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

This dispute came from the mobile technology market, where products depend on industry standards and patent owners often promise to license essential patents on FRAND terms. Optis held patents declared essential to those standards. Apple sold devices that used them.

By the time of this appeal, at least some UK patents in the Optis portfolio had already been found valid and essential. So the case was not mainly about whether Apple needed a licence. It was about what the licence should say and, above all, what Apple should pay.

The court described its role as settling the terms of a FRAND licence after negotiations failed. In practical terms, that meant deciding the commercial terms the parties could not agree for themselves.

The sums at stake were large because even small changes in royalty method can produce very different outcomes when applied across global sales and multiple years.

Negotiations started in 2017 but did not end in a deal. In February 2019 Optis began proceedings in England for patent infringement and for a declaration about FRAND terms. At almost the same time, it also sued Apple in the Eastern District of Texas on US patents from the same wider portfolio.

The English proceedings were managed in stages. Technical patent trials dealt with validity and essentiality. A later FRAND trial was then used to settle the terms of a global licence.

Meanwhile, the US case moved ahead. After a retrial, the Texas court issued a verdict in September 2021 finding five US patents valid and infringed and awarding Optis a lump sum of $300 million for past and future sales.

That created a difficult overlap. If the English court later determined a global licence covering the same period and portfolio, how should that sit with the US judgment? That question became one of the most important practical issues on appeal.

The Court of Appeal’s discussion shows that timing and procedural choices matter. A cross-border patent dispute is not just about who is right on valuation. It is also about what has already happened in other courts and whether one court should interfere with a foreign judgment.

The first issue was valuation. The High Court had fixed a lump-sum licence fee based on an annual figure of $5.13 million over eleven years, making $56.43 million before interest. Optis said that result was far too low.

The valuation fight was really a fight about method. Both sides relied on comparable licences and on a process called unpacking. That means taking a lump-sum licence and estimating the per-unit royalty it effectively represents.

The parties disagreed about which licences were the best comparables, how unpacking should be done, and whether the right way to compare royalties was by a fixed dollar-per-unit figure or by a percentage of device price. Those choices mattered because Apple sold high-priced products, and a percentage-of-price model could produce a much larger royalty than a fixed per-unit model.

The evidence also showed that some licences covered both future use and releases for past unlicensed acts. That made the exercise more complex because the court had to consider how much value related to future sales and how much related to the past.

The second issue was cross-border. The court had to decide how the English-determined global licence should interact with the US proceedings and the US judgment, while avoiding unfairness, double counting and unnecessary interference with a foreign court’s decision.

A further issue was interest. If royalties should have been paid earlier for past sales, should interest continue until payment is actually made, or stop at an earlier date chosen by the court? That point had real financial significance because the licence included a release for past years.

The judgment explains that both sides used comparable licences as a starting point. That is common in FRAND disputes because the court wants evidence of what commercial parties have actually agreed in the market.

But comparable licences are rarely neat. Some are expressed as per-unit royalties. Others are lump sums. Some include cross-licensing features. Some include discounts, regional features, or releases for past conduct. So the court often has to translate one deal into a form that can be compared with another.

That is where unpacking comes in. If a licence is a lump sum, the parties may try to estimate the per-unit royalty hidden inside it. The judgment notes that this can involve assumptions about unit volumes, whether the deal is a cross-licence, whether the royalty is ad valorem or DPU, and whether part of the payment is really for a release of past infringement rather than future use.

The parties also disagreed about whether ad valorem or DPU was the better way to compare value. Ad valorem means a percentage of the selling price. DPU means a fixed amount per unit. The difference matters a lot where one implementer sells premium devices at higher prices than others.

Apple argued that an ad valorem approach could over-reward the patent owner where the product’s high retail price was not caused by the licensed technology. Optis relied heavily on its own licensing material, including the Google licence, to argue for a much higher result than the judge had reached.

The Court of Appeal’s judgment also discusses wider valuation ideas such as top-down cross-check reasoning. In simple terms, that means testing whether the royalty being proposed would imply a sensible overall burden for all patents essential to the standard. The court treated that kind of reasoning as a cross-check rather than the main engine of valuation.

The Court of Appeal said the English court’s jurisdiction in this area is well established. It is based on the contractual effect of the undertaking given by the patentee to ETSI under the ETSI IPR policy.

The judgment also repeated the wider FRAND structure recognised in earlier authority. An implementer is entitled to insist that the terms offered by the SEP holder are FRAND. If the SEP holder offers FRAND terms and the implementer refuses to enter into the licence, the SEP holder may seek a FRAND injunction restraining infringement in the territory for patents found valid and essential.

In this case it was not disputed that the FRAND licence in issue would be global in nature. So the appeal was not about whether a global licence could be relevant in principle. It was about the terms of that licence and the consequences of overlapping foreign proceedings.

The court also referred to the policy concern behind this framework. It is meant to reduce both hold up by SEP holders and hold out by implementers. For businesses, that means courts are alert to unreasonable pressure from either side.

This matters commercially because a FRAND dispute is not only a patent dispute. It can also become a contract dispute, a valuation dispute and a multi-jurisdiction strategy dispute at the same time.

The Court of Appeal confirmed the established position that the English court can determine the terms of a global FRAND licence in a case of this kind. It then dealt with specific disputed consequences of the High Court’s order.

On interest, Optis succeeded. The Court of Appeal held that the judge’s chosen date for interest to stop was unsupported. Referring to later authority, the court said that, in principle, interest should be paid on royalties for past sales so that the passage of time between when payment should have been made and when it is actually made is cost-neutral.

On the US proceedings, the court rejected the idea that the English order should require the US judgment to be vacated or neutralised. The court said that comity mattered where there was a regularly obtained foreign judgment from a competent court.

The decisive point was procedural fairness. The court held that Apple was responsible for the US proceedings having continued to judgment before its later change of stance. In those circumstances, the least-worst solution was to treat the final US judgment as a floor for the royalties payable under the English licence.

The court also considered that ongoing supervisory or interfering terms aimed at the foreign proceedings were not justified. It criticised provisions that created unclear or continuing court supervision over what happened in the US case.

So the appeal did not simply produce a fresh royalty number in public view. It clarified how interest should work in principle and how an English FRAND licence should accommodate an existing US judgment without requiring improper interference with it.

Most businesses will never run a FRAND appeal in the Court of Appeal. Even so, the case gives useful process lessons for businesses using standards-based technology.

First, identify standards exposure early. If your products depend on mobile, wireless or similar standards, licensing risk should be checked before launch, not only after a claim arrives. Once products are in market, past sales and interest can become part of the dispute.

Secondly, keep negotiation records. The judgment shows how much weight can fall on licensing conduct, comparable deals and the commercial logic behind offers and counteroffers. A clear written record helps if the dispute later turns on whether your position was commercially reasonable.

Thirdly, align cross-border strategy. If there is litigation or negotiation in more than one country, inconsistent positions can create expensive overlap. This case shows that one court may have to work around the consequences of what happened elsewhere.

Finally, be realistic about supply chain assumptions. Importers and brand owners should not assume a manufacturer has already solved all SEP licensing issues. Check who holds licences, who bears infringement risk and whether contracts allocate that risk clearly.

The practical lesson should be kept modest. This is a specialist patent appeal. For most businesses, the reusable point is not how to price a telecoms SEP portfolio. It is how to run a licensing process carefully enough that valuation, past sales and foreign proceedings do not spiral into a much larger problem.

If your business receives a licensing approach involving standards-based technology, the most useful practical step is often disciplined record-keeping. Courts in these disputes look closely at what was offered, how the other side responded and whether positions were commercially coherent.

That does not mean you must accept the first demand made. It means you should review it properly, respond in a measured way and document the basis for any counterproposal.

Where there are overseas proceedings, keep one internal owner for the dispute and one shared chronology. Finance, product, procurement and legal teams should all work from the same assumptions about sales, product scope, timing and settlement options.

If you rely on suppliers, ask direct questions. What standards does the product implement? Is there already a licence in place? Does the contract allocate infringement risk and indemnities clearly? Those checks are often more useful than trying to master the finer points of FRAND valuation law.

If a dispute is already live, avoid fragmented decision-making. A licensing position taken in one country may affect settlement leverage, credit arguments or procedural fairness points somewhere else. This case is a good example of how those issues can become central later.

FRAND means fair, reasonable and non-discriminatory. In practice, it is the licensing standard often used for patents declared essential to an industry standard.

Parallel proceedings means related court cases in more than one country at the same time. That matters because one judgment can affect the commercial room for manoeuvre in another case.

A global FRAND licence is a licence whose scope is not limited to one country. In this case, it was common ground that the FRAND licence in issue would be global in nature.

For most smaller businesses, the practical relevance is narrow but real. If you sell standards-based products, the key lesson is to sort out licensing responsibility early and keep your records straight.