Saturday, January 19, 2019

Qualcomm temporarily switched into settlement mode on Day 7 of FTC trial with unhelpful arbitration proposal

Yesterday (Friday, January 18) was the first full day of Qualcomm's case-in-chief as it is seeking to defend itself against the FTC's apparently very strong antitrust case. Of course, Qualcomm's legal team consisting of one of the most prestigious firms of the United States (Cravath Swaine & Moore), a firm very well-respected for its representation in Bay Area jury trials (Keker, van Nest & Peters), and a firm with a particular strength in patent matters (Morgan, Lewis & Bockius), previously had other opportunities to make its point: in an opening statement, through cross-examination, and the first few hours of Qualcomm's case-in-chief (on Tuesday).

Good lawyers--and these are partly absolutely great lawyers--can sow the seeds of doubt about pretty much anything; if necessary, even about the Earth being round. It's not that they don't score any points or have nothing to say that might give the court some pause, but let's always remember the legal standard as well as who the fact finder is. In a criminal defense trial with a layperson jury and the beyond-reasonable-doubt standard, Qualcomm would have a decent chance of finding at least some jurors who would vote for acquittal, even if only because it's easy to create smokescreens in a context that requires a certain degree of expertise to understand.

But the fact finder here is none other than Judge Koh. On the first six trial days, she didn't ask questions during testimony except in one case where it wasn't clear to her how Qualcomm was seeking to impeach a witness with a certain cross-examination strategy. Yesterday, however, Qualcomm was at some point basically trying to relitigate through testimony some subissues of her summary judgment on rival chipset licensing (though, to be fair, the related subissues are technically relevant again in connection with the FTC's position that any FRAND licensing obligation for SEPs gives rival chipset makers a right to a license). It was the first point--and so far still the only one--in this trial where Judge Koh indicated she didn't see the point in something that was said.

In an ideal scenario for Qualcomm, they wouldn't merely sow seeds of doubt but dismantle the FTC's case to such a degree that no reasonable fact finder could decide the case against them. We have three more trial days (not counting closing arguments on February 1), so anything could still happen, but if Qualcomm had some silver bullets of the kind it had against the "Lazynski" report, why haven't we seen them, after 7 of 10 trial days? This is again just an alternative-universe analogy, but if this were a FRAND determination case at the end of which a licensee would have to pay precisely the rate determined by Mr. Lasinski and Qualcomm would have to content itself with the resulting amount, then there would be strong reasons to side with an appellant because the report just isn't anywhere near precise enough (it may still be just about good enough to confirm what common sense says, which is that Qualcomm's rate is out of line in light of such data as this chart). But that's just one part of the web that the FTC has woven, and like the Internet, where data can be routed differently if a node fails, the FTC has very cleverly provided the court with so much evidence and testimony in its favor that the FTC's case is pretty resilient. There isn't a single point of failure of the kind that Qualcomm could easily exploit.

Qualcomm's case-in-chief generally appears to be falling short of what they need to turn this around. Surprisingly, the FTC even managed to get more mileage out of Qualcomm's CTO's testimony, through a short but highly effective cross-examination, than Qualcomm itself. After some lengthy tech talk that had little bearing on the issues in the case, the FTC got Dr. Thompson to confirm an internal email in which he had suggested that Qualcomm use its leverage over Apple, while Apple still needed CDMA in parts of the Northern American, Chinese and other markets, before Apple would challenge, or "chip away at," Qualcomm's licensing business.

Some of the videotaped depositions were presented after Judge Koh sealed the courtroom. Qualcomm played additional segments of the depositions of witnesses we heard on previous days. What was shown in public included some answers by Lenovo's Ira Blumberg that make his testimony a bit weaker with respect to "No License-No Chips," but not to the extent that the related claim by the FTC would suddenly make no sense anymore. The same applies to a Qualcomm lawyer and licensing executive, Fabian Gonell, claiming--in an effort to defuse the "final hand grenade" at the end of the FTC's case-in-chief--that an internal emali exchange according to which some shipments to Sony Mobile were put on hold over licensing disagreements resulted from a misunderstanding. We'll see how credible that evasive explanation is in Judge Koh's eyes. Even if one buys Mr. Gonell's explanation, it still means that Qualcomm's chip division apparently understood the situation to be one in which a licensing negotiation affected chip shipments.

Generally, the FTC is in good shape with respect to No License-No Chips, but my unsolicited advice to them is to develop a good response to Qualcomm's argument that companies accepted consistent licensing terms even without that kind of pressure (such as because they signed deals before Qualcomm had such leverage, or because they simply weren't Qualcomm customers). I think the FTC has some critical homework to do: it must show that those deals were in many cases simply not comparable. That even includes Qualcomm's new deal with Samsung, where patent licensing just part of a broad, multifaceted commercial partnership involving chips and manufacturing services and the parties could have made concessions in one area to optimize a different part of their story. Moreover, the FTC can argue that even if some accept a rate, the rate still isn't necessarily FRAND, especially when there's so much industry testimony that Qualcomm's rates are considered too high, not only by numerous licensees but even by other licensors such as Ericsson. If some overpay voluntarily, it doesn't mean anything; and if Qualcomm was able to negotiate with those not buying their chips but point to what large players had paid, it doesn't make a supra-FRAND rate more reasonable. It just makes Qualcomm's strategy mor successful.

Qualcomm called an Apple witness, chip engineer Matthias Sauer, who originally worked for Infineon's mobile chipset division in Germany, which got acquired by Intel, and later got hired by Apple, and showed a videotaped deposition of an ST Ericsson employee. The purpose of that part of Qualcomm's case-in-chief was to show that the exclusive arrangements with Apple under the 2011-2012 and 2013-2016 agreements didn't cause actual anticompetitive harm because companies like ST Ericsson and Intel couldn't have met Apple's requirements during that period anyway. In a fast-paced business like cellular baseband chipsets, six years is a long time, and as we heard from witnesses (most vividly from Intel's Aicha Evans), suppliers and OEMs work hand in hand on a daily or at last weekly basis, so I don't believe no one could have met Apple's requirements during all of that time if Apple hadn't faced contractual penalties ("clawback" etc.) that made it commercially prohibitive to source baseband chips elsewhere. However, Qualcomm will continue to argue that the FTC bears the burden of proof regarding actual anticompetitive harm, not just potential/hypothetical harm.

After so many trial hours I've formed my own opinion on how important the different ones of the FTC's four core issues in the case are. I've believed all along that simply requiring Qualcomm to license its SEPs on FRAND terms at the chipset level would solve a lot of problems at the same time. It would mean that there are licensees who cannot be threatened with "No License-No Chips" because they make, but don't buy, chips. It would up the ante considerably for Qualcomm's ability to extract vastly supra-FRAND royalties. And by strenthening the competitiveness of more chipset makers in all segments (particularly premium baseband chips), device makers would be less susceptible to pressure form Qualcomm and, conversely, less receptive to incentives.

"No License-No Chips" is an issue, and the FTC has a solid case there on the merits, with Qualcomm presumably focusing already on the question of the appropriate remedy, arguing that it no longer has monopoly power.

The exclusive deal with Apple was a problem at the time, but it's not as easy to prove actual anticompetitive harm as it would be in a case involving commodities like paper or oil.

After closing argument we can all think and talk about those issues again. For the remainder of this post I'll focus on two fallacies espoused by Qualcomm yesterday:

  • One of Qualcomm's own witnesses was asked about the percentage of the end price of a $1,000 phone (obviously alluding to the iPhone for the most part) that Qualcomm's current SEP royalty demand represents. Qualcomm says it wants 3.25% of the end price of the device, but capping the royalty base at $400, which means that $13 is the maximum per-device cost. Qualcomm's witness gave counsel for Qualcomm the mathematically correct answer (1.3%), but that's not going to impress Judge Koh. It's a rather iPhone-centric perspective, but the biggest problem here for Qualcomm is that Judge Koh ruled in GPNE Corp. v. Apple a couple of years ago that the royalty base for wireless SEPs is the smallest salable patent-practicing unit, which is the modem chip--and what Judge Koh found in FTC v. Qualcomm on summary judgment makes clear that, by inference, the same rule applies here as well.

  • Since a joint motion (filed in October) to delay a summary judgment ruling (an invitation Judge Koh declined immediately), we haven't heard about settlement in this context. We've heard that Apple hasn't discussed a potential settlement with Qualcomm in a number of months, but the FTC case is separate from that.

    Cravath's Gary Bornstein, who may be the thought leader among Qualcomm's external lawyers on the case, brought up the FTC-Motorola consent decree on standard-essential patents when questioning Qualcomm's Mr. Gonell. That was the closest thing to a settlement proposal that one could imagine in that situation (a trial, not a mediation).

    He elicited testimony according to which Qualcomm would have been, and remains, willing to arbitrate license terms with Apple, Huawei, and others. Apart from not being nearly as good an idea as it may seem at first sight, it's simply unrelated to the merits of the FTC's claims. That's why I interpret it as a settlement proposal in open court (I did so immediately on Twitter).

    My opposition to coerced arbitration is longstanding, and I most recently voiced it in connection with Huawei v. Samsung.

    Qualcomm also mischaracterized the FTC-Motorola consent decree (PDF on FTC website) as a settlement that favors arbitration. I've re-read it now (after several years), and a licensee who doesn't want to go to arbitration could simply obtain a judicial FRAND determination instead. Qualcomm said yesterday that Apple told the United States District Court of the Southern District of California that such a determination (between Apple and Qualcomm) was "illegal," but it was just about subject-matter jurisdiction. Again, reading the FTC-Motorola consent decree is instructive. That decree made it clear that licensees are free to challenge the validity and infringement of patents. The disagreement between Apple and Qualcomm in the San Diego case was not about a FRAND determination for SEPs that a court of law would have held valid and infringed by Apple, but about Qualcomm wanting a portfolio determination for tens of thousands of SEPs without the court evaluating a single one (!) of its SEPs. Whenever Apple challenged any Qualcomm patents in that case, Qualcomm did whatever it took such as a covenant not to sue) to deprive the court of subject-matter jurisdiction and Apple of a justiciable controversy.

    The FTC case is about a lot more than just rate-setting. No rate-setting mechanism would get other chipset makers (such as Intel) as license; no rate-setting mechanism would preclude Qualcomm from leveraging "No License-No Chips" before the rate-setting process would even begin; and it's entirely unrelated to "incentives" paid for exclusivity or for badmouthing competing technologies.

    But even in a rate-setting context, coerced arbitration is not an option. The solution is to firstly establish the merits (infringement of valid, non-exhausted patent) in court and to then determine a rate in a court of law unless both parties voluntarily (without the threat of an injunction or a disruption of chip supplies or whatever other evil) agree on arbitration.

    I generally wouldn't recommend to licensees to arbitrate. Let's assume a hypothetical scenario in which one would assume that it's 80% likely that a court of law sets a royalty between $1 and $2 (so let's focus on the middle, $1.50), but the licensor wants $15. Knowing that arbitration gravitates toward the middle (a bell curve with a peak in the middle), the licensor could go into arbitraiton with a $30 demand--but the licensee would have to propose a negative royalty (-27%) in order for $1.50 (the likely outcome in a court of law) to be the middle between the two. A negative royalty is impossible, so the licensee could at best argue that the correct price is somewhere around $0.50 or $0.75, but in most cases, the arbitration would then result in something far above the actually fair amount of $1.50 only because of the licensor having taken an outrageous position in the beginning.

Qualcomm's white flag is yet another sign of weakness, but not a roadmap to a real solution. The trial will continue on Tuesday (the 21st is a public holiday in the U.S., MLK Day, thus Monday will be trial-free).

Share with other professionals via LinkedIn: