Thursday, February 2, 2017

Qualcomm's two mutually-reinforcing monopolies: SEP thicket, baseband processor chipsets

In late December, the Korea Fair Trade Commission held Qualcomm in violation of antitrust laws, and after a reader pointed me to what appears to be Qualcomm's unofficial translation of the decision, I promised "further discussion here at a later stage." Then, a couple of weeks later, the United States Federal Trade Commission sought injunctions against Qualcomm (a complaint was filed in the Northern District of California). A little later, Apple brought its own lawsuit against Qualcomm (see PatentlyApple's post, which contains the complaint, according to paragraph 4 of which Apple claims to have been "overcharged billions of dollars on Qualcomm's illegal scheme" and now "seeks to recover its damages").

It appears that Samsung played a more active role in Korea, as the KFTC decision lists an expert who participated in the investigation on Samsung's behalf (Professor Sang-Seung Yi of the School of Economics of Seoul National University) but not one for Apple. In the U.S., Apple has now been more proactive as it filed its own complaint against Qualcomm, though I wouldn't be surprised to see a similar one by Samsung.

While we're on the subject of Apple and Samsung, or #appsung, just a quick update. If you don't care about design patents, please just click here.

I've been watching the dispute between those two companies for almost six years and most of the time I agreed with whomever was the defendant. Since Samsung dropped its own standard-essential patent (SEP) assertions against Apple, the main concern has been about Apple's assertions of design patents and non-standard-essential utility (technical) patents against Samsung. In December, Samsung prevailed in the Supreme Court, though the ultimate outcome of that dispute remains to be seen. Apple filed a statement in late December asking the Federal Circuit to uphold the original judgment, while Samsung ideally wants a new trial or, at a minimum, new appellate proceedings (briefings and oral argument).

I wouldn't want Apple to win affirmance of the original judgment against Samsung. Even though the Supreme Court ruling should enable future defendants to avoid a total disgorgement of infringer's profits over a design patent infringement and Apple's argument is now all about Samsung having allegedly failed to present evidence regarding the "article of manufacture" question, the signal would be that draconian damage awards are still available, and that would have chilling effects and encourage extortion. I'm fairly optimistic that Samsung will get a remand to the district court simply because there will have to be a damages retrial anyway (with the same design patents at issue again), so the Federal Circuit may very well consider it more efficient to just have the design patent damages question addressed there.

An industry issue

For now, there is every indication that Apple and Samsung, despite Apple's still-disappointing design patent damages claims and Samsung's once-disconcerting royalty demands and injunction requests over SEPs, agree on the huge Qualcomm issue--and so many others agree with them, including Intel and the Brussels-based Fair Standards Alliance, whose members include (among others) Google and several automotive and other Internet-of-things companies. Antitrust enforcers in multiple jurisdictions agree as well, but they don't always act forcefully enough to bring about change.

Over time the different Qualcomm proceedings will make plenty of news and bring interesting facts to light. I'll write about this from time to time, depending on how interesting the revelations are. For now, while one can already find a number of interesting tidbits in the various complaints and decisions (also including the Chinese NDRC's Qualcomm order), let's focus on the biggest issue:

Qualcomm has a patent-based monopoly as well as monopoly power (one might argue it's not a monopoly in a strict sense but "merely" market superdominance) in the baseband processor chipset market, and beyond appearing to abuse such monopoly power to extract supra-FRAND royalties and anticompetitive concessions, Qualcomm is (based on what certain regulatory agencies have held) leveraging each monopoly to reinforce the other.

This is a vicious circle for the mobile industry, and I hate to even think about all the money that I have indirectly paid to Qualcomm as I've bought roughly half a dozen Samsung Galaxy phones and almost as many iPhones over the years. Qualcomm deserves to get fair compensation for its contributions to innovation, but the vicious circle I just mentioned must be broken, lest it turn into a spiral with Qualcomm potentially gaining control over additional smartphone components. Monopolies are always a problem, but monopolies that spawn additional ones are a worst-case scenario for competition and innovation.

SEPs are monopolies by definition--but the problem is even bigger here

Courts and regulatory agencies in different jurisdictions (including the U.S. FTC and the KFTC) have recognized that even a single patent, if it is truly standard-essential (meaning you can't implement the standard without licensing or infringing the patent), confers monopoly power on its holder. A very few outlier opinions that denied this on the basis of more than one patent holder owning SEPs to a specific standard were plain ridiculous as a patent is not a right to do something but a license to sue someone, so even if you own one patent or have a license to one patent, it won't serve as a defense in case you infringe someone else's patent.

One SEP can be powerful (as a Motorola expert once said, "it only takes one bullet to kill"), but if your only concern as an implementer of a standard was a single patent, you would at least have the chance to evaluate the patent, identify prior art to get it invalidated if necessary, and analyze whether the patent actually is essential to any implementation of the relevant standard.

Not so with an entire patent thicket. In 2010, told Samsung that it owned an entire "thicket of patents" around the iPhone. Depending on your definition of a patent thicket, it did, but the portfolio Apple was referring to back then is dwarfed by those thousands of Qualcomm patents declared (by Qualcomm itself, which means court would still have to verify) essential to wireless standards.

Baseband processor chipsets: Qualcomm's competition is somewhere between a heavily-endangered and an extinct species

What regulators have found is that Qualcomm's position in the market for baseband processor chipsets--the basic component that handles the low-level (i.e., basic) communication of the device with the base stations of mobile network operators--is extremely dominant. Without a baseband processor chipset, it's impossible to make a phone call, but even the dumbest dumbphone has one, which shows that it takes Samsung a whole lot more to build a Galaxy smartphone and Apple a whole lot more to build an iPhone.

It's obviously not impossible from an engineering point of view to build chipsets that implement different cellular standards. The market is huge. So why is there a lack of competition? After reading multiple decisions and complaints, I believe it's because of the way Qualcomm has made it next to impossible for others to compete in certain market segments. If a company like Intel struggles to stay in this business at all, despite its ability to develop and manufacture even the most complex chips, the answer must be non-technical.

Qualcomm holds patents related to different wireless standards and makes chips implementing different standards, but its position in the baseband processor market is particularly strong with respect to CDMA (code division multiple access), a standard that the existing network infrastructure of Verizon and Sprint, a large chunk of the market in the U.S., supports. It's a 2G (GSM-level) standard, but backwards compatibility is still key. This gives Qualcomm particular leverage--and an incentive to prevent CDMA from being replaced by competing standards that might work over the same existing network infrastructure.

No license, no chips--no chips, no affordable license

One of the antics that the FTC and apparently also other regulators don't want to put up with anymore is that Qualcomm has a "no license, no chips" policy: unless you take a patent license from them, you won't get to buy their chips. This is a huge issue given that the principle of patent exhaustion says a maker of a product can't allege infringement of its patents by any direct or indirect ("downstream") customer. Basically, Qualcomm wants to get paid for patents that have already been exhausted. (According to what I read in a complaint, Qualcomm also wants to get paid for patents that have expired...)

The even bigger problem is that Qualcomm allegedly charges far higher license fees if a company doesn't exclusively use its chips. Apparently, the way it works is that Qualcomm charges fairly high fees but pays some of the money back (Apple's complaint describes this as "rebates" in spite of Qualcomm apparently rejecting the term) in exchange for exclusivity commitments and other concessions that may be similarly anticompetitive.

But the biggest issue in my view is this: Qualcomm refuses to grant licenses to other baseband processor chipset makers. How could such conduct not be discriminatory, i.e., the opposite of FRAND (fair, reasonable and non-discriminatory)? Qualcomm is not the only SEP holder to refuse to license chipset makers. Ericsson even explained publicly why it likes that strategy. The Qualcomm cases in different parts of the world may be the most splendid opportunity in a long time to have that practice be declared anticompetitive and non-FRAND. By the way, as I already discussed here almost five years ago, Qualcomm once changed its corporate structure just in an effort to avoid patent exhaustion.

A cornerstone of Qualcomm's strategy is that it has positioned itself as the primary clearing house for cellular patents. It requires its licensees to enter into a covenant not to sue Qualcomm's customers. Apple benefitted from this in the early stages of its dispute with Samsung (patent exhaustion), but patent pools with FRAND licensing terms are a much better vehicle to achieve this than a private, allegedly predatory enterprise that has simply leveraged monopoly power to elevate itself to this.

For device makers, this means that if you don't buy from Qualcomm, you're exposed to patent infringement claims that Qualcomm may bring against you; but if you do buy from them (which includes that you need to take a license), you'll be protected from infringement assertions by a long list of cellular SEP holders.

SEPs should be licensed on FRAND terms, and chipset makers should compete on the technical and economic merits. Regulators, customers, the sole remaining competitor (Intel) and industry associations are now trying to get Qualcomm to offer licenses on FRAND terms to everyone, including its competitors and its competitors' customers.

This blog has been pro-FRAND since the year I started it (2010). While I'm not as prolific a writer nowadays as I used to be, I'll try to follow the various Qualcomm cases. In light of Apple's latest financials (with record iPhone sales), I also intend to talk about the economic dimension of this issue based on the information available to me, i.e., regulatory decisions, publicly-accessible complaints, and SEC filings.

Share with other professionals via LinkedIn: