Thursday, January 10, 2019

Seeking to defend "no license-no chips" in FTC trial, Qualcomm says it never actually cut off chip supply

When a witness testifies in FTC v. Qualcomm, questions related to different ones of the Federal Trade Commission's antitrust allegations and to different ones of Qualcomm's defenses are asked. But yesterday, on Day 3, two topics got particular attention: Qualcomm's "no license-no chips" policy (which this post is about) and its refusal to license rival chipset makers (which the next post will focus on).

While Qualcomm can't deny that it has a policy of supplying chips only to companies licensed to its patent portfolio, its lawyers have nevertheless come up with a multiplicity of attack vectors to make the case against the case against Qualcomm in this context:

They try to define the relevant market broadly enough to be able to claim they didn't have a monopoly.

They deny that the "no license-no chips" policy resulted in the acceptance of supra-FRAND patent royalties by licensees. Qualcomm argues that its standard 5% rate is FRAND anyway. The FTC will have a licensing expert testify to the opposite, and there's Huawei's testimony that Qualcomm's royalties alone account for 80%-90% of its total patent licensing cost, as well as similar testimony from other licensees. In an effort to argue that its royalty rate is accepted regardless of "no license-no chips," Qualcomm points to companies that, at different points in time, accepted that rate despite not being chipset customers at the time (in some cases because Qualcomm wasn't even selling chipsets yet).

One of the problems here is comparability, especially of very old license deal that involved foundational patents that have expired since. Foundational patents can be so valuable that even a small number of them gives a patent holder more leverage than a larger number of patents building on breakthrough innovations later (which makes them relatively narrow). And even if Qualcomm had (as it said it might have) already made FRAND pledges with respect to certain patents at the time of concluding a given license deal, the FRAND licensing obligation actually doesn't kick in until the standard has been adopted and someone requests a license in order to implement the standard.

Qualcomm's last line of defense is to claim that there was no bottom-line (rule of reason) anticompetitive harm. First, they categorically deny that there would have been anticompetitive harm even if some companies had accepted supra-FRAND royalties; but that's not really a mainstream position under the case law, no matter how badly Assistant Attorney General Makan Delrahim would like things to work that way. Second, they stress that they never actually did cut off, or would ultimately actually have cut off, chip supplies. Third, they say that whoever threatened anyone with a cutoff of chip supplies if someone didn't take a patent license (particularly, Eric Reifschneider) wouldn't have had the authority anyway to make a decision on chipset shipments. That point was stressed in particular by Qualcomm president Cristiano Amon.

At this point, I'm unconvinced of those defenses being strong enough to avoid a holding that at least some of Qualcomms behavior in the "no license-no chips" context was an antitrust violation and indeed caused anticompetitive harm (in the form of consumers ultimately paying higher prices for their wireless devices). I said "at this point" because there will be seven more trial days, and some testimony was sealed.

Let's start with the last point: the claim that people like Mr. Reifschneider didn't have the authority to carry out the threat. It's obvious that in a large organization like Qualcomm, and especially when there are formally separate entities in place (QCT for the chips and QTL for patent licensing), someone working in division A can't just make decisions for division B. However, that doesn't mean the guy from division A couldn't just call someone in division B, or take the matter to a higher level of the organization, and chipset supplies would be cut off just because someone didn't agree to Qualcomm's licensing terms.

The FTC showed a passage from a Qualcomm chip supply agreement that its trial brief (significant parts of which have been unsealed by Judge Lucy H. Koh) also mentions:

"For example, Qualcomm's supply agreement with LG Electronics provides that Qualcomm 'may terminate this Agreement if [LGE] is in default under the License' [...] Other supply agreements contain similar terms."

Footnote 4 even says that "[a] term to this effect appears in virtually all of Qualcomm's Component Supply Agreements."

Such a cross-over clause that creates a hard legal connection between a license agreement with one entity and a supply agreement with another entity makes it rather clear that those entities (QCT and QTL) did not operate totally indepenently, but very much kept the overarching interests of the parent company in mind.

Imagine you had an agreement with a telco and one with the postal service, and the postal agreement would say that if you're ever in default under your telephone contract, you might no longer get your mail. Would you then believe that those are independent entities? Obviously not. Nor would you feel safe because your account manager on the telco side isn't in charge of mail delivery.

With different witnesses having confirmed "no license-no chips" threats and even some Qualcomm-internal documents referring to such threats (also including a handwritten note such as "Eric [Reifschneider] constantly threatening [Motorola]"), Qualcomm's last-ditch defense is to argue that the threat was never actually carried out, nor would there have been a will to carry it out.

Mr. Amon, who was more forthcoming and likable than most of the other Qualcomm witnesses heard so far, explained the negative repercussions it would have had, such as that carriers relying on a particular Qualcomm-powered mobile device would have lost faith in Qualcomm if a termination of chipset supplies had happened. Also, other chipset customers might have drawn their conclusions (though they might not have had much of a choice anyway, at least in the premium segment)

I don't doubt that there would have been a very significant downside involved for Qualcomm. Mr. Amon explained that part very well. But the threats are now an established fact, and in the absence of cases in which Qualcomm would have continued chipset supplies for a long period of time despite someone refusing to take or renew a patent license, or despite someone actually challenging patent licensing terms (such as by seeking a judicial FRAND determination), the answer simply appears to be that Qualcomm never had to carry out the threat because it served its purpose.

To sum it up, Qualcomm's lawyers defend in multiple ways against the "no license-no chips" tying allegation, but no single one of those attack vectors appears strong at this stage. The parties' expert witnesses haven't been heard yet, and the economists will talk about market definition and FRAND royalty levels. What really doesn't look like it could solve the problem for Qualcomm is the notion that only because a threat was so effective that it never actually had to be carried out, or that those who made the threat lacked the authority to carry it out, or that there would have been some backlash if Qualcomm had done so.

Based on what's been said and presented so far, it's my impression that stories of licensees having accepted more or less the same royalty rates even without buying chipsets at the time (or even without Qualcomm offering any at the time) are something the FTC must counter effectively. And I believe it will be able to do so. Also, even if some people accept exorbitant royalty rates, those rates still aren't FRAND.

I particularly liked one thing the FTC said in its opening statement last Friday. The FTC's lead counsel in this case, Jennifer Milici, said that Qualcomm may well point to companies that don't respect intellectual property and refuse to take a license, but if Qualcomm's patents are as valuable as Qualcomm says they are, Qualcomm need not be afraid of proving the value of those patents through infringement litigation in cases where it may be necessary. The FTC doesn't deny that it isn't always easy to successfully enforce patents, but--the FTC argues--Qualcomm can't violate the antitrust laws only because it seeks to avoid what every other patent holder must do.

What is not at issue in this trial, but what I'd like to add nonetheless: it's absolutely key here to let patent holders prove the value of their patents through litigation. Arbitration is not a substitute for litigation unless both parties voluntarily (without undue pressure of the "no license-no chips" kind) choose arbitration, under mutually agreed-upon parameters, over litigation in Article III courts and the ITC. A few months ago I explained this in the context of Huawei v. Samsung.

Share with other professionals via LinkedIn: