Wednesday, October 3, 2018

Qualcomm's Mannheim Steamroller is stuttering: patents too young to defeat Apple in tech-savvy court

When Qualcomm started filing patent infringement actions against Apple (and, by practical extension, Intel) in Mannheim and Munich last year, it hoped to gain leverage through injunctive relief over non-standard-essential patents in the largest EU economy. In February, Qualcomm's German counsel exuded confidence and predicted an injunction by the summer. Summer is gone, and based on where things stand after all the hearings and trials held so far (up to yesterday's Mannheim trial, which this post is mostly about), Qualcomm's next chance to make serious headway against Apple in Germany may not even come until next summer (or, maybe, spring). In the meantime, Qualcomm's FTC trial in San Jose and possibly also the Apple trial in San Diego (a pretrial conference will be held this year, but no firm trial dates are on PACER) will go forward.

Each patent is different, though Qualcomm's asserted patents do have some overlaps (several of them are asserted against an envelope tracker made by Qorvo), but there is a pattern here. The first Qualcomm v. Apple (and, by extension, Intel) Mannheim trial was already symptomatic: Qualcomm felt forced to stipulate to a stay over a huge validity issue. This was a first win for Apple, Intel, their attorneys at law, and especially their patent attorneys from Samson & Partner. At the second Mannheim trial, held the week before last, the court didn't propose a stay, but took note of an independent opinion handed down by the Swedish patent office, according to which that patent should never have been granted in the first place. At yesterday's trial over EP3036768 on a "layout construction for addressing electromigration", Presiding Judge Dr. Holger Kircher again suggested a stipulated stay and reminded Quinn Emanuel's Dr. Marcus Grosch that the rather strong presumption of validity that issued patents normally enjoy in German infringement proceedings does not apply to patent claims that the patent holder feels forced to modify substantively in response to prior art presented by a defendant. At that stage, it's a new game and simply remains to be seen whether an examiner will or will not be persuaded of patentability.

This time around, Qualcomm declined because it firstly wants Apple to provide more specifics about whether a Qorvo 8081 chip was publicly-available prior art on the patent-in-suit's summer 2013 priority date. Other invalidity contentions don't turn on that question, but Qualcomm wouldn't give up easily. A decision has been scheduled for mid February, and in the meantime the parties will make additional filings. A stipulated stay is still a possibility. Judge Dr. Kircher told Qualcomm that the way they're now trying to defend a narrowed patent claim (more about that in a moment) was tantamount to "cherry picking" (in this case, tailoring a claim to the accused Qorvo chip). The three-judge panel is furthermore concerned that the imported and newly-created claim limitations through which Qualcomm seeks to differentiate its claimed invention from the prior art are "devoid of technical substance" and merely relate to an arrangement of circuitry that is simply not a means of achieving the stated objective of the patent (addressing electromigration).

I don't want to go into too much technical detail here (at the risk of someone saying that this post is equally devoid of technical substance). If you do intend to, I have a recommendation: copy the text from the patent document, paste it into a new document, and when you see all those long sequences of sentences that have more or less the same structure, highlight the parts that change from one sentence to the next (boldface and/or italics). Otherwise it's very hard to read.

The bottom line is that Qualcomm's patentability theory fails (even over other prior art than that older Qorvo chip than the one Qualcomm is currently suing Apple over) if one assumes that a person of ordinary skill in the art would simply know that orthogonal and parallel arrangements are viable alternatives.

But it may also fail, regardless of prior art, because of an impermissible broadening or extension. Qualcomm's new claim is a combination of claim 1, claim 3 (dependent on claim 1), claim 5 (dependent on claim 3), claim 7 (dependent on claim 1, which means it actually wasn't meant to be combined with claims 3 and 5 that are on a different branch of the tree), and a new claim limitation referred to as "7.3" (unfortunately neither the court nor counsel explained what exactly it was about), which may not even have adequate support in the specification.

What is undoubtedly creative, inventive and potentially innovative here is how Qualcomm's lawyers created this unique blend of claims, including claims that apparently were never meant to be combined with each other in the first place. However, the court's concern is about a lack of technical substance.

What's behind all of this, well, contortion? Qualcomm's counsel explained that the new claim--which is the one over which Qualcomm is seeking an injunction, and it's a fallback position as an amendment (if needed) in the parallel nullity proceedings--was meant to address the prior art presented by Apple and the Qorvo 8081 chip from two angles: prior art and (private) prior use. The prior use defense didn't get much traction, by the way, but it most likely won't be outcome-determinative. A stay looms large, and this patent is most likely going to go down the proverbial, non-electrical drain.

Why is Qualcomm having such problems with its patents-in-suit (its Munich patents are also of dubious validity)? One common characteristic is that they're fairly recent patents. The younger a patent is, the more prior art, and the greater the potential for prior-use rights. Conventional wisdom in German patent infringement lawsuits is that you want to have a certain enforcement window left, but even if it's less than a year, that may be all it takes to get the upper hand. The likes of Nokia typically assert patents that are fairly close to expiration, as did Microsoft against Motorola. Only Apple was different because its iPhone patents were, obviously, not very old at the time it started enforcing them against Android device makers.

At those German Qualcomm trials, there's usually a lot of talk about Apple's license-based defense. Sometimes one of Apple's contract manufacturers, Pegatron, is licensed, while Qualcomm claims the others are not. Yesterday, Qualcomm argued that the patent fell outside the capture period of all of its license agreements with Apple's contract manufacturers. I have no idea whether the capture period relates to the priority, filing, publication or grant dates of those patents. At any rate, in order to avoid being defeated on licensing grounds, Qualcomm needs to assert young patents--patents that were filed, published, or issued after the relevant capture periods. And then those patents are inherently far more suspectible to invalidity and prior-use defenses.

In the prior-use context, it became clear once again that the Mannheim court is actually, like many patent-specialized courts, quite sympathetic to the interests of patent holders in enforcing their rights. They don't take stays over validity concerns lightly. But Qualcomm has no one--certainly no judge--to blame but its own web of agreements and the lamentable state of the European Patent Office, which simply grants way too many junk patents because its leadership treats high-volume applicants as key account customers whose demand for more and more patents must be served. SUEPO, the trade union of EPO examiners, has warned against the consequences of this many times. The worst consequence we can all observe here is legal uncertainty.

While I doubt that Apple's antitrust defense will be reached in this particular case, it's worth noting that Judge Dr. Kircher acknowledged substantial progress made by Apple's counsel (yesterday's lead counsel on technical issue was Klaus Haft, but the antitrust defense is argued in all those cases by Dr. Frank-Erich Hufnagel) since the "grand evil scheme" allegations made the week before last. Not only does it appear that Apple has made quite some effort to address the court's concerns over divergent markets and subjective theories, but in the case tried yesterday there was a much closer technical nexus between the asserted patent and Qualcomm's desire for a baseband chipset monopoly because the Qorvo envelope tracker at issue is directly relevant to the operation of the Intel baseband chip used in the German iPhone 7 and 7 Plus devices at issue in that case.

However, the court also noted that Apple seemed to contradict itself by arguing that the chip layout at issue involved a trivial design question while arguing in the antitrust context that short of switching to Qualcomm chips and a different envelope tracker, its products would have to be removed from the German market for a long (by the standards of this industry) period of time if Qualcomm enforced the patent-in-suit. Apple's counsel explained the differences between triviality at the engineering stage versus all the development, testing and manufacturing steps and logistics involved before a modified product can be sold. Also, let's face it: parties are perfectly allowed to raise actually or seemingly contradictory defenses because if one of them succeeds, the other one won't be reached anyway.

Apple also leveraged the final initial determination on Qualcomm's first ITC case, which came down last Friday. There was a bit of a terminology dispute between Dr. Hufnagel and Dr. Grosch. While Dr. Hufnagel correctly stated that Administrative Law Judge Pender arrived at the unusual conclusion (by ITC standards) that Qualcomm should not obtain an exclusion order (U.S. import ban) for public-interest reasons, he also described this as an antitrust finding. Dr. Grosch temporarily even wanted a particular representation to be put down in writing (in the official minutes of the court session), which German lawyers do when they believe their adversary is making an untruthful representation.

Dr. Grosch is absolutely right that the ITC's "public interest" determination does not have an equivalent in German patent law. But in this particular case, the distinction is academic at best because the ITC's concern over Qualcomm seeking to get rid of its sole major competitor, Intel, is substantively and effectively a competition concern. It is not the kind of concern that some companies (including at least three Quinn Emanuel clients) have raised, such as potential implications of patent enforcement for emergency responders or rural communities.

Qualcomm's counsel also wondered why Apple would on the one hand say that the final initial determination wasn't available yet but would quote from it. The explanation is simply that a notice--like an executive summary--was published on Friday, but it takes a couple of months before complete rulings (often about 100 pages long) have been redacted.

It's still not like an antitrust ruling in Apple and Intel's favor was looming large, but step by step, with guidance from the judges, the case is getting more interesting.

Yesterday's trial also raised interesting legal questions in connection with prior use, and I'd be happy to talk about them some more--maybe in connection with some other case, or in case it does become outcome-determinative in this case, contrary to where things appear to be heading.

Share with other professionals via LinkedIn: