Wednesday, May 22, 2019

BREAKING NEWS: Federal Trade Commission wins antitrust case against Qualcomm in Northern District of California

The Federal Trade Commission (FTC) of the United States has won the first round of litigation against Qualcomm. Judge Lucy H. Koh of the United States District Court for the Northern District of California has just found that the San Diego-based chipset maker violated the FTC Act, and has ordered the following remedies:

(1) Qualcomm must not condition the supply of modem chips on a customer’s patent license status and Qualcomm must negotiate or renegotiate license terms with customers in good faith under conditions free from the threat of lack of access to or discriminatory provision of modem chip supply or associated technical support or access to software.

(2) Qualcomm must make exhaustive SEP licenses available to modem-chip suppliers on fair, reasonable, and non-discriminatory ("FRAND") terms and to submit, as necessary, to arbitral or judicial dispute resolution to determine such terms.

(3) Qualcomm may not enter express or de facto exclusive dealing agreements for the supply of modem chips.

(4) Qualcomm may not interfere with the ability of any customer to communicate with a government agency about a potential law enforcement or regulatory matter.

(5) In order to ensure Qualcomm's compliance with the above remedies, the Court orders Qualcomm to submit to compliance and monitoring procedures for a period of seven (7) years. Specifically, Qualcomm shall report to the FTC on an annual basis Qualcomm’s compliance with the above remedies ordered by the Court.

Even though some of the remedies requested by the FTC were found too vague, the FTC has practically prevailed on all counts. This is a resounding victory for the U.S. competition enforcement agency over Qualcomm.

The judgment in the narrowest sense is just one paragraph (PDF):

On May 21, 2019, the Court entered its Findings of Fact and Conclusions of Law, and found that Defendant violated the Federal Trade Commission Act. ECF No. 1490. Accordingly, the Clerk shall enter judgment in favor of Plaintiff. The Clerk shall close the file.

Judge Koh's underlying findings of fact and conclusions of law, however, span 233 pages (this post continues below the document):

19-05-21 FTC v. Qualcomm Ju... by on Scribd

Judge Koh is amazing. Just like I noticed at the January trial, she's totally focused on the facts and not interested in lawyers' rhetoric. And her ruling explains the complex commercial and technical issues in this case as well as applicable antitrust case law in a very understandable form (though it's obviously still a complex matter).

I am now still studying the findings of fact and conclusion of law in detail, and will continue to update this post accordingly, but this is a resounding victory for the FTC--and totally consistent with my coverage of and commentary on the January trial in San Jose (here and on Twitter). And a defeat for Assistant Attorney General Makan Delrahim, whose subordinates made a last-minute submission that failed to persuade Judge Koh.

What was an even greater failure for Qualcomm was the extreme degree to which its senior executives' testimony contradicted their own handwritten notes, emails, and presentation slides, including but not limited to the question of whether Qualcomm explicitly threatened device makers with a disruption of chipset supplies unless they agreed to certain patent licensing terms. As a result, "the Court largely discounts Qualcomm's trial testimony prepared specifically for this litigation and instead relies on these witnesses' own contemporaneous emails, handwritten notes, and recorded statements to the IRS." For Qualcomm's statements to the IRS, let me refer you to a previous blog post. Here's what Judge Koh noted beyond those contradictions:

"In addition to giving testimony under oath at trial that contradicted their contemporaneous emails, handwritten notes, and recorded statements to the IRS, some Qualcomm witnesses also lacked credibility in other ways. For example, Dr. Irwin Jacobs (Qualcomm Co-Founder), Steve Mollenkopf (Qualcomm CEO), and Dr. James Thompson (Qualcomm CTO) gave such long, fast, and practiced narratives on direct examination that Qualcomm's counsel had to tell the witnesses to slow down. [...] By contrast, when cross-examined by the FTC, each witness was very reluctant and slow to answer, and at times cagey."

This discrepancy between their answers to their own counsel as compared to the FTC's questions is something I also observed in my trial commentary on Twitter and on this blog. One detail that I noted was that Mr. Mollenkopf generally wears glasses, but at the trial he took his glasses off when the FTC had questions, only to complain that he couldn't quickly read some text that appeared in a smaller font on a screen.

The way Judge Koh explains Mr. Amon's self-contradiction even suggests between the lines that witnesses testifying under oath shouldn't do what he did.

Judge Koh then discussed market segmentation and agrees with the FTC that the CDMA modem chip markets (CDMA was invented by Qualcomm and is key to its leverage) and the premium LTE modem chip markets are appropriate antitrust markets. Based on this definition of markets in which Qualcomm held monopoly power during the relevant period, Judge Koh analyzes the evidentiary record, which is unusually rich for an antitrust case. These are the OEMs with respect to which Judge Koh analyzes Qualcomm's conduct:

  1. LG Electronics

  2. Sony

  3. Samsung

  4. Huawei

  5. Motorola

  6. Lenovo

  7. BlackBerry

  8. Curitel

  9. BenQ

  10. Apple

  11. VIVO

  12. Wistron

  13. Pegatron

  14. ZTE

  15. Nokia, and

  16. smaller Chinese OEMs.

In numerous cases Qualcomm threatened with a disruption of chipset supplies unless OEMs accepted its patent licensing terms, and there were various agreements under which OEMs paid a higher patent royalty when using third-party modem chips than Qualcomm's products (one example: "In 2003, Qualcomm charged Huawei a reduced royalty rate of 2.65% if Huawei purchased 100% of its CDMA modem chips for use in China from Qualcomm, but a 5-7% royalty rate if Huawei purchased CDMA modem chips from a Qualcomm rival."). This is the problem of Qualcomm's two mutually-reinforcing monopolies (patents and chips) that I first explained two years ago. Qualcomm even refused to provide samples for technical integration and testing purposes without a patent license in place.

With such a rich and powerful body of evidence, it's going to be hard for Qualcomm to persuade the appeals court (the Ninth Circuit) that the facts are favorable to its defenses. And the above list of industry players shows that Qualcomm-aligned op-ed authors, analysts and organizations completely overstated the weight that Huawei (the Chinese bogeyman) and Apple really have here. It's about an issue affecting an entire industry, not just two companies at the heart of conspiracy theories Qualcomm's allies planted in the media.

There is only one respect in which Qualcomm's dealings with Apple play a central role here: exclusive dealing. Qualcomm originally accomodated Apple's requests for lower effective patent royalties through a "Variable Incentive Fund" (VIF) contingent on volume commitments:

"Apple's receipt of the hundreds of millions of VIF funds also depended on the volume of chips Apple purchased from Qualcomm. [...] If Apple purchased more than 115 million Qualcomm modem chips from October 1, 2011 to September 30, 2012, Apple received the full amount of VIF funds for that year. [...] If Apple purchased fewer than 80 million Qualcomm modem chips in that period, Apple received no money. [...] In future years, Apple needed to increase purchase volumes to 125 million annual units and then to 150 million units to receive the full amount of VIF funds for that year."

Later, Apple had to agree to total exclusivity, where any shipment of a non-negligible quantity of devices with non-Qualcomm modem chips on board would have made them lose certain benefits going forward and entitled Qualcomm to a clawback, and that is the basis for one of the FTC's monopolization claims. But other than that, and Apple's importance to competition (because chipset makers gain a lot of credibility from being chosen by Apple), there is nothing special about Apple that wouldn't just be consistent with how Qualcomm treated everyone else. Arguably, Qualcomm treated some others a lot worse, and Apple wasn't the only company with which Qualcomm had exclusive agreements in place. The ruling also mentions similar arrangements with such companies as BlackBerry and Motorola. This is Judge Koh's summary of Qualcomm's anticompetitive conduct vis-à-vis Apple:

"In sum, Qualcomm engaged in anticompetitive conduct with respect to Apple by (1) refusing to sell Apple modem chips or even share sample chips until Apple signed a license; (2) eliminating a competing standard supported by Intel; (3) attempting to require Apple to cross-license its entire patent portfolio to Qualcomm; and (4) and using Qualcomm’s monopoly power to enter exclusive deals with Apple that foreclosed Qualcomm’s rivals from selling modem chips to Apple from 2011 to September 2016."

After the analysis of those dealings with device makers, Judge Koh's ruling addresses the problem of Qualcomm refusing to grant exhaustive (meaning that the downstream is protected) SEP licenses to rival chipset makers. Last year she entered summary judgment in the FTC's favor with respect to Qualcomm's obligations under FRAND commitments made to two U.S. standard-setting organizations (ATIS and TIA). But the new decision is now broader because it is based on antitrust law as opposed to specific agreements.

The obligation to license its SEPs at the component level is likely going to be Qualcomm's #1 priority with respect to the appeal, and it then won't help Qualcomm that the decision points to Qualcomm's own understanding of FRAND and the obligation to license all comers:

"Then, in 2000, Steve Altman (then a Qualcomm lawyer, and later Qualcomm President) complained in a letter to Motorola that Motorola was not licensing its modem chip SEPs to Qualcomm despite 'Motorola's commitment to the industry to license its essential patents.' [...] More recently, Qualcomm has repeated that understanding of FRAND. During a 2012 meeting with the IRS, Eric Reifschneider (QTL Senior Vice President and General Manager) explained to the IRS that when SEP holders participate in SSOs, 'as part of that you often have to make commitments that you will, you know, make that technology available to people who want to make products that practice the standard.' [...] Eric Reifschneider explained that refusing to license a rival modem chip supplier is 'not a great, you know, position to be in in terms of defending yourself against, you know, claims that you’ve broken those promises to make the technology available.'"

Qualcomm's actions may speak even louder than its words:

"Moreover, in a 1999 email, Steve Altman (then a Qualcomm lawyer, later Qualcomm President) stated to Marv Blecker (QTL Senior Vice President) that Qualcomm had licensed modem chip suppliers: 'ASIC licensees pay royalties to QUALCOMM at 3% with no minimum dollar amount.' [...] As the Court will explain below, Qualcomm later stopped licensing rivals because Qualcomm decided that it was more lucrative to license only OEMs."

And, as everyone in the industry knows, Qualcomm's inbound licensing differs from its outbound licensing in this regard: they do require their own licensees to grant exhaustive licensees that benefit Qualcomm's customers. For an example, "[Qualcomm witness Fabian] Gonell conceded that Qualcomm has an existing license from Ericsson, [...] which Christina Petersson (Ericsson Vice President of Intellectual Property) confirmed."

In the chipset-licensing context, testimony from Nokia and Ericsson ultimately didn't benefit Qualcomm here, simply because those companies used to take the very opposite position originally when they were dealing with Qualcomm:

"Nokia and Ericsson's contemporaneous documents and statements contradict Nokia's and Ericsson's self-serving and made-for-litigation justifications for refusing to license modem chip suppliers."

That is Judge Koh's polite and professional way of saying, "I don't buy that BS."

After a thorough analysis of the anticompetitive impact of Qualcomm's exclusive agreements with Apple and others, Judge Koh finds that "Qualcomm’s royalty rates are unreasonably high" (Section G). That holding will hurt Qualcomm, though it's common sense. They simply charge a lot more than the rest of the industry--from the perspective of some major OEMs, more than the rest of the industry combined.

What's particularly important to mention in the royalty context is that Judge Koh disagrees with Qualcomm's royalty base (the entire phone, though capped at $400). She notes that it's not modem chips that drive the value of smartphones, and she sees Qualcomm's conduct fundamentally at odds with the Entire Market Value Rule (EMVR): "Qualcomm's use of the handset device as the royalty base is inconsistent with Federal Circuit law on the patent rule of apportionment." Just like she did in GPNE Corp. v. Apple a couple of years ago (a fact I repeatedly mentioned, though hardly anyone else pointed to it), she stresses the need to base a royalty on the smallest salable patent-practicing unit (SSPPU).

As I predicted during and after the January trial, it was really industry testimony more so than expert testimony that mattered in the royalty context. And Judge Koh wasn't impressed by Qualcomm's economic experts Professor Nevo and Snyder ("It makes little sense to evaluate whether conduct 'reasonably appears capable' of causing anticompetitive harm [...] by ignoring evidence of that conduct altogether.") or with one of the FTC's experts, Mr. Lasinski ("the Court does not rely on Lasinski's testimony") any more than I was, as it appears. Interestingly, the FTC's economic expert Professor Carl Shapiro (Berkeley) isn't even mentioned once. I liked his testimony in January, but ultimately, industry testimony trumps expert testimony in this case. Professor Shapiro's testimony may have had some persuasive impact nonetheless, but apparently isn't needed here from an evidentiary point of view.

In the analysis of actual anticompetitive harm caused by Qualcomm's business practices, Judge Koh shows a very telling Qualcomm-internal slide according to which the company was well aware of antitrust implications of its behavior (click on the image to enlarge; this post continues below the image):

Patent exhaustion is mentioned at multiple points. Qualcomm's separation of patent royalties from chipset sales, and the games Qualcomm plays on that basis, are key, and the Supreme Court's Lexmark ruling on patent exhaustion (which came down a few months after the FTC filed its complaint) ups the ante for Qualcomm.

This is Judge Koh's summary of her conclusions on anticompetitive conduct and harm:

"In combination, Qualcomm’s licensing practices have strangled competition in the CDMA and premium LTE modem chip markets for years, and harmed rivals, OEMs, and end consumers in the process. Qualcomm's conduct 'unfairly tends to destroy competition itself.' [...] Thus, the Court concludes that Qualcomm's licensing practices are an unreasonable restraint of trade under § 1 of the Sherman Act and exclusionary conduct under § 2 of the Sherman Act. [...] Therefore, Qualcomm's practices violate § 1 and § 2 of the Sherman Act, and that Qualcomm is liable under the FTC Act, as “unfair methods of competition” under the FTC Act include 'violations of the Sherman Act.'"

This decision, which Qualcomm will obviously appeal (and they will try to get the injunctive remedies stayed, especially with respect to the renegotiation of existing license agreements), is an unbelievable success for the FTC's litigation team led by Jennifer Milici and Daniel Matheson. They did some amazing work and put on a really strong case. And they've prevailed to an extent that I would describe as "100% minus a rounding error."

Judge Koh's ruling is a piece of art. I enjoyed reading it, though it's a bit hard to read a few pages, then add details to a blog post, and then resume reading. So I'll re-read it, and probably do some follow-up posts in the coming weeks and months--and I'll follow the appellate proceedings, of course.

(By the way, this blog was--to the best of my knowledge--the first medium worldwide to report on and publish the decision.)

Share with other professionals via LinkedIn: