Optis Cellular Technology LLC & Ors v Apple Retail UK Ltd & Ors

Factual and Procedural Background

This appeal concerns a dispute over the terms of a licence for a portfolio of standards essential patents (SEPs) related to mobile telecommunications technology. The Appellant, a SEP holder, had declared its patents essential to standards set by a standards setting organisation. The Respondent, an implementer of these standards, manufactures products incorporating the technology. Prior to this appeal, the High Court had declared certain licence terms to be fair, reasonable and non-discriminatory (FRAND). The parties could not agree on the terms of the licence, particularly the valuation of the licence fee. The High Court judge, Mr Justice Marcus Smith, determined the licence fee and other terms, which led to this appeal.

The litigation originated after failed negotiations, with the SEP holder initiating patent infringement proceedings in the UK and parallel proceedings in the United States. Multiple technical trials were held in the UK to determine validity and essentiality of patents, followed by a trial to fix the FRAND licence terms. The judgment under appeal rejected the expert accountancy evidence presented by both parties and adopted a novel approach devised by the judge to determine the licence fee. The appeal challenges the judge’s rejection of expert evidence, the methodology used to determine the licence fee, and certain licence terms, including those relating to parallel US proceedings.

Legal Issues Presented

1. Whether the High Court judge was entitled to reject the expert accountancy evidence in its entirety.
2. Whether the judge’s novel lump sum averaging approach to determine the FRAND licence fee was correct in principle and fact.
3. Whether the correct approach to valuation is a comparables-based approach selecting the best comparable licences rather than averaging all.
4. The appropriate treatment of hold up and hold out in assessing FRAND rates, including whether the judge erred in distinguishing between “legitimate” and “illegitimate” hold out.
5. Whether a retrial is necessary or whether the Court of Appeal can determine the FRAND rate on the existing evidence.
6. The appropriate terms of the licence relating to foreign proceedings, specifically the interplay between the UK and parallel US litigation.
7. Whether the judge’s order imposing ongoing court supervision and restrictions on the parties’ conduct in foreign proceedings was justified and consistent with principles of comity and FRAND.
8. The appropriate date for cessation of interest on unpaid royalties.

Arguments of the Parties

Appellant's Arguments

• The judge erred in rejecting the accountancy experts’ evidence wholesale, as their expertise was within their proper scope and they provided necessary tools for the court’s valuation.
• The judge’s lump sum averaging methodology was flawed, lacked evidential basis, and contradicted established principles requiring selection of the best comparable licences rather than averaging all.
• The judge failed properly to account for sales volumes and other relevant factors in his averaging approach.
• The distinction drawn by the judge between “legitimate” and “illegitimate” hold out was erroneous and led to incorrect conclusions regarding the Apple licences.
• The judge’s approach to foreign proceedings, including requiring Optis to consent to dismissal of US proceedings and imposing ongoing court supervision, was procedurally unfair, contrary to comity, and not FRAND.
• The interest stop date chosen by the judge was arbitrary and unjustified.
• Optis contended no retrial was necessary; the Court of Appeal could correct errors and determine the FRAND rate based on existing evidence.

Respondent's Arguments

• Supported the judge’s lump sum approach with the important modification of excluding the Optis-Google licence from the averaging.
• Contended that the judge was entitled to reject the expert evidence and adopt his own approach.
• Argued that the Apple licences were reliable comparables and did not evidence hold out sufficient to depress rates below FRAND.
• Maintained that the judge’s order relating to foreign proceedings was necessary to prevent double recovery and preserve the integrity of the English court’s determination.
• Accepted the point on interest stop date but contended it was a trade-off with the interest rate chosen.
• Asserted that if the appeal succeeded on valuation grounds, a retrial might be necessary if the Court of Appeal could not determine the FRAND rate itself.

Table of Precedents Cited

Court's Reasoning and Analysis

The Court of Appeal began by considering whether the High Court judge was entitled to reject the expert accountancy evidence. The judge had found the evidence unreliable due to three "subjectivities" inherent in unpacking licences—differences in rate types (lump sum, ad valorem, DPU), cross-licensing effects, and treatment of past releases—and because the experts had largely followed their clients’ instructions without applying independent judgment. The Court of Appeal found that the judge’s wholesale rejection was unjustified and unfair, as the criticisms were not put to the experts in cross-examination and the experts were properly within their accountancy expertise. The court distinguished between comparability (a matter for the judge based on factual context) and reliability (a matter for experts), concluding that the experts’ role was to assist with reliability, not comparability.

The Court then examined the judge’s novel lump sum averaging approach to valuation. The judge had placed 19 licences on a common scale normalized by patent stack share and taken a simple average to derive a total stack price, then applied the SEP holder’s share to determine the licence fee. The Court found the approach flawed for several reasons: it lacked evidential or principled basis; it effectively adopted an exclusionary approach despite the judge’s stated rejection of one; the averaging ignored important factors such as sales volumes; and the inclusion of the Google licence on this basis was illogical because it related to a different volume of units, distorting the results.

The Court held that the correct approach is a comparables-based method selecting the best comparable licences and excluding others, consistent with established legal principles. The judge’s method of averaging all licences was contrary to the guidance in Cimetidine and prior FRAND cases. The Court further found that the judge erred in rejecting the DPU (dollar per unit) approach, which was supported by both parties’ experts and consistent with the implementer’s circumstances.

On hold up and hold out, the Court rejected the judge’s distinction between “legitimate” and “illegitimate” hold out as unhelpful and erroneous. Hold up and hold out are mischiefs the FRAND system aims to prevent, and the FRAND rate should reflect the outcome of negotiations between willing licensors and licensees, free from such distortions. The Court found that the Apple licences showed evidence of hold out depressing rates, and thus the judge erred in placing weight on all Apple licences collectively.

The Court considered whether to order a retrial but concluded it was unnecessary. Given concessions by the parties and the Court’s ability to correct errors, the Court was able to determine an appropriate FRAND rate itself based on the best comparables and the expert DPU data previously rejected by the judge.

In valuing the FRAND rate, the Court examined the DPU data derived from the Google licence and selected Apple licences (notably Ericsson, InterDigital, Nokia, and Sisvel). The Google licence implied a higher DPU rate than the Apple licences, but both sets of data indicated that the judge’s figure of approximately $0.02 per unit was too low. The Court concluded that a DPU of $0.15 per Apple unit was FRAND, balancing the higher Google rate and the Apple comparables, and consistent with reasonable aggregate royalty stack percentages.

The Court converted this DPU rate to an indicative lump sum licence fee, estimating approximately $502 million for the period from 2013 to 2027, a figure significantly higher than the judge’s $56.43 million lump sum but lower than the appellant’s highest estimate.

Regarding the non-royalty terms of the licence, the Court reviewed the procedural history of the consequential matters hearings and found some procedural irregularities in the judge’s imposition of certain licence terms, but noted that many issues could be resolved by the parties or on paper.

On the interplay with foreign proceedings, the Court analysed the parallel US litigation and the UK court’s jurisdiction to determine global FRAND terms. The Court held that the licence must be worldwide, including patents subject to US proceedings, rejecting any carve-out proposed by the appellant. The Court emphasised that the parties could dispose of US proceedings by agreement, and the English court’s determination was a contractual remedy rather than a competing jurisdictional claim.

The Court found that the judge erred in imposing ongoing court supervision and restrictions on the parties’ conduct in foreign proceedings via paragraph 6(2) of the order and related licence clauses, which were procedurally unfair, lacked proper justification, and were inconsistent with comity principles. The Court invited the parties to agree on outstanding licence terms, with the Court to decide any unresolved points.

On the interest stop date, the Court allowed the appellant’s appeal, finding the judge’s chosen date arbitrary and unsupported, consistent with recent authority that interest should run until payment.

Holding and Implications

The Court of Appeal ALLOWED the appeal on grounds relating to the rejection of expert evidence, the valuation methodology, hold out, averaging, sales volume considerations, and interest stop date. The Court set aside the judge’s lump sum licence fee and determined that a FRAND rate of approximately $0.15 per unit (DPU) was appropriate, leading to a lump sum licence fee around $502 million for the relevant period.

The Court rejected the judge’s novel lump sum averaging approach and reinstated the principle that valuation should rely on selecting the best comparable licences rather than averaging all. It clarified that hold up and hold out are mischiefs to be guarded against in FRAND valuation, and that evidence of hold out in the Apple licences justified discounting their influence.

The Court declined to order a retrial, concluding it could determine the FRAND rate on the existing evidence with appropriate corrections. It also held that the FRAND licence must be worldwide, including patents subject to parallel US proceedings, but found that the judge’s order imposing ongoing court supervision and restrictions on foreign litigation was unjustified and inconsistent with comity.

The direct effect of this decision is to set a revised FRAND licence fee and to remand non-royalty licence terms for resolution, either by agreement or on paper. The Court’s reasoning refines the approach to valuation in SEP licensing disputes, emphasizing proper use of comparables and expert evidence, and clarifies the treatment of parallel foreign proceedings. However, the decision does not establish new binding precedent beyond the application of existing principles to the facts of this case.