Warner Bros Discovery Inc & Ors v Nokia Corporation & Anor

This is my judgment on the interim payments to be made by Warner Bros Discovery (“WBD”) and by Paramount in these RAND proceedings. The need for a decision of the court on an interim payment arises because, as reflected in an order of 16 March 2026, the parties came to an accommodation (referred to as the “Agreed Mechanism”) whereby they agreed (1) to withdraw all parallel litigation in this and other fora and refrain from any new claims, (2) to enter into a global RAND licence covering the Nokia Video Portfolio (“NVP”) on terms to be settled by the Court at trial, and (3) for WBD and Paramount to make any interim payment that the Court may determine, pending the final RAND licence. The NVP covers encoding and decoding patents, and various NEPs. NOTE: this version of my judgment is public and has redactions to remove confidential information. At the form of order hearing, on a date to be set, I will consider the parties’ submissions and submissions from affected third parties or members of the media or public as to which if any of the redactions should be maintained.

The Agreed Mechanism came about because Nokia decided, following a hearing before me in January 2026, that a single global RAND determination in the UK was an acceptable way forward to resolve the parties’ global disputes, and preferable to worldwide litigation. So it made a proposal on 19 February 2026 which in due course became the Agreed Mechanism. In circumstances described in an oral judgment that I gave on 20 May 2026, Nokia’s conduct in the way that its offer became public was unsatisfactory, but that is now history and of no relevance to determining the appropriate interim payment amounts, and Nokia’s approach in making its offer in February and progressing it with WBD and Paramount was constructive and welcome, leading to the elimination of extensive unnecessary litigation between the parties. The Agreed Mechanism will efficiently lead to a RAND result, ensuring Nokia gets paid properly, and with no risk of supra-RAND rates arising from injunction pressure.

An important part of the Agreed Mechanism was the opportunity for Nokia to seek an interim payment from WBD and Paramount, and that is what I now have to address.

Mr Saunders KC argued the application for Nokia and Mr Piccinin KC (for WBD) and Ms Pickard KC (for Paramount) divided the advocacy in opposition. I am grateful to all of them.

I recently considered the applicable principles in TP-Link v. Huawei [2026] EWHC 179 (Pat), with reference to the earlier Court of Appeal decisions where interim licence (F)RAND terms had been set. I will take my decision in TP-Link as read for the purposes of this judgment but will refer to various of the principles from it where they are important.

The hearing before me on this issue took place over two days, on 7 and 19 May 2026. The reason for the split hearing was that although (contrary to my expectation from reading the skeletons) one day would have been enough for the argument, as is desirable for an exercise which ought to be quick, simple and low cost, there was one significant issue (the Nokia Lump Sum Offer, “NLSO”) which for reasons that will become apparent was potentially subject to privilege/confidentiality issues, and in relation to which WBD’s advisers were in the dark at the time of the 7 May hearing. This was resolved in time for the resumed hearing on 19 May.

It will help at the outset to summarise the parties’ positions (in doing so I will refer to various bilateral licence agreements that Nokia has made with other parties by code letters, the identities of the other parties being asserted to be confidential; a key to the code letters is a confidential Annex to this judgment):

I have given the sums involved by reference to ranges and approximate amounts because of confidentiality issues. The exact numbers do not matter at this stage. It should also be noted that I have referenced amounts in both euros and US dollars, which the parties had done and which unfortunately is not very convenient (and contributed to an error I made in the draft judgment supplied to the parties).

It is a major feature of the arguments that WBD and Paramount have been using the relevant streaming standards for a long time, since 2011. Nokia says that on the basis of TP-Link and the earlier Court of Appeal case law cited in it, the final RAND amount will require payment for all use back to then, with interest, and that the interim payment should be assessed on the same basis. This makes a major difference to the amounts payable. For example, the NLSO did not go back nearly as far as 2011, and the additional payments back to that date with compound interest increase the amount greatly.

The parties are all agreed that WBD and Paramount should each make a non-refundable interim payment (“non-refundable” in the sense that it would not be adjusted after the final RAND decision at trial), to reflect the irreducible minimum that Nokia will be entitled to, based on their pool scaling cases. Thereafter their positions diverge hugely.

WBD and Paramount first say that there should be no adjustable interim payment at all, primarily because Nokia does not have a real use for the money given that it says that it cannot recognise it as revenue in its accounts.

Second, they say that Agreements A, B and C are all completely unreliable because the absolute amounts payable by the licensees were so small that they did not reflect real economic deals for use of the technology concerned, but rather the counterparties paying what was asked because e.g. the costs of litigating would outstrip the price. They say agreements A to C are so unreliable for this reason that using them as part of a mid-point approach should be rejected altogether.

Third, they rely on Agreement D which they said should be used instead of Agreements A to C.

Fourth, they rely on the NLSO.

Fifth, they say that payment for the past should be strictly limited because Nokia did not approach them until more recently, because there was an industry practice of not requiring past payment, because Nokia did not even have a video licensing program until much more recently, and because (while Nokia stood by) the industry adopted and locked itself into the relevant video standards in the understanding that the technology would be paid for by device manufacturers and not the streaming companies. Although their opposition to paying for all past use is general, they do present scenarios based on Agreement D for consideration which cover all past use (amounts are given for all past use, for past use back to 2023, and the average of those two; the amounts for all past use are over double the amounts for use going back to 2023).

Nokia’s position is as follows.

First, it says there should be a substantial adjustable part of the interim payment and that the argument based on the limited use to which it could put the money is not relevant, having been rejected by the Court of Appeal in e.g. Ericsson v. Lenovo.

Second, it says that Agreement A, or Agreements A to C combined cannot be simply rejected at this stage but are properly arguable at trial, and so are a usable and appropriate starting point for a mid-point approach. It says that uncertainties about Agreement A to C are all factored into taking the mid-point.

Third, it says that Agreement D is not a meaningful comparable and that Dr Lopez’s analysis of it is obviously unreliable (I explore the reasons below, but they are confidential in any case).

Fourth, it says that if Agreement D is to be considered at all then account must be taken of a non-FRAND factor, significant value which it received not reflected in the agreement itself. It provides an alternative analysis on that basis leading to much higher figures than WBD and Paramount advocate. Again, the details are confidential.

Fifth, it says that the NLSO is also unreliable and inappropriate to be used, but that if it is used then it should be adjusted to cover the period all the way back to 2011, with appropriate compound interest, and again it provides an analysis leading to much higher numbers.

Nokia emphasises that while the absolute amounts sought are large, that is because the periods of unlicensed use are very long (with the effect of compound interest on top) and it emphasises that the businesses of streaming companies are built on codec technology. I thought the second factor was rather overplayed: the streaming companies’ businesses are in content creation and the video standards in this case are only a means to that end, albeit an important one, but it does not ultimately matter to my decision and the real point founded in the authorities as to why the amounts sought are potentially large is the first one.

I have no hesitation in rejecting the argument by WBD and Paramount that there should be no adjustable portion at all. As I have mentioned, the basis for this contention is that “[t]here is no discernible benefit to Nokia of a large adjustable interim payment”, given that the money could not be recognised as revenue in Nokia’s accounts. Nokia accepts the fact that an interim payment will not as a matter of accounting practice be recognisable as income, as it did in Alcatel/Nokia v Amazon [2025] EWCA Civ 43, when it was resisting any interim licence declaration.

However, I am not satisfied that this means that the money is of no use at all to Nokia. It might be put in an interest-bearing account, or be used to offset some debt.

Furthermore, the argument now made by WBD and Paramount was rejected in terms by the Court of Appeal in Lenovo v Ericsson [2025] EWCA Civ 182 at [119-122], partly on the evidence and partly as a matter of principle (it should be noted that the argument was being made by the patentee to resist an interim licence declaration being made at all, but that makes no difference to my reasoning or decision).

WBD and Paramount argued that there would be real world prejudice to them in making a significant adjustable payment because they would not have the use of the money in the period from now until the RAND trial, when they might get some or all of it back. The evidence on this was not at all convincing since there was nothing specific that they said they could not do. It was in the end largely just the reverse of Nokia having or not having the money.

At a more fundamental level, I think that arguments about the usefulness of the adjustable payment to the patentee are likely to be beside the point. The court is trying to reflect what willing parties would agree at an interim stage. They would accept that there was uncertainty about the final RAND amount and that either of them could turn out to be (more) right or (more) wrong in the end, but that what was certain was that the implementer was going to use the technology in the meantime, and had already been using it, and ought to pay an amount which reflected the possible range of outcomes. It is not only patentees in pressing need of money who ought to be entitled, as a matter of what is fair and reasonable, to meaningful interim payments.

In TP-Link I noted that the mid-point approach has been consistently used but is not mandatory. It is worth noting that it has been applied as the mid-point between the parties’ most recent offers, generally in circumstances where the parties’ offers have converged towards each other once litigation is underway or imminent. This does make the interim payment reached somewhat time-dependent, as evidenced by the further offer made by Ericsson during the pendency of the appeal in Panasonic v Xiaomi and although it was not said to matter greatly there, I think the court should be astute not to let timing make the amount too adventitious.

I noted in TP-Link at [33] that the benefits of the mid-point approach are that it is fair, simple and robust and can be done quickly and cheaply without a mini-trial. But, as I said already, it is not mandatory.

The present case is somewhat unusual with respect to using the mid-point approach. Although it is common for the parties to be far apart at the start of (F)RAND negotiations (as I noted in TP-Link at [27]) there is usually a degree of convergence over time, as I have referred to above. In the present case the difference between the camps is unusually large, however, and (leaving aside for one moment the NLSO and focusing on what has happened during the course of the litigation) there has not been any meaningful convergence between Nokia’s position based on Agreements A to C and WBD’s and Paramount’s based on pool scaling. The two approaches are radically different in concept; this is not a case where both sides are using bilateral comparables and their positions start to converge after disclosure of additional comparables, or of more context to the comparables already in issue. It is in a sense difficult for WBD and Paramount to move at all in a meaningful way while maintaining the primary argument for pool scaling, since even a significant tweak to the details of the pool scaling approach would leave it, effectively, just as far from Nokia’s starting position. Put another way, while the pool scaling method remains in play the mid-point approach to the adjustable component will always be half of Nokia’s position.

I think it is also important to note that if WBD’s and Paramount’s pool scaling approach is rejected at trial, the effect will not just be that matters shift a little in Nokia’s favour. Such a rejection would at a stroke move the amount payable to, probably, at least the Agreement D amounts or NLSO amounts, a step change of tens of millions of dollars and multiple orders of magnitude.

These considerations do not necessarily make the mid-point approach inappropriate, (and Mellor J went ahead and used it in a somewhat similar situation in Acer and others v Nokia [2025] EWHC 3331 (Pat), albeit from what I can tell on the basis of less evidence and argument, for a shorter period, and in the context of a smaller ratio between the pool scaling rate and the mid-point) but they are signs that some additional thought is needed, at least.

There was some discussion during submissions about circumstances in which the mid-point approach might reflect uncertainties in the parties’ cases twice over. In particular, is it unfair on the patentee to scale down its current offer to reflect uncertainty over the recoverability of royalties going back into the past and then to apply the mid-point? Does this involve two deductions for the same uncertainty? I think that the function and effect of the mid-point approach is to reflect the uncertainty over which side in the dispute is right about the underlying method used (here, pool-scaling versus bilateral comparables), and if there are uncertainties within the underlying method used (on either side, or both) it may be appropriate to reflect that separately. So two adjustments of those different kinds are doing different things. Although not articulated in this way, I took a similar approach in TP-Link in relation to the argument about payment for the past, right back to the start of use of the technology: having concluded that Huawei was clearly right on the law I used its rack rate approach with full past payment as the top end of the mid-point, but had I thought past payment was more arguable I could have considered allowing for that and then using the mid-point approach. But I do not pretend that this is an exercise in mathematical precision, or a rigid rule.

I said in TP-Link that (F)RAND interim payments should be assessed without a mini-trial. I think there are two aspects to this.

The first is that the court should not try to resolve issues which are arguable either way. That is for trial. It may decide issues where one side is very clearly right, as I did in TP-Link in relation to past payment.

The second is that the exercise should not be allowed to become burdened with too much detail, of the kind that may be appropriate for a trial but is not appropriate for a short interim hearing. The parties have a role to play in this, but with a lot of money at stake it is predictable that they will not skimp on detail, and so it has proved in the present case: the parties have gone into every small point. The court has a role to play, in trying to control how much detail comes in, but also, and pertinently for present purposes, in not allowing itself to be sucked into trying to factor in every small detail to what should be a simple and robust decision.

WBD and Paramount invited me to reject Agreements A to C entirely on the basis that they are clearly artificial given their low absolute value, and given (it was said) the disproportionate and widely varying amounts in absolute terms that result from their use as starting points. Nokia was similarly dismissive of the pool scaling approach, for different reasons.

I do not consider there is enough to reject either approach entirely at this stage, applying the quasi-summary judgment standard that I referred to in TP-Link, and to attempt to do so would involve a mini-trial. For example, WBD and Paramount made specific arguments about the kind of business carried on by the counterparty to Agreement C as it compared to those of WBD and Paramount, the effect of which I am not in a position to assess with sufficient certainty.

However, I do consider there is real force in the point that Agreements A to C vary very significantly from one to another in the amounts they imply for WBD and Paramount (by a factor of about [REDACTED]) and it looks arbitrary for Nokia simply to choose the best Agreement. I do not think Nokia can reasonably invoke the principle that the patentee can choose any of multiple approaches which are all (F)RAND when this variability itself gives rise to a tenable argument that not all of the agreements are (F)RAND in the first place. Nokia in a sense recognised this by its fallback averaging approach applied to Agreements A to C, which gives a significantly lower adjustable amount but still is affected by the possibility that the most favourable Agreement(s) in particular is/are not FRAND.

[REDACTED].

[REDACTED].

[REDACTED].

[REDACTED].

[REDACTED].

[REDACTED].

The NLSO has the advantages of being (a) relatively simple compared with all the other approaches, (b) an actual offer, and (c) between Nokia and one of the parties before me for the same (or at least very similar) portfolio. As to (c) it makes really no difference that it was only made to Paramount and not to WBD since no one says there is a material difference between them other than subscriber numbers, which is just a question of applying the agreed ratio.

In my overall assessment, and taking into account the complexities that I will come on to consider, these positives are sufficiently significant that the desire for a simple and robust approach at this interim stage means that I should give very significant weight to the NLSO in preference to the other possibilities put forward by the parties. However, I do not think it would be right to entirely ignore Agreements A to C and use the pool scaling approach as something akin to the bottom end of a mid-point approach. That could be unfair to Nokia as it would involve completely rejecting the possibility that it might succeed on Agreements A to C at trial, while continuing to factor in WBD’s and Paramount’s best case. I return to this below. In addition, the NLSO, while an offer, is not Nokia’s latest offer in the course of converging towards the implementers’ in the same sense as Ericsson’s most up to date offer at the date of the Court of Appeal judgment in Panasonic v Xiaomi. Nokia has not given up on Agreements A to C and retreated to the NLSO; the NLSO came first and was made in a different context for a different purpose.

The complexities that the NLSO introduces are all to do with the time period it covered. It covered from [REDACTED] to [REDACTED] and was for [REDACTEDper annum]. So it did not go back nearly as far as 2011 when WBD and Paramount started streaming. WBD’s and Paramount’s answer to this is simply that I should take the period which it did cover, use it as the top end of a mid-point approach (subject to adjusting it to reflect the conventional approach that an interim licence should be until the expected date of judgment in the RAND trial, in early 2027), and scale as appropriate for WBD. Nokia responds, as mentioned above, that it is entitled to payment right back to 2011, in particular on the basis of the decision of the Court of Appeal in InterDigital v Lenovo [2024] EWCA Civ 743. WBD and Paramount retort that Nokia specifically proposed the NLSO in 2024 when that decision had been made, so it can be inferred that Nokia thought it was only entitled to go back a short time, despite it.

In the event that I were persuaded that the NLSO should be used but adjusted to cover a longer period in the past, then it would be necessary to assess what WBD’s and Paramount’s usage would have been in earlier years, and to decide what to do about interest for those earlier years. For the resumed hearing Nokia provided an analysis which assumed that WBD’s and Paramount’s usages had tapered up linearly from their respective inceptions, and applied interest at the rate for which Nokia contended accordingly. Paramount criticised this on the basis that it overstated the total number of subscribers for 2011-2021 (which it does) but Nokia pointed out that if Paramount’s numbers were used then the effect would be to shift the balance of usage back in time quite heavily, which would lead to a lot more interest being due. I think Nokia’s point is correct, but one cannot tell from the information provided whether it is enough to balance out the “additional” subscribers.

None of these complexities arise unless Nokia is entitled to payment going back to 2011, or at least considerably earlier than the period provided for by the NLSO. So that is the issue to which I now turn.

The Court of Appeal in InterDigital v Lenovo gave a robust decision that in general all past use should be paid for in a licence on FRAND terms, but it did admit of a potential exception if the implementer could show an industry practice of forgiving past sales (at [198]). TP-Link tried to rely on that in TP-Link v Huawei, and I rejected it at [39]-[41] on the basis that it was not enough to show merely that past sales were sometimes released, there had to be an industry practice.

In the present case, WBD and Paramount can again show that Nokia at least sometimes does not explicitly recover for all past sales, but, again, I do not think the evidence rises to the level of an industry practice. As I said in TP-Link, that position may change at trial with further evidence and/or disclosure.

As to the fact that InterDigital v Lenovo had been decided when Nokia made the NLSO, the existence of a decision, even an important one, in the UK alone did not mean that the industry approach changed overnight. Clearly there was (and probably continues to be) a period of adjustment to the UK decision, and the fact that Nokia did not immediately insist in the NLSO (which was not made in UK litigation or any litigation specifically) on payment right back to 2011 cannot be taken to imply that it was consciously passing it up. It may well just have been accepting the reality that to strike a deal it would have to give up on past sales, as has often happened in the (F)RAND world.

Therefore, taking into consideration only the matters which have arisen in previous cases I would hold that Nokia is entitled to payment right back to 2011. But this case has some unusual features which I think differentiate it: I have referred already to WBD’s and Paramount’s position that the industry developed on the assumption that the technology would be funded by royalties on streaming devices and the contention that companies such as Nokia did not make clear for many years that they would seek substantial royalties for streaming activities. They say it is material that Nokia did not try to monetise its video patents until much later than 2011 so it is not realistic to say in the present situation (Cf. the arguments in InterDigital v Lenovo) that WBD and Paramount’s proper course was to seek out Nokia at the outset of using the technology and try to arrange a licence. Nokia was not in the business of offering such licences then, they say.

For these unusual reasons I think there is some doubt over whether Nokia is entitled to recover all the way back to 2011. Not being entitled to recover all the way back would not necessarily mean that it is only entitled to the past period covered by the NLSO and I acknowledge that there is a significant degree of uncertainty over what the appropriate period is if it is not right back to 2011. In any event, although it is a subjective exercise I must make allowance for the doubt.

Whatever the past period to be covered, compound interest ought to be payable in respect of it. I see no reason to depart from the general position on that in addition to notionally adjusting the past period. And I accept Nokia’s contentions as to the rate.

Nokia’s revised calculations based on the NLSO were done by taking the annual rate offered in US dollars, applying the linear taper described above, and then converting to euros at the rate applicable when the NLSO was made in 2024, which was 1.0583. The more contemporary rate used in the pleadings is about 1.18, so the difference is appreciable, amounting to about a 12% uplift for Nokia. I heard only brief argument about this, without the citation of any authority, but my conclusion based on it is that it would be unprincipled to accept Nokia’s position. The offer was made in US dollars and WBD and Paramount will be paying now, not in 2024. I am going to make my decision in US dollars and Nokia can convert the money to euros if it wants to. With fuller argument the position at trial could turn out to be different.

According to the NLSO-based linear-tapering calculations done by Nokia (Darvill 11 paragraph 14), converted back to US dollars at the rate of 1.0583 referred to above, the full mid-point based amount for the period back to 2011 with interest on the basis contended for by Nokia is [REDACTED] for Paramount. Applying the 4:3 ratio that is [REDACTED] for WBD. By working in US dollars like this I have taken out of account Nokia’s argument on exchange rates. I should mention that in the draft version of this judgment I made a rather rudimentary error in this conversion and these are the corrected numbers as agreed between the parties. Of course I should not have made the error but in fairness to myself it was partly attributable to the confusing situation of both currencies being used in the arguments and evidence, and sometimes in the same table.

Paramount’s position on the NLSO is that if I were to use it as a basis for an interim payment, I ought simply to take the annual amount and apply it to the period of the interim licence (i.e. through until the likely time of judgment from the RAND trial), then take the mid-point between that and the pool scaling based amount. The evidence in Pereira 9 was that this was [REDACTED] which would scale to [REDACTED] for WBD. Essentially what this does is to treat the NLSO as giving an annual rate, giving nothing for the disputed period of prior use (or interest in respect of it).

What I need to do is to assess how to reflect my view that WBD and Paramount may have to pay for use all the way back to 2011 or may not, and to do so in the overall context of the various arguments and approaches to assessing the right refundable amount.

It is impossible to be quantitative about the numerical effect of the uncertainty over payment for the past. I cannot meaningfully assign a percentage chance to Nokia’s obtaining payment for earlier use at trial, or to the chance of WBD and Paramount fending it off altogether, and there are possibilities that payment might be due for the past but not all the way back to 2011. So this assessment can only be a subjective one.

It is tempting to take the average of the midpoints, which would be [REDACTED] for Paramount and [REDACTED] for WBD (these figures are corrected to allow for the point referred to in paragraph 58 above). In the overall context of the application before me and the TP-Link approach I would regard this as being at a broadly appropriate level, but for the points I mentioned early in this judgment, that (a) if the pool scaling approach is rejected then the amount payable by WBD and Paramount would jump dramatically, and (b) I should not unfairly go for a method which gives no weight to the chance of Nokia winning on Agreements A to C while at the same time using the pool scaling approach of WBD and Paramount as the minimum. As to (a) I am not dealing with a situation where one might notionally say there is a range of equally likely answers between Nokia’s top position based on the NLSO and the pool scaling results. As to (b), though a more minor point since I have concluded that Agreements A to C (and Agreement D) are less reliable for the task I face at this stage, working to a mid-point between NLSO-based numbers and pool scaling numbers gives effect to there being some possibility of WBD and Paramount succeeding completely but assigns no weight to the possibility of Nokia doing considerably better than the NLSO by prevailing to some degree on Agreements A to C.

Balancing all this up, I think it is fair and reasonable to set the refundable amounts at [REDACTED] for Paramount and [REDACTED] for WBD, i.e. adjusting up somewhat from the average of the NLSO midpoints. I regard this as true to the principle that I should not conduct a mini-trial and generally faithful to the mid-point approach, although I have modified it in an admittedly subjective way to the unusual range of possible methods for reaching an outcome, and the potentially large and abrupt impacts of choosing between them that this case presents. These numbers are also adjusted from those in the draft judgment to reflect the point at paragraph 58 above. In submissions after the provision of my judgment in draft when pointing out the issue, Nokia submitted that the adjustment should be a simple pro-rating from the numbers in the draft judgment, which would give higher numbers than those above. WBD and Paramount submitted that I should not change the ultimate numbers given in this paragraph at all, on the basis that my intention was to reach a round number (pre-adjustment, [REDACTED] for Paramount) reflecting my overall assessment and not a strictly mathematical exercise from the average of the midpoints. There is force in both submissions and indeed I was not trying to undertake a strictly mathematical approach here, but I do not think it would be intellectually honest not to acknowledge that my miscalculation understated the numbers to Nokia’s disadvantage, even if thereafter there was a subjective step. So I have increased the number somewhat from the draft judgment, though not as much as Nokia submitted, and still arrived at round numbers.

There was a complicated argument about how to deal with channel partner subscribers, who get WBD and/or Paramount content but from a third party. This will be an issue at trial, but a further advantage of the NLSO is that it is not affected by the point since it scales from a set amount per year for Paramount as of 2024.

The parties were agreed about the non-refundable amount of the interim licence. I assess the refundable amounts as [REDACTED] for Paramount and [REDACTED] for WBD.