal Trade Commission's antitrust remedies, Qualcomm has taken this matter to the United States Court of Appeal for the Ninth Circuit (this post continues below the document):

19-07-08 Qualcomm Motion to... by on Scribd

I just wanted to be of service and share the document with you, though I lack the time right now to go into detail (I may do so later, most likely after Qualcomm's reply brief). Therefore, just a very few observations for the time being:

  • Originally, Qualcomm wanted Judge Koh to stay all aspects of the injunctive relief obtained by the FTC. Now Qualcomm focuses just on two parts: (i) the requirement to extend exhaustive SEP licenses to rival chipset makers, and (ii) the "No License-No Chips" part, which includes an obligation to renegotiate existing licenses.

  • On LinkedIn, the chairman of Orrick Herrington Sutcliffe's antitrust practice group, John J. "Jay" Jurata (whom I previously mentioned in connection with a paper on the UK Unwired Planet global FRAND rate case), noted an irony:

    "Even Qualcomm believes patent hold-up is real! 'If this Court does not grant a stay, Qualcomm will be forced to negotiate under the cloud of an injunction requiring it to accept terms to which it would not otherwise agree.'"

    I obviously remember Qualcomm's various policy statements and amicus curiae briefs arguing that SEP holders should be entitled to injunctions. The most outrageous one of those bashed Apple at a time when Apple was just a customer--not an adversary--of Qualcomm's, and was subsequently withdrawn. So Qualcomm has no qualms about others being "forced to negotiate under the cloud of an injunction requiring [them] to accept terms to which [they] would not otherwise agree."

  • What Qualcomm's attorneys, now led by Goldstein & Russel's Tom Goldstein, put front and center is that the FTC brought the lawsuit with only two commissioners voting in favor (at the time, there were only three commissioners), and that former Qualcomm attorney (in terms of his positions, forget the "former") and now-Antitrust Assistant Attorney General Makan Delrahim and FTC commissioner Christine Wilson disagree. So they're trying to discredit the case and the ruling, and I have my doubts that this will impress the Ninth Circuit, especially given Judge Koh's stellar reputation throughout and beyond that circuit as well as the fact that her judgment is simply in the global antitrust mainstream in light of other decisions in the EU and in Asia (with a second EU antitrust hammer having been unofficially announced by DG COMP to come down in the months ahead, possibly just at a time when the Ninth Circuit will be working on a decision on this motion).

Share with other professionals via LinkedIn:

Sunday, July 7, 2019

Nokia privateer Conversant keeps on losing: UK court declares patent invalid due to added matter

Conversant, a privateer asserting former Nokia patents against different industry players (though Nokia denies any economic interest or revenue share in Conversant), has suffered some significant setbacks lately in France (against LG) and the United States (against Apple). On Thursday (Independence Day), it also lost a UK case (this post continues below the document):

19-07-04 Conversant v Huawe... by on Scribd

Justice Richard Arnold of the England & Wales High Court (EWHC)--see this post on a panel speech he gave in Munich a few months ago--has ruled that EP1797659 on a "slow Mac-E for autonomous transmission in High Speed Uplink Packet Access (HSUPA) along with service[-]specific transmission time control" is invalid due to added matter. HSUPA is a 3G protocol. Huawei's primary invalidity contention was that the claims of the '659 patent were amended pre-grant so as to read onto a particular technology in HSUPA called Uplink DRX (DRX = discontinuous reception) by adding matter into the claim beyond the content of the application as filed. Under Art. 123(2) of the European Patent Convention, such additions are not allowed, and patent claims containing such added matter are invalid.

According to the judgment, Huawei "dispute[d] essentiality, and hence infringement, and counterclaim[ed] for revocation on the grounds of added matter, obviousness and insufficiency." With the added-matter allegation having been found to be meritorious, it's game over for this Conversant patent in this UK court.

In an attempt to piggyback on that terrible Unwired Planet precedent (which Justice Birss, not the far more balanced Justice Arnold, is responsible for), which the Supreme Court of the UK agreed to hear, Conversant is seeking a global FRAND rate determination against Huawei in the UK. But apart from whether the related appeal (which Justice Arnold was well aware of, but which did not prevent him from holding a purely technical trial last month anyway) will succeed (I hope and believe it will), Conversant needs to prevail on the merits of at least one patent claim in order to have a basis for claiming anything, be the regional scope global, multinational, national, regional, or local.

So where does Conversant's overall patent assertion campaign against Huawei stand?

It all stated in July 2017. The original claim form alleged that Huawei was infringing four Conversant patents:

  • EP1031192 on "packet radio telephone services";

  • EP0978210 on "connecting a multimode terminal to the network in a mobile communication system";

  • EP'659, which the judgment shown and discussed above has disposed of; and

  • EP1878177 on a "fixed HS-DSCH or E-DCH allocation for VoIP (or HS-DSCH without HS-SCCH/E-DCH without E-DPCCH)".

The first decision had to be made on Huawei and ZTE's jurisdiction challenge. Justice Henry Carr heard that challenge in early 2018. He held that there was no irreparable harm that Huawei and ZTE would suffer from the denial of a stay, however, provided that Conversant would be precluded, by way of a stipulation, from later claiming that Huawei and ZTE submitted to UK jurisdiction by continuing on the technical side. Conversant agreed, so the case continued.

In April 2018, EP'210 expired, and EP'192 had only six months left. Against that background, Justice Carr scheduled two technical trials: a "Trial A" for EP'177, and a "Trial B," further to which the judgment by Justice Arnold that is shown and discussed above came down.

Back then, Justice Carr already had a certain premonition. He stated that in the event of Conversant losing both of those trials (A and B) over allegedly standard-essential patents that had not yet expired, "then it [would be] very unlikely that a FRAND hearing [would] proceed."

With Conversant having lost Trial B (over EP'659), its fate now depends on Trial A. In fact, Conversant even agreed to delay entry of partial judgment, which means it can't immediately appeal the added-matter ruling.

Trial A was originally going to be about just one patent--EP'177--but Conversant amended that one further and injected two divisionals (EP3267722 and EP3197206) into the Trial A proceedings. In order to do so, Conversant had to bring the UK equivalent of a U.S. motion for leave to amend a complaint, which Justice Birss (who held a case management conference) granted.

An interesting detail regarding the three Trial A patents is that Conversant claims EP'177 and EP'722 are essential to LTE, while EP'206 is allegedly essential to UMTS. Originally, Conversant's UMTS patent in this dispute was EP'659, but that's the one that was just held invalid due to added matter.

Conversant is running out of ammo. Time is not on its side either. How come it doesn't fare better with those former Nokia patents? Maybe Conversant's management and investors simply lacked the knowledge they'd have needed to understand what they were buying. But it's also possible that Nokia itself overrated the relevant portfolio.

Share with other professionals via LinkedIn:

Thursday, July 4, 2019

Judge Koh (not unexpectedly) denies Qualcomm's motion to stay enforcement of FTC antitrust remedies

Yesterday, on the eve of the Fourth of July, Judge Lucy H. Koh of the United States District Court for the Northern District of California ruled on Qualcomm's May 28 motion to stay the enforcement of the Federal Trade Commission's antitrust remedies as well as on a motion by the FTC to strike material that Qualcomm sought to inject into the NorCal record from its settled SoCal litigation with Apple (this post continues below the document):

19-07-03 Order Denying Qual... by on Scribd

Like in that Bruce Springsteen song, "Just say goodbye it's Independence Day." Goodbye to the Northern District, I mean. Now there's nothing left to do in San Jose, and on to the Ninth Circuit.

The only thing that surprised me in this context was when someone told me on Twitter Qualcomm's stock was down 3% after hours on this news. At least the investors who talk to me, individually or on group calls, perfectly knew that this was the most likely outcome, especially since Qualcomm for the most part would have required Judge Koh to contradict her own ruling (a true opus magnum), at least implicitly. It wasn't 100% unthinkable that maybe some irreparable-harm argument might have gotten some traction, but I couldn't really find a totally pressing reason for a stay, for the reasons I explained in previous posts.

What's going to be different now before the Ninth Circuit is that there will be a panel of new judges. Judge Koh is still sitting on the district court, but only because Donald Trump's 2016 victory derailed her already fairly advanced nomination process. Frankly, it might even have been good for her in the short term because she then got this historic FTC case just a couple of months later. On the Ninth Circuit, she most likely wouldn't have had a similar opportunity, and she doesn't seem to care about titles or salaries as much as about doing high-quality meaningful work. But I really do wish her the best for being promoted as soon as possible at this stage. With so much partisan divide it may be difficult, and I totally agree with those Republicans who would like to bring more political balance to the Ninth Circuit (which has already happened under this President to some extent), but Judge Koh is so obviously a perfect choice for the next Ninth Circuit nomination and certainly not the kind of ideological judge because of whom Rush Limbaugh calls the Ninth Circuit the "Ninth Circus." She would be a great consensus candidate, and that would be a politically smart move for Republicans (not with a view to California, which they obviously won't carry anytime soon, but from a broader perspective). But that's a separate story, though it is related to this process at this procedural juncture.

The order doesn't specifically discuss the parties' (or the amici's) arguments. The decision is what it is.

As a secondary item, Judge Koh granted an FTC motion I hadn't previously reported on. The FTC moved to strike Qualcomm's opening slides from the Apple v. Qualcomm antitrust and FRAND trial in San Diego. While Qualcomm's lead counsel in that case, Evan "Fire!" Chesler, was delivering his opening statement and showing those slides to the jury, the settlement was already a done deal. In fact, the settlement was already announced on Apple's websites when he still had about 20 minutes left for that opening statement.

Antitrust AAG Delrahim's folks also tried to leverage that material in a footnote of their filing asking Judge Koh to hold a separate hearing on remedies. They're obviously Qualcomm's best friends in DC--they and FTC commissioner Christine Wilson, I mean.

So what is that about? It's all about diverting attention away from the real issues surrounding Qualcomm's business practices by trying to blame it all on a huge conspiracy between Apple and the FTC. Those opening slides contain statements that, if taken out of context, may be understood by many people to believe Apple was basically just an evil empire trying to squeeze a highly innovative supplier and licensor. So there were some Apple internal documents about a strategy designed to "hurt" Qualcomm and to "devalue SEPs." Like I said, if taken out of context, one may interpretin a certain way. Some important context was, however, given in January. Qualcomm treated Apple in such a way that Apple, in the aftermath of a "watershed moment" as an Apple executive described it, determined it had to fact because otherwise it was going to be at Qualcomm's mercy and Qualcomm was going to shamelessly exploit that leverage. (The question now is again to what extent Apple was at Qualcomm's mercy with a view to 5G modems that it had to settle despite having brought some fairly strong claims against Qualcomm's conduct.)

Procedurally, the problem there is that neither the FTC nor Apple ever got to react to that material, at least not in the FTC case. No discovery, no cross-examinations, nothing. That's why the FTC moved to strike those opening slides. The material quoted in them was available to Qualcomm long before, and if they had wanted, they could have discussed any of that in January (when the FTC trial took place).

As for the term "devalue SEPs," it's something that I've been accused of contributing to by a formerly significant smartphone maker in a private conversation outside the Mannheim court. The way I see it, "devalue" can have a nonjudgmental meaning and a slanted one. The nonjudgmental one is to bring a price down, and if you do that because the original price is outrageous, then there's nothing objectionable or reproachable about it. Of course, there is that common connotation that it's about bringing the price of something down below its fair value. That's the way Qualcomm would obviously like to interpret that quote from some Apple-internal document. But they apparently preferred to do so without testimony on what was really meant and why and how.

Share with other professionals via LinkedIn:

Wednesday, June 19, 2019

Qualcomm elaborates on its theories of irreparable harm from immediate enforcement of FTC's antitrust remedies

Originally, Qualcomm asked Judge Lucy H. Koh of the United States District Court for the Northern District of California to set an extremely tight briefing schedule for the San Diego chipmaker's motion to stay the enforcement of the victorious Federal Trade Commission's antitrust remedies pending an appeal to the Ninth Circuit. If the court had adopted that schedule, Qualcomm would even have been prepared to waive its right to file a reply brief in a support of its motion.

But Judge Koh declined to give the FTC only a very few days for its opposition filing--and Qualcomm, though it still could have waived its right to file a reply brief, elected to file a reply brief yesterday (not a single day before the deadline) and counter not only what the FTC wrote but also the amicus curiae briefs submitted by LG Electronics and ACT | The App Association (this post continues below the document):

19-06-18 Qualcomm Reply Iso... by on Scribd

That reply brief is far more interesting--because it is a lot more specific especially on the question of irreparable harm--than the original motion. While I would categorize some of Qualcomm's reply arguments as non sequitur material and consider some others outright smokescreens, that reply brief does contain food for thought.

The first footnote is, however, almost a concession that a stay of the entirety of the FTC's remedies may have been too much to ask for:

"Should the Court determine that the irreparable harm Qualcomm would suffer as a result of provisions (1) and (2) of the injunction does not warrant a stay of the entire Order, Qualcomm respectfully submits that the Court should enter a partial stay of only provisions (1) and (2) of the injunction."

Accordingly, the reply brief focuses on the requirements to (re)negotiate agreements, to sell chips to unlicensed customers, and to extend exhaustive SEP licenses to rival chipset makers. Those requirements flow from the first two provisions of the injunction the FTC secured--the only ones with respect to which Qualcomm, according to the FTC and also in my observation, even attempted to make an irreparable-harm argument in its original motion.

As for any potentially renegotiated license agreements, Qualcomm insists that any harm could not be undone should Qualcomm prevail on appeal. Qualcomm labels as a "red herring" the FTC's argument that Qualcomm would still obtain "fair value" for its SEPs. Qualcomm says the problem is not that those new agreements would be negotiated without the well-known "No License-No Chips" kind of leverage, but "because of the need to negotiate in the shadow of an Order that declares—erroneously, in Qualcomm's view—that Qualcomm's typical licensing terms are unreasonable." The latter is, by the way, what I told CNN within less than 24 hours of the ruling.

Qualcomm is concerned about licensees stopping royalty payments "under valid contracts" (though Judge Koh's order obviously serves to invalidate many such agreements), "even if temporarily," and mentions Huawei as an example.

Then Qualcomm attacks the FTC's suggestion that Qualcomm could, after a successful appeal, seek "damages for any past infringement" against those who (at least temporarily) stopped royalty payments. Qualcomm states something indisputable: a license is a defense to an infringement claim as 35 U.S.C. § 271(a), the statutory definition of infringement of a U.S. patent, applies only to those who practice a patented invention "without authority."

Qualcomm is furthermore worried that the non-discrimination provision of its FRAND licensing commitment or "most favored" provisions in some of its license agreements could ultimately drive Qualcomm's royalties to the "lowest common denominator."

Qualcomm, with a declaration by one of its licensing executives (John Han), alleges that it would be impossible to prevent irreparable harm by agreeing only on "short-term or interim licenses" (as the FTC called them) or (again quoting the FTC) "contractual provisions that would mitigate or eliminate any long-term adverse consequences to Qualcomm").

All of that needs to be considered, but none of it is convincing. The appeal will take time, but we're talking about roughly a year. Negotiations, however, also take a fair amount of time. The FTC v. Qualcomm decision doesn't require Qualcomm to accept a specific set of terms within a short time frame lest the company be held in contempt. I believe it just comes down to how much of a risk Qualcomm is willing to take when betting on a successful appeal. If Qualcomm really was sure that it would prevail, then it could insist on contractual provisions that protect its rights, and there would not be any contempt sanctions before the Ninth Circuit decides. And what's practically totally impossible is that Qualcomm could not only be forced to enter into a license agreement on unfavorable terms with a first licensee and that a second licensee would then have enough time to also get a better deal (or an adjustment under a "most favored" clause) before the Ninth Circuit renders a decision. And even if the appeal surprisingly took longer, anything that happens as a result of a decision that is being appealed would hardly serve as a FRAND benchmark in any other litigation. I would expect any other case to simply be stayed in that event.

What Judge Koh could do now is provide some guidance to Qualcomm on what structural options it does have even if a stay is denied (such as that Qualcomm would not act in contempt of the injunction if it insisted on contractual provisions that protect its rights, including retroactive adjustments, in the event of a successful appeal). Maybe Qualcomm is primarily hoping to obtain such clarification while formally moving for a stay.

With a view to LG Electronics, which supports the FTC's opposition to Qualcomm's motion, Qualcomm contradicts LG's description of the state of negotiations between the parties. Qualcomm notes that it "continuously supplied chips to LGE, without any interruption, throughout the negotiations"--but if it intended to do so going forward, the injunction against its "No License-No Chips" tactics wouldn't make a difference. Qualcomm mentions "a written offer to enter into binding FRAND arbitration with LGE to try and resolve the dispute, which included an express guarantee of chip supply during the pendency of the arbitration, but LGE declined that offer." As I've explained on various occasions, including my FTC v. Qualcomm trial coverage, arbitration isn't necessarily fair. Without the right parameters, it can give an extremely unfair advantage to a patent holder making out-of-this-world royalty demands since the licensee can't counterbalance a supra-FRAND royalty demand by proposing a negative royalty.

Qualcomm also argues that a requirement to sell chips to unlicensed OEMs would have profound implications due to patent exhaustion. Qualcomm rejects the FTC's argument that Qualcomm should simply "price[s] its modem chips to reflect the fair value of its patents" (which, by the way, pretty much every other chipset makers does, and even Qualcomm does so in some other business areas than cellular modems). Qualcomm argues that it would either lose chip sales or its ability to fully monetize its SEPs "so long as other chip makers are not licensed and thus do not price into their chip offerings the cost of Qualcomm's patents, this would leave Qualcomm in the untenable position of either charging much more for its chips than do its competitors (and therefore likely losing the sales), or reducing its chip prices so that, once again, they do not reflect the fair value of Qualcomm's patents."

The passage I just quoted may very well be the weakest one in that entire reply brief. If its chipset makers aren't licensed, then this is an apples-to-bananas comparison as device makers (as many of them--and competing chipset makers such as Intel--testified in the FTC litigation) would still have to factor in the cost of licensing Qualcomm's patents.

Finally, and as we all know, Qualcomm dreads the notion of having to extent exhaustive SEP licenses to rival chipset makers. But Qualcomm's argument for a stay of that particular provision comes down to what Judge Koh has already rejected at the merits stage: the question of efficiency of multi-level licensing (licensing cellular SEPs to baseband chipset makers and other patents to device makers) versus component-level licensing. Qualcomm says that neither OEMs nor chipset makers would jump to pay royalties, so there would be disputes over whether certain patents are practiced by particular products or not. Qualcomm describes this as a risk of "obstruction and delay," but that holdout-style argument does not appear to be sufficient to make a case for irreparable harm. Other patent holders face those challenges all the time (as the FTC noted in its opening statement back in January).

In that context, Qualcomm also demands "adjudicative comity" (respect for other jurisdictions) because its settlements with China's NDRC and the Taiwan Fair Trade Commission did not involve a requirement to extend licenses to rival chipset makers. I don't understand the comity issue here: if additional parties get licensed, it doesn't frustrate any policy or decision by other antitrust agencies. If the NDRC or the TFTC settled their cases without such a requirement, why would it hurt them if, say, MediaTek obtained a license as a result of the U.S. FTC case? I can't think of any negative effects on competition in those markets.

Part B of Qualcomm's reply brief addresses the likelihood of success of its appeal, with an emphasis on the need for a court to determine that violations are "likely to reoccur" before the FTC can be granted the injunctive relief it successfully sought here. What is missing here is a proposal for how an antitrust offender in a dynamic market could ever be enjoined from some behavior since the court would always have to set some sort of cutoff date.

Part C, finally, makes a national security-centric public-interest argument, with a particular focus on the decision by the Committee on Foreign Investment in the United States (CFIUS) to block Broadcom's attempted acquisition of Qualcomm. The FTC's position was that the CFIUS decision was unrelated to the behavior at issue in the antitrust case, but Qualcomm points to the CFIUS's position that "[c]hanges to Qualcomm's business model [as a result of an acquisition by Broadcom] would likely negatively impact the core R&D expenditures of national security concern." I tend to agree with Qualcomm that the content of the CFIUS letter is significantly more helpful in the public-interest context of the motion to stay than the FTC acknowledged. However, the CFIUS did not block Broadcom's hostile takeover of Qualcomm in order to enable continuing antitrust violations.

What I found interesting is that, with reference to ACT | The App Association's amicus brief, Qualcomm's reply brief also refers to the European SEP policy debate and, specifically, the two competing CWAs (CEN-CENELEC Workshop Agreements). One of the exhibits attached to Qualcomm's reply brief is CWA1, which Qualcomm and similarly-minded companies support, while there is far broader and stronger support for CWA2.

Share with other professionals via LinkedIn:

Tuesday, June 18, 2019

Anatomy of a patent case gone awry: appeals court's order to stay enforcement of Munich fake injunction published

This is a long-overdue follow-up to a post of two months back on an order by the Oberlandesgericht München (Munich Higher Regional Court) granting a motion by Apple to stay the enforcement of Qualcomm's illegitimate (for multiple reasons) Germany-wide injunction over EP2724461 on a "low-voltage power-efficient envelope tracker"--a patent that an opposition panel of the European Patent Office revoked last month because it shouldn't have been granted in the first place, not even in a narrower form (Qualcomm can and likely will appeal that decision). And Judge Lucy H. Koh's landmark FTC v. Qualcomm antitrust ruling came down that same week.

Of all the cases I've watched since I started this blog nearly a decade ago, what went wrong in this Munich case makes it the worst non-standard-essential patent case by a wide margin, just like the district court's Oracle v. Google rulings were the worst in any software copyright case and the Mannheim Regional Court, in 2012, set a negative example for how to handle a standard-essential patent (SEP) case when it totally failed to recognize Motorola Mobility's blatant antitrust violation by seeking to enforce SEPs after initially making bad-faith out-of-this-world royalty demands (a royalty on computers that was effectively more than a 100% royalty rate since Microsoft would have had to pay Motorola more than it typically earned per copy of Windows sold to an OEM). Apart from that, I've certainly seen--and keep seeing--very bad stuff coming out of the Eastern District of Texas on various occasions, but those weren't cases I followed closely.

The regional government of the state of Bavaria published the December 2018 fake injunction ruling, but I still haven't been able to find a public redacted version of the appeals court's order that tears the fake injunction into pieces, so I'm going to publish it here and now (this post continues below the document):

19-04-09 OLG München Ausset... by on Scribd

There's a host of reasons any single one one of which would have been sufficient for the appeals court to conclude that Apple was more likely than not to get the lower court's ruling overturned at the end of the appellate procedings (which were cut short by the Apple-Qualcomm settlement). The likelihood of success of an appeal is also extremely key to obtaining a stay in the U.S., but even more so in Germany, where irreparable harm is just a requirement and given less weight. When Apple brought its motion in December (one day after the lower court's ruling), there would have been a stronger irreparable-harm argument than after a workaround involving Qualcomm's baseband chipsets became known, but again, what mattered here was just the likelihood of success of the appeal.

Apple requested not only a stay for the duration of the appellate proceedings (which was granted) but also a stay for the period during which the appeals court considered the motion for a stay. The latter was denied. The order explains that such a short-term stay is granted only under the most egregious of circumstances, and the unavailability of products due to the enforcement of a patent injunction is insufficient to justify such a short-term stay in its own right.

The order does not make reference to antitrust or invalidity arguments. I suspect that Apple did pursue its antitrust defense (which the appeals court didn't have to reach, however, after granting the stay on an infringement-related basis), while it is possible that the invalidity defense was not stressed in the motion for a stay. At the time the lower court handed down the injunction, the EPO's preliminary opinion was that the patent was valid, though the preliminary opinion did mention that certain issues would have to be discussed (and indeed, the patent got invalidated in May).

As the injunction had been granted on an agnostic basis (without an actual infringement having been ascertained), the appeals court's order to stay enforcement does not--because it does not have to--provide clarity as to whether or not there was an infringement. The stay was ordered because the appeals court concluded that the original reasoning of the Landgericht München I (Munich I Regional Court) could not stand, and in that situation the patent holder no longer has a legitimate interest in enforcement regardless of whether (such as on remand) an infringement might be identified later on. In other words, there was no right-for-the-wrong reasons analysis. But it's highly unlikely at any rate that there was an infringement, given that the United States International Trade Commission (first its most experienced Administrative Law Judge and then the Commission, which is the six-member decision-making body at the top of the U.S. trade agency) and the United States District Court for the Southern District of California (on summary judgment, which is relatively unusual on the factual aspects of an infringement) had all the evidence and testimony before them and agreed with Apple's primary non-infringement argument.

The standard of review for the different parts of the lower court's decision to deem Apple's non-infringement defense insufficiently substantiated (which is why it allowed Qualcomm to prevail even without actual proof of infringement) is either the equivalent of "clear error" or "abuse of discretion" under U.S. law--and since the appeals court adjudicated a motion for a stay as opposed to a full-blown appeal, the hurdle for Apple was to show clear error or an abuse of discretion in such a convincing way that the appeals court was enabled to reach such conclusion on a basis comparable to a summary judgment.

The combination of the standard of review and the limited scope of the analysis is the reason why some of Apple's attack vectors didn't succeed. They might have succeeded under the framework of full-blown appellate proceedings, but not at this stage.

While I can understand that the appeals court afforded the lower court some deference, there is one argument that Apple made in its motion that really should have succeeded even under the given circumstances. The lower court should not have condoned Qualcomm's German lawyers' refusal to submit to the protective order that the same firm had negotiated with Qorvo, the maker of the accused envelope-tracker chip, in the Middle District of North Carolina. The appeals court cites to an unpublished ruling it made in an earlier case, and apparently the Munich Higher Regional Court is uncomfortable with the implications of a plaintiff's lawyers being precluded from discussing a piece of evidence (here, the chipset schematics obtained through a U.S. discovery proceeding for the stated purpose of using them--though it never happened due to Qualcomm's about-face--in the Munich litigation) with their clients' engineers. There wouldn't have been a contempt-style sanction if Qualcomm had refused to do so, but the lower court could and in my view should have drawn an adverse inference.

Between the lines of the related passages the Munich appeals court appears to be concerned about U.S. discovery rules and the related opportunities for access to documents, but also the restrictions that are imposed by protective orders, playing too big a role in German patent infringement proceedings. What I find disappointing is that the appeals court would even have that concern in a situation in which the relevant party (here, Qualcomm) itself initiated a U.S. discovery proceeding under 28 U.S.C. § 1782 for the specific purpose of presenting the material so obtained in Munich, and negotiated a protective order there, but then deems it opportune to keep such critical evidence out of the German case by way of the German lawyers declining to enter into the protective order. This is a matter of legal culture, of course, and I could name a number of federal judges in the U.S. (outside the Eastern District of Texas, obviously) of whom I'm sure they would never ever tolerate such extreme and obvious gamesmanship. Again, the appeals court might have decided differently at the end, but my reading of the passage relating to Qualcomm's about-face (regarding the presentation of the chipset schematics) is that there was, at a minimum, a rather strong tendency on the appeals court's part to accept Qualcomm's conduct.

Nevertheless, the outcome was the right one: the injunction got lifted. The key to this result was that the appeals court concluded Apple's initial defense (we're basically talking about the answer to the complaint) was good enough that Qualcomm would have had to do more to prove an infringement. The appeals court identified two errors on the lower court's part in this regard:

  • The court-appointed expert, the narrow scope of whose engagement was just a kind of plausibility check, had opined in the November 2018 trial that an alternative solution (without an "offset," which was the claim limitation at the heart of Apple's primary non-infringement contention) would be less efficient than the method covered by the patent. But inferior or suboptimal efficiency is separate from whether Apple met its burden of showing that an "offset" wasn't necessarily required. The lower court, however, erroneously concluded from the expert's opinion on inferiority that Apple had failed to show that the desired effect could possibly be achieved in the absence of an "offset."

  • Qualcomm's infringement theory, based on an inherently unreliable teardown report, involved the presence of a digital-analog converter (DAC). While Apple conceded that there was one around, its lawyers also stressed that it was not used in the mode of operation relevant to the alleged infringement. The appeals court has thankfully corrected the lower court's plaintiff-friendly error that consisted in a misreading and misapplication of the Rangierkatze (Shunting Trolley) ruling by the Bundesgerichtshof (Federal Court of Justice; the highest court in Germany for the substantive aspects of patent infringement and patent validity cases). In September 2018 I discussed Shunting Trolley in connection with another Munich Qualcomm v. Apple infringement case.

    In 2016, Judge Dr. Matthias Zigann (for whom this envelope tracker case here is really an outlier!) gave a great presentation (PDF) at a Federal Patent Court conference on predicting and applying the guidance provided by the Federal Court of Justice.In that one he discussed the difficulties involved, including threshold questions where the top court agreed with a defendant in one case and with a plaintiff in a structurally similar case. No doubt about those troublesome issues--but when it comes to Shunting Trolley, I really think it's a pretty straightforward one to understand to apply:

    In the narrowest sense, Shunting Trolley said that a manual containing a recommended mode of operation cannot cure an infringement that occurs regardless, be it by a user choosing a different mode, by environmental conditions, or even happenstance. The guidance resulting from that narrow issue is admittedly a bit broader. But even the German equivalent of a SCOTUS syllabus here still makes it clear that an infringement requires each claim limitation to be practiced. So whenever a patentee or a court suggests (as Qualcomm did in a different Apple case involving the Spotlight search) that Shunting Trolley provides a substitute for the infringement of each and every claim limitation, I would like to ask them: "Which part of 'each' don't you understand?"

    In the envelope-tracker case (unlike the Spotlight case), the issue was not the "each" but the question of whether a user could modify the product as to render it infringing. The Munich Higher Regional Court convincingly explains that Shunting Trolley does not have scope for an infringement in that Qualcomm v. Apple case, where the DAC would only come into play in the relevant mode of operation if an iPhone user modified the firmware (program code) executed by Intel's baseband chip, which then controls the Qorvo envelope-tracker chip in a certain way. Unlike the "each" question, this isn't a binary one. However, Shunting Trolley was about the manual operation of a brake, or even totally automatic events, and, therefore, lightyears away from reverse engineering and tampering with the firmware of a baseband chip (which presumably consists of millions of lines of code, by the way).

Against the background of those errors, the appeals court found that Apple's non-infringement argument was sufficiently substantiated; that Apple's request to reopen the proceedings should have been granted; and that the lower court should have allowed Mike Kay, the chief designer of the accused Qorvo chip (who waited outside the courtroom for 12 hours, in vain and possibly even in pain), to testify.

I submitted information on this extreme case of wrongful enforcement of a Germany-wide patent injunction to the government officials working on what may be an injunction-centric German patent reform bill, and I was positively surprised that they acknowledged my unsolicited input (I was not among the recipients of a related questionnaire) with a rather thoughtful answer.

Share with other professionals via LinkedIn: