Yesterday evening by Central European Time, the European Commission's Directorate-General for the Internal Market (DG GROW) published (expand the part entitled "Group of experts on licensing and valuation of Standard Essential Patents") the long-awaited standard-essential patent (SEP) expert group report.
When the expert group was formed in 2018, I was extremely skeptical. It looked to me like the deck was stacked against SEP implementers. I was very vocal about that concern, but the report that has been published is all about pluralism, not a particular agenda. It reflects a diversity of views and ideas, and an Ericsson executive wrote an official dissent, which I regards as a very positive sign as I tend to disagree with Ericsson on SEP matters. Maybe I was wrong in 2018 to suspect an effort to rubberstamp pro-SEP-holder policies--or, which I can't rule out for lack of knowing what exactly happened in all those expert group meetings, the dynamic changed at some point. One way or the other, my concerns were unfounded, a fact that I simply have to acknowledge in all fairness.
On a similar note, no matter how much I may disagree with EU internal market commissioner Thierry Breton on SEP matters, I like his Digital Markets Act proposal (a bill that will hopefully be improved in the further legislative process, but which is a great starting point and tackles the most important tech policy issue at this time). I hope he'll fight for that potential game changer of historic proportions even harder than he's been fighting for Nokia and Ericsson's patent monetization interests.
That said, I had to read only one sentence of the part I'm most interested in--licensing in the value chain--to find myself in some fundamental disagreement. This is just my first post on the SEP expert group report (there'll likely be more). Part 3.2 is found on pages 35-54 of the full-length version of the report. Since the FTC v. Qualcomm and Apple v. Qualcomm complaints were filed about four years ago, that topic has come up again and again on this blog, and I even organized a conference on it in 2019.
While the expert group report is merely meant to juxtapose different schools of thought, the Licensing in the Value Chain section uses Nokia/Ericsson terminology that I reject because it serves to frame the debate. Terminology is so crucial in political and legal contexts. I call on those advocating SEP implementers' interests never to adopt the terms "license to all" and "access to all"--at least not without clarifying that you actually reject them.
In the first paragraph of Part 3.2, the expert group report actually defines the terms accurately and pragmatically:
"license to all": "SEP holders are under an obligation to grant FRAND licences to entities at any level of the value chain requesting such licences"
"access to all": "[SEP holders] can select the level in the value chain where they grant FRAND licences" (followed by a "while" subclause I don't endorse)
SEP holders have many shrewd policy experts working for them. They know how to frame a debate, and they understand the role terminology plays. If you win the war over terminology, you can make the fight over substance an uphill battle for your adversaries. Implementers have their strengths, too, but if they acted as strategically as their counterparts, they'd adamantly refuse those misnomers "license to all" and "access to all":
"license to all": That term unrealistically implies that everybody--from baseband chipset maker to network access device (NAD) maker to (in the automotive industry) telematics control unit (TCU) maker to end-product maker--would conclude a license with a given SEP holder. But that's a purely theoretical outcome and won't happen in practice thanks to patent exhaustion: if a license is taken high up in the value chain, the downstream is fine (example: Huawei's license deal with Sharp, through exhaustion, covers most Mercedes cars). Those vertical supply chains are like pyramids, so it would just take a few license deals at the thin top to take care of everyone.
"access to all": This incorrectly sounds like everyone's taken care of, without everyone having to sign a license agreement.
In the context of value-chain licensing, the term "access" is nebulous at best and mendacious at worst. I wrote above that I rejected a "while" subclause. That one says "other actors in the value chain only have the right to access the technology." What an empty word that has no basis in patent law:
You can access any patent by downloading the relevant document from a patent database. But once you practice a patent, such as by making a product that practices a product claim or contributes to the subsequent practicing of a method claim, "access" is simply not a legal concept. You can have a license agreement; the patent rights may have been exhausted; or there may be other contractual arrangements either between yourself and the patentee (a covenant not to sue, which in the U.S. triggers exhaustion) or yourself and a third party (in which case some jurisdictions will make it harder than others for you to enforce your rights as a third-party beneficiary).
"Access" in this context is simply a euphemism for what is more accurately described as "significant legal uncertainty for one or more levels of the value chain." It means that you may be theoretically covered by an agreement, the terms of which may not even be known to you due to confidentiality clauses, and which you can't enforce. It means that you don't have freedom to operate.
Part 3.2 of the expert group report more or less acknowledges that if SEP holders choose the licensee, they'll go after the bottom of the value chain, and if the implementer side has the choice, the license is likely to be taken higher up in the value chain. But that's just what tends to happen, not what necessarily happens. There may be situations in which an end-product maker considers it efficient to take a license at the end-product level, and others in which a SEP holder would rather manage fewer license deals with chipset makers than many more deals with end-product makers.
So what is it really about?
I propose that we non-judgmentally and truthfully--and in perfect consistency with the expert group report's own definition--distinguish from now on between "implementer's choice" and "patentee's choice" in this context. "Implementer's choice" means any implementer, at whatever level of the value chain, is entitled to a license. Of course, "any" may also mean licensing requests from more than one level of the supply chain, but realistically those who are covered by patent exhaustion wouldn't have a need to sign a license agreement. "Patentee's choice" steers clear of suggesting legal certainty by "access to all" when there are serious issues at unlicensed levels of the supply chain, as patent exhaustion works only downstream, not upstream.
Back in 2018, I was profoundly concerned that a predominantly Qualcomm-aligned SEP expert group would endorse and promote patentee's choice. Presumably, some members of that group sought to accomplish just that. Fortunately, they met resistance. They couldn't impose their will, and the outcome is now an "agree to disagree" chapter spanning 20 pages that serve to crystallize the issues.
One of many things I like about the chapter on component-level licensing is the acknowledgement that different licensing levels are prevalent in different industries, and that there are exceptions (such as component-level licenses covering smartphones--for instance, even Ericsson granted Qualcomm a chipset-level SEP license--or end-product-level licenses for cars).
It's also true that some standard-setting organizations required the participants in their standard-setting efforts to make a clearer commitment to the implementer's choice approach than others, though it's debatable whether ETSI's FRAND pledge could be interpreted in any number of ways. I've heard some very compelling argument from French law scholars that on the bottom line even ETSI's IPR policy entitles component makers to an exhaustive license.
On the technical side, I'm unconvinced of claims--which are often heard from SEP holders and their allies and also found in the expert group report--according to which it would be hard to determine at what level of the value chain a SEP is actually practiced. The baseband chipset is the master brain of cellular communications. About a year ago I showed that Nokia's patents-in-suit against Daimler are easily identifiable as baseband chipset-level patents. If one focuses on where the allegedly inventive step in any of those claimed inventions resides, the answer is simply the baseband chip. A patent litigator who frequently represents both Daimler and Qualcomm confirmed this.
The chapter on value chain licensing does make a number of valid points. For example, it might indeed be desirable to enable the companies in a given value chain, or even companies at the same level of such a chain, to engage in some coordination in order to determine the most efficient approach to licensing (Proposals 29-33, pages 46-49; I don't mean to endorse any particular one of those, but they raise important questions). Concerns over potentially violating competition law through illegal forms of cooperation appear to complicate coordination especially in the automotive industry, and in some situations the upside of allowing them to work out the most efficient solution in a given context might outweigh the downside.
Proposal 35 on page 53, Limited exhaustion -- in rem, comes down to limiting exhaustion to a field of use. This would require new patent legislation. So far, patent exhaustion is based in case law, but legislation would have more flexibility. The EU's Single Market is large enough that an EU-wide patent exhaustion regime would be relevant, and might lead other jurisdictions to take similar initiatives. However, exhaustion always has a territorial aspect. Given the chaos surrounding antisuit (and especially "multiple-anti"-suit) injunctions and extraterritorial royalty determinations (Unwired Planet), what I would consider most helpful would be a new international patent law treaty, ideally at the WHO level or at least between the U.S., the EU and key Asian jurisdictions, that would address exhaustion, antisuit injunctions, and extraterritorial patent-related determinations (such as patent valuation).
Finally, Proposal 37 on page 54, Different connectivity rates (so as to enable a patentee to "different royalties for the different chips depending on the connectivity rates of these chips"), is also very interesting. Many Internet of Things products actually just need minimal data throughput: they must be connected, but the amount of data they send may be limited to reporting a temperature once every five minutes. I'm not against price differentiation if it's based on objective criteria. I just can't see why a high-end smartphone that makes use of, say, 4G connectivity just like any low-end "dumbphone" should entitle SEP holders to a higher per-unit royalty only because of larger and higher-quality screens, more memory, and other technical characteristics that are unrelated to connectivity.
In practical terms, it's much more likely, however, that the Dusseldorf Regional Court's referral of certain component-licensing questions to the Court of Justice of the EU will bring clarification than that any policy initiative in this context is going to make an impact.
Share with other professionals via LinkedIn:
