Shenzhen court issues written judgment in Huawei v Samsung case

In January this year the Shenzhen Intermediate court held a live broadcast of its announcement of the judgment on two SEPs infringement cases filed by Huawei against Samsung (reported here). We reported the court’s expectation that the written decisions would follow a few days later. That proved to be somewhat over-optimistic: the final decision was issued on 21 March. 

The delay was caused by disputes over redactions in the final written judgment. In retrospect that may have been predictable: the same problems arose in the UK's Unwired Planet decision (see here). To be fair to the Shenzhen court, the delay in issuing the final written decision in Unwired Planet whilst the parties agreed confidentiality was significantly longer, at over seven months. 

Each of the Shenzhen court's two written judgments is substantial, at more than 350 pages. We summarise the findings below:

FRAND binds the affiliates of the SEP holder and implementer

The question of whether the FRAND commitments bind affiliates in a group of companies is surprisingly topical at present. It was relevant in the Shenzhen case because the Defendants were Samsung China, Samsung Huizhou and Samsung Tianjin. The issue arises because the acts that Huawei complained about were not carried out by these companies. They were carried out by their parent company, Samsung Electronics Co Ltd. 

The Shenzhen court found that the parent company and its Chinese affiliates formed a community of shared economic interests. Samsung Electronics gains its economic benefits through the three Chinese affiliates' manufacturing and sales activities in China. The court found that where a parent company acts on behalf of itself and its affiliates to negotiate a licence, the parent company's acts should be deemed as the affiliates' acts. The obligations that arose from the FRAND commitment therefore bound the defendants in this case. 

This is an interesting step because although the parties were negotiating a cross licence, in this case Samsung was acting as an implementer. In other words Samsung was a beneficiary of Huawei's licensing commitments to ETSI. But, as in the Huawei v ZTE case in the CJEU, the Chinese court recognised that the obligations flowing from the FRAND commitment are not purely one way from patent owner to implementer: the implementer seeking a licence must take certain steps too. In determining whether to grant an injunction against these Defendants, the court would take into account the acts or omissions of their parent company.

FRANDly negotiations?

The court found the following behaviours were not consistent with Samsung’s FRAND obligations:

  1. In their cross- licence negotiations, Samsung insisted upon discussing a licence covering both SEPs and non-SEPs (a practice known as "bundling"). The court found that this caused significant delay to the negotiation process.

    The court based this finding on a term sheet sent by Samsung to Huawei in September 2012. In this document, Samsung said that that they would like the licensing scope to include all patents owned by each party, because that would ensure world peace. In October 2012, Huawei requested that the licensing scope was restricted to SEP patents only. In subsequent meetings the parties continued to argue about the licensing scope as to whether it should cover LTE, UMTS, OMA/Wifi, or be broader. Huawei's six offers only covered LTE at first, but later included UMTS. Corresponding claim charts were exchanged but the parties made no substantial progress.

    In March 2015, the GC of Huawei emailed Samsung management claiming that Samsung had never provide¬d Huawei with a standalone offer for Samsung's SEP portfolio, and this was forcing Huawei to negotiate SEP and non-SEP licensing in parallel.  Samsung replied that it was being reasonable in voluntarily offering to negotiate non-SEPs, but this was never a pre-condition to agreeing a licence in respect of the SEP portfolio.  In the mediation meeting, Samsung told the court that the lack of progress in confirming a definite offer to Huawei arose because the parties had not agreed on the licensing scope.  

    The Shenzhen court found that according to the recognized international practices, parties can negotiate a licence covering SEPs and non-SEPs, but only if both parties agree to discuss on that basis. If one of the parties was only willing to discuss a licence only covering SEPs, the other party should not insist in bundling the licence of SEPs with non-SEPs.  It found that Samsung, by its actions, had only offered to discuss a licence to both portfolios. 

    This is interesting because until this decision the bundling of patents had been an allegation levelled at SEP owners rather than implementers. The Shenzhen court recognises that an implementer may not be negotiating in good faith if he demands a licence to non-SEPS. Although the implementer may simply be seeking patent peace in demanding such a licence, the court recognised that the patent owner has no obligation to license its non-SEPS, and may legitimately want to retain or exclusively license these assets for product differentiation purposes. Allowing an implementer to demand a licence would illegitimately extend the FRAND obligation to assets that it was never intended to cover.

  2. The court found that Samsung employed delay tactics in technical discussions – it failed to reply or failed to reply in a timely manner to the claim charts provided by Huawei.

    Samsung explained that their technical team could not attend the technical discussions suggested by Huawei because their team was too busy with handling the litigation cases against Apple, Ericsson and IDC.  The Shenzhen court found that a multi-national company like Samsung should be capable of conducting technical analysis and licensing negotiations against several patent owners at once.

    The court also noted that in the 6-year negotiation process, there had been a year (during Feb 2014 to Jan 2015) that Samsung failed to provide any response to Huawei's requests of continuing the technical discussions. It regarded this as an unacceptable delay by an implementer.

  3. The court found that Samsung failed to provide their licensing offers or respond to Huawei's licensing offers in a proactive manner. From 2011 to 2017, Huawei proposed six licensing offers. Samsung only provided one counteroffer after Huawei proposed its fifth offer. Samsung did provide a second offer during the court-organised mediation (see below), but the court considered that the second offer was not substantially different from the first offer.

  4. The court found that Samsung rejected Huawei's arbitration proposal without justified reasons. The court referred to the EU Commission's anti-trust investigation against Samsung in 2014. As part of the resolution of that investigation, Samsung committed that it would negotiate on FRAND terms with a potential licensee for a period of a year.  If no agreement could be reached during this period, the parties were allowed to opt for submitting the dispute to court or arbitration. The court found that Samsung had thereby accepted the idea of court adjudication or arbitration as an effective way to solve a licensing dispute.

    Huawei made three proposals to submit the licensing dispute to arbitration: on 27 May 2013, 21 May 2014, and 28 Oct 2015. In its 28 Oct 2015 proposal Huawei also offered to disclose the terms of the Huawei-IDC licence and Huawei – Apple licence in the arbitration if confidentiality measures were in place.  

    Samsung argued that it was unacceptable for Huawei to push for arbitration while in parallel Huawei continued to keep their Chinese lawsuits for injunction moving. So on 8 Aug 2016, Huawei provided detailed arbitration terms for resolution of the dispute at the International Center for Dispute Resolution.  If Samsung agreed, Huawei would withdraw all of their cases currently seeking injunctions. 

    Samsung rejected this offer. The court found that this rejection was not consistent with Samsung’s commitment to arbitrate, and was inconsistent with Samsung’s obligations as an implementer. 

  5. The court found that Samsung failed to provide a substantive solution proposal in the court-organized mediation. 

    In order to help the parties to reach a licence, the Shenzhen court had organized two mediation meetings. The court requested the parties to provide their licence offers 40 days before the first mediation meeting. Huawei made their offer within the designated period, but Samsung failed to do so. Samsung argued that it was not yet the right time to make an offer. When the court asked for an explanation, Samsung was unable to give one that satisfied the court.

    The court then indicated that it would take into account acts in the mediation in assessing the parties' compliance of FRAND. This may seem surprising to UK or US readers but illustrates the rather different approach to mediation that is taken in China. On receiving this warning, Samsung responded to Huawei's offer, but the decision suggests that it had not made any significant move from its previous offer. The court found that Samsung lacked good faith in negotiations.

In contrast, the court found little basis to criticise Huawei's negotiation approach. There was one point that attracted attention: during the licensing negotiation, Huawei initially claimed that it had acquired 189 LTE families from Sharp. Further investigation revealed that only 39 LTE families were in fact acquired.  The court considered that this misleading information had a negative impact on the progress of the negotiation.  But because Huawei clarified the issue later in the negotiations, the court found that this was not a significant fault.

Calculating a FRAND royalty offer: strength of the portfolios

The Shenzhen court compared three measures of the strength of the parties’ portfolios:

  1. The court compared the number of technical proposals accepted by 3GPP. To support this Huawei submitted the Huawei-Ericsson arbitration ruling to show that this was an approach used by those parties in that arbitration proceeding. The court agreed that the number of technical proposals being accepted by international SSOs to some extent reflected the proportion of SEPs owned by a patentee, and so this was a relevant indicator for assessing the strength of patent portfolios.

    This is a further contrast to Judge Selna in TCL v Ericsson (report here). Judge Selna dismissed Ericsson’s approach to valuation based on SSO contributions, because contributions do not necessarily give rise to patents. That may be true at the patent by patent level, but when valuing a portfolio it is clearly arguable that a party who has made more and better contributions to a standards setting process is likely to create a more valuable patent portfolio. There are refinements to the contribution analysis that could be made in future cases that may improve it: not all contributions are equal. But this is the first time that a court has accepted that contributions can be one measure of portfolio strength, and in doing so the Shenzhen court has taken a step beyond the US and UK in terms of sophistication.

  2. The court compared the number of declared SEPs and the number of truly essential patents owned by the parties. In doing so, it looked into a number of third party reports submitted by the parties. These included:

    • three ABI Research Reports about LTE SEPs in year 2011, 2013, 2016 (submitted by Huawei);
    • a Research Report about LTE SEPs issued by Cyber Creative Institute from November 2012 (Huawei);
    • a Research Report on WCDMA SEPs issued by Fairfield Resources International (Huawei);
    • an Intellectual Property Office Report (Samsung); and
    • a Research Report issued by Ministry of Industry and Information Technology on LTE standard essential patents (requested by the court).

    Third party reports are often criticised in these exercises because researchers looking at large portfolios inevitably cannot spend much time on each patent. But in applying these reports to the patents under consideration and to the rest of the industry, the Shenzhen court has at least applied a consistent measure to all patents. The application of different standards of essentiality testing to a party's patents and the industry's patents is one of the strongest criticisms levelled at the TCL decision. It was a problem that Mr Justice Birss recognised and sought to correct in Unwired Planet.

  3. The court took into account the results of the invalidation proceedings in China. Huawei and Samsung respectively asserted eight SEPs against each other before the Shenzhen court. Four of the eight SEPs asserted by Samsung were invalidated by the Patent Re-Examination Board of the Chinese State Intellectual Property Office (China has a bifurcated system in which invalidity is heard by the patent office, not the local courts). In contrast, only one of the eight SEPs asserted by Huawei was invalidated by Patent Re-Examination Board.  The court took this to mean that Huawei's patent portfolio is of a higher quality. The difficulty with this decision is that it is unsafe to draw conclusions about a large portfolio from a very small sample size. The result may simply mean that Huawei was better at selecting patents for assertion than Samsung. Or it could be sheer chance: eight patents is not a statistically significant sample for portfolios the size of either Samsung or Huawei.

The Court also had the benefit of two Huawei economic expert reports and one Samsung expert report. It concluded from all of the above information that both portfolios were similar in strength, although Huawei's Chinese SEP portfolio was stronger than Samsung's Chinese portfolio.

The court did not take into account a Thomson Reuters report submitted by Samsung and a forward citation analysis in Samsung's expert report. These were disregarded as their analysis was only based on US patents and patent applications. The court found these unhelpful in China.

Finding that Huawei’s offer was FRAND

The royalty offers by both parties are redacted in the written judgment. From page 19 of the judgment we learn that one of Huawei’s initial offers for its own portfolio was 1.5%.

To compare the offers, the Shenzhen court carried out a “top down” analysis. It found that Huawei held 5% of the total numbers of 3G SEPs. It found an aggregate royalty rate for 3G of 5%. For 4G, it found that Huawei held about 10% of the total number of 4G SEPs, and the aggregate royalty rate for 4G would be 6%-8%.

The methodology then used by the court to assess Huawei's offers is redacted.  But the court found that Huawei's royalty offers were made based on the considerations of the strength of its global portfolio, the above aggregate royalty rates for 3G and 4G essentials, and the market sales information of Samsung products. It found that Huawei's offers were reasonable as they gave room for Samsung to negotiate a suitable price. This again is interesting: it suggests that the court was not looking to see if Huawei’s offers were exactly on the range that it calculates as FRAND. But it suggests that if the offers were not excessively far above the range, then that would not “breach” FRAND or competition law. That is consistent with the approach taken in the UK and US. 

Samsung submitted a Strategy Analytics report to argue that it would be unable to generate a reasonable sales profit if it accepted Huawei's royalty offer.  The court accepted that whether the industry could maintain a reasonable profit margin is one of the factors for assessing a FRAND royalty.  However whether an individual manufacturer could make a profit would also depend on their efficiency and other factors in the market. Against this, the court also had to ensure that the SEP holder could get reasonable return for their investments.

Huawei submitted its licence agreements with Ericsson, Teltronic, Baicells and Zhi You Tong as comparable licences to demonstrate that its offers were FRAND. The court did not comment on these licence agreements in their reasoning.
Finding that Samsung's offer is not FRAND

Samsung only made one offer on 20 July 2015. The amount is redacted. The court found that the parties’ portfolios are of equal strength, but Samsung's offer to license Samsung patents to Huawei was 3 times higher than Huawei's offer to license Huawei's patents to Samsung. 

The royalty rate proposed by Samsung for Huawei’s portfolio was based on the FRAND royalty rate determined by the Guangdong High Court in the Huawei v. IDC case. Samsung used the IDC case royalty rate as the basis to estimate the royalty rates for Samsung and Huawei's global portfolios.  By applying the estimated rates to the sales revenue of both parties, Samsung further calculated the royalty payments for Samsung/Huawei and the difference that one should pay to the other.  The court considered that the rate determined in the IDC case was not comparable because: (1) the royalty rate in the IDC case only applies to IDC's Chinese portfolio, while in the current case the parties are negotiating a global cross licence; (2) IDC was an NPE; and (3) because Huawei and IDC entered into a global licence agreement after the conclusion of the Guangdong High court case, the rate determined by the Guangdong High court was superseded by the global licence agreement. 

In finding that royalty rates in China are not good comparisons for global royalty rates, the Shenzhen court reaches the same conclusion as Mr Justice Birss in Unwired Planet and Judge Selna in TCL. That point seems to be universally accepted. The rejection of the Guangdong High court's decision as a benchmark echoes Mr Justice Birss' rejection of the results of an arbitration as a benchmark: rates freely agreed between parties are a better benchmark.
Finding that the two patents-in-suit are essential to China's 4G standards

Huawei argued the essentiality of the patents by reference to published 3GPP Technical Specifications. It argued that the 3GPP Technical Specifications are the de facto 4G/LTE standard and have been incorporated into the official Chinese telecom YD/T standards. The technical details of the essentiality determination are beyond the scope of this article, but the court accepted essentiality. Validity was not before the court (because of the Chinese bifurcated system).

No Patent exhaustion defence

Samsung argued that it received a licence to the Huawei patents through its use of Qualcomm chipsets. Huawei submitted a copy of their licence with Qualcomm. The court found that there was no exhaustion on the wording of the agreement. This is similar to the findings of Mr Justice Arnold in Nokia v HTC [2013] EWHC 3247 (Pat), who also dismissed an argument by HTC that in using Qualcomm chipsets it obtained a right to use Nokia’s patents under the patent licence agreement between Qualcomm and Nokia. 


The Shenzhen court granted the injunction, but noted that Huawei and Samsung can continue their negotiations after the judgment comes into force. If they reach an agreement or Huawei agrees not to enforce the injunction, then the enforcement of the injunction would not take place. 


Whether or not one agrees with each conclusion reached by the Shenzhen Court, there is no doubt that this is a highly detailed and reasoned decision. The eventual approach: the determination ultimately of a simple yes/no question about whether to grant an injunction based on party negotiation conduct, is closer to the German court's approach in applying Huawei v ZTE. But the much more detailed consideration of what is a FRAND rate reflects the greater economic depth of analysis seen in the UK and US court's recent decisions. In some areas, in particular in moving beyond simple patent counting and allowing evidence relating to the rate of acceptance of contributions to standards-setting organisations, this decision is more sophisticated than the UK and US cases. We predict that we may see more portfolio determination cases coming from China.