This article was first published in Competition Law Insight, December 2018
On 23 October 2018, the Court of Appeal handed down its judgment on Huawei’s appeal of Unwired Planet v Huawei, the ground-breaking High Court decision on FRAND licensing of standard essential patents (“SEPs”). This article provides a summary of the Court of Appeal’s decision and considers its implications. It then discusses, in the context of the Internet of Things, some of the issues in FRAND licensing that remain unresolved following the judgment.
Unwired Planet v Huawei
Unwired Planet (“UP”) is a non-practising entity in the business of patent licensing. After acquiring over 2,000 patents from Ericsson in 2013, it sued Huawei and others for patent infringement in the English High Court. Five of the patents in suit had been declared as essential to ETSI telecommunications standards (2G/GSM, 3G/UMTS and 4G/LTE). Huawei claimed that UP had failed to offer a licence on fair, reasonable and non-discriminatory (“FRAND”) terms as required by its obligations to ETSI. Following a series of patent trials, UP and Huawei contested a lengthy FRAND trial lasting for much of October and November 2016.
The judgment of Mr Justice Birss in April 2017 was the first in which FRAND issues were considered in detail by the English High Court. It was also the first judgment globally in which a full FRAND determination was carried out for a portfolio of ETSI SEPs. In June 2017, Birss J issued a remedies judgment, granting an injunction prohibiting Huawei from selling mobile phones in the UK owing to its failure to take a global licence to UP’s SEP portfolio. The injunction would cease to have effect once Huawei entered into the FRAND licence settled in that judgment and was stayed pending Huawei’s appeal.
Huawei appealed the High Court judgment on three main grounds:
1. Global licensing – that the High Court was wrong to hold that there should only be one set of FRAND terms and that those terms should be global. Huawei claimed it should have been able to enter into a national licence on terms set by the Court.
2. Non-discrimination (“ND”) – that the High Court was wrong to decide that the ND limb of FRAND allows a SEP owner to charge similarly situated licensees such as Huawei and Samsung substantially different royalty rates for the same SEPs (provided neither rate is higher than the benchmark fair and reasonable rate).
3. Huawei v ZTE – that the High Court should have found that Huawei had a defence to UP’s injunction claim under Article 102 TFEU, as UP had failed to comply with the steps set out in Huawei v ZTE by the CJEU.
UP cross-appealed the findings that: (i) it held a dominant position, (ii) its licence with Samsung was comparable to one with Huawei; (iii) it needed to meet the Huawei v ZTE criteria at all (as its injunction claim included a caveat limiting its effect).
The Court of Appeal’s judgment
The Court of Appeal rejected Huawei’s first ground of appeal. It held that a SEP owner with a global portfolio can in principle meet its FRAND obligations by offering a worldwide licence. If the implementer refuses that offer, then it may be subject to an injunction preventing further infringement in the UK. In reaching this finding, the Court commented on the inefficiency of country-by-country licensing and noted that it may be ‘prohibitively expensive’ for a SEP owner to litigate in every country in which its rights subsist; requiring it would be in effect a ‘blue print for hold-out’.
The Court of Appeal did state that Birss J had fallen into error by holding that there was only one true FRAND rate. It explained that that it was possible for a tribunal (whether a court or arbitrator) to find that two different sets of terms are FRAND, and that the SEP owner would have satisfied its obligation to ETSI if it offered either one of them. However, the Court concluded that this error had no material effect on Birss J’s conclusion. It agreed with Birss J’s factual assessment that in the circumstances of this case only a global licence would be FRAND, noting that the judge’s descriptions of country by country licensing as ‘madness’ that ‘no rational business’ would do if it could be avoided was not language that suggested a national licence could be FRAND between these parties.
The Court of Appeal agreed with Birss J that the ND limb of FRAND imposes a general non-discrimination obligation. This requires a SEP owner to offer all potential licensees a rate which reflects the proper valuation of the portfolio (i.e. the FRAND benchmark rate). However, a SEP owner is not prevented from charging less than that benchmark rate if it wishes to do so. There is no ‘hard-edged’ discrimination obligation requiring a SEP owner to offer similarly situated licensees exactly the same royalty rate. The Court accepted that the potential for discrimination below the benchmark rate did exist, but suggested that this could be dealt with using competition law if necessary. It rejected this ground of Huawei’s appeal.
In considering non-discrimination, the Court also assessed UP’s cross-appeal: UP claimed that its licences with Samsung and Huawei were not comparable and, as such, that the ND obligation did not arise. The Court held that the focus should be on the transactions themselves, rather than the circumstances in which they are entered into. It agreed with Birss J that the licences were comparable, noting that UP’s financial circumstances and PanOptis’ subjective reasons for offering Samsung a lower rate were not relevant features of the transaction itself.
Huawei v ZTE
The Court of Appeal confirmed that Birss J was entitled to find that UP was dominant. It agreed that there was a presumption that a SEP owner with a 100% market share holds a dominant position, and that this presumption can be rebutted. It recognised that the FRAND undertaking and potential for hold-out may constrain a SEP owner’s market power, but concluded that UP had not provided sufficient evidence to interfere with the High Court’s conclusion on dominance.
The Court of Appeal also determined that UP had not abused its position of dominance. It interpreted the CJEU’s Huawei v ZTE judgment not as laying down mandatory conditions required to avoid committing an abuse, but as providing a protocol that, if followed, provided safe harbour for a SEP owner. The only aspect of the protocol that is mandatory is for a SEP owner to provide notice prior to commencing proceedings; the nature and extent of the notice will depend upon the circumstances of the case. In the High Court, Birss J had held that Huawei had sufficient prior notice of UP’s SEPs (despite not having been presented with claim charts or a written offer) and the Court of Appeal saw no reason to interfere with the High Court’s factual assessment.
The Court of Appeal also indicated (without expressing a concluded view) that it may have been disproportionate to refuse an injunction at the stage of English proceedings when relief was being considered, i.e. after the trials of patent infringement and validity, and a detailed assessment of FRAND by the court. The focus of the Huawei v ZTE framework is on whether the commencement of proceedings seeking an injunction amounts to an abusebecause it may put pressure on a potential licensee to pay higher than FRAND rates to avoid an injunction. The view of Birss J and of the Court of Appeal appears to have been that in in this case the fear was unfounded as the Court was willing to set FRAND terms and an injunction was available only if those FRAND terms were rejected by the licensee.
The Court did not need to decide upon UP’s cross appeal about the wording of the injunction UP sought, but it indicated that the caveated wording UP used to try to avoid the Huawei v ZTE criteria would not have prevented it from committing an abuse, if seeking an injunction would have been abusive in the circumstances.
Implications for SEP licensing
The Court of Appeal’s judgment is effectively an announcement that English Courts are open for global FRAND licensing business. However, it is important to note that the judgment does not rule out other more territorially limited licences; the scope will depend upon factors such as the geographic spread of a SEP owner’s portfolio and on commercial practice, so the facts and evidence will be important in future cases. It is possible that, for example, we may see regional licences in the future if that reflects the business of the licensee and the patent coverage of the licensor.
Despite this, the Court of Appeal’s conclusion that more than one set of terms might be FRAND and that the SEP owner meets its FRAND obligation by making one compliant offer does give significant power to the SEP owner to propose the scope of the relevant licence. It seems unlikely that a global licence would not be at least one of the FRAND options in most disputes between a large multinational with global sales and another company with a significant, worldwide SEP portfolio. In those circumstances, there would be little incentive for the licensor to offer anything other than a global licence. In any global licence, the importance of the licensor owning SEPs in the country in which the licensee manufactures its products should not be underestimated. These SEPs ensure effective coverage in other countries where the SEP owner does not own patents, because royalties on sales will still be payable at the rate applicable for the country in which the products are manufactured.
The findings on non-discrimination leave open the possibility for SEP owners to offer bespoke deals to particular implementers without breaching their FRAND undertaking, and without being forced to offer the same deal to all comers. However, any bespoke deal is still likely to be factored into the assessment that establishes the FRAND benchmark rate for a portfolio as a comparable licence. SEP owners will need to take this into account when considering whether to make a below benchmark offer.
The Court’s interpretation of Huawei v ZTE focusses on whether a SEP owner has given notice before commencing proceedings. This view gives SEP owners considerable discretion in deciding when to issue proceedings as part of a licensing negotiation. However, it would still be a high risk strategy for a SEP owner to depart completely from the Huawei v ZTE protocol; go too far and the SEP owner will fall outside the safe harbour and may be denied injunctive relief, limiting its ability to force a licensee to enter into a settled FRAND licence. Despite that, the flexibility offered by the Court of Appeal on this point, combined with the availability of the ‘FRAND injunction’, the chance to secure a global resolution to a licensing dispute and the fact that the UK is often an important market to implementers (one that they might be unwilling to pull out of if faced with an injunction or accepting a licence they dislike) means that litigation in the English Courts is an attractive prospect for SEP owners.
However, there is an unresolved question regarding the jurisdiction of the English Courts to grant global FRAND licences that could affect the ability of SEP owners to bring proceedings in England. None of the defendants in Unwired challenged jurisdiction when the claim was first issued, and so, although the Court of Appeal noted that some of Huawei’s submissions on the global licensing ground of appeal suggested that England was not the appropriate or natural forum in which to litigate this dispute, it held that it was far too late for Huawei to raise points of this nature.
However, the Court of Appeal will have the chance to consider the jurisdiction arguments inherent in global FRAND licensing disputes in full in December 2018. In another case, Huawei and ZTE are appealing the first instance judgment of Carr J that the High Court did have jurisdiction to hear proceedings brought by Conversant, a patent licensing company based in Luxembourg. The appeal is based on grounds of non-justiciability and forum non conveniens.
In Unwired, the Court of Appeal did make clear that it rejected any submission that in granting a global licence, Birss J had in some way usurped the right of foreign courts to decide issues of infringement and validity of patent rights subsisting in their respective territories. If the judges hearing the appeal in Conversant are inclined to take the same view, then the appeal on the grounds of non-justiciability will face significant challenges. However, Huawei and ZTE are also claiming that China is the more appropriate forum to hear the dispute, particularly given the fact that only 0.07% of ZTE’s turnover comes from the UK. If the appeal is dismissed, circumstances in which the English courts will not have jurisdiction for any FRAND dispute where a SEP owner can assert UK SEPs are likely to be limited.
What about other FRAND issues?
The Court of Appeal could only comment on the issues before it. As a result, although the Unwired Planet judgments offer a clear route for the determination of a global FRAND licence, they do not solve all FRAND issues. There are a number of further complications that may be still be raised in future cases, particularly as FRAND and SEP licensing is likely to become increasingly important with the rollout of 5G and the continuing evolution of the Internet of Things.
These developments will mean that a large number of new companies will have to engage in SEP licensing to an extent that they are unlikely to have experienced before. Start-ups developing new, connected products may have had no previous experience in licensing negotiations (or even much experience in any commercial negotiations) and yet if their products make use of 5G they will need licences. Even more established companies may find themselves in difficult positions. Car manufacturers racing to compete with the connected cars being developed by companies such as Google’s Waymo, or Uber, will need to negotiate SEP licences. However, given that car manufacturers do not tend to have large 2G/3G/4G patent portfolios, they may be forced to negotiate at a disadvantage. Alternatively, they may seek to put the licensing burden upon their suppliers, demanding that they sell chips that are already licensed. Whether chip manufacturers will be able to do so is a FRAND issue that is yet to be fully resolved.
Licensing to all
It is common practice in the mobile phone industry for SEP owners to seek to license handset manufacturers rather than chipset manufacturers. SEP holders claim that this is more efficient; that some of their SEPs read only upon the handsets and not the chipset itself, meaning that they would not be able to license their entire portfolio portfolios in one transaction if they had to do so at the chipset level – they would still have to try and offer a more limited licence to handset manufacturers as well. Others suggest that the significantly higher price of a handset as opposed to a chipset is the reason for SEP holder seeking to secure royalties (as a percentage of the net selling price of each device sold) on handsets rather than chipsets.
There has been significant debate as to whether FRAND requires SEP holders to grant a licence to any company that asks for one. The answer is potentially of wide significance. If all chipset manufacturers were licensed the manufacturers of smartphones may not require licences at all (depending on laws relating to pass-through and exhaustion). In those circumstances, the price of chipsets themselves might need to rise significantly to account for the increased IPR costs. A half-way house might be manufacturers seeking to tailor licences, perhaps splitting value along different parts of the supply and distribution chain.
The first significant judicial authority on this issue was published in November 2018. In FTC v Qualcomm, Judge Koh granted partial summary judgment in favour of the FTC against Qualcomm, making an order that Qualcomm must license its SEPs to rival chipset manufacturers (such as Intel). This case involved the IPR policies of ATIS and TIA rather than ETSI as in Unwired Planet, but they imposed RAND obligations very similar to the FRAND obligations imposed by ETSI. The District Court dismissed Qualcomm’s argument that the IPR policies of those standard setting organisations contain limitations and that Qualcomm was not required to license its SEPs to applicants like chipset manufacturers that only produce components of devices. Whether courts in Europe will take the same approach remains to be seen and it is worth being aware that Judge Koh based her reasoning squarely on the particular facts of the case.
An issue which is related to the identity of licensees, and where in the supply chain a licence should be granted, is what the basis of royalties for that licence should be. Some commentators have proposed the concept of ‘use-based’ licensing, where a SEP owner could charge different royalty rates for different uses. For example, commentators point out how important communications technology is to a connected car and the extent to which the technology will be used every time the car is driven, compared to how such technology might be used only once every week or two by a smart fridge to automatically order new groceries. Given the high cost of products like connected cars, there is considerable value at stake. It seems plausible that this issue may give rise to litigation in the coming years.
The Unwired Planet judgments offer a method for calculating a global FRAND licence that is likely to be replicated in future cases in England and Wales. They make clear that the English courts are an attractive venue for SEP owners to initiate FRAND litigation against companies with which they are struggling to agree licence terms. However, the judgments are not a panacea, and there remain a number of unanswered questions in FRAND that will become increasingly important as new companies and new products require licences.
Disclaimer: Bristows LLP acts for a number of clients involved in SEP licensing and in ongoing SEP litigation with a variety of views and interests. This article is written from the personal perspective of the authors and does not represent the views of Bristows LLP or any of its clients.
 UP sued Samsung and Google as well as Huawei, but both companies settled with UP before trial.
 The European Telecommunications Standards Institute.
  EWHC 711 (Pat), later replaced with a less redacted version:  EWHC 2988 (Pat).
  EWHC 1304 (Pat).
  EWCA Civ 2344.
 Ibid, paragraphs 55, 88.
 Ibid, paragraph 125.
 Ibid, paragraph 128.
 Ibid, paragraph 200.
 Ibid, paragraphs 172-176.
 Ibid, paragraphs 269, 282.
 Ibid, paragraph 271.
 Ibid, paragraph 283.
 Ibid, paragraph 89.
 That is, an injunction that ceases to have effect once the licensee enters into a FRAND licence, as awarded by Birss J in the ‘remedies’ judgment,  EWHC 1304 (Pat).
 Supra n6, paragraph 112.
  EWHC 808 (Pat).
 Supra n6, paragraph 81.
 The involvement of car companies in SEP licensing disputes is foreshadowed by BMW joining the Avanci patent pool in December 2017, and the recent actions brought by Broadcom against Audi and Volkswagen in Germany (Az. 7 O 174/17).
 See for example, Why the ETSI IPR Policy Requires Licensing to All by Karl Rosenbrock, and Why the ETSI IPR Policy Does Not and Has Never Required Compulsory “License to All”: A Rebuttal to Karl Heinz Rosenbrock, by Dr. Bertram Huber.
 Northern District of California, Case No. 17-CV-00220-LHK.
 The Alliance for Telecommunications Industry Solutions (ATIS), and the Telecommunications Industry Association (TIA).
 However, note that there is no globally accepted single method of calculating a FRAND royalty rate. For example, in TCL v Ericsson (Case 8:14-cv-00341-JVS-DFM), Judge Selna adopted a slightly different approach, taking into account expired patents and the months remaining on patents that had not yet expired in calculating a FRAND royalty rate.
This article was first published in Competition Law Insight, December 2018