Getty Images (US), Inc & Ors v Stability AI Ltd [2025] EWHC 3419 (Ch)

[2025] EWHC 3419 (Ch)Case No IL-2023-000007
IN THE HIGH COURT OF JUSTICE
BUSINESS AND PROPERTY COURTS OF ENGLAND AND WALES
INTELLECTUAL PROPERTY LIST (ChD)
Venue The Rolls Building, 7 Rolls Buildings, Fetter Lane, LondonDate 17 December 2025
EC4A 1NL
Before
MRS. JUSTICE JOANNA SMITH DBE
Between
STABILITY AI LTDDefendant
Ms Lindsay Lane KC, Ms. Jessie Bowhill and Mr. Joshua Marshall (instructed by Fieldfisher LLP) for ClaimantsMr. Hugo Cuddigan KC and Mr. Edward Cronan (instructed by Bird & Bird LLP) for Defendant
Approved Judgment(On Costs)Transcript of the Stenograph Notes of Marten Walsh Cherer Ltd.,2 nd Floor, Quality House, 6-9 Quality Court, Chancery Lane, London WC2A 1HP.Telephone No: 020 7067 2900. DX 410 LDEEmail: info@martenwalshcherer.comWeb: www.martenwalshcherer.com
[1]I must now determine the question of costs arising in light of my judgment in this matter, handed down on 4 November 2025 ([2025] EWHC 2863 (Ch)).[2]Both parties have served evidence in connection with this issue, which I have read. I have taken all of that evidence and all of the parties' helpful submissions into account in arriving at this judgment.[3]I need not rehearse the outcome of the trial at any length. In short, the claimants abandoned very substantial parts of their claim shortly before closing submissions at trial, specifically claims referred to in the judgment as:(1) the Training and Development Claim;(2) the Outputs Claim; and(3) the Database Rights Infringement Claim. These claims had involved reliance by both parties on significant quantities of evidence and submissions, much of which was rendered irrelevant by reason of the abandonment. These claims have now been dismissed by the order I have made at this hearing.[4]There remained only three main claims for final determination: the Trade Mark Infringement Claim, the Passing Off Claim (which was closely tied to the Trade Mark Infringement Claim and had no real life of its own) and the Secondary Infringement of Copyright Claim, which turned on an issue of statutory construction. The claimants succeeded in part on the Trade Mark Infringement Claim. It was unnecessary to determine the Passing Off Claim. The defendant succeeded on the key issue of construction in the Secondary Infringement of Copyright Claim.[5]My decisions on the main issues are summarised at [757]-[758] of the judgment.[6]Against that short background, the parties are now at loggerheads over the approach the court should take to costs.[7]In brief summary only, the claimants contend that if there is an overall winner of the litigation, it is they. The claimants point out that they have obtained substantive relief in the form of a finding of trade mark infringement in relation to Stable Diffusion Models v1.x and v2.x. They say that this outcome was robustly resisted at all times by the defendant, that they could not have obtained this relief without fighting the trial and that it is plainly of considerable value to them. They point to the undertaking now given by the defendant in respect of Models v1.x and v2.x, an undertaking not offered previously, together with the defendant's agreement to the principle of a damages inquiry. They say it can be inferred that on such an inquiry money will be moving from the defendant to the claimants.[8]By way of alternative, the claimants contend: first, that if there is no overall winner then the costs, including general costs, should be apportioned between the different claims or; second, that even if the defendant is the overall winner, the claimants should be entitled to recover the costs attributable to the Trade Mark Infringement Claim.[9]Ultimately, therefore, as I understand the claimants' submissions, it matters not which approach the court takes because, with one wrinkle to which I shall return, each produces much the same result. This is because, as I understand it, whatever view the court takes of these arguments, the claimants accept that they must pay the defendant’s costs of “the causes of action on which they did not win”, i.e. everything except the Trade Mark Infringement Claim. Thus, even on their best case, the claimants accept that the net result will be that the costs of the Trade Mark Infringement Claim should be their costs, while the costs of all other causes of action should be the defendant’s costs. They also accept that this means that a net payment of costs will ultimately be made by them to the defendant.[10]The one wrinkle to which I have referred relates to the incidence of general costs in respect of which it is accepted by the claimants that, if the court were to find the defendant to be the overall winner, they would not be entitled to recover those costs. I shall return to the submissions on general costs and what the defendant has referred to as “marginal costs” in a moment.[11]The defendant contends that it is the overall winner and should be identified as such, that there should be no reduction in its costs to reflect the fact that it lost on some issues, and that having regard to the various aspects of the Trade Mark Infringement Claim on which the claimants lost, the proposition that the claimants won on that claim as a whole and that the defendant lost is unarguable. Even if the court were to decide to treat the Trade Mark Infringement Claim as a standalone claim for the purposes of costs, the defendant contends that it was still the overall winner of that claim. It accepts that the Trade Mark Infringement Claim could form the basis for an issue-based deduction if the claimants had won it in its entirety, but it says that is not the case. It rejects the suggestion that it should be liable to pay any of the claimants’ costs in respect of that claim. Overall, therefore, the defendant seeks payments of its costs in full.[12]Notwithstanding their very different perspectives on costs, each side has provided evidence as to the amount of time spent on the Trade Mark Infringement Claim as a percentage of the litigation as a whole. Their analysis has produced very different results. Neither side has attempted to apportion the costs of the Trade Mark Infringement Claim to reflect issues within that claim on which the claimants won and issues on which the claimants lost. Essentially, both sides appear to acknowledge that the issues in that claim overlap to such an extent that any such exercise would be extremely difficult, if not impossible.

The relevant legal principles

[13]There is little between the parties on the legal principles to be applied. The dispute is as to the application of those principles to the very unusual facts of this case.[14]The court has a wide discretion on costs which is regulated by CPR Part 44.2. The general rule is that the unsuccessful party will be ordered to pay the costs of the successful party, but the court may make a different order. The successful party is the party who is successful in the litigation and not a party who is successful on a particular issue, see Kastor Navigation Co Ltd v Axa Global Risks (UK) Ltd [2004] EWCA Civ 277 at [143].[15]In deciding what order to make about costs, the court will have regard to all the circumstances, including whether a party has succeeded on part of its case, even if that party has not been wholly successful (CPR 44.2(4)(b)). The various orders available to the court are identified in CPR 44.2(6) and include the making of an order that a party must pay a proportion of another party's costs (CPR 44.2(6)(a)) and that a party must pay costs relating only to a distinct part of the proceedings (CPR 44.2(6)(f)). The overarching aim is always to make an order which reflects the overall justice of the case.[16]The parties each referred me to various authorities. The principles to be extracted are largely common ground. I start by observing that it appears to be standard practice in intellectual property cases to consider an issue-based approach to costs orders following on from the starting position identified in CPR 44.2. This is clear from Lifestyle Equities CV v Royal County of Berkshire Polo Club Ltd [2023] EWHC 2923 (Ch) [2024] Costs LR 449 (“Lifestyle Equities”) per Mellor J at [35]. This approach appears to have originated in patent cases, but it has since been adopted as a standard approach in other IP cases. Accordingly, both parties invited me to adopt this approach, although the defendant acknowledged that it should not represent a straitjacket for the court. I agree with that submission. The fact that a particular approach to the exercise of the discretion on costs may very often be convenient and appropriate in a particular type of case does not mean that it is the only proper approach. It is, as Mellor J observed in the patent case DSM IP Assets B.V. v Algal Omega 3 Ltd [2025] EWHC 1514 (Pat) (“DSM”) at [32], a convenient structure by which to analyse costs arguments, but it is not the only approach. Everything must depend, as CPR 44.2 makes clear, on the circumstances of the particular case and the need to achieve justice between the parties.[17]The parties agree that the issue-based approach involves asking the three questions identified by Arnold J (as he then was) in Hospira v Novartis [2013] EWHC 886 (Pat) at [2]: “... first, who has won; secondly, has the winning party lost on an issue which is suitably circumscribed so as to deprive that party of the costs of that issue; and thirdly, are the circumstances (as it is sometimes put) suitably exceptional to justify the making of a costs order on that issue against the party that has won overall”.[18]As to the first question, the identification of the winning party, the claimants referred me to the well-known test articulated by Sir Thomas Bingham MR in Roache v News Group Newspapers Limited [1998] EMLR 161, at page 168:
“The judge must look closely at the facts of the particular case before him and ask: who, as a matter of substance and reality, has won? Has the plaintiff won anything of value which he could not have won without fighting the action through to a finish? Has the defendant substantially denied the plaintiff the prize which the plaintiff fought the action to win?”
[19]The claimants also referred me to the decision of the Court of Appeal in AL Barnes Ltd v Time Talk (UK) Ltd [2003] EWCA Civ 402 (“AL Barnes”) at [28]:
“In what may generally be called commercial litigation ... the disputes are ultimately about money. In deciding who is the successful party the most important thing is to identify the party who is to pay money to the other. That is the surest indication of success and failure.”
[20]The claimants pointed out that AL Barnes is cited in the intellectual property context in the case of Illuminav Premaitha [2018] EWHC 180 (Pat) where Henry Carr J took account of the fact that an inquiry as to damages had been awarded and, indeed, was the only relief realistically available to the patentee (see [8] and [13] of Illumina).[21]As to the second and third questions identified by Arnold J (as he then was), Mellor J has summarised the relevant principles recently in two judgments to which I have been referred. The first, Lifestyle Equities, at [33]-[41], and the second DSM, at [10]-[21].[22]In the most recent of these cases, DSM, Mellor J identified the approach to the second question (at [15]) as follows:
“As to the second question and what is meant by 'an issue' in this context, in Unwired Planet v Huawei [2016] EWHC 410 (Pat) at [5] Birss J (as he then was) made some comments in the context of patent liability trials but which nonetheless illustrated that an 'issue' may be considered at any suitable level of generality / granularity: ‘One issue is: What is a suitably circumscribed issue? Or in other words, at what level of generality or granularity is that matter to be decided? Plainly it will vary from case to case. Often in patent cases one kind of suitably circumscribed issue and appropriate level of granularity is taking things at the level of individual cited items of prior art, but that is not a hard and fast rule. It is possible for a suitably circumscribed issue to arise within a broader category. An example of this was the Court of Appeal in ConvaTec Technologies Inc v Smith & Nephew plc [2015] EWCA 803 (Civ). Here, instead of dealing with the costs at the level of the issue of infringement as a whole, the court made a special order relating to experiments which formed part of theinfringement case.’”
[23]I pause there to note that it is therefore clear that an issue for these purposes may be considered at any level of granularity as long as it is suitably circumscribed. Any significant degree of overlap with other issues would appear to take an issue out of this category.[24]At [16]-[17], Mellor J went on to say this:
“16. In relation to the third question, there is no specific requirement of exceptionality. The question is whether it is appropriate in all the circumstances of the individual case not merely to deprive the winning party of its costs on an issue in relation to which it has lost, but also to require it to pay the other side's costs (see Arnold J. In Hospira v Novartis at [4]). In Unwired Planet Birss J. explained: ‘The second issue is the role of reasonableness and culpability. Clearly if the winning party has behaved unreasonably or is culpable in relation to an issue on which it lost, then that is a factor which can be relevant to the question of whether to deprive the winner of its costs or award those costs to their opponent. But the law is clear that even if the overall winner was reasonable to pursue an issue on which it lost, it does not follow that one should not deprive the winner of those costs nor does it follow that one should not award those costs to the opponent, merely because the overall winner behaved reasonably in relation to that issue.’ 17. After referring inter alia to Arnold J. in Novartis at [4], Birss J. concluded on this point at [8]: ‘Therefore, in my judgment, using the expression “suitably exceptional” is not really an appropriate way of describing the test today. The test today is whether it is appropriate, in all the circumstances in an individual case, to make whichever order one is thinking about. The test is appropriateness, not exceptionality.’”
[25]At [19], Mellor J referred to the oft-cited summary of Stephen Jourdan QC on the approach to be taken to issue-based costs in Pigot v Environment Agency [2020] Costs LR 825 at [6], as follows: “(1) The mere fact that the successful party was not successful on every issue does not, of itself, justify an issue-based cost order. In any litigation, there are likely to be issues which involve reviewing the same, or overlapping, sets of facts, and where it is therefore difficult to disentangle the costs of one issue from another. The mere fact that the successful party has lost on one or more issues does not by itself normally make it appropriate to deprive them of their costs. (2) Such an order may be appropriate if there is a discrete or distinct issue, the raising of which caused additional costs to be incurred. Such an order may also be appropriate if the overall costs were materially increased by the unreasonable raising of one or more issues on which the successful party failed. (3) Where there is a discrete issue which caused additional costs to be incurred, if the issue was raised reasonably, the successful party is likely to be deprived of its costs of the issue. If the issue was raised unreasonably, the successful party is likely also to be ordered to pay the costs of the issue incurred by the unsuccessful party. An issue may be treated as having been raised unreasonably if it is hopeless and ought never to have been pursued. (4) Where an issue based costs order is appropriate, the court should attempt to reflect it by ordering payment of a proportion of the receiving party's costs if that is practicable. (5) An issue based costs order should reflect the extent to which the costs were increased by the raising of the issue; costs which would have been incurred even if the issue had not been raised should be paid by the unsuccessful party. (6) Before making an issue-based costs order, it is important to stand back and ask whether, applying the principles set out in CPR rule 44.2, it is in all the circumstances of the case the right result. The aim must always be to make an order that reflects the overall justice of the case”.[26]While an issue-based approach will often produce a fair answer, it may necessarily be somewhat crude. The process of estimation may work both ways, sometimes favouring one and then the other party. It is for this reason that it is ultimately necessary to stand back and consider whether any particular analysis has produced the right result and reflects the overall justice of the case (DSM at [20]).[27]The costs of proceedings as a whole include the general costs, i.e. those unattributed to any specific issue. These may be substantial. This is why it is so important for the court, if it can, to identify the overall winner of the proceedings, because the overall winner “is likely, save in the most exceptional circumstances, to be entitled to payment of all his costs which are not or cannot be allocated to a particular issue” (Monsanto v Cargill [2008] FSR 16 (“Monsanto”), per Pumfrey LJ at [7]).[28]Mr. Cuddigan KC, on behalf of the defendant, emphasises the words “are not or cannot be allocated” in this passage from Monsanto in support of the proposition that Pumfrey LJ was referring both to general costs, i.e. costs which are not allocated to a particular issue, and also to what he calls marginal costs, i.e. costs which cannot be allocated to a particular issue because they relate to matters which overlap. I did not understand this reading of Monsanto to be disputed by the claimants.[29]Finally, I should observe that it is accepted by both parties that there may be cases in which there is no overall winner. The claimants drew my attention to a passage to this effect in Omnipharm v Merial [2012] EWHC 172 (Pat) per Floyd J (as he then was) at [19] (approved by the Court of Appeal in that case ([2013] EWCA Civ 2 at [132]), to the effect that:
“... there are some cases ... where the court has come to the conclusion somewhere on the spectrum of possible results between the two extremes contended for by the parties where the court has recognised that there is no overall winner.”
[30]One such case was the case of Vringo Infrastructure, Inc. v ZTE(UK) Ltd. (“Vringo”) [2014] EWHC 4475 (Pat), in which there was effectively a score draw in respect of three patent claims such that Birss J (as he then was) decided, at [11], that:
“The fair approach, although it involves a degree of imprecision, is to apportion the general costs relative to the three patents in suit.”
[31]While such an approach was justified on the particular facts of that case, I do not understand it to be appropriate where the court can identify an overall winner. In such circumstances, and consistent with the observations of Pumfrey LJ in Monsanto, the overall winner will recover all of its costs which are not or cannot be allocated to a particular issue.

Decision

[32]I start by asking who has won the litigation? Contrary to the submissions of Ms Lane KC on behalf of the claimants, I do not consider it to be the claimants. I say that for the following main reasons.[33]The claimants ran the three very significant claims to which I have already referred at trial, only to abandon them in their entirety prior to closing submissions, leading now to their dismissal. A very substantial amount of the evidence and opening submissions was directed to those claims, together with a significant amount of the case management. Of the remaining claims, the claimants won the Trade Mark Infringement Claim in part. They lost the Secondary Infringement of Copyright Claim.[34]Accordingly, looked at somewhat crudely and setting the claim in passing off to one side, the claimants won in respect of one claim but lost on four claims. In respect of each of the four claims on which they lost, the claimants accept without argument that they must pay the defendant’s costs.[35]Against that background, it is very difficult to see how the claimants can sensibly be regarded as the overall winners. The defendant has substantially denied the claimant the success that it sought in these proceedings. Looking simply at the claim for relief in the particulars of claim, the defendant has denied the claimant: the injunctive relief it sought in relation to Copyright Works (a claim fought by reference to a sample but ultimately said to apply to millions of Copyright Works); the injunctive relief it sought in relation to the importation of Stable Diffusion into the United Kingdom and the possession of Stable Diffusion in the course of its business; the injunctive relief it sought in relation to the extract and/or re-utilisation of a substantial part of the contents of the claimants' database, together with the additional damages that the claimants sought by reason of their allegations of blatant and widespread IP infringement. The defendant has also denied the claimants the opportunity to seek the damages that would have been said to flow from findings of liability in these claims.[36]I reject the claimants’ arguments that, in reality, the relief they have now obtained is not really significantly different from the relief that would have been available to them had they succeeded in their other claims or indeed that it represents ‘the real prize’ in the litigation. That is certainly not consistent with my understanding of the way the case was advanced in the lead up to, and at the outset of, the trial.[37]I acknowledge, of course, that the claimants have established an entitlement to substantive relief in the Trade Mark Infringement Claim, that they have obtained an undertaking in relation to Models v1.x and v2.x, such that the defendant is no longer able to host, distribute or otherwise deal in or make those models available for use or download via any platform, and that the defendant has submitted to an order for an inquiry as to damages. This is undoubtedly a measure of success, and it is an important one. I bear in mind that a key indication of success will always be the transfer of money from one party to another or the potential for that transfer. But, this is not the only measure of success and, in my judgment, it would be erroneous in all the circumstances of this case, including the nature and scope of the claims on which the claimants lost, to accord it primacy. I do not agree with Ms Lane’s submission that it is obviously ‘the touchstone’ in assessing the winner on the facts of this case.[38]Given what I have already said, it is not necessary to address in any detail the arguments made by the parties and the evidence served as to the potential value of the Trade Mark Infringement Claim at the inquiry into damages. The defendant sought to persuade me that I should not close my eyes to what it described as an absence of any articulated case on the part of the claimants as to the quantification of their alleged damages, suggesting this was relevant to my overall assessment of the winning party. The claimants rejected this characterisation of their position, pointing to their own evidence.[39]However, I remind myself that the amount of any financial relief to which the claimants may be entitled was not an issue at trial and will be a matter for determination in due course. I must assess the overall winner at the liability trial based upon the issues that were raised in the course of the litigation leading to that trial. The figure that may be recovered at the inquiry is neither here nor there and I express no view on it.[40]It follows from what I have already said that this is not a case in which it is impossible to identify an overall winner. This is not, in my judgment, a score draw of the type identified in Vringo and it is therefore not a case in which the costs which are not, or cannot, be allocated to particular issues should be apportioned by reference to the approach taken in that case.[41]As the overall winner, the defendant must have its costs of all claims that failed or were abandoned (as in fact the claimants already recognise) together with its costs of all matters which are not or cannot be specifically attributed to a specific issue. As overall winner, it is plainly entitled to these costs.[42]The next question I must consider, applying the three-question approach, is whether there is a suitably circumscribed issue in respect of which the defendant lost, such that it should be deprived of the costs of that issue or (even) so as to require a payment of costs going in the other direction to the claimants. The claimants say that the Trade Mark Infringement Claim is just such an issue. I agree. The Trade Mark Infringement Claim involved standalone causes of action and, although the claimants did not succeed on every aspect of that claim, they did succeed in respect of Models v1.x and 2.x. They succeeded, as I have already said, in obtaining undertakings to remove those models from the market and they succeeded in obtaining an order for an inquiry into damages.[43]Had the Trade Mark Infringement Claim been a standalone claim, as it very well could have been, I consider that the claimants' success in obtaining this relief would have rendered it the overall winner of that claim, consistent with the principles to which I have already referred, even though it failed on a number of (the defendant would say many) important issues. In that scenario, and given the nature of the issues on which the claimants failed, I consider that the court would have been looking to award costs in favour of the claimants, but after the application of an appropriate percentage deduction designed to reflect the fact that the claimants also lost on a number of significant issues.[44]Such an approach is of course permitted by CPR 44.2(6)(a) and is designed to reflect the fact, albeit in an inevitably broadbrush manner, that the winning party has not prevailed on every issue. Indeed, although the parties did not refer my attention to any of the authorities dealing with this approach, it is well-recognised that in many cases the court can and should seek to reflect the relative successes of the parties at trial on different issues by making a proportionate costs order (see the White Book, Volume 1, at 44.2.8).[45]Of course, the Trade Mark Infringement Claim is not a standalone claim. It is part of much wider litigation in respect of which I have found that the defendant is the overall winner. However, to my mind, there is certainly no question that it is suitably circumscribed so as to enable me, if appropriate, to deprive the defendant of its costs of that claim, or to order the defendant to pay some or all of the costs of that claim to the claimants.[46]What then of the answer to the third question? In considering the approach to take to the Trade Mark Infringement Claim, I consider the claimants’ status as the overall winner of that claim (viewed as a standalone claim, as I have just discussed) to be important. Once that is recognised, it appears to me that the appropriate approach is not merely to deduct some costs from the defendant to account for the claimants’ success in that claim, but that the fair approach must involve a payment by the defendant to the claimants of costs in relation to that claim. The mere deduction of costs from the defendant's overall recovery would not, in my judgment, do proper justice to, or reflect, the claimants' commercial success in the Trade Mark Infringement Claim.[47]The fact that I have already found that the defendant is the overall winner of the litigation means that the general costs of the Trade Mark Infringement Claim, i.e. the costs that are not or cannot be allocated to that claim, must of course be paid to the defendant. However, my starting point is therefore that the claimants have an entitlement to the remainder.[48]Applying the three-question test, the claimants submit they should get the entirety of that remainder. They say the three-question test does not permit deductions to be made to reflect the fact that they failed on some of the issues because it is concerned only with circumscribed issues and neither side has identified any circumscribed issues at a granular level in the Trade Mark Infringement Claim. Essentially, the claimants say the issues in the Trade Mark Infringement Claim are all overlapping so the appropriate order is for them to be awarded all of their costs.[49]While it might be tempting to take this approach on the grounds that it has the advantage of simplicity, on analysis, I do not consider that it would be correct to do so. Standing back, the trouble with it, and thus the danger of a straightforward application of the three-question test in the very particular circumstances of this case, is that in my judgment it does not do real justice between the parties. Specifically, an order for all of the claimants' costs on the Trade Mark Infringement Claim would fail to acknowledge the extent of the issues on which the claimants failed in that claim. I accept, of course, that there are different ways of reflecting the extent of the claimants' failure and that the parties have each illustrated (in the so-called ‘green and red’ tables produced for the court) the scope for potential subjectivity in identifying failures.[50]However, what is indisputable is that the claimants lost in relation to trade mark infringement in respect of SD XL, XL Turbo and v1.6, all more modern versions than v1.x and v2.x. They also lost on infringement of the GETTY IMAGES mark in respect of model v1.x and they lost on infringement of the iSTOCK mark in respect of model v2.x. Their main economic case at trial, that average consumers used verbatim or reworded Getty Images captions as prompts as a means of obtaining a Getty Images’ image without paying for it, also failed. So did their entire case under section 10(3) of the Trade Marks Act 1994, notwithstanding the significance that was accorded to that aspect of the case at trial. Ultimately, the findings that I was able to make in the claimants’ favour are recorded in the judgment at [757], as “historic and extremely limited in scope.”[51]Owing to the way in which the parties have approached their costs arguments, however, I have no evidence whatever to assist me on how these failures might be reflected in a costs order. I asked the parties to address me on the point, but the claimants continued to view the issue of costs through the prism of an issue-based approach, maintaining that the making of a proportionate costs order in relation to the Trade Mark Infringement Claim was not an approach the court should take absent evidence; while the defendant did not address the level of any deduction that might be appropriate in respect of an award of costs in the claimants' favour as a means of reflecting the issues in the Trade Mark Infringement Claim on which the claimants had failed.[52]Realistically, however, Mr. Cuddigan acknowledged that rigidly following the three questions approach has the potential to introduce unfairness into the analysis, what he described as “a slightly perilous flop one way, or flip the other”, which may not be consistent with the overall outcome of the dispute. Ms Lane did ultimately accept that it may be appropriate to deduct a very small percentage from the claimants’ costs, but she did not accept that it would be of any significance.[53]In all the circumstances, it is inevitable that trying to reflect the justice of the outcome of the Trade Mark Infringement Claim will inevitably be a rough-and-ready exercise. But I do not accept that the absence of evidence from the parties addressing the point means that I should not undertake the exercise. I have sat through the trial, I have seen the evidence on which the parties relied and I have a real sense of what is proportionate and fair in those circumstances.[54]Accordingly, doing the best I can, bearing in mind the issues at trial in the Trade Mark Infringement Claim, the way in which they were developed, the available evidence and the extent of the overlaps, I consider that a 25% reduction should be made to the claimants' costs of that claim, to reflect the fact that they failed on a number of significant issues.[55]That then brings me to the question of what percentage of the overall costs of the litigation can be attributed to the Trade Mark Infringement Claim and thus to what figure that 25% reduction must be applied.[56]As I have said, the parties have each conducted exercises designed to assess the percentage of the costs spent in the proceedings on the Trade Mark Infringement Claim alone. Broadly, the claimants have conducted a page-turning exercise in respect of each phase of the proceedings identifying an appropriate percentage for each phase, while the defendant has counted paragraphs in documents, identified a percentage for each phase of the proceedings and then calculated the mean. Each criticises the other's approach for various reasons and there is plainly something in those criticisms. Ultimately, however, the analysis is bound to be rough-and-ready, as both sides acknowledge.[57]The claimants originally calculated that the total costs in the Trade Mark Infringement Claim represented 26.3% of the overall costs of the litigation. The defendant calculated that the costs of the Trade Mark Infringement Claim represented 10.9% of the overall costs.[58]During the course of the hearing, there was some movement in those positions. The claimants' percentage takes general costs into account and (as was accepted) these must be stripped out in circumstances where the defendant is the overall winner. If this is done, I understand the claimants’ revised percentage to be 20.3%. The defendant accepted in submissions that taking a mean percentage in respect of each phase of the litigation may not have been the fairest approach and suggested an alternative approach of taking a weighted mean resulting in an increase to about 15%, albeit Mr. Cuddigan points out that this includes what he referred to as the “overlap issues”.[59]Doing the best I can, and bearing in mind that I cannot sensibly test where the merits of these respective positions lie, but where it is clear that points can be made on each side, I am going to take 17.5% as representing the costs of the Trade Mark Infringement Claim for the purposes of my assessment, which falls roughly between the two percentages identified by the parties. I shall leave the parties to calculate the resulting figure.[60]The application of the further 25% to which I have already referred to that figure will go some way towards addressing the overlap issues and ensuring that, in so far as is possible, given the circumstances, my order reflects the fact that the claimants did not succeed across the board in the Trade Mark Infringement Claim. I accept that this approach is inevitably somewhat imprecise, but it is impossible to arrive at an entirely accurate result in a situation such as this. All I can do is to try to reflect the overall justice of the case.[61]The result of the analysis that I have conducted is that there will be a deduction of 17.5% from the defendant's overall costs, to reflect the costs that it cannot recover in respect of the Trade Mark Infringement Claim, together with a further deduction of approximately 13% to reflect the award to the claimants of 75% of their costs of that claim. This means that there will be a 30.6% reduction in the defendant’s overall costs.[62]Standing back, and always appreciating that this approach is, of necessity, as I have made clear, somewhat crude, I must ask myself whether this analysis has produced the right result and whether it reflects the overall justice of the case. In my judgment, having regard to all the features that I have identified already, it does and I shall make an order accordingly.[63]I turn then to consider some remaining issues on costs. It is agreed between the parties, consistent with authority, that the defendant is entitled to an interim payment on account of costs, pursuant to CPR 44.2(8). The parties are also agreed that incurred costs merit a 70% interim payment, while budgeted costs merit a 90% interim payment. There remains however a dispute over:(1) some costs claimed by the defendant which fall outside the budget; and(2) costs incurred by the defendant by way of overspend above its budget.[64]As to costs falling outside the budget, I understand the dispute to have narrowed to the question of whether the defendant’s costs of out of budget interim applications should be subject to an interim payment at 70%, as the defendant contends, or at 50% as the claimants contend.[65]Given that the claimants agree to the principle of 70% for incurred costs, I cannot see the justification for suggesting a lower figure for the costs of these interim applications. Accordingly, the defendant shall have 70% of its costs for these applications by way of interim payment.[66]As to costs incurred by way of overspend and the approach to be taken, I was referred to the summary of Mellor J in the Lifestyle Equities at [57]. I was also reminded that CPR 3.15A (1)-(2) provides that “A party ... must revise its budgeted costs upwards or downwards if significant developments in the litigation warrant such revisions” and that a revised costs budget must be submitted promptly, but that it is not open to a party to seek to amend its costs budget to provide for costs that have already been incurred.[67]CPR 3.18(b) provides that a budget may be departed from if there is good reason to do so. In Harrison v University Hospitals Coventry & Warwickshire NHS Trust [2017] EWCA Civ 792, at [44], the Court of Appeal held that this places “a significant fetter on the court having an unrestricted discretion: it is deliberately designed to be so”. The question of what constitutes a good reason is a matter for the costs judge based on the circumstances of the case and bearing in mind the Denton principles of relief from sanctions.[68]In the case of Walton International Ltd v Verweij Fashion Ltd [2018] EWHC 2315 (Ch), Arnold J (as he then was) was invited to depart from the costs budget, but refused to do so, identifying that his role was simply to take a view as to:(1) whether the defendant was likely to establish that there was a good reason to depart from the costs budget; and(2) what would be a reasonable sum for interim payment. He noted that he had to be cautious as to the sum that would ultimately be recovered and he awarded 50% of the total unbudgeted costs.[69]Finally, I was taken by both parties to Montres Breguet SA v SamsungElectronics Limited [2022] EWHC 1895 (Ch) at [39]-[52] per Falk J (as she then was). My attention was drawn by the claimants to the fact that Falk J awarded only 20% of the unbudgeted costs as an interim payment in that case to reflect the fact that the increase in the budget should have been sought and that she had limited evidence on which to determine that a higher percentage could be justified.[70]The defendant seeks an interim payment in respect of five areas of overspend: statements of case; case management conferences; disclosure; witness statements; and trial preparation/trial, the total sum of the overspend being £394,985.31. This is dealt with in two statements from Mr. Aries. Mr. Rose responds to the defendant’s case on this in his evidence, to which Ms Lane took me during her submissions. Although he makes specific points of detail in relation to each area, his overarching point is that in many respects the defendant should have anticipated the need to make provision for various matters in its costs budget or that there is no good reason advanced by the defendant to depart from its budget.[71]Doing the best I can, and having regard to the available evidence, while it would appear to be the case that, as Mr. Rose suggests, the defendant could have applied to amend its costs budget in relation to various of the matters it now raises, I nevertheless accept that the nature of this case, the unprecedented levels of case management, the number of issues, the frequent amendments to the claimants' pleadings, and the somewhat moving targets make it likely that the defendant will establish a good reason to depart from its costs budget in various respects. These factors may well support Mr. Aries’ evidence that there was, in reality, no time for preparation of costs budgets or for there to be a hearing of any dispute in relation to them, although I cannot, of course, make any final decision about that.[72]I also accept Mr. Cronan’s submission on behalf of the defendant that not everything in litigation of this type will necessarily be regarded as a significant development at the time, such that an application to amend the costs budget is required. Costs of this nature may nevertheless add up (particularly in litigation of this scale) and there may be good reason why they have been incurred. Of course, I am obviously not in a position to determine whether this applies to the costs overspend sought by the defendant and nor should I.[73]On the available evidence, specifically that in Mr. Aries’ seventh statement, I consider that good reason is likely to be established on the grounds of the defendant having to undertake material work that was not expected and which resulted in material additional costs. On balance, I think it is more likely to be established in relation to the pre-trial phases, particularly the statements of case, case management conferences and some witness statements, than in relation to the disclosure phase or the trial phase, although I obviously make no final determination on the point for present purposes; that will be a matter for the costs judge.[74]Overall, given that I cannot determine definitively whether there were good grounds not to apply to vary the costs budget and given that the evidence suggests a better case in relation to some of the phases identified by the defendant than for others, I consider that I should take a similarly cautious approach to that taken by Falk J in Montres Breguet, albeit for slightly different reasons. I shall award 20% of this overspend by way of interim costs. (For continuation of proceedings: please see separate transcript) - - - - - - - - - - -