Reasserting the UK’s FRAND framework: implications of the Court of Appeal’s decision in Optis v Apple

Nearly a decade after establishing the UK’s FRAND framework in Unwired Planet v Huawei (2017), Lord Justice Birss returned in Optis v Apple (2025)—this time in the Court of Appeal—to steer the law back to his original approach.¹ That means identifying the closest comparator(s), adjusting the rates to reflect party-specific differences, and conducting a high-level cross-check based on the top-down method.

The Court of Appeal’s decision in Optis v Apple reinforced several principles that are likely to shape litigation strategies in upcoming FRAND disputes in the UK and elsewhere. This article examines some of the key takeaways, including the factors that are likely to influence the courts’ choice of closest comparators, the weight given to implementers’ licences, and the evolving role of the top-down method.²

Comparable licences method

The comparable licences method has underpinned several UK FRAND decisions of the last decade—namely, Unwired Planet v Huawei, InterDigital v Lenovo and Optis v Apple. From an economics perspective, this method is appealing because it relies on ‘market-tested’ evidence on the rates agreed between commercial parties through negotiation. Despite the broad consensus on the usefulness of this method, there is often significant debate over (1) which licences are most comparable; (2) whether the agreed terms are likely to reflect an element of hold-up or hold-out; and (3) how to unpack and/or adjust the rates from those licences. These questions are discussed below.

1 Assessing comparability

In Optis v Apple the question of which licences are most comparable was hotly debated before the High Court in the 2022 trial.³ The trial judge resorted to an ‘averaging’ approach instead of identifying the most comparable licences. However, the Court of Appeal firmly rejected this approach and reiterated the importance of excluding licences that were not good comparators instead of ‘using less good comparables to modify a result from better ones’.⁴ This suggests that assessing which licences are most comparable is likely to remain of central importance in the UK courts’ determination of FRAND royalties for standard essential patents (SEPs).

In assessing comparability, the Court of Appeal drew a distinction between the ‘comparability’ and the ‘reliability’ of a licence. Comparability, as the Court put it, ‘is primarily concerned with the situation of the parties and the subject matter being licensed’,⁵ while reliability ‘is concerned with the quality of the information derived from a given licence’.⁶ What does this distinction mean for the use of licences as comparators in a FRAND analysis?

As an initial point, it suggests that the SEP owner’s licences for the same portfolio (patents-in-suit) with other implementers are likely to be the best place to start although, as the Court cautioned, a starting point is all they are.⁷ Indeed, the implementer’s licences—in this case, Apple’s licences with other SEP owners—played an important role in the Court of Appeal’s determination of the FRAND rate, as discussed later in this article.

More significantly, the distinction means that licences cannot be dismissed as potential comparators simply because their rate structure differs from that being sought in a dispute. Specifically, a licence with a lump-sum royalty payment must be converted, or ‘unpacked’, into a running royalty rate, which can then be used as a benchmark. This unpacking requires assumptions about past and projected sales, whether the implicit running royalty rate is percentage-based (ad valorem) or a dollar-per-unit (DPU) rate, and other terms that are often contested between parties in a dispute. Such disagreements, however, concern the issue of reliability, not comparability. As the Court of Appeal clarified in Optis v Apple, unpacking ‘is not concerned with making a licence more or less comparable [but] about improving the reliability of the comparison’.⁸ The Court thus viewed comparability as a distinct step in the analysis, determined by whether the parties to the disclosed licences were similarly situated to the parties in dispute.

When assessing whether the parties were ‘similarly situated’, the Court of Appeal did not set out a check-list of factors. However, based on the recent FRAND decisions, the following factors seemed to play some role:

• the overall size and financial strength of the parties involved;
• the volume of devices being licensed;
• the market share and position of the parties involved;
• the sales mix in terms of handset segments and/or geographies.

In InterDigital v Lenovo, for instance, the UK High Court considered InterDigital’s licence with LG to be the most appropriate comparator because, among other things, ‘LG’s sales were very similar to those of Lenovo’.⁹ The same reasoning did not apply to Optis’ licence with Google in Optis v Apple, since Google’s handset sales were far lower than Apple’s. Nevertheless, the Court of Appeal considered the Google licence to be one of the best comparators,¹⁰ emphasising that Google ‘is obviously not a small entity’—a point underscoring the licensee’s overall size and financial strength.¹¹ This factor also feeds directly into the next question of whether a licence contains elements of implicit hold-up or hold-out.

2 Assessing hold-up vs hold-out

In Optis v Apple, the Court of Appeal noted that ‘[t]he extent to which licences are affected by hold up or hold out will also have an impact on their comparability’.¹² Hold-up by a SEP owner refers to the scope to charge excessive royalties once the technology becomes essential to the standard. Hold-out by an implementer refers to the tactics used to unduly delay or avoid paying royalties while using the technology.

As a matter of principle, the Court of Appeal emphasised that parties under the FRAND system must be protected from the ‘mischiefs’ of hold-up and hold-out.¹³ Accordingly, if a court finds that some degree of hold-up or hold-out occurred in a real-world negotiation, the outcome of that negotiation cannot be treated as FRAND.¹⁴ This marked a departure from the trial judge’s approach. The trial judge had included 19 licences of Apple in his analysis, reasoning that these licences reflected not ‘illegitimate’ hold-out,¹⁵ but ‘legitimate’ hold-out—which he described as ‘normal manifestations of parties seeking to negotiate an outcome that serves their own interests’.¹⁶ The Court of Appeal found this notion of ‘legitimate’ and ‘illegitimate’ hold-up and hold-out to be irrelevant and, indeed, unhelpful.¹⁷

That said, as a matter of practicality and given the evidence before it, the Court of Appeal determined a FRAND royalty rate based on Optis’ licence with Google and a subset of Apple’s licences, while observing that ‘there would appear to be degrees of hold up and hold out involved’.¹⁸ The subset of Apple’s licences comprised four licences with the highest implied rates for Optis—namely, Apple’s licences with Ericsson, Nokia, InterDigital and Sisvel.¹⁹ The Court concluded that a rate for Optis below the rates implied from these licences would be too low.²⁰ The Court also concluded that a rate derived solely from Apple’s licences would be too low in light of the implied rate from Optis’ licence with Google. It thus determined a FRAND rate that lay between the rates derived from the subset of Apple’s licences and Optis’ licence with Google.²¹

The Court of Appeal’s treatment of Optis’ licence with Google is particularly interesting in light of Apple’s allegations that this licence reflected ‘nuisance value’. In InterDigital v Lenovo, the trial judge had warned that ‘where the costs of litigation would be around or greater than the total sum payable under a licence, it is far more likely that the implementer has little choice but to accept what the licensor is demanding’.²² In Optis v Apple, Apple had argued that Optis’ licence with Google was not FRAND ‘because the lump sum paid was comparable to the cost of litigation with Optis and so Google were in effect paying Optis off based on the nuisance value of the dispute’.²³ The Court of Appeal rejected this argument, citing the trial judge’s finding that ‘Google would scrutinise the sum in question with some care’^.24 Although much of the trial judge’s reasoning is redacted, this is likely to reflect the fact that Google—given its size and financial strength—does not necessarily fit the image of an implementer with ‘little choice’ in a negotiation.

More broadly, when assessing alleged hold-up or hold-out, economic evidence on the parties’ relative bargaining positions can be valuable, particularly where disclosure offers little or no insight into the negotiations underlying the disclosed licences. In Optis v Apple, the economic experts commented on factors affecting hold-up and hold-out such as the role of the FRAND commitment, financial pressures faced by Optis, and the implications of Apple’s market-leading position in mobile handsets. This evidence played a central role in relation to Apple’s allegation that, under competition law, Optis had abused a dominant position. The trial judge ultimately concluded that the FRAND commitment limited Optis’ market power and, as such, it did not consider Optis to be a dominant undertaking.²⁵ This finding was not challenged on appeal.²⁶

3 Assessing reliability

As an initial point, the Court of Appeal sought to address the trial judge’s reasons for rejecting the evidence of the FRAND experts. In doing so, the Court seemed to suggest that FRAND experts can help to improve the reliability of comparators—for example, by addressing differences in rate types, cross-licences and past releases, and by presenting licences on a like-for-like basis, although ‘accountancy experts are not best placed to express a view about […] comparability’.²⁷

In principle, FRAND experts can provide meaningful insights into comparability from an economics perspective. While there is no doubt that the court ultimately decides which licences are most comparable, economic experts are able to assess factors such as the parties’ relative sizes, business operations, downstream market positions and sales mix—considerations that can help to determine whether parties to different licences are ‘similarly situated’. Indeed, such analysis forms the basis of market-based valuation approaches, such as the multiples-based method, which are commonly used by economic and/or valuation experts in commercial disputes.

Returning to the question of reliability, an important issue is how lump-sum payments are allocated across past and future sales. Historically, industry practice may have excluded past sales from royalty negotiations but, following recent UK decisions allowing royalties to be awarded on past sales, this practice may evolve. Nevertheless, caution is warranted when unpacking licences that pre-date this legal development. In Optis v Apple, the Court of Appeal noted that the difference between assuming a free release (where no royalty applies to past sales) and a simple unpacking (where the same rate applies to past and future units) was ‘not that large’.²⁸

A further reliability issue concerns whether licences should be unpacked on an ad valorem or a DPU basis. In Optis v Apple, both the trial judge and the Court of Appeal recognised that a FRAND rate could take the form of an ad valorem rate, a DPU rate, or a lump sum. However, both judgments somewhat mischaracterised the evidence of Optis’ economic expert on this point. The expert explained the basic economic logic that ad valorem rates, being expressed as a percentage of sales, tend to capture more accurately the economic value to downstream users than DPU rates or lump sums. That does not mean—as the judgments seem to suggest—that economists regard ad valorem rates as the only appropriate form of royalty rates. There may be practical or commercial reasons for adopting other types of rate.

Top-down method

The top-down method is motivated by concerns over ‘royalty stacking’—the risk that cumulative royalties paid by an implementer for all SEPs that are relevant to a standard could become too high, given the large number of SEPs that an implementer needs to license from different SEP owners. In practice, the top-down method involves estimating the aggregate royalty burden (ARB) that could be reasonably charged for all SEPs relevant to a specific standard, and then apportioning this ARB across different SEP owners. This methodology was followed by the US court in TCL v Ericsson (2017),²⁹ and by the Chinese court in Huawei v Samsung (2018).³⁰ In Unwired Planet v Huawei, the UK court employed it as a cross-check, calculating the ARB implied from the rates derived from comparable licences to assess their reasonableness.

One of the key questions under this method is what level of ARB is reasonable: high enough to reward SEP owners for continued innovation and participation in the standard, yet not so high as to discourage implementers from investing in product development. Striking this balance between the incentives of SEP owners and implementers is a fundamental principle of the FRAND commitment, which is explicitly recognised by various courts, most notably the CJEU in Huawei v ZTE (2015).³¹

In Interdigital v Lenovo, the top-down method appeared to fall out of favour with the UK courts. The FRAND rate set by the trial judge implied an ARB of approximately 1% of a handset’s average selling price (ASP), which was significantly below the ARBs considered reasonable in Unwired Planet v Huawei (8.8%) and TCL v Ericsson (6 to 10%).³² The trial judge rejected the use of a top-down cross-check as it considered the results from the comparable analysis to be more reliable. The Court of Appeal in that case seemed to agree with this position, satisfying itself with the observation that its ‘conclusion as to the correct rate is less inconsistent with the top-down analysis’ when compared with the rate set by the High Court.³³

In Optis v Apple, the top-down cross-check regained prominence in the Court of Appeal’s reasoning. The Court used it to comment on the reasonableness of the rate derived from Optis’ licence with Google.³⁴ It also calculated the implied ARBs to evaluate the reasonableness of the rates lying between the figures derived from Optis’ licence with Google and the subset of Apple’s licences.³⁵ Specifically, the Court rejected a rate of $0.20 per unit, which implied an ARB of 10.6% of Google’s ASP, but found a rate of $0.15 per unit to be FRAND, which implied an ARB of 8.4% of Google’s ASP.³⁶ The Court did not explain why 8.4% was more reasonable than 10.6%, although Birss LJ may have had in mind the 4G multi-mode ARB of 8.8% that he previously deemed reasonable in Unwired Planet v Huawei.

When assessing reasonable ARBs, it is also worth highlighting the trial judge’s comment in Optis v Apple that any ARB should at least be ‘consistent with a sensible business model’—that said, he could not explore this further due to lack of information on the costs of handset implementers in that case.³⁷

Looking forward: implications of the judgment and open questions

The Court of Appeal’s decision in Optis v Apple offers several practical lessons for parties engaged in, or contemplating, SEP litigation in the UK.

First, given the Court’s emphasis on comparability, parties may be tempted to limit disclosure to those licences that they view as most comparable, focusing resources on justifying their relevance rather than unpacking numerous agreements. This strategy, however, carries certain risks. Invariably, there will be allegations of implicit hold-up and hold-out affecting the rates set out in the disclosed licence(s). Against this backdrop, licences involving major players—whether implementers or SEP owners—may still provide valuable context for the court’s assessment, even if those players are not perfectly ‘similarly situated’ to the parties in dispute.

Second, parties may want to re-examine the role of FRAND experts. As this article has explored, while the ultimate decision about which licences are most comparable rests with the court, economic evidence can play a pivotal supporting role. A well-grounded economic assessment—rooted in the factual matrix of the dispute, informed by the commercial realities of the sector, and tested against relevant market data—can help the court to evaluate how closely a given licence mirrors the circumstances of the parties before it.

Finally, the Court of Appeal’s revival of the top-down method reopens an important debate: what constitutes a reasonable 5G multi-mode ARB? Here, too, economic evidence—both on the economic value added by 5G technology and on the economics of handset production—can assist the court in calibrating the FRAND rate.

It remains to be seen whether these strategies will be adopted—and how they will fare—in upcoming FRAND disputes.

¹ Unwired Planet v Huawei [2017] EWHC 711 (Pat); Optis v Apple [2025] EWCA Civ 552.
Dr Gunnar Niels, Managing Partner at Oxera, acted as the economic expert on behalf of Unwired Planet and Optis in these proceedings.

² The views expressed in this article are solely those of the authors, and do not necessarily reflect those of Oxera. The authors are grateful to Dr Gunnar Niels for his valuable feedback. Any remaining errors are the authors’ own.

³ Optis v Apple [2023] EWHC 1095 (Ch).

⁴ Optis v Apple [2025] EWCA Civ 552, para. 114.

⁵ Optis v Apple [2025] EWCA Civ 552, para. 90.

⁶ Optis v Apple [2025] EWCA Civ 552, para. 91.

⁷ Optis v Apple [2025] EWCA Civ 552, para. 95.

⁸ Optis v Apple [2025] EWCA Civ 552, para. 91.

⁹ Interdigital v Lenovo [2024] EWCA Civ 743, para. 76.

¹⁰ Optis v Apple [2025] EWCA Civ 552, para. 145.

¹¹ Optis v Apple [2025] EWCA Civ 552, para. 18.

¹² Optis v Apple [2025] EWCA Civ 552, para. 90.

¹³ Optis v Apple [2025] EWCA Civ 552, para. 118.

¹⁴ Optis v Apple [2025] EWCA Civ 552, para. 119.

¹⁵ Optis v Apple [2023] EWHC 1095 (Ch), para. 481.

¹⁶ Optis v Apple [2023] EWHC 1095 (Ch), para. 371.

¹⁷ Optis v Apple [2025] EWCA Civ 552, paras 118 and 120.

¹⁸ Optis v Apple [2025] EWCA Civ 552, paras 137 and 145.

¹⁹ Optis v Apple [2025] EWCA Civ 552, para. 132.

²⁰ Optis v Apple [2025] EWCA Civ 552, para. 137.

²¹ Optis v Apple [2025] EWCA Civ 552, para. 139.

²² Interdigital v Lenovo [2023] EWHC 539 (Pat), para. 288.

²³ Optis v Apple [2025] EWCA Civ 552, para. 18.

²⁴ Optis v Apple [2025] EWCA Civ 552, para. 34.

²⁵ Optis v Apple [2023] EWHC 1095 (Ch), paras 387–388.

²⁶ Optis v Apple [2025] EWCA Civ 552, para. 46.

²⁷ Optis v Apple [2025] EWCA Civ 552, para. 90.

²⁸ Optis v Apple [2025] EWCA Civ 552, para. 135.

²⁹ TCL v Ericsson [2017] Case No. 8:14-cv-00341-JVS-DFM (C.D. Cal. Dec. 21, 2017).

³⁰ Huawei v Samsung [2018], Intermediate People’s Court of Shenzhen.

³¹ Huawei Technologies Co. Ltd v ZTE Corp., ZTE Deutschland GmbH, Case No. C-170/13.

³² Interdigital v Lenovo [2024] EWCA Civ 743, para. 285.

³³ Interdigital v Lenovo [2024] EWCA Civ 743, para. 286.

³⁴ Optis v Apple [2025] EWCA Civ 552, para. 140.

³⁵ Optis v Apple [2025] EWCA Civ 552, para. 144.

³⁶ Optis v Apple [2025] EWCA Civ 552, para. 145.

³⁷ Optis v Apple [2025] EWCA Civ 552, para. 60.