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: