Objective PAtent Landscape (OPAL)
5G (3GPP Zone)
The question of who owns 5G standard-essential patents is an issue that many institutions and companies are grappling with as this important technology gets deployed. Many have claimed through declarations that they own 5G essential patents, but there is currently no economically sensible way to evaluate these claims. At least when companies self-declare their 5G patents, they also generally agree to license them on a fair, reasonable, and non-discriminatory (FRAND) basis, often limiting their use for securing injunctions and disproportionately high royalties. Unfortunately, other standards teach that there are many potentially essential patents that are never declared and that are not encumbered by any FRAND obligations. Unified Patents' 5G landscape, called OPAL, not only identifies self-declared 5G patents but also these undeclared and FRAND-unencumbered patents.
3GPP Patent Declarations - Standards Participation vs Declarations
According to OPEN, Unified Patents’s 3GPP standards submissions database, over 100,000 5G technical contributions have been submitted. Of those, only 812 technical specifications were agreed, approved, or endorsed for the 5G standard with about half of them submitted by Ericsson, Huawei, and Nokia.
As of March 2020, these technical contributions have led to over 25,000 patent families (over 96,000 patents and applications) being self-declared as essential as shown in Unified Patents’s AI-based landscaping tool OPAL. As a rule, participants in ETSI 5G standardization work are required to declare in a timely fashion any patents essential to a 5G technical specification, particularly any standard or specification where they participate. In their declarations, ETSI participants are expected to specify the technical specifications to which the declaration is made. While this provides a wealth of data for evaluating potential essential patent ownership, the data is lacking on a number of fronts:
- There is no independent review that declared patents and applications are actually essential or that the correct technical specifications and their parts are referenced.
- Participants are not required to conduct an exhaustive search for potentially essential patents.
- The data can lag behind the official release of the referenced standard specification.
- As discussed later, the data does not capture any potentially essential patents that were not declared by 3GPP standardization participants or which belong to entities that do not participate in 3GPP standardization work.
Problems with Existing SEP Landscapes
A landscape inherently should include or model for all patents related to a technology area in order to understand what the market share is of each company. A number of companies have released 5G tools, but none of these tools can answer this question since their data is incomplete. Most provide a comprehensive database of only self-declared patents and do not provide any assessment of essentiality or invalidity. Others filter for essentiality based on organic or court adopted essentiality reviews but again these only take into account the self-declared patents and applications. The essentiality filters used by some of these tools are usually based on a potentially biased sampling of SEPs declared to other 3GPP standards rather than 5G. The percentage results of these reviews are also sometimes applied indiscriminately across all 3GPP participants’ portfolios without any regard to each participant’s declaration and patent prosecution processes.
The percentage of self-declared patents estimated by courts and experts to be actually essential ranges from 50% to lower than 20%. In 2005, Goodman and Myers found 21% of self-declared patents in 3G were essential. In 2010, Fairfield Resources estimated 50% of self-declared patents in LTE were essential. In the remanded US TCL vs Ericsson case, the court’s calculation resulted in about 40% of all 2G, 3G, and LTE self-declared patents being essential. Finally, in the UK Unwired Planet vs Huawei case, the court used 16.6% to calculate the number of essential LTE patents from all self-declared ones. If we are to believe that 5G will have similar results compared to previous 3GPP standards, the actual landscape of essential patents will be much smaller than the number which was declared and the share of each company might change significantly.
As pointed out above, a landscape should not just take into account patents owned by 5G contributors under an obligation to self-declare but also those potential patents owned by non-contributors to 5G who have no obligation to self-declare. This could result in a much larger 5G landscape than the current studies that only include self-declared patents. Unfortunately, until now no 5G tool has addressed these undeclared patents.
Using AI to Objectively Assess Essentiality
OPAL can truly be considered a 5G landscaping tool since it measures 5G essentiality for all patents rather than just those self-declared to 5G. OPAL uses machine learning to objectively evaluate the semantic similarity of any patent or application to all of those that were self-declared. Each patent or published application receives a semantic similarity score on a scale of 0 to 100. The higher the score, the closer a patent’s correlation to the semantics of self-declared 5G patents. This allows users to identify potentially essential patents that would never be found otherwise.
The machine learning algorithm that provides this semantic similarity score is trained by representative sets of true positives (i.e. patents that are essential or likely to be) and true negatives (i.e. patents that are not essential or likely not to be). Specifically, the positive training set consists of all declared 5G patents as of March 2020, which amounts to 133,665 patents and applications. The negative training set consists of all patents declared to previous 3GPP standards (e.g. 2G, 3G, and LTE) but not declared to 5G, amounting to 162,675 patents and applications.
For practical purposes, OPAL does not evaluate every known patent in the universe but only a subset of patents which are relevant or at least tangentially related to 5G technology. This subset consists of all patents declared in ETSI, as well as any patent having a CPC code that belongs to the top 100 CPC codes corresponding to each 3GPP standards (i.e. 2G, 3G, LTE, and 5G). After including all patent family members, the number of 3GPP-relevant or tangential patents evaluated by OPAL is roughly 3 million.
In sum, after training the algorithm on nearly 300,000 patents from the positive and negative sets, OPAL evaluates and assigns a semantic similarity score to all 3,000,000 patents in the landscape universe.
The overall distribution of scores is plotted in the histogram below:
As expected, the patents belonging to the positive labels (i.e. the patents declared to 5G), shown in blue below, tend to have higher scores than the rest of the universe:
Exploring the 5G Landscape
In OPAL, by setting the baseline at 62, the landscape of 5G patents consists of only those patents with semantic similarity scores greater than or equal to 62. At this baseline, the top 5G patent holders are listed below:
The size landscape can expand or contract depending on where the baseline is, which also affects the ranking of top patent owners. For example, when the baseline is at 75, the rankings are as follows:
A powerful aspect of OPAL is the ability to identify 5G patents that have not been declared in ETSI. For example, at the 75 baseline, OPAL identified more than 21,000 families (76,050 patents and applications). Almost 95% belong to existing 5G declarants. Of those, almost 30% have not been declared and may not be subject to FRAND. OPAL further found 4,131 patents held by entities that did not participate in 5G standardization or self-declared few (if any) patents that appear to be essential to 5G. OPAL sheds light on these unseen patents, and if they are truly unencumbered by any FRAND obligations can significantly increase the cost of deploying 5G.
The chart above lists the top 18 entities identified by OPAL as holding 5G patent families and the jurisdictions in which their applications have been filed. Altogether, these entities hold 362 families (647 grants and applications). The 8 Chinese entities own 28% while the 6 US entities own 21%. These patents could be valuable targets for acquisition or licensing in order to establish a reasonable royalty rate,and some of the patent holders could be open to sales or licensing. R&D entities such as Shanghai Langbo, Ofinno, and Innovative Technology Lab hold about half of these patents, and they may need financing for their R&D activities.
Please note that the results described above correspond to the current version of the 5G landscape. The landscape is subject to change in future versions, and results may change accordingly.
- Version 0
- For the training set, the set of true positives consists of declared 5G patents and the negative labels consist of patents declared to 2G, 3G, and LTE but not 5G.
For more information about Unified's OPAL Reports, or to learn more about Unified's 3GPP Zone, please contact us at firstname.lastname@example.org.