A technical examination of over a thousand patents in 5G found Huawei had the strongest claims. The patent system at every level is distorted, especially in standards-essential patents (SEP.) Companies have claimed over 10,000 patents are essential to 5G, looking for a share of over one hundred billion in royalties.
34% of Huawei’s declared patents are “essential,” according to very experienced analysts at Amplified and GreyB. 34% qualifying is a disappointing result, but better than Qualcomm (30%,) Nokia (27%) or Ericsson (22%.) I believe many of these patents will be rejected in court as obvious or not sufficiently original,
The Western world was shocked when Huawei emerged on top of the international counts of patents a few years ago. Innuendos suggested that Huawei was filing bad patents rather than original innovative work. Some came from commercial envy.
Huawei has been spending billions on R&D for a decade. Its 2020 R&D budget is US$20 billion. It has research centers in many countries and has hired thousands of engineers outside China. They include many IEEE Fellows and hundreds of PhDs (or more.) It would amazing if Huawei weren’t among the leaders.
Jacob Schindler of IAM is more polite than I am. “Standards-developers are incentivised to err on the side of over-declaration, leading to a situation where only a fraction of declared SEPs are truly essential to a standard’s operation. His colleague, Joff Wild, adds, “There is no organisation whose role is to sort the truly essential wheat from the non-essential chaff.”
Race is involved, of course. I remember when one of the best switch designers in France told me, “The Chinese can’t design something this advanced.” I looked around his engineering team and saw half were Asian and several Chinese. Their cousins back home were just as sharp.

The report is only 13 pages but loaded with good information about standards and patents. The study was led by Matt Luby of Amplified as well as Muzammil Hassan and Aman Kumar of GreyB. They consult with many in the industry.