There was a lot of discussion about use cases during the Licensing Executive Society – Silicon Valley Chapter (“LES-SVC”) program titled “The Ubiquitous World of Internet of Things (IoT) Connected Devices – and the Coming Tsunami of Patent Licensing,” (“Program”).[1] You might reasonably ask me, what is a use case? As far as I can tell, it is something that SEP licensors have made up to justify charging more money, sometimes much more money, to license the same patents for different types of products.
Let me explain it with this example. You go to the pet food store to buy dog food. As you are waiting in line, the pet shop owner quizzes the person at the counter about what kind of dog they own and where they got the dog. It turns out that they own a small mutt which they got at the SPCA. The pet shop owner tells them that they owe $3 for the dog food. The next person in line wants to buy the very same amount of the very same type of dog food. The pet shop owner asks the second person what kind of dog they have and where they got it. It turns out the second person has a Labrador which they also got from the SPCA. The second person hands $3 to the pet shop owner but the pet shop owner says the second person owes $5 for the very same amount of the very same dog food the first person purchased for $3.
Now it is your turn - you want to buy the very same amount of the very same dog food as the other two people. You live on a ranch. You tell the pet shop owner that you have a working cattle dog which you got from (an ethical) breeder and trained yourself so that it could work on the ranch. The pet shop owner tells you the pet food costs $20. When you ask why it is so much more than the others paid, the pet shop owner tells you that you need to pay more money for the very same dog food because your dog is more useful and thus more valuable. Because you are using that dog food to feed a more useful and valuable animal, the dog food is worth more to you. The pet shop owner does not acknowledge that one of the reasons your dog is more useful and hence more valuable is because you trained it up yourself, expending a lot of your own time and energy.
This is obviously not how we do business in the real world. But, in the world of patent licensing, particularly SEP licensing, somehow the idea of a use case, i.e. that it is reasonable to charge a differential in royalty amount based on the value of the end product, has taken hold among some people.
During the Program, the Avanci representative talked quite a lot about use cases. I did not pluck the numbers above entirely from thin air. These are the numbers Avanci is seeking for licenses to the 4G/5G patents in its licensing program depending on whether the product being sold is (a) a smart meter in which case a 4G license is $3; (b) an EV charger in which case the license fee varies between $5 - $13 for a 4G license depending on the output power, whether there is an infotainment systems and whether it also has WiFi enabled connectivity; or (c) a connected car in which case the Avanci rate for a 4G license is $20.
I always understood that the baseline for valuing any patent is that the patent holder is entitled to a royalty only on the value of its contribution to the technology and is not entitled to a royalty on anyone else's contributions to that technology. Thus, any methodology for calculating a reasonable royalty should distinguish between the value contributed by the patent at issue and the value contributed by all other features, functions and attributes of the product. For a SEP, this should be a two-step process. First, one should determine what the standard as a whole, i.e. the “technology”, is worth (often called the aggregate value). Second, one should determine the value of a given SEP within the value of the standard as a whole. Each SEP licensor is entitled to no more than the portion of the value of the standard which has been contributed by its SEPs.
By introducing this concept of “use cases,” Avanci and certain other SEP licensors have abandoned this common sense, real-world way of doing business. It is no wonder they are trying to push for a royalty based on the value of the end product – it nets them a lot more money than would a royalty on the implementing chip itself. At one stage during the Program, the moderator asked the panel to address where in the value chain licensing should occur. The moderator stated that licensors want to license downstream, but the chipmakers cannot afford that. The panelists then got into a discussion of indemnities and how it does not make sense for the chipmakers (or for Continental, a component supplier to the vehicle industry) to agree to indemnify their downstream customers. But this was in my view an off-track discussion because indemnity or not, absent very extenuating circumstances for which there is no evidence in the 4G/5G space, the value of the standard as a whole can be no more than the value the market has attributed to a working version of that standard, e.g. to a 4G/5G chip.
The former Samsung in house counsel made this point several times during the Program. I’ll try to summarize what he said. First, he mentioned that connected products use lots and lots of embedded chips. Implementing different standards into those chips takes a great deal of research and development and investment by the chipmakers.
Chip implementers should be adequately rewarded for undertaking that research and development. 5G chips cost from $15-50 per chip. They are not all the same which is why the cost varies. Some of these chips have more bells and whistles and so cost more. But that additional value is because of all of the other features, functions and attributes of the chip, not because of the standards they implement. In the real world, the most sensible and efficient way to collect SEP royalties for a given standard would be to collect them from the chipmakers based on the cost of the chips, not from each end user based on the cost of their end product.
The panelist from Hitachi piped up to say that the chips, and the modules into which they are incorporated, do not know the end use cases. In other words, the first line implementers, the chipmakers, do not make more money based on “use cases.” Chipmakers sell the same chips at approximately the same price no matter who is the ultimate end user of those chips and no matter what is the ultimate end use.[2] The chips, and the 4G/5G modules, Hitachi uses are off the shelf (my words not his) and not "high-end."[3] Hitachi can use the same chips in their nuclear power plants and in their health care products. Obviously each nuclear power plant costs much more than a single health care product. But, the value of that nuclear power plant is based on so much more than whether it implements 4G/5G. The overall value of the standard should be determined at the base level, e.g. the chips or modules, not at the level of the nuclear power plant in which it can be used.
He also noted that the chipmakers do not get additional value from the ultimate end use. The chipmakers do get additional value when they make a high-end chip but that’s because of the extra bells and whistles, not because of the standard. So, if the chipmakers, the ones who initially implement the standard into their products, do not get higher revenue based on ultimate end use, it does not make any sense that those who contributed only a portion of the technology to develop that standard should be entitled to get revenue based on that use. It also does not make any sense to pay more to license the standard than the chipmakers get for a working product that implements that standard.
One of the panelists pointed out that Sisvel is demanding $15 per connected car for its 4G portfolio. If you add up just the Avanci and Sisvel demands for connected cars, that is already $35 per car in order to use a chip that cost as little as $15 to purchase. That is a lot of money in total before even considering the other SEPs in the 4G stack.[4]
Several commentators also noted that basing royalties on usage makes for a very complicated quantification. A connected car might use its connectivity to talk to a base station, to connect a laptop in the car to the cellular network or Internet or to connect to other infrastructure (for example a smart road). A connected car might use its connectivity to talk to pedestrians or other vehicles. Should each of these use cases be charged different amounts? And if so, how can you accurately quantify what that amount should be? What if a connected car starts out only able to talk to a base station but a later software upgrade allows it to talk to the cars around it - would an additional royalty then be owed based on this new use case? At some stage, this reliance on use cases is not only fundamentally unsound as a methodology, but it becomes nearly impossible to do as a real-world exercise.
There also was a discussion of the fragmentation of FRAND laws with Germany, the UK, the US, China and, if the Proposal passes, the EU taking different stances, particularly around injunctions. Up until now, the SEP licensing world has been populated with huge companies (SEP holders) against huge companies (handset and car companies). As one panelist pointed out, the end game of current litigation for the licensors is to sue in injunction-friendly jurisdictions (Germany) in order to get an injunction and wrest (unduly) favorable licensing terms from an enjoined company. The panelists also pointed out that SEP litigation has been hugely expensive to date, but at least these huge companies on each side have the resources to engage in it. IoT is different – it is made up primarily of small to medium businesses.
In conclusions, there was a lot of disagreement among the panelists about how to change the system to address the unique needs of IoT companies. But, one thing all of the panelists agreed, the old way of SEP licensing will not work well. There need to be new solutions to reduce litigation and increase transparency and efficiency of licensing in the IoT space.
[1] You can read more about what was said at the meeting on my other post about it: IoT - They Are Coming for You Next (sepessentials.com).
[2] Chip prices can vary based on volume discounts and similar business type issues, but these variations are small, particularly compared to, for example, the difference Avanci charges different end uses for the same technology which can be an entire order of magnitude difference.
[3] In his words, Hitachi “purchases cheap modules, not Snapdragon by Qualcomm.” To level set, Snapdragon is an expensive, system-on-chip/system-in-package semiconductor product that implements 4G/5G and has a lot of other bells and whistles. Snapdragon System-in-Package | Qualcomm
[4] It really is too bad that the representative from Ericsson was unable to make the Program. It would have been interesting to hear from them about why Ericsson is participating in two different patent licensing schemes (both Sisvel and Avanci) as well as whether Ericsson also intends to engage in its own IoT campaign.