Optis v Apple: Court of Appeal redefines FRAND principles for global SEP licensing

In a landmark judgment handed down on 1 May 2025, the Court of Appeal overturned the High Court’s decision on the FRAND (fair, reasonable and non-discriminatory) terms for a global patent licence between Optis and Apple. The Court of Appeal’s ruling not only resets the specific outcome of this long-running dispute but also offers substantial guidance for parties engaged in global SEP (standard-essential patent) licensing disputes. The decision reaffirms core principles of UK FRAND case law and has likely cemented the UK’s status as a leading global venue for such matters.

Comparables-Based Valuation Prevails

The Court of Appeal reinstated the primacy of a comparables-based approach to FRAND valuation, decisively rejecting the High Court's averaging methodology. The trial judge’s attempt to generate a royalty rate by averaging unpacked lump sums across a broad range of licences was found to lack economic rigour. Instead, the Court adopted a pragmatic yet principled approach, setting a FRAND rate of $0.15 per unit.

This $0.15 rate sits between the unpacked figures from the Optis-Google licence and those derived from four of Apple’s own licences (with Ericsson, InterDigital, Nokia and Sisvel). The Court applied what Birss LJ described as a "broad axe" to arrive at a figure it considered commercially fair, evidentially grounded, and within a reasonable band of reliability. It accepted that perfect precision was neither possible nor required. The Court also noted that grossing up the $0.15 rate to represent a full 100% royalty stack implied a total royalty burden of approximately 6.3% of Apple’s historical ASP ($625), 3.9% of a more current ASP ($1000), and 8.4% of a Google ASP ($470)—all commercially reasonable figures.

Expert Evidence and Procedural Fairness

A critical issue on appeal was the treatment of expert evidence. The trial judge had summarily rejected the evidence of both parties’ accountancy experts without giving them an opportunity to respond to allegations of bias or overreach. The Court of Appeal found this to be procedurally unfair.

It clarified that the role of experts in FRAND disputes is not to determine comparability, which remains a matter for the court, but to improve the reliability of financial data extracted from licence agreements. Both experts had appropriately sought to "unpack" lump sums to produce useful DPU (dollar-per-unit) equivalents for analysis. A key conceptual error by the trial judge was to conflate the notions of comparability and reliability—two distinct components in the FRAND valuation process.

Rejection of Averaging Methodology

A significant part of the judgment critiques the trial judge’s method of averaging figures from 19 licences (18 of which were Apple’s) to estimate a stack royalty, from which a portion was attributed to Optis based on its estimated 0.38% share. This method, while superficially equitable, was described as economically unjustifiable.

The Court found that the divergence in the figures made averaging inappropriate. Not all licences were plausibly FRAND, and equal weighting obscured differences in bargaining power and commercial context. The Court preferred to select the best comparables and rely on robust unpacking analysis. It criticised the inclusion of the Google licence in the table of lump sums without adjusting for unit volume, noting that this materially underweighted its importance.

Additionally, the Court rejected Apple’s attempts to salvage the methodology by excluding Google or focusing on internal consistency among Apple’s own licences. The judge’s dismissal of volume considerations was a fatal flaw, rendering the averaged result unreliable. The methodology was described as an example of the dilution problem warned of in Cimetidine—where stronger comparables are obscured by the inclusion of weaker ones.

Addressing Hold-Out

Another core issue was the treatment of hold-out. The Court rejected the trial judge’s attempt to distinguish between legitimate and illegitimate forms. Drawing from the Supreme Court’s Unwired Planet ruling, the Court of Appeal confirmed that any form of hold-out or hold-up undermines the FRAND process.

In Apple’s case, its insistence on patent-by-patent licensing, which the judge had even described as "indefensible," was viewed as a clear form of hold-out. The Court noted that larger licensors achieved better terms, suggesting Apple used its market power to suppress rates from smaller licensors like Optis. The Court clarified that the correct test is the "willing licensor/willing licensee" benchmark, excluding hold-out and hold-up distortions.

No Retrial Needed

Apple argued that Optis’s revised reliance on the Google licence warranted a retrial. The Court disagreed, pointing out that the Google licence had featured prominently at trial. Further, both parties had converged on many key assumptions by the time of the appeal. A retrial would add little and delay finality unnecessarily. The Court therefore determined FRAND terms itself.

Citing Simetra v Ikon, the Court reiterated that a retrial is a last resort and should only be ordered where no other just course is possible. Given the narrowing of disputes and shared inputs (e.g., stack share and use of Innography data), a retrial was unnecessary.

Determination of the Royalty Rate

The Court analysed unpacked data from the Google licence (adjusted for Optis’s stack share) and the four strongest Apple comparables. It concluded that the trial judge’s original $0.01–$0.02 per unit rate was too low. Conversely, the rate implied solely by the Google licence appeared too high when benchmarked against ASPs (average selling prices).

Applying the $0.15 DPU rate to historical and projected device sales (2013–2027), the Court arrived at a lump sum of $502 million. With interest, this figure is likely to exceed $700 million. Notably, the Court rejected the trial judge’s approach of deriving lump sums via average stack rates. The $0.15 DPU was chosen for being well-grounded in unpacked data and commercially realistic when cross-checked through a top-down lens.

Both experts had calculated similar unpacked rates using a "free release" method, producing results broadly in line with the simple DPU figures. This gave the Court further confidence that the chosen rate was fair and reliable.

Treatment of US Proceedings and the EDTX Judgment

Optis also challenged provisions in the High Court’s order that constrained its conduct in US litigation, particularly in respect of the EDTX judgment awarding damages for infringement of five US patents. The Court of Appeal found paragraph 6(2) of the trial order procedurally unfair, likening it to an anti-suit injunction imposed without proper process.

The Court also addressed how to reconcile the FRAND royalty with the existing US damages award. It accepted Optis’s third proposed solution: treat the US judgment as a "floor" within the total FRAND sum. This avoided double recovery while respecting comity, acknowledging that the US judgment was lawfully obtained and should not be disturbed by the English court.

The Court noted that Apple’s late-stage change in position—agreeing to take the UK-determined licence—had created the conflict. Optis had not acted inconsistently or sought double recovery, but instead offered pragmatic solutions. The Court rejected Apple’s arguments that Optis needed expert evidence to justify its proposal, finding the legal principles sufficient.

Broader Implications for FRAND Jurisprudence

The decision reinforces the UK’s central role in resolving global SEP licensing disputes. It provides clarity on key points:

• Comparables remain king: Judicial reliance should focus on the most probative licences rather than averaging disparate figures.

• Hold-out analysis matters: Courts must scrutinise how bargaining power and market conduct shape licence terms.

• Expert evidence has a place: Unpacking techniques remain valuable where properly conducted.

• Top-down remains relevant: Though not primary, it can be a useful cross-check.

The adoption of a broad-axe approach also reflects a more commercially nuanced style of judgment, mirroring the pragmatic reality of licensing negotiations. While not mathematically exact, the Court’s $0.15 DPU outcome is grounded in a synthesis of reliable inputs and judicial discretion.

The ruling may also act as a benchmark in future UK FRAND disputes, particularly in determining when foreign awards should be incorporated into global licences and how to approach conflicting litigation outcomes without resorting to retrials or anti-suit orders.

Conclusion

This Court of Appeal judgment in Optis v Apple will have wide-ranging consequences for SEP disputes, reinforcing English law's position on key FRAND issues. It strikes a careful balance between economic evidence, procedural fairness, and legal principle. More than anything, it provides a clearer roadmap for how courts will determine FRAND terms in a world of increasingly complex and international patent licensing battles.