Epic Games, Inc v Google LLC

The Nature of the Respondents' Application 2 The Applicants are Epic Games, Inc ('Epic Games') and Epic Games International S.à.r.l ('Epic International') (collectively, 'Epic'). The Respondents are Google LLC, Google Asia Pacific Pte Ltd ('Google Asia') and Google Payment Australia Pty Ltd ('GPAL') (collectively, 'Google'). Epic sues Google over the exclusion of its popular game, Fortnite, from the Android platform, and Google's earlier - and continuing - refusal to permit Epic to offer players of Fortnite the ability to make purchases within the game other than via Google Play. Fortnite was excluded from the Android platform on 13 August 2020 when, on the same day, Epic made its own in-game payment system available to players of the game. Epic alleges that Google's conduct was in breach of Pt IV of the Competition and Consumer Act 2010 (Cth) ('CCA') and that it was unconscionable within the meaning of s 21 of the Australian Consumer Law ('ACL', being Sch 2 to the CCA). The present proceeding was commenced on 8 March 2021. There are parallel proceedings in the United States before the District Court for the Northern District of California (Oakland Division) concerning the same events. 3 By an interlocutory application filed on 9 August 2021, Google seeks orders dismissing Epic's proceeding or its permanent stay. Various alternative orders are also sought. At the heart of the application is an exclusive jurisdiction clause in a contract between, on the one hand, Epic Games and, on the other, Google LLC and Google Asia, in which it was agreed that any dispute between them would be governed by the law applying in California and that any differences between them would be resolved by the courts in Santa Clara County in California. In effect, Google seeks to enforce this provision. 4 A similar case is also pending in this Court: Epic Games, Inc & Anor v Apple Inc & Anor (NSD 1236 of 2020). In that case, Epic also seeks redress for the removal of Fortnite from the Apple App Store. There is a similar, although materially different, exclusive jurisdiction clause involved in that case. Apple applied to stay Epic's proceeding against it on a similar basis to that which Google now advances. At first instance I granted that stay on a temporary basis: Epic Games, Inc v Apple Inc (Stay Application) [2021] FCA 338; 151 ACSR 444 ('the Apple Decision'). That decision was reversed by the Full Court which refused the stay application: Epic Games, Inc v Apple Inc [2021] FCAFC 122; 392 ALR 66 ('the Full Court Apple Decision'). An application for special leave to appeal was refused by the High Court on 2 December 2021: [2021] HCASL 234.

The Evidence 5 The evidence consisted of affidavit evidence and some exhibits. Affidavits from the following witnesses were read on the application: Mr Simon Walter Johnson: Mr Johnson is a partner of Corrs Chambers Westgarth, the solicitors for Google in this proceeding. He explained the operations of each of the Google and Epic entities, the various agreements between Epic and Google and the circumstances surrounding the removal of Fortnite from the Google Play Store. He also gave evidence about the related proceeding which is presently pending before the United States District Court in the Northern District of California (Oakland Division). Mr Johnson prepared an affidavit dated 27 August 2021 and was not cross-examined. Judge Jeremy Fogel: Judge Fogel is a retired judge of the United States District Court of California on which he served between 1998 and 2011 in its San Jose Division. As a judge of that court he heard many cases involving technology as well as cases involving anti-trust claims. From 1981 to 1998 Judge Fogel was a judge of the Santa Clara County Municipal and Superior Courts. The docket of those courts was broad and included cases involving technology and anti-trust. Judge Fogel gave expert evidence about the operation under Californian law (including federal law) of the aforementioned agreement between Google Asia, Google LLC and Epic Games. He also gave evidence about how a United States Court would handle Epic's Australian suit in the event that the present proceeding was stayed and Epic commenced an equivalent proceeding in the United States. Judge Fogel prepared an expert report dated 27 August 2021. He also prepared a further report dated 30 September 2021. This was given in reply to a report prepared by Epic's expert, Judge Stephen Larson (see below). Judge Fogel was not required for cross-examination. Mr David Poddar: Mr Poddar is a partner of Clifford Chance, the solicitors for Epic. Like Mr Johnson, Mr Poddar gave some background to the proceedings. This included evidence about the Epic entities, the game Fortnite and the role of GPAL together with some evidence about the various contentions which have passed between the parties during the pendency of the litigation. Mr Poddar prepared two affidavits dated 24 September 2021 and 8 October 2021 and was not required for cross-examination. Judge Katherine Bolan Forrest: Judge Forrest's role is as the attorney for Epic in the United States proceedings. Judge Forrest is a retired judge of the United States District Court for the Southern District of New York. It is the practice in the United States to use the honorific 'Judge' for retired judges. I will use that honorific even though Judge Forrest's role is that of an attorney. She gave evidence about the current status of the proceedings against Google in the United States including its consolidation with several other suits concerned with the competitive qualities of the Android platform. She also gave evidence as to what the likely outcome might be if Epic sought to pursue the present proceeding in the United States as well as whether the pursuit of anti-trust claims based on foreign law was common in the United States. Judge Forrest prepared an affidavit dated 24 September 2021 and was not required for cross-examination. Judge Stephen Larson: Judge Larson is a former judge of the United States District Court for the Central District of California on which he served between 2000 and 2009. He gave evidence on the same issues as Judge Fogel. He prepared a report dated 1 October 2001 and was not required for cross-examination. 6 The parties raised many objections to each other's lay evidence. With one exception these were worked out consensually between them. I was provided with a record of the agreed rulings and these were applied to the Court Book containing the affidavits on each of which the agreed rulings have been annotated. The one exception to this concerned s21 of Mr Johnson's affidavit. By agreement the first sentence of this paragraph was not read. Epic objected to the remaining two sentences on the basis that they were hearsay or, if admissible, should nevertheless be excluded under s 135 of the Evidence Act 1995 (Cth) ('Evidence Act') since they were unfairly prejudicial. I accept that the two sentences are hearsay but on an interlocutory application hearsay is admissible if the source of the hearsay is identified: s 75 Evidence Act. In the case of s21, the requirements of s 75 were met and the objection fails. I reject Epic's parallel attempt to exclude this material under s 135 of the Evidence Act. The exclusion of evidence under that provision will generally only be appropriate where there is a risk that the tribunal of fact may handle the evidence irrationally or where the admission of the evidence would give rise to procedural prejudice: Australian Competition and Consumer Commissioner v Ramsay Health Care Australia Pty Ltd [2020] FCA 308; 380 ALR 300 at [282]-[283] per Griffiths J. Insofar as irrational behaviour by the tribunal of fact is concerned, the provision has little application where the tribunal of fact is a judge sitting alone and, in that regard, it is generally accepted that the provision is aimed principally at jury trials. Since Epic did not identify any procedural prejudice which might be relevant in a judge-alone hearing, s 135 has no application. The two sentences were therefore admitted into evidence. 7 Insofar as the expert evidence was concerned, a number of objections were taken by Epic to the admissibility of Judge Fogel's evidence. Most of these were resolved consensually and marked up in the Court Book. Of those which remained unresolved, all were on the basis that Judge Fogel had expressed an 'inadmissible foreign opinion'. The first of these, which may serve as an exemplar, was to the last sentence of s81 of Judge Fogel's report in chief which was to the effect that he thought it very unlikely that a US Court would stay a proceeding in California on forum non conveniens grounds where the parties had selected California as the forum for their dispute. Whilst the content of foreign law is a question of fact usually established by expert opinion evidence, it is for this Court to apply the foreign law once it has been identified. As such, as I understand it, Epic's objection is that Judge Fogel may not give evidence as to the ultimate issue which this Court must determine. If this be the objection, then I reject it on the basis of s 80 of the Evidence Act which abolishes that rule. If the objection is to Judge Fogel's qualifications to express the opinion, then I reject it on the basis that he is plainly qualified to express the opinion. If some other objection is secreted within the words 'inadmissible foreign opinion' then I reject it on the basis that it has not been sufficiently explained to permit its adjudication. 8 Similar objections were taken to Judge Fogel's report in chief at ss117, 119, 125, 127 and 141 and to his report in reply at ss6, 14, 15, 16, 20 and 22. These objections fail for the same reason as does Google's identical objection to the last sentence of s95 of Judge Larson's report. 9 In addition to the affidavit evidence there were five exhibits tendered on the application. I will refer to these when their relevance arises. The hearing took place on Wednesday 13 October 2021.

Epic's Amendment Application 10 Towards the end of the oral argument, Epic applied to amend its Amended Originating Application and its Amended Concise Statement. This proposed amendment was sought in response to a submission that Google made about the state of the pleadings. I declined to deal with the application instanter and made directions for Epic to make that application formally which it did on 21 October 2021. This application was supported by a further affidavit of Mr Poddar dated 21 October 2021. I deal with this application at the end of these reasons. The parties were content for that application to be dealt with on the papers. For the reasons given below, the amendments will be allowed.

The Structure of These Reasons 11 As I have mentioned, Google pursues its stay application on two bases: (a) the exclusive jurisdiction clause in the agreement between Google Asia, Google LLC and Epic Games; and (b) forum non conveniens grounds. In response, Epic submits that the application for a stay on the basis of the exclusive jurisdiction clause should be refused because the validity of that clause has itself been directly challenged in this proceeding. There are two conceptual paths to this submission. The first assumes that the clause is enforceable but designates the challenge to its validity as something which should be taken into account in deciding whether to enforce it by granting a stay. The second is conceptually equivalent to a notional application for an interlocutory injunction by Epic seeking to restrain Google from enforcing the clause pending a trial of Epic's allegation that the clause is invalid. Although somewhat different in terms of analysis, the difference between these two paths may, in practice, turn out to be more theoretic than real. As will be seen, in these reasons I have treated the argument as involving the second path. For reasons which will become apparent, the difference in this case between the two paths does not matter. 12 The conclusions which I have come to are: (a) assuming that the challenge to the exclusive jurisdiction clause is not in itself sufficient to defeat the stay application, the stay application should in any event be refused; (b) Epic's challenge to the exclusive jurisdiction clause in fact defeats the stay application so that the stay application should be dismissed for that reason too; and (c) no stay should be granted on forum non conveniens grounds. 13 I will deal with these issues in the above order.

The Stay on the Basis of the Exclusive Jurisdiction Clause 14 In this section, I assume that Epic's challenge to the validity of the exclusive jurisdiction clause does not in itself defeat Google's application for a stay. 15 As I have explained above, Google's application for a stay rests upon an agreement between Epic Games and two of the Respondents, Google Asia and Google LLC. The agreement is entitled the 'Google Play Developer Distribution Agreement' ('the DDA'). Neither the Second Applicant, Epic International, nor the Third Respondent, GPAL, are parties to this agreement. The relevant part of the DDA is cl 16.8 which provides: All claims arising out of or relating to this Agreement or Your relationship with Google under this Agreement will be governed by the laws of the State of California, excluding California's conflict of laws provisions. You and Google further agree to submit to the exclusive jurisdiction of the federal or state courts located within the county of Santa Clara, California to resolve any legal matter arising from or relating to this Agreement or your relationship with Google under this Agreement, except that you agree that Google will be allowed to apply for injunctive relief in any jurisdiction… 16 For present purposes this clause has three significant components. The first identifies the claims to which the clause is to apply. These are 'all claims arising from or relating to' the DDA or Epic Games' relationship with 'Google' (which is defined in cl 1 to include Google LLC, Google Ireland Ltd, Google Commerce Limited and Google Asia). The second is a choice of law clause which records the parties' agreement that any such claim will be governed by the laws of the State of California but subtracting from those laws any of California's conflict of laws rules. The third is an exclusive jurisdiction clause by which the parties have agreed that any claims to which the clause applies will be submitted to the federal or state courts in Santa Clara County in California and to no other courts. The parties on the present application are in disagreement about the operation of each of these three aspects of cl 16.8 although they agreed that these disagreements were to be resolved by reference to the laws of California.

The Issues Which Arise in Relation to Cl 16.8 17 The debates between the parties were substantively as follows: The Proper Construction Issue: Google submits, and Epic denies, that the present dispute before this Court is a legal matter 'arising from or relating to this Agreement' within the meaning of cl 16.8. If it is not, then Google's application for a stay on the basis of cl 16.8 falls away since the clause has no application; The Choice of Law Clause Issue: Epic submits that its allegations in this proceeding involve claims under mandatory laws of the forum (i.e. under Australian law), that is to say, claims which in their nature are unable to be contractually bargained away. It says that the effect of the choice of law clause is that any court sitting in California will be obliged to apply Californian law excluding its conflict of laws rules to the dispute. The effect of the choice of law clause will be to exclude its ability to rely upon Pt IV of the CCA and s 21 of the ACL in any suit brought in Santa Clara County. It is not in dispute that as a matter of Australian law, the parties to the DDA are not at liberty to agree that Pt IV of the CCA and s 21 of the ACL cannot be invoked. Epic says that in such a situation the burden falls on Google to demonstrate that, despite the terms of cl 16.8, a court sitting in Santa Clara County would apply Pt IV of the CCA and s 21 of the ACL to the dispute if Epic is now forced to bring this case in that place. Based on the expert testimony of Judge Larson it submits that in light of cl 16.8 it is unclear whether such a court would in fact apply Australian law to the dispute. Consequently, Epic submits that Google has failed to show that Australian law would be applied and that as such, this Court should not give effect to cl 16.8 by granting a stay. To do so would, on Epic's submission, undermine the principle that Pt IV of the CCA and s 21 of the ACL cannot be contractually bargained away. In response, Google submits that it has proffered an undertaking to this Court not to rely upon this aspect of cl 16.8. Epic denies that this Court can accept such an undertaking and, even if it can, denies that it is sufficient to satisfy the Court that Google will not rely upon cl 16.8 to defeat any suit in California; The Fragmentation Issue: Here the issue is whether the refusal of the stay will result in fragmentation of the proceedings into multiple international jurisdictions and, if it will, the extent of that fragmentation (there being pending in California related proceedings about the removal of Fortnite from the Android platform); The Parties Issue: The parties to this litigation include entities which are not parties to the DDA, specifically, GPAL and Epic International. Epic submits that these parties are not bound by cl 16.8 since they are not parties to it and that this means that: (i) Epic International's suit against the Google parties (which is no different to Epic Games' suit) cannot be stayed under cl 16.8; (ii) Epic Games' suit against GPAL cannot be stayed; and (iii) in that circumstance, even if cl 16.8 permits a stay of Epic Games' proceeding against Google Asia and Google LLC, there are good discretionary reasons not to require parts of this suit to be determined by the courts of Santa Clara County whilst requiring other parts to be determined by this Court. For its part, Google submits that GPAL is a named beneficiary of cl 16.8 under the DDA and hence that it is entitled to rely upon cl 16.8. In relation to the suit brought by Epic International, it submits that Epic International is bound by cl 16.8 even though it is not a party to it under a principle of law applying in California known as the Manetti-Farrow rule. Alternatively, it submits that even if cl 16.8 does not bind Epic International nevertheless its proceeding should be stayed on discretionary grounds if Epic Games' proceeding is stayed; The Remedies Issue: Whether a court sitting in California would be able to grant similar relief to Epic to that which this Court could grant; The Platform Provisions Issue: The significance of ss 83 and 87(1A) of the CCA. These provisions entitle third parties to rely upon the findings of this Court in any subsequent proceedings which may be brought against Google. If the proceeding is heard in California, the effect of these provisions will be evaded for they do not apply to the findings of a foreign court. There is a related issue about the right of the Australian Competition and Consumer Commission to intervene in proceedings before this Court and whether this is adequately protected by a right to file a brief in US proceedings as an amicus curiae; The Remaining Issues: The parties joined issue on a number of smaller arguments not usefully collected under any other heading. 18 It is useful to deal with these in the order set out above.

The Proper Construction Issue 19 The question is whether the proceeding in this Court is a 'legal matter arising from or relating to this Agreement or your relationship with Google under this Agreement'. If it is not, the debate need proceed no further because the clause will have no application. 20 In my view, it is clear that the present proceeding is a legal matter of the kind referred to in cl 16.8. Epic Games is the developer of Fortnite. That game is available on a number of platforms including, until 13 August 2020, mobile devices installed with the Android operating system. The Android operating system is owned by Google LLC. One of the apps which is available on Android devices is the Google Play Store which is, in effect, a market place for other apps. Developers of apps wishing to distribute apps via the Google Play Store enter into a developer distribution agreement with one or more Google entities depending on the country in which the developer wishes to distribute the app. In the case of North America, such an agreement is with Google LLC and in the case of Australia, with Google Asia. It is for that reason that Epic Games' DDA is with both Google LLC and Google Asia. 21 In the present case, Epic Games entered into the DDA on 8 November 2016. As I explain a little later, the agreement has been amended since that time. The current version of the DDA upon which Google relies contains a term, cl 4.1, which requires Epic Games to adhere to Google's 'Developer Program Policies'. One of these policies requires app developers who wish to offer for sale in-app content to use only Google's payment system, Google Play Billing. Another clause in the DDA (cl 4.5) prevents developers from making available any product that has the purpose of facilitating the distribution of apps outside the Google Play Store. 22 Epic Games began to distribute Fortnite from the Google Play Store on 21 April 2020. On 13 August 2020 it made available within the Fortnite app a facility which permitted players to pay for in-app content by means of a payment system which was not Google Play Billing. On the same day, Fortnite was removed from the Google Play Store. 23 In this proceeding, Epic Games alleges that the way in which Google LLC conducts the Android operating system makes it practically impossible for an app developer to distribute its app to Android users other than via the Google Play Store. Hence it says that app developers are in practice forced to use the Google Play Billing facility. It alleges that the manner in which the Respondents conduct the Google Play Store is anti-competitive and contrary to several provisions of Pt IV of the CCA. It also alleges that the same conduct is unconscionable within the meaning of s 21 of the ACL. In this case, it seeks declarations of the contravening conduct, injunctions to restrain the contravening conduct, and an injunction requiring Google to restore Fortnite to the Google Play Store with its own in-app payment system intact. 24 The question then is whether that dispute is a 'legal matter arising from or relating to' the DDA or Epic Games' relationship with Google under it. The parties agree that the meaning of this expression is governed by the law of California. I accept the expert opinion of Judge Fogel that under the law applying in California, a dispute will 'relate to' an agreement containing a forum selection clause if it has a logical or causal connection with the agreement: Yei A. Sun v Advanced China Healthcare, Inc 901 F3d 1081 (9th Cir 2018) at 1086. I reject the Applicants' submission based on the expert opinion of Judge Larson that the decisions in In re Orange SA 818 F3d 956 (9th Cir 2016), Yan Guo v Kyani, Inc 311 FSupp3d 1130 (CD Cal 2018) and Intel Corporation v Tela Innovations, Inc (ND Cal, 18 September 2018) 2019 WL 5697922 establish that for a dispute to 'relate to' the agreement it must involve the interpretation of the contract. They do not so hold. In any event, I accept Judge Fogel's evidence that the allegations that the Applicants make in this proceeding under Pt IV will involve the interpretation of the DDA: Cung Le v Zuffa, LLC 108 FSupp3d 768 (ND Cal 2015) at 776. 25 The question then is whether there is a logical or causal connection between the allegations made by the Applicants in this case and the DDA. In my view, there is. It is cl 4.1 of the DDA which required Epic Games to comply with Google LLC's Developer Program Policies and it is those policies which required Epic Games to use only Google Play Billing for the purchase of in-app content. It is also those same policies which prevent it from distributing its app outside the Google Play Store. But for that clause, the Applicants' complaints would not have been made in this proceeding. Consequently, there is both a causal and a logical connection between the DDA and this litigation. Clause 16.8 is engaged and the parties have therefore agreed that this dispute should be heard by the federal or state courts sitting in Santa Clara County. 26 I reject Epic's alternate submission that even if the dispute was within the scope of the clause a Californian court would nevertheless decline to give effect to it. Judge Larson gave evidence that under Californian law no effect would be given to a clause such as cl 16.8 if it was 'the result of overreaching or the unfair use of unequal bargaining power'. Whilst I accept that this might be so in relation to other developers, I do not accept that any relevant imbalance of bargaining power has been demonstrated on this application as between Google and Epic.

The Choice of Law Issue 27 The allegations made by the Applicants are made under Pt IV of the CCA and s 21 of the ACL. Both are 'mandatory laws of the forum' in the sense that the parties are unable to contract out of their application either directly or indirectly by means of a choice of law clause. For present purposes, this is relevant because, unless this Court can be satisfied that the mandatory law of the forum will be applied by a foreign court in the event that a stay is granted, this will generally be a strong reason not to stay the local proceeding: Akai Pty Ltd v People's Insurance Co Ltd (1996) 188 CLR 418 ('Akai'); Full Court Apple Decision at [87]. 28 In this case, cl 16.8 contains a choice of law clause in which the parties have indicated that the disputes covered by the clause should 'be governed by the laws of the State of California, excluding California's conflict of laws provisions'. If Epic now commenced a suit in Santa Clara County, it is possible (although unclear) that under Californian conflict of laws principles a state or federal court sitting in that place might apply Australian law, including Pt IV of the CCA and s 21 of the ACL. However, the explicit exclusion by cl 16.8 of those conflict of laws principles means that, on the face of things, Pt IV of the CCA and s 21 of ACL could not be applied. If that be so, the suit before this Court cannot be brought in Santa Clara County because cl 16.8 will prevent the Australian laws invoked by Epic being applied to the dispute. 29 It becomes necessary, therefore, to be precise about the operation of the choice of law clause in any such proceeding. There was agreement on both sides that a state or federal court sitting in Santa Clara County would have jurisdiction to entertain Epic's suit so any concerns about want of jurisdiction may be put aside. The evidence about how the choice of law clause would be applied under the law applying in California was as follows: 30 First, in the first of his reports, Judge Fogel gave evidence on behalf of Google that a US court would apply the substantive provisions of the CCA and ACL which Epic claims have been violated although it would apply the lex fori to the question of remedies: First Report of Judge Fogel at s78. In reaching this conclusion, however, Judge Fogel did not address the fact that cl 16.8 excludes Californian conflict of laws principles. In that circumstance, I do not regard this aspect of Judge Fogel's evidence as compelling. I should say in Judge Fogel's defence that he was not asked by Google's solicitors to express a view on the choice of law clause. 31 Secondly, Judge Larson, who gave evidence on behalf of the Applicants, addressed the issue in more detail. He thought that Judge Fogel's failure to address the conflict of laws clause was a critical gap in his analysis: Report of Judge Larson at s81. Subject to my previous observation that Judge Fogel is not responsible for this critical gap, I agree with Judge Larson - the operation of the choice of law clause is central to understanding whether a court sitting in Santa Clara County would apply Australian law to the present dispute. 32 Judge Larson gave evidence of three different approaches which a court sitting in California might take to cl 16.8. The first was that it might give effect to the clause and not engage in any conflict of laws analysis: ss73-74. If this approach were adopted it would lead the court to apply cl 16.8 on its own terms. I would infer from this that the effect would be that the dispute would fall to be determined under the substantive law applying in California which would not include Pt IV of the CCA or s 21 of the ACL. 33 A second approach would involve the court in engaging in a conflict of laws analysis prior to enforcing the choice of law clause: s75. In doing so it would ask whether California had a substantial relationship with the parties or their transaction or whether there was any other reasonable basis for the parties' designated forum: Nedlloyd Lines B.V v The Superior Court of San Mateo County 3 Cal4th 459 (Cal 1992) ('Nedlloyd') at 464-465. In Nedlloyd the Court observed that 'if one of the parties resides in the chosen state, the parties have a reasonable basis for their choice' (at 467). In this case, Google LLC has its principal place of business at 1600 Amphitheatre Parkway, Mountain View in California. 34 As part of this second approach, however, Judge Larson also opined that the court would additionally ask whether there was a conflict between the law required by the choice of law clause to be applied and a fundamental policy of the local forum. In the present case, that situation does not arise because the law sought to be applied by cl 16.8 is the law of the local forum. Consequently, if this second approach were taken it would lead a court sitting in California to apply the substantive law applying in California to the dispute which would not include Pt IV of the CCA or s 21 of the ACL. 35 Judge Larson thought that a third approach could be taken which would involve asking whether Epic's claims under Pt IV of the CCA or s 21 of the ACL involved non-waivable rights. If any claim did then the court 'may shift the burden to the party trying to enforce the choice of law clause to demonstrate that applying California law would not diminish the non-waivable rights under Australian law': s79. If this burden was not met or there was a risk that applying California law may not protect the non-waivable rights under Australian law 'the court may decline to enforce the choice of law clause and apply Australian law even without a conflict of laws analysis': s79; see also s47. 36 Judge Larson did not express a view on which of these three approaches was the most likely to be adopted (although he did venture into that territory in parts of his report which were not read by agreement between the parties: see for example, s72). The first two approaches both lead to cl 16.8 being applied on its own terms and to the exclusion of Pt IV of the CCA and s 21 of the ACL. As to the third approach, making the assumption that a court sitting in California would conclude that Epic's claims involved non-waivable rights under Australian law, it would appear to follow that this would shift the burden to Google to prove that a court sitting in California would be satisfied that the application of Californian law would not diminish Epic's Australian law claims. 37 In his reply evidence, Judge Fogel did not take issue with Judge Larson's analysis because, by the time his reply evidence was prepared, Google's solicitors had instructed him to proceed on the basis that Google would undertake not to rely on the choice of law clause. The evidence of Judge Larson therefore stands uncontradicted on this point, at least by Judge Fogel. 38 There is, however, a wrinkle. The cases cited by Judge Larson at ss79-80 for the third approach are Hall v The Superior Court of Orange County 150 CalApp3d 411 (Cal 1983) ('Hall') and Verdugo v Alliantgroup, L.P 237 CalApp4th 141 (Cal 2015) ('Verdugo'). It is not self-evident to me that those decisions establish that a Californian court would decline to enforce a choice of law clause selecting Californian law as the applicable law where to do so would diminish the non-waivable rights of another jurisdiction. Both decisions appear to involve situations where non-waivable rights under Californian law would be lost if the plaintiff was forced to litigate in the jurisdiction selected by the choice of law clause. The principle in play in Hall and Verdugo appears similar to that established in this country by Akai. 39 In my view, it would be unwise to embark on an inquiry into the correctness of Judge Larson's views on the third approach. Although Mr Hutley did take issue at T15.5-11 with Judge Larson's views about non-waivable rights, this was on a different branch of the argument, specifically that concerned with the forum selection clause rather than the choice of law clause. At ss47-48 Judge Larson had expressed the view that a court sitting in California might decline to enforce the forum selection clause because of the presence of non-waivable rights citing Wimsatt v Beverley Hills Weight etc International, Inc 32 CalApp4th 1511 (Cal 1995) ('Wimsatt'). Mr Hutley submitted that Wimsatt did not establish that proposition. However, Mr Hutley left untouched Judge Larson's parallel opinion about non-waivable rights and choice of law clauses appearing at ss79-80. There the focus was on Hall and Verdugo which Mr Hutley did not criticise. 40 If I were to reject Judge Larson's views about Hall and Verdugo on the basis of my own reading, this would be procedurally unfair to Epic. For example, Mr Hutley's remarks about Wimsatt were made in chief and Epic was afforded an opportunity to respond to them. My untested views about Hall and Verdugo have not been exposed to that rigour. Further, even assuming that Hall and Verdugo do not go as far as Judge Larson suggests, it does not follow that his opinion is incorrect, just that the decisions cited do not seem to support it. 41 In that circumstance, I will proceed on the basis that I should accept Judge Larson's opinion and conclude that a court sitting in California might ask whether: (a) Epic's rights were non-waivable; (b) whether, if so, the burden ought to shift to Google to show that those non-waivable rights would not be diminished by the application of California law; and (c) if it did, whether Google could discharge that burden. 42 As to question (a), I conclude that a court sitting in California would find that Epic's claims in this proceeding are non-waivable in the relevant sense. 43 As to question (b), I am unclear what the position of a court sitting in California would be. Judge Larson said that the court 'may' shift the burden but did not say when this would occur. 44 As to question (c), as I explain later in these reasons, I am satisfied that Epic's rights under Californian law in relation to this dispute are significantly less than they are under Australian law. As to the substantive rights, the need to bring this suit under statutes such as the Sherman Act is encumbered by the need to prove an impact on United States commerce and the need to show monopolisation of the relevant markets. No such impact need be proven in this proceeding and it will suffice to demonstrate a substantial lessening of competition in the relevant markets which falls well short of the monopolisation requirement under the Sherman Act. So far as remedies are concerned, as I discuss later at [115]-[146], the remedies available in this Court are more generous to Epic than the corresponding remedies that could be granted by a court sitting in California. 45 In that circumstance, the evidence does not permit me to make a finding as to what a court sitting in California would do if confronted with the choice of law clause. I do not know which of the three approaches adumbrated by Judge Larson would in fact be applied by a court sitting in California. Two of the approaches would certainly lead to Pt IV of the CCA and s 21 of the ACL not being applied in the proceeding. Whilst it is possible that the third approach might lead to Pt IV of the CCA and s 21 of the ACL being applied, I have no metric to measure when the burden shifting referred to by Judge Larson would take place. 46 The evidence therefore does not establish how a court sitting in California would approach cl 16.8 and, in particular, it does not establish on the balance of probabilities that if this suit were pursued before a court sitting in California that that court would apply Pt IV of the CCA and s 21 of the ACL. 47 As foreshadowed in Judge Fogel's reply evidence, Google sought to outflank this conclusion by relying upon an undertaking it proffered to this Court not to rely upon the choice of law clause in any proceedings brought by Epic in Santa Clara County. The undertaking (Exhibit 1) was in these terms: Respondents' undertaking to the Court Each of the respondents, Google LLC, Google Asia Pacific Pte. Ltd. and Google Payment Australia Pty Ltd, undertake to the Court as follows: For the strictly limited purpose of responding to the claims advanced by the applicants pursuant to the Competition and Consumer Act 2010 (Cth) in Federal Court of Australia proceeding NSD 190 of 2021 as presently pleaded in the Amended Concise Statement filed 26 July 2021 (Claims), the respondents undertake that they will neither: (a) oppose the federal or state courts located within the county of Santa Clara, California hearing and determining the Claims; nor (b) rely on the Choice of Law clause contained in cl 16.8 of the Google Play Developer Distribution Agreement as having the effect of excluding the Claims. 48 Mr Young submitted that the Court could not accept this undertaking citing Australian Competition and Consumer Commission v Pacific National Pty Ltd [2020] FCAFC 77; 277 FCR 49 ('Pacific National'). He submitted that that case established that there were very limited circumstances in which this Court could accept an undertaking and referred me to [326]-[327] and [334] (per Middleton and O'Bryan JJ). I do not accept this submission. Pacific National was concerned with the grant of substantive remedies. It says nothing about the ability of the Court to accept undertakings in relation to interlocutory matters. 49 There is no direct evidence of how a court sitting in California might respond to the existence of such an undertaking to this Court. There is, however, evidence of how such a court might approach an inter partes undertaking. At the time that Judge Fogel prepared his report, it is apparent that the undertaking which Google then had in mind took only the form of an undertaking by Google to Epic. Judge Fogel's evidence, whilst admittedly very brief and contained in an en passant footnote in his reply report, is to the effect that Google would be bound by such an inter partes undertaking. Because this evidence appeared in Judge Fogel's evidence in reply, Judge Larson was not afforded an opportunity to respond to it although I do not propose to discount Judge Fogel's evidence on that basis since Epic did not submit that I should. 50 This situation then brought forth from Epic a written submission that Google could not unilaterally alter the bargain embodied in cl 16.8 by proffering an inter partes undertaking of the kind mentioned by Judge Fogel: Applicants' Submissions at ss44-50. However, that debate has become irrelevant since the undertaking now proffered is one which is proffered to the Court and not to Epic. In that circumstance, the relevance of the earlier inter partes undertaking under Californian law has receded. The fact that the undertaking to this Court is not an inter partes undertaking also appears to remove the facts upon which Judge Fogel's footnote opinion was based. 51 It may be that a Californian court would approach an undertaking to this Court in the same way as it would approach an inter partes undertaking but whether that is so would be speculation on my part. I do not find that it has been proved that a Californian court would enforce such an undertaking. 52 This is not the end of the matter, however. If in any proceeding commenced by Epic in Santa Clara County Google were now to seek to rely upon the choice of law clause in cl 16.8, this would be, on the face of it, a contempt of this Court. I do not agree with Epic's submission that no contempt proceeding could in that circumstance be commenced because these proceedings would by then have been stayed. I do not think that a stay of this proceeding would be any impediment to the pursuit of a contempt charge since it would not involve the prosecution of this proceeding. If that were wrong, I think it likely that the stay would be lifted sufficiently to permit the prosecution of the contempt. If that too were wrong, there is no reason why a fresh contempt proceeding could not be started. Regardless of the means by which this was achieved, if Google procured a stay of proceedings in this Court by proffering an undertaking which it then breached it would face contempt proceedings. 53 On the other hand, neither Google LLC nor Google Asia is resident in Australia and Google did not point me to any evidence that either had assets in the jurisdiction. Accordingly, it is possible that Google LLC and Google Asia could seek to rely on the choice of law clause in any proceeding in California whilst GPAL (which is resident here) might remain silent. In that case, there would be no means of enforcing any fine imposed for the contempt. Google LLC and Google Asia would lie beyond the reach of this Court's contempt remedies and GPAL would not have committed contempt. I accept that in that scenario, the contempt remedy would be ineffective to secure compliance with the undertaking. I am unable to be precise about the probability of Google acting in this way but I do not think I can disregard it as trivial. 54 The Google parties submitted that an obvious step which could be taken if they breached the undertaking would be for Epic to apply to lift the stay of this proceeding. I accept this is possible. But it would then leave on foot two parallel and identical proceedings, one in this Court and one in Santa Clara. The existence of such twin identical suits in New South Wales and California is undesirable. Indeed, it would give rise to procedural possibilities the precise outcome of which I do not feel can be confidently assessed in advance. For example, Epic might choose to seek to discontinue the proceeding in California. However, it is not clear to me what the terms (if any) which might be applied to such a discontinuance would be. It is also unclear to me how doctrines such as lis alibi pendens or forum non conveniens might apply or whether the spectre of anti-suit injunctions might intrude. On balance, I am not sufficiently confident that the lifting of the stay would be something which would compel Google to cease its reliance upon the choice of law clause in the Californian proceeding. 55 There is an additional reason I do not think the undertaking is sufficient. As Epic correctly submitted, the Respondents in terms only undertake not to raise cl 16.8 against the claims 'presently pleaded in the Amended Concise Statement filed 26 July 2021'. This does not extend to any other Australian law allegations which may be made subsequently. Indeed, at the end of these reasons I conclude that Epic should be granted leave to amend the Amended Concise Statement. This new pleading raises a distinct challenge under Pt IV to the choice of law clause. The undertaking in Exhibit 1 does not extend to this allegation. Further, if a proceeding is now commenced in Santa Clara County with the same content as the Amended Concise Statement I think it very likely that there will be future amendments to it as the litigation progresses to trial. In complex competition cases such as the present, frequent amendments are the norm. It is likely that the market definition allegations will be amended as the evidence for both parties is filed. I think it is also likely that the case on substantial lessening of competition and Google's purpose will be heavily influenced by the filing of evidence on both sides. Google's undertaking will not extend to any such allegations. 56 This problem with the undertaking will leave it open to Google to rely upon the choice of law clause in relation to all allegations freshly raised after the Amended Concise Statement. There are obvious practical difficulties with that course. Although Epic raised this problem at the hearing, Google did not seek to prepare an undertaking that would address the problem but made its stand on Exhibit 1. In particular, it did not seek to proffer an undertaking not to rely upon cl 16.8 as an answer to the allegations made in the Amended Concise Statement or to any other allegations reasonably relating to those allegations or arising out of the same subject matter. 57 In that circumstance, I reach two conclusions: first, that I am not satisfied on the balance of probabilities that a court sitting in California would apply Pt IV of the CCA and s 21 of the ACL in any proceeding commenced by Epic there; and second, that Google's undertaking to this Court is insufficient to dispel that uncertainty.

The Fragmentation Issue 58 Here the issue is whether the refusal of the stay may lead to the undesirable outcome that the same issues get tried in two different courts. There are several problems which arise when such a circumstance occurs. The first is that it is inconvenient to the witnesses and the parties and more expensive for all concerned. Witnesses will be forced to travel to California and Sydney even allowing for an increased likelihood that the evidence of some witnesses may be elicited by means of a virtual platform. Even in the case of virtual platforms, however, the time differences between California and Sydney are large. Giving evidence at night time or early in the morning to a courtroom on the other side of the Pacific is undesirable. Further, four sets of attorneys will need to be retained. 59 Secondly, in complex litigation of the present kind there will be practical problems concerning how documents discovered in one proceeding are to be discovered in the other. Where confidentiality orders are involved this can be especially burdensome as applications may need to be made for confidentiality regimes to be varied in order for the attorneys in one case to access that which the attorneys in the other have. These kinds of problems have arisen before in cases in this Court where litigation has occurred here and in the United States: see, for example, just in my docket: Motorola Solutions Inc v Hytera Communications Corporation Ltd & Anor (NSD 1285 of 2017) and Capic v Ford Motor Company of Australia Pty Ltd (NSD 724 of 2016). These problems are surmountable and are usually surmounted, but they are time consuming and difficult. 60 Thirdly, witnesses cross-examined once in the first trial tend to know what is coming when cross-examined again in the second trial, potentially degrading the forensic quality of the second trial. 61 Fourthly, there is the risk that the judges trying the two cases will make inconsistent findings of fact about the same matters, in part because different judges have different attitudes to evidence and in part because of the fact that witnesses in the second trial have already been around the track once. The spectre of inconsistent findings of fact tends to degrade confidence in the judicial systems of both of the countries involved. There is an attendant risk that the conduct of the two trials will give rise to spot fires in the second trial concerned with issue estoppel, Anshun-type estoppel and abuse of process. In the present case, one may foresee debates about the extent to which findings about markets in one case may have forensic consequences in the other (although not what the outcome of those debates will be). 62 Fifthly, there is a considerable wastage of governmental resources in making two public court systems available for one dispute. Fragmentation of this case will tie up two judges for several months in hearing and then determining the same case. This is wasteful. 63 A fragmentation of a dispute into multiple forums is therefore regarded as undesirable. 64 To gauge the probability of any of these risks materialising in the present case, it is necessary to identify the two sets of proceedings and the issues in them.

The Australian Proceeding 65 Epic alleges that Google LLC and Google Asia have hindered and prevented Epic Games from distributing its apps to Android devices in Australia in any other way than through the Google Play Store. It nominated a number of mechanisms which it says Google has employed to achieve this end. These include: Imposing and enforcing contractual and technical restrictions which stifle or block the ability of consumers to download apps (including apps which provide competitor app stores to the Google Play Store) from the developers' own websites. For example, Epic alleges that in order for an original equipment manufacturer ('OEM') - that is, a company which designs and sells smart mobile devices - to obtain Google mobile services it must enter into a Mobile Application Distribution Agreement. Under this agreement, the OEM is required by Google to pre-install a number of proprietary Google apps including the Google Play Store and it must place the icon for the Google Play Store on the Android device's home screen. Consequently, so Epic alleges, the Google Play Store is often the first or only app store that consumers ever see. Further, it is a term of the DDA that if a developer seeks to distribute an app through the Google Play Store it may not make available any product that has the purpose of facilitating the distribution of apps outside the Google Play Store. Counselling consumers against downloading apps directly from developers. Although it is possible for a consumer to download an app directly from a developer's website, a consumer who attempts to do so is met with a series of warnings which advise them against doing so and they must then change their settings to permit the download to occur. If the consumer persists, further warnings then appear such as 'This type of file can harm your device. Do you want to keep EpicGamesApp.apk anyway?'. Not automatically updating apps which are downloaded in other ways, unlike apps which are acquired through the Google Play Store. In the case of apps downloaded from a developer's website, for example, the consumer must approve each update manually and, in so doing, is confronted with further security warnings. Obliging developers under the DDA only to use Google's in-app payment solution (Google Play Billing) for in-app purchases of digital content where an app is distributed through the Google Play Store. When this system is used the developer must pay a fee which was, until 30 June 2021, 30% of the earnings of the developer. On that day, it was reduced to 15% in relation to the first US$1 million of earnings but 30% thereafter. According to Google, this change resulted in 99% of developers seeing a 50% reduction in the fees they had been paying. 66 Epic Games alleges that this (and other) conduct constitutes an infringement of ss 46(1), 47 and 45 of the CCA. Section 46(1) prohibits conduct by a corporation which has a substantial degree of market power from engaging in conduct which has the purpose or has or is likely to have the effect of substantially lessening competition in that market or other related markets. Section 47 prohibits a corporation from engaging in exclusive dealing. Section 45 prohibits corporations from entering into or giving effect to contracts, arrangements or understandings which have the purpose or have or are likely to have the effect of substantially lessening competition. The two markets in which these deleterious competitive effects are alleged to occur are the Android App Distribution Market and the Android In-App Payment Processing Market. 67 In addition to these competition counts, Epic also relies upon s 21 of the ACL. Section 21 of the ACL prohibits a corporation from engaging in unconscionable conduct. Broadly speaking, the same kind of conduct as underpins the competition case also underpins the s 21 case although there is also some emphasis on the way in which Google secures the agreement of developers to cl 16.8 of the DDA based on the inequality of bargaining power which developers have when dealing with Google. Epic seeks extensive declaratory relief, an injunction to compel Google to restore Fortnite to the Google Play Store with the ability to use its in-game payment system and a raft of other injunctions restraining Google from engaging in the conduct which it says infringes Pt IV of the CCA and s 21 of the ACL. 68 The evidentiary footprint for this proceeding is largely likely to concern the nature of the Android App Distribution Market and the Android In-App Payment Processing Market. Some of this will be in the nature of expert evidence but it may be assumed that there will also be a great deal of documentary evidence about these markets too and some oral testimony. In relation to the central technical facts of what actually happened in terms of the removal of Fortnite and the various agreements entered into there is likely to be little controversy although how all these matters actually operate is likely to be contested. The question of Google's purpose under ss 45 and 46 potentially will involve contested oral testimony.

The US Proceeding 69 There is presently pending in the United States District Court for the Northern District of California a suit entitled Epic Games Inc v Google LLC No 3:20-cv-05671 ('the Epic v Google complaint'). As in the case before this Court, Epic alleges against Google the existence of two markets, the Android App Distribution Market and the Android In-App Payment Processing Market. It alleges that largely the same conduct constitutes breaches of the Sherman Act, the Californian Cartwright Act and the Californian Unfair Competition Law. These are by no means the same as Pt IV of the CCA and s 21 of the ACL but the evidentiary ambit is likely to be similar although with different emphases especially on the nature of the markets concerned. The competition issues posed by the Sherman Act, the Cartwright Act and the Unfair Competition Law are not identical to those posed by Pt IV of the CCA and s 21 of the ACL. Whilst this is no doubt an oversimplification, it is generally easier to establish liability under Pt IV of the CCA than under the Sherman Act since the former usually focuses on a substantial lessening of competition in a market whilst the latter has as its focus the question of whether a defendant has monopoly power. In Epic Games, Inc v Apple Inc (ND Cal, 10 September 2021) 2021 WL 4128925 ('the US Apple Proceedings'), where injunctive relief of the type sought in these proceedings was granted to Epic in respect of similar claims that it made against Apple, Judge Gonzalez Rogers noted that the threshold market share for finding a prima facie case of monopoly power was generally no less than 65%: see 135. A defendant to a claim under s 46, however, can be found liable without being shown to be a monopolist in that sense. Further, under the Sherman Act, the 'rule of reason' which is applied in respect of s 1 requires countervailing pro-competitive effects to be brought into account as an offset: see US Apple Proceedings at 140-141. No such principle exists in this country outside the context of authorisations under Pt VII of the CCA. Finally, whilst purpose is relevant to some claims under Pt IV of the CCA, it is not directly relevant under Sherman Act style claims although in practice it may be collaterally relevant as part of a circumstantial case. 70 Notwithstanding those differences in emphasis, I accept that in a perfect world all of these allegations ought to be tried together and that if they are heard separately the risks which attend the fragmentation of proceedings will be likely to eventuate.

The Practicalities of the US Proceedings 71 Judge Fogel's evidence was that if Epic were now to commence a proceeding in federal court (presumably in Santa Clara County) then it would be assigned to the San Francisco Division of the US District Court for the Northern District of California as a related case to Epic's case against Google which is presently pending there: First Report of Judge Fogel at s17. However, the evidence of both Judge Forrest and Mr Johnson indicates that the picture is more complicated than Judge Fogel's statement suggests. On 5 February 2021, the United States Judicial Panel on Multidistrict Litigation created a centralised action of In re Google Play Store Antitrust Litigation, No 3:21-md-02981 ('the MDL Action'). It presently comprises: (1) The Epic v Google complaint (i.e. the US version of this suit); (2) Two developer class actions which were previously consolidated (Pure Sweat Basketball v Google LLC and Peekya App Services v Google LLC); (3) 14 consumer class actions; and (4) A complaint made by 37 State Attorneys-General against Google. 72 An order made on 9 September 2021 by Judge Donato indicates that the parties to the MDL Action were instructed to plan for the first trial to commence on 6 September 2022. It is unclear which of the 18 actions will be the first to be tried. The evidence of Mr Johnson (for Google) was that document production in the MDL Action was substantially completed on 31 July 2021: Affidavit of Mr Johnson at s59. The evidence of Judge Forrest (for Epic) was that if Epic now commenced a proceeding in the US Federal Court it would most likely be related to and coordinated with the existing MDL Action for pre-trial purposes: Affidavit of Judge Forrest at s15. She thought that the knock on effect of that would be that the trial date for the MDL would be pushed back. I accept this evidence. I also accept Judge Forrest's evidence that from a practical perspective, the substance of Epic's Australian proceeding would be different to the other actions in the MDL Action inasmuch as it would be based on foreign competition law with all of the evidentiary and procedural baggage that would entail. 73 Further, as I discuss later, whilst Google has demanded a right to a trial by jury in the MDL Action, it has now undertaken to this Court (Exhibit 5) to seek a bench trial for this proceeding if Epic is forced to pursue it in California. Just how Judge Donato will react to this development is unclear to me. Because Exhibit 5 came only at the end of the argument in this case I do not have the benefit of any evidence from either party as to what impact this would have on the question of whether this action would be consolidated with the MDL Action. I do not know whether it will be left to make its own way in a separate docket or whether, if joined to the MDL Action, the judge trying the MDL Action will try the bench trial in this case at the same time as the jury trial. 74 For completeness, Epic did not submit that Google's undertaking to seek a bench trial was practically unenforceable (as it did in relation to Exhibit 1). On the other hand, Google did not submit that I was bound to approach this undertaking consistently with the manner in which Exhibit 1 was approached although it was plainly open to it to do so. I do not think that it would be procedurally fair on Epic now to conclude that Google's undertaking to seek a bench trial is unenforceable when Google did not advance such an argument (for obvious reasons). The parties conducted this branch of the debate on the assumption that the undertaking to seek a bench trial would be honoured by Google. I do not think that for me now to conclude that the undertaking was not enforceable would be consistent with the manner in which the hearing was conducted.

Conclusions on Fragmentation 75 In light of these matters, it is unclear to me just what the fragmentation position will therefore be. The most favourable view to Google is that the judge presiding over the jury trial of the MDL Action will also be the judge hearing this action. The least favourable view is that it is decided that this action must make its way separately from the MDL Action. Either results in fragmentation. In the former case, there will be the risk that the trial judge reaches different conclusions to the jury hearing the same evidence. However, there will be no inconvenience to witnesses and only two sets of attorneys. In the latter, there is the risk of inconsistent fact finding and the witnesses will need to give their evidence twice, albeit only in California. 76 If this proceeding is not stayed, there will be the same risk of inconsistent fact finding, a certainty that the witnesses will need to give their evidence twice, four sets of attorneys and the additional expense and inconvenience which I have already noted. I conclude therefore that if I do not stay this proceeding, some - but not all - of the risks associated with fragmentation will come to pass. 77 There is a related fragmentation issue which arises in relation to the position of Epic International which I discuss below.

The Parties Issue 78 The parties to the DDA are Epic Games, Google LLC and Google Asia. On the other hand, the parties to this suit include two other parties, Epic International and GPAL. Where litigation involves only the parties to an exclusive jurisdiction clause the Court starts with a strong disposition in favour of staying a proceeding commenced in contravention of the clause: Australian Health & Nutrition Association v Hive Marketing Group [2019] NSWCA 61; 99 NSWLR 419 ('Australian Health') at [90] per Bell P; Full Court Apple Decision at [69]. But where the interests of other parties are involved this disposition is not generally appropriate: Australian Health at [90] per Bell P. There may be cases, such as Global Partners Fund Ltd v Babcock & Brown Ltd (in liq) [2010] NSWCA 196; 79 ACSR 383, where it may be said that although in form one of the parties to the suit is not a party to the exclusive jurisdiction clause, in substance this is not so ('there are non parties and non parties' as Spigelman CJ drily observed at [74]).

GPAL 79 The issues concerning GPAL are twofold. First, Google submits that the case against GPAL is so insubstantial that the Court should not regard its presence as a third party as anything more than the ornamental addition of a third party for the purpose of evading the effect of cl 16.8. Secondly, if that is not so, Google submits that GPAL is entitled to the benefit of cl 16.8 although not a party to it. Ornamental joinder 80 In the Apple Decision, I reached the conclusion that the joinder of an Australian member of the Apple group of companies, Apple Pty Ltd, was an example of an ornamental joinder since it had no substantive role in the allegations which Epic made against Apple. However, this conclusion was reversed by the Full Court which concluded that Epic's concise statement disclosed actual allegations against Apple Pty Ltd: see the Full Court Apple Decision at [74]-[75]. 81 The same debate now takes place in relation to the role of GPAL. The allegations relevant to GPAL in Epic's Amended Concise Statement are at ss11, 23, 25, 44(b), 51 and 52(b). These are as follows: 11. The Third Respondent (Google Australia) is a subsidiary of Google LLC. It enters into contracts with app developers for the processing of Google's payment transactions in Australia, (including purchases through Google Play Billing) and requires that app developers pay Google Australia a commission of typically 30% on payments for apps sold through the Google Play Store in Australia, and in-app purchases of digital content consumed within such apps. 23. In order to distribute their Android OS apps through the Google Play Store, developers must enter into the Google Play Developer Distribution Agreement (DDA). The DDA is a standard form, non-negotiable contract. It requires developers to submit every app that they wish to be distributed through the Google Play Store to Google for review and approval, and permits Google to disable and remove apps that violate the DDA. For apps distributed in Australia, Google Asia Pacific is a contracting entity with app developers under the DDA. 25. The DDA requires app developers to enter into the Google Payments - Terms of Service - Seller Agreement (Payments Agreement) with Google Australia in order to receive payment for apps distributed through the Google Play Store in Australia and for in-app purchases in Australia of digital content consumed within those apps. The Payments Agreement requires that app developers pay Google Australia a commission of typically 30% on payments for apps sold through the Google Play Store in Australia, and in-app purchases of digital content consumed within such apps. 44. Further, by reason of the matters referred to in paragraphs [9]-[10] and [34]-[41] above, Google has a substantial degree of power in the Android In-App Payment Processing Market, including in Australia. Further, Google has engaged, and continues to engage, in conduct that has the purpose, effect or likely effect of substantially lessening competition in the Android In-App Payment Processing Market, including in Australia, by the following means: … (b) (paragraph [25] above) Google LLC and Google Asia Pacific require, through the DDA, app developers to enter into the Payments Agreement with Google Australia in order to receive payment for apps distributed through the Google Play Store and for the in-app purchase of digital content consumed within such apps in Australia; 51. Further or alternatively, by reason of the matters referred to in paragraphs [23]-[27] above, Google has made, and continues to make, contracts or arrangements, or has arrived at or continues to arrive at, understandings with app developers through the DDA, and/or through the Google Policies, and/or through the Payments Agreement, which contain provisions that: (a) restrain app developers from using any in-app payment processing system, other than Google Play Billing, for the purchase of digital in-app content by Android device users, including in Australia; (b) restrain app developers from distributing their apps to Android device users, including in Australia, other than through the Google Play Store; (c) permit Google to remove from the Google Play Store apps that violate the DDA; and (d) permit Google to charge a supra-competitive price, of typically 30%. 52. By reason of the matters referred to in paragraphs [42]-[45] above, the provisions referred to at paragraphs [50] and [51] above have, individually and/or cumulatively, the purpose, effect or likely effect of substantially lessening competition in the Android App Distribution Market and/or the Android In-App Payment Processing Market, including in Australia. In addition: … (b) by Google Australia's conduct at paragraphs [11] and [44] above, Google Australia has given effect to the provisions referred to at paragraph [51] above, including in Australia or in relation to Australian users of Android devices. 82 In its originating application, Epic seeks declarations that this conduct contravened ss 45 and 46(1) of the CCA and s 21 of the ACL. 83 The gist of the case pleaded against GPAL seems to be this: for apps distributed in Australia, the app developer is required to enter into a DDA with Google Asia (s23). It is a term of the DDA that the app developer will then enter into a payment agreement with GPAL (s25). It is a term of any payment agreement with GPAL that the app developer will pay a commission (typically of 30%) on payments for apps sold through the Google Play Store in Australia and on in-app purchases of digital content (s11). It is then alleged that Google Asia has engaged in conduct which has the effect or likely effect of substantially lessening competition in the Android In-App Payment Processing Market by requiring, through the mechanism of the DDA, that a developer enter into a payment agreement with GPAL (s44(b)). Finally, it is alleged that by entering into the payment agreement with its 30% commission entitlement, GPAL has given effect to an arrangement or understanding which substantially lessens competition in the Android In-App Payment Processing Market (ss51-52). 84 This appears to be a real case against GPAL in the sense that the allegations coherently disclose allegations under Pt IV of the CCA. Google sought to demonstrate that the case had no substance by showing that Epic Games had not entered into a payment agreement with GPAL: Respondents' Submissions at s10. Whilst this is not a summary judgment application it seems to me that the Court is being invited to conclude that the case I have just described has no reasonable prospects of succeeding. 85 As finally developed in Mr Hutley's submissions in reply, Google's contention appeared to be that there was no evidence that Epic Games had entered into a payment agreement with GPAL: T54.35-55.6. 86 The evidence about this is somewhat complex. The DDA was entered into on 8 November 2016: Affidavit of Mr Johnson at s28. In correspondence, Google indicated that it was relying on the most recent version of the DDA which was located at pages 146-152 of Exhibit SJW-2 to Mr Johnson's affidavit and which is dated 17 November 2020. The relevant clauses are first, cl 2.1 which provides inter alia that 'You are contracting with the applicable Google entity based on where You have selected to distribute Your Product (as set forth here)'. By clicking on this hyperlink, one is taken to a webpage that bears the title 'Supported locations for distribution to Google Play users' (this webpage is extracted at pages 3-12 of Exhibit SJW-2). On that webpage is a series of tables purporting to show 'app availability information, along with supported currency and price range information for Google Play users'. At page 10 of SJW-2 is this statement above a table applying to Australia and a number of other countries: Commercial role between Google and You. You appoint Google Asia Pacific Pte. Ltd. as Your marketplace service provider to make Your products available in Google Play to users in the following countries… 87 Consequently, the relevant contracting party to the DDA insofar as Australia is concerned is Google Asia. Clause 3.1 provides: 3.1 You hereby appoint Google as Your agent or marketplace service provider as outlined here to make Your Products available in Google Play. 88 Clause 3.2 provides: 3.2 This Agreement covers both Products that users can access for free and Products that users pay a fee to access. In order for You to charge a fee for Your Products and to be paid for Products distributed via Google Play, You must have a valid Payment Account under a separate agreement with a Payment Processor, be approved by a Payment Processor for a Payment Account, and maintain that account in good standing. If there is a conflict between Your Payment Processor agreement and this Agreement, the terms of this Agreement will apply. 89 So Epic Games was required to have a separate agreement with a 'Payment Processor' and a valid 'Payment Account'. By cl 1 a Payment Processor is an entity authorised by Google Asia 'to provide services that enable Developers with Payment Accounts to be paid for Products distributed via Google Play'. Clause 4.1 provides that Epic Games 'must adhere to the Developer Program Policies'. The link to which that hyperlink went was located in Annexure DP-9 to an affidavit of Mr Poddar sworn on 7 April 2021 and was to a page entitled 'Developer Program Policy'. This affidavit was tendered as Exhibit 3. Clause 1 of the policy, under the heading 'Payments', says that 'Developers charging for apps and downloads from Google Play must use Google Play's billing system as the method of payment'. 90 Clause 8.1 of the DDA relevantly provides: ….You agree to comply with this Agreement and the Payment Processor's Payment Account terms of service… 91 I accept that the effect of cl 4.1 of the DDA (with the Developer Program Policies) and cl 8.1 of the DDA is to require Epic Games to comply with whatever the terms of service are which govern the relationship with the relevant Payment Processor. Epic also relied on cl 3 of the Google Play Terms of Service which requires the party to that agreement to have a Google Payment account and to agree to the Google Payments Terms of Service. I do not accept that this document is relevant. It governs the position of users accessing the Google Play service not the position of developers. 92 There is no direct evidence that the Payment Processor for Epic Games is GPAL. However, Epic relied upon a document entitled 'Product Disclosure Statement (Google Payment Australia Pty Ltd)' (located at page 91 of Exhibit SJW-2). In its electronic form it bears these words on its front page: 'By clicking "Accept" below, I confirm that I have read and accept the terms of the Product Disclosure Statement/Terms of Service Document and agree to receive them online'. It is open to infer on the basis of this document that the Google Payments Service is operated by GPAL in Australia. Clause 3.1 provides: 3.1 The Google Payments Service is an online payment processing service that is designed to facilitate the processing of payments by a valid payment method accepted by GPAL, between purchasers ("Buyers") and participating merchants ("Sellers"). 93 Clause 4.1 provides: 4.1 On registering to use the Google Payments Service, both Buyers and Sellers will be bound by this PDS and the relevant Terms of Service as they apply to either a Buyer or Seller using the Google Payments Service. 94 On the other hand, cl 2.1 provides: 2.1 This PDS only applies to you if you receive it in Australia. 95 There is no evidence that Epic Games received the PDS in Australia. The PDS also contains at Schedule 2 a series of terms entitled 'Sellers Terms of Service'. Under cl 6.2 of those terms GPAL is obliged to transfer funds due to the seller to 'the Settlement Account'. This is defined in cl 1 of Schedule 2 as meaning 'The deposit account of Seller maintained at a financial institution located in Australia that is designated by Seller and approved by GPAL for receipt of funds from the processing of Payment Transactions'. Mr Young conceded that Epic Games did not have any such bank account: T49.27-28. 96 The fact that Epic Games has not complied with cl 6.2 of the Sellers Terms by having an Australian bank account is an indicator that it does not use GPAL as its Payment Processor. If it did, it is difficult to understand how it receives funds from GPAL. The absence of any evidence that Epic Games entered into the PDS with GPAL is evidence which tends in the same direction. The fact, as I have accepted, that Epic Games is obliged to comply with the terms of service with its Payment Processor does not assist in identifying who that Payment Processor is. 97 There is other evidence bearing on this issue. Mr Johnson gave evidence on information and belief that GPAL does not pay any funds to Epic Games associated with in-app purchases made by Australian users on the Fortnite app: Affidavit of Mr Johnson at s21. This is consistent with Mr Young's concession that Epic Games does not maintain an Australian bank account. 98 Further, there is evidence that Epic Games did enter into a payment agreement with a separate Google entity, Google Payment Corp. This agreement is entitled 'Google Payments Terms of Service - Seller (US)' (Exhibit SJW-2 at page 160). It identifies Google Payment Corp as the contracting Payment Processor. Mr Johnson gave evidence in his affidavit at s27 that he had been shown an extract from an internal Google database that records the terms and conditions that Epic Games had accepted with Google LLC through its developer account (Exhibit SJW-2 at 137). This extract shows that an entity identified as the customer: ACCEPTED terms and conditions for document 0.buyertos for country US, version 5.5, legal_document_id: 16222, Customer has ACCEPTED terms and conditions for document 0.sellertos for country US, version 1.9, legal_document_id: 13872. 99 It is open to infer from Mr Johnson's evidence that the customer is Epic Games. The terms of service referred to by Mr Johnson at page 160 of Exhibit SJW-2 bears the document ID 13872. It is open to infer that the document is an agreement that Epic Games reached with Google Payment Corp. Section 12 is headed 'Special Terms for Google Play and Other Google Marketplaces' and appears to regulate payments made when a seller operates in more than one Google marketplace. Clause 12.1 provides: 12.1 Google Play and other Google Marketplaces. Notwithstanding the first sentence of Section 3.1, Seller may use the Service to process transactions on a Google Marketplace. To use the Service on a Google Marketplace, a Seller must separately agree to the terms of service applicable to such Google Marketplace (the "Google Marketplace TOS") and have the Seller's Service account linked to the Seller's account for that Google Marketplace. In the event of any conflict between this Agreement and the applicable Google Marketplace TOS, the terms of the applicable Google Marketplace TOS will control. 100 This appears to contemplate that Epic Games might use Google Payment Corp to collect payments from the Australian Google Play Store. However, it also appears to require that Epic Games have a 'Seller's account for that Google marketplace'. Provisions exist in cl 12.5 dealing with currency differences. As it happens there is a largely identical cl 12.5 in the GPAL terms of service. 101 The two payment processing agreements contemplate that a Seller will be able to sell in a different Google Play Store to the jurisdiction in which the Payment Processor is located. However, there is no satisfactory explanation of how these cross-border arrangements work. Whilst it is easy to think that what occurred in this case is that sales of Fortnite were collected by Google Payment Corp I do not think that the complete picture has yet been presented. In particular, the requirement in cl 12.1 of the Google Payment Corp agreement that Epic Games should have a Sellers Account for the Australian Google Play Store is evidence from which one could infer that there is a contractual relationship between GPAL and Epic Games. It is also open to infer that this would be on the terms of the GPAL agreement. The fact that Epic Games has no Australian bank account is certainly inconsistent with Epic Games using GPAL as its primary Payment Processor but it is not inconsistent with it using Google Payment Corp as its primary processor and having a Sellers Account with GPAL as required by cl 12.1. 102 In my view, it would be inappropriate to reach any firm conclusions about this prior to trial. I do not accept therefore that it has been sufficiently shown by Google that Epic Games is not a party to a contract with GPAL. Is GPAL a third party beneficiary? 103 Google submits that by reason of cl 16.4 of the DDA, GPAL is a third party beneficiary of cl 16.8. Clause 16.4 provides: 16.4 You acknowledge and agree that each member of the group of companies of which Google is the parent shall be third party beneficiaries to this Agreement and that such other companies shall be entitled to directly enforce, and rely upon, any provision of this Agreement that confers a benefit on (or rights in favor of) them. Other than this, no other person or company shall be third party beneficiaries to this Agreement. 104 Epic submits that cl 16.4 only applies to the Google companies defined in cl 1.1 (which do not include GPAL). However, that is not what cl 16.4 says. It confers upon each member of the group of which Google is the parent the status of a third party beneficiary. GPAL is a member of that group and I therefore accept that GPAL is a third party beneficiary under cl 16.4. 105 California Civil Code s 1559 provides that a contract made expressly for the benefit of a third party may be enforced by that party at any time before the parties to the contract rescind it. I therefore accept that as a matter of Californian law GPAL is entitled to enforce cl 16.8. Epic disputed that the DDA had been made expressly for the benefit of GPAL but this appears to me to be untenable in light of cl 16.4. 106 In relation to GPAL I therefore conclude that it is a properly joined respondent against whom a real case is put but that it is, in fact, entitled to rely upon cl 16.8. It is not therefore a stranger to the exclusive jurisdiction clause and its presence in the suit provides no reason to step outside the usual strong reasons approach to exclusive jurisdiction clauses.

Epic International 107 Google submitted that Epic International was bound as a non-party by cl 16.8 of the DDA by the application of the Manetti-Farrow rule. That rule was described by Judge Fogel at s58 of his first report. At ss65-66 Judge Fogel thought that the rule would apply on the assumption that Epic International was the agent of Epic Games for the distribution of apps through the Google Play Store. 108 Neither party took me to any evidence proving or disproving that Epic International was the agent of Epic Games for the purpose of app distribution. It would appear therefore that the facts underlying this aspect of Judge Fogel's evidence have not been proven. 109 In its submissions, Epic denied that Epic International was involved with apps being provided to Google for distribution through the Google Play Store: s41. This submission was based on the information and belief evidence of Mr Poddar at s7 of his first affidavit. That evidence was based on statements that Legal Counsel at Epic Games had made to him to the effect that Epic International's role was to enter into end user agreements with players of Fortnite and that it had no role with Google in relation to the distribution of the Fortnite app. 110 Mr Poddar's evidence removes the premise upon which Judge Fogel's evidence rested. I therefore do not accept the submission that Epic International is bound by cl 16.8 because of the Manetti-Farrow rule. In any event, it is doubtful whether that rule would have any application in the present case. The rule is one which holds that a non-party to a forum selection clause may assert the clause where the non-party's conduct is sufficiently related to the contractual relationship. It takes as its point of departure the non-party's status as a defendant in a US proceeding asserting an entitlement to rely upon the clause. It is not self-evident to me that this is a principle of Californian contract law but rather seems to be part of the law applied by courts in California to the enforcement of forum selection clauses. As such I doubt that the principle has any application in an Australian proceeding where a defendant seeks to hold a non-party plaintiff to a forum selection clause. The relevant principle is most likely the domestic one referred to by Spigelman CJ in Global Partners that there are non-parties and non-parties; that is to say, the question is really one of Australian law and its willingness to extend the burden of a forum selection clause to a party before it who is not a party to the forum selection clause. Further, Judge Fogel's views of the Manetti-Farrow rule do not explore whether it would be excluded by the choice of law clause in cl 16.8 as part of the conflict of laws principles applying in California. In any event, since these questions do not arise it is not necessary to pursue them further. 111 Google did not pursue an argument that Epic International was an express beneficiary of cl 16.8 by reason of cl 15.4 of the DDA (as I have held GPAL to be). Had such a submission been advanced I would have rejected it on the basis that Epic International is not a member of the Google group of companies. 112 In those circumstances, I do not accept that Epic International is bound by cl 16.8 by reason of the Manetti-Farrow rule. An argument based on what was said by Spigelman CJ in Global Partners was not advanced and any consideration of it by me would lie outside the way in which the parties conducted the case.

Conclusion on third parties 113 I conclude that GPAL is entitled to rely upon cl 16.8 against Epic Games and that none of the Respondents may rely upon it against Epic International. Since Epic International's suit against the Respondents is essentially identical to Epic Games' suit against them it is necessary to attend to what the consequences of any stay would be. It seems to me that if I granted a stay on the basis of cl 16.8 this would leave the suit brought by Epic International on foot. This would in turn create a trifurcated proceeding where Epic Games' proceeding would need to be heard as a bench trial in the United States (in light of Google's undertaking to consent to a bench trial of this action), the MDL Action would be heard by a jury and Epic International's case would be heard by this Court. This would appear to be an undesirable outcome. 114 Two solutions present themselves both of which turn on the exercise of discretion. One would be to refuse the stay of the proceeding brought by Epic Games leaving the whole suit in this Court on foot. The other would be to grant the stay against Epic Games and then grant a discretionary stay against Epic International. Google submitted that this might be done on the basis articulated by Lockhart J in Sterling International Pty Ltd v The Boots Company (Australia) Pty Ltd (1992) 34 FCR 287 at 290-291. I accept that the Court has the power to do as Google submits (stay Epic International's suit) just as it has the power to do as Epic submits (refuse the stay application). I return to how that discretion should be exercised below.

The Remedies Issue 115 The relief sought by Epic has four elements. The first consists of extensive declarations that the Respondents have engaged in conduct in breach of Pt IV of the CCA and s 21 of the ACL. Secondly, it seeks injunctions restraining the same conduct. Thirdly, it seeks an injunction that Google immediately restore Fortnite to the Google Play Store with Epic's in game direct payment system. Fourthly, it seeks orders pursuant to s 87 of the CCA and s 237 of the ACL declaring that certain contractual provisions are void and varying certain other contracts. 116 Google submitted that if Epic's proceeding were to be pursued in Santa Clara County in accordance with cl 16.8 it would be able to obtain substantially the same relief as it could obtain from this Court. Epic disputed that this was so both in relation to the injunctions it sought under CCA s 80 and ACL s 232 and to orders it sought under s 87(2) of the CCA revising the terms of the DDA. In relation to the injunctions it contended that a stricter test would be applied for the grant of an injunction by a US court than this Court under s 80 of the CCA. In relation to orders under s 87(2) of the CCA, it submitted that no such relief would be available from a US court. 117 Because Epic's argument extends only to the position of injunctions under s 80 and the revision of contracts under s 87(2), it is not necessary to consider the position of the declaratory relief it seeks.

The injunctions sought by Epic 118 It was submitted by Google that there were two bases for concluding that a state or federal court in California could grant injunctions of this kind. The first was that whilst remedies are generally governed by the lex fori (here Californian law), a court would depart from that principle where it was satisfied that the remedy afforded by foreign law was so inseparable from the cause of action that it must be enforced to preserve the character of the cause of action. The second was that if local remedies were to be applied, the injunctions sought were well within the scope of the equitable relief which US federal or state courts are authorised to grant. 119 As to the first basis, Judge Fogel and Judge Larson agreed that if a remedy afforded by foreign law was inseparable from the cause of action which it vindicated then notwithstanding the general rule that remedies were governed by the lex fori, a court sitting in Santa Clara County would nevertheless apply the foreign law remedy. Judge Larson did not feel that any of the remedies sought by Epic were of this nature. Judge Fogel did not deal with that question in his evidence in chief although in his evidence in reply he said that he questioned Judge Larson's conclusion. His thinking was that the remedial features of Australian law which led Judge Larson to the conclusion that the forum selection clause would not be enforced would actually 'seem to make it more likely that the Relevant U.S. Court could conclude, to the extent that such provisions might be necessary to support the remedies sought by Epic, that some or all of them are in fact inseparable from Epic's substantive rights under Australian law'. 120 I do not accept Judge Fogel's opinion on this matter. In his report in chief, Judge Fogel identified two decisions in which the exception had been applied: Black Diamond S.S Corporation v Robert Stewart & Sons (Norwalk Victory) 336 US 386 (1949) and Matter of Bethlehem Steel Corporation 435 FSupp 944 (ND Ohio 1976). Both of these cases were concerned with the question of whether a time bar in a foreign statute went to the existence of the cause of action or were mere procedural time bars. Judge Fogel was explicit that he was unaware of any decision of a federal court within the Ninth Circuit or of a state court in California where the exception had been applied. 121 I do not think that the relationship between an injunction under CCA s 80 and ACL s 232 and the causes of action in Pt IV and s 21 resembles the relationship between a substantive time bar and the cause of action in which it is embedded. This is not to say that examples where the remedy and the right cannot be disaggregated are confined to time bar situations. I would accept that in a number of situations, particularly in equity, there may be significant difficulties in distinguishing between a right and the remedy which vindicates it. For example, the action in devastavit is conceptually intertwined with the action in chancery for the taking of accounts and there may be difficulties in distinguishing a constructive trust from the remedy which establishes it. In the statutory field, a court's power to appoint receivers may fall in the same category. 122 But accepting all that to be so, I do not think that the statutory causes of action erected by Pt IV of the CCA and s 21 of the ACL are intertwined with the remedy of injunction in this fashion. A plaintiff does not have to seek an injunction under s 80 for a contravention of Pt IV (or under s 232 for a contravention of s 21 of the ACL) and may instead merely seek damages. If the causes of action can be vindicated without an injunction it is difficult to see how it can be said that the injunction remedy is inseparable from the causes of action. 123 Mr Hutley submitted that a different conclusion was required because of s 4L of the CCA. Section 4L provides: 4L Severability If the making of a contract after the commencement of this section contravenes this Act by reason of the inclusion of a particular provision in the contract, then, subject to any order made under section 51ADB or 87, nothing in this Act affects the validity or enforceability of the contract otherwise than in relation to that provision in so far as that provision is severable. 124 He submitted that the effect of this provision was that if the relevant provisions of the DDA (or the agreements relied upon by Epic) breached Pt IV then s 4L had the substantive effect of rendering those provisions void. So much may be accepted. The operation of s 4L has been determined by the High Court in SST Consulting Services Pty Ltd v Rieson [2006] HCA 31; 224 CLR 516 ('SST'). It reverses the ordinary rule that a contract which is illegal will not be enforced. Instead, the High Court held that s 4L affirms the 'central proposition' that a contract the making of which contravened the CCA is valid and enforceable: at [34]. It renders unenforceable an offending provision of the contract in so far as the provision is severable: at [35]. 125 Consequently, s 4L would be applied by any court sitting in California as part of the substantive law of the CCA. This mattered, so the argument ran, because ss 80 and 87 are 'tied up with' the operation of s 4L: T20.38. By the phrase 'tied up with', the submission should be understood to refer to the test of 'inseparability' described by Judge Fogel and Judge Larson. In respect of s 80, for example, Mr Hutley submitted that 'the injunction, properly understood, is merely the external manifestation by the court to the relevant party to prevent it from, as it were, going forward on the basis that the provision could be valid because the application of Australian law, if it's achieved, the relevant provision will be invalid by virtue of 4L': T21.40-44. The effect of s 4L was therefore that the injunction remedy was to be seen as 'supportive of the operation of the law, which is a substantive operation' and not a matter of remedy but as an 'entailed consequence of a breach': T21.44-46. On this basis Mr Hutley submitted that the exception under the law applying in California would be applied. Both s 80 and s 87(2)(b) were by reason of s 4L part of the underlying causes of action in Pt IV and as part of the causes of action any court sitting in California would apply them. 126 I do not accept this submission, for two reasons. First, the concept of the remedial provisions as 'external manifestations' of s 4L is difficult to square with the fact that the orders which may be made under them are much broader in effect than s 4L. Section 87(2) extends to orders varying the whole or any part of a contract. Likewise, an injunction under s 80 is not limited to restraining the contravention of the CCA which has occurred but extends to the grant of an injunction 'in such terms as the Court determines to be appropriate': Pacific National at [327] and [433]. Nothing of such breadth can be achieved by s 4L. Consequently, I do not accept that s 4L is to be interpreted as secreting within it the whole of ss 80 and 87(2). The premise for Google's argument is not made good. 127 In any event, even if the effect of s 4L and ss 80 and 87(2) were the same, it would not follow that the latter are part of the causes of action conferred by Pt IV. Rather, what would follow would be that ss 80 and 87(2) were otiose to the operation of the CCA. Thus, if Google's submission should be understood to be that ss 80 and 87(2) are secreted within s 4L this would appear to leave them with no work to do. That conclusion would be inconsistent with the principle of statutory construction which counsels against interpretations leading to redundancy: Western Australian Planning Commission v Southregal Pty Ltd [2017] HCA 7; 259 CLR 106 at [55]. 128 Secondly, I am not satisfied that orders under ss 80 or 87 are necessary for s 4L to be given effect to and no support for such a proposition can be found in the High Court's decision in SST. If ss 80 and 87 are not necessary for the operation of s 4L, it is difficult to see how they can be seen to be inseparable from it, let alone from ss 45-47. 129 I therefore do not accept that s 4L advances Google's position. 130 Independently of Mr Hutley's argument based upon s 4L, Judge Fogel also thought that the remedies for breaches of Pt IV and s 21 might well be inseparable from the cause of action. He reasoned that these remedies were somewhat different to those which were available and that this lent credence to the idea that they were part of the substantive causes of action. I do not accept this contention. The fact that Australian remedies are different to the remedies in California says nothing about whether those remedies are part of the causes of action. I therefore reject this aspect of Judge Fogel's evidence. 131 Consequently, I do not accept that an injunction under s 80 or other remedies under s 87(2) are inseparable from the causes of action in Pt IV (or under s 232 in relation to s 21 ACL). I do not accept that if the present suit were brought in Santa Clara County that a court sitting there could grant an injunction under s 80 or other relief under s 87(2). 132 For completeness, although Judge Fogel suggested that Epic might be able to obtain injunctive relief under the Sherman Act, the Cartwright Act and the Unfair Competition Act this evidence was not relied upon by Google in its submissions. Whilst I would accept that if Epic commenced a proceeding in Santa Clara County it could add to its Australian law claims fresh claims under these US statutes, in each case it would be necessary to show that the impugned conduct impinges on US commerce and to meet the monopolisation test. No such requirement attends the allegations it presently makes. I do not think therefore that the fact that Epic could seek US statutory remedies if it brought a case alleging an impingement on US commerce says anything useful on the question of the equivalence of the US and Australian remedies on the case it does actually advance. 133 It is therefore necessary to consider the circumstances under which a US court may grant an injunction in its equitable jurisdiction. There is no doubt that a state or federal court sitting in Santa Clara County has an equitable jurisdiction to grant injunctions. Both Judge Fogel and Judge Larson agreed on that. 134 Judge Larson gave evidence at ss107-114 about the test for the grant of an injunction under the law applying in California. Both federal and state law in the US require a plaintiff to show that it has suffered irreparable damage or will do so if the injunction is not granted. Consequently, before such an injunction will be issued it must be shown that damages are not an adequate remedy. I did not apprehend Judge Fogel to dissent from this proposition. 135 By contrast, an injunction may be granted under s 80 even if there is no threat of substantial damage (s 80(4)(c)) and even if the defendant is not threatening to repeat the act in respect of which the injunction is sought (s 80(4)(a)). I therefore accept Epic's submission that, in principle, there is a difference between the circumstances in which an injunction may be granted under s 80 and the circumstances in which an injunction may be granted by a court in Santa Clara County. 136 Google did not pursue a submission that this difference might turn out to be illusory in the practicalities of this litigation. For example, it did not submit that it was obvious that any breaches by it of Pt IV would give rise to irreparable damage if not restrained so that the distinction between the requirements for an equitable injunction in the courts in Santa Clara and the requirements under s 80 were immaterial in fact even if distinct in theory. The closest Google came to this was to point out that an injunction had been granted by Judge Gonzalez Rogers in the US Apple Proceedings. But I do not extract from that passing observation the proposition that the practicalities of this litigation make any difference between the two standards illusory. Had the submission been made I may well have concluded in light of the US Apple Proceedings that the injunction that Epic seeks requiring Fornite's restoration to the Google Play Store might well have been granted under both standards. However, even if that conclusion were reached it would throw no light on the injunctions also sought by Epic to restrain further contraventions of Pt IV (i.e. the second set of injunctions it seeks).

Revision of contracts under s 87(2)(b) 137 This Court may make orders under s 87(2)(b) of the CCA varying the terms of a contract constituting or resulting from a breach of Pt IV of that Act (this language is loose but will do). As mentioned above, Epic seeks relief of this kind in relation to a number of agreements. Epic submits that relief of this kind is not available from state or federal courts sitting in Santa Clara County and it will therefore lose a juridical advantage if required to litigate its suit before the courts of that place. 138 Judge Larson gave evidence supporting this submission at ss115-116 of his report. In his view, whilst it was possible for a state or federal court in Santa Clara to rescind terms in a contract, it was not open to it to rewrite a contract. Judge Fogel took issue with this opinion in his reply report. In his opinion US courts have the power to reform contracts and sever unconscionable provisions from them. 139 Judge Fogel did not explain what reforming a contract entailed although he did refer me to three decisions dealing with the concept: CIGNA Corporation v Amara 563 US 421 (2011) ('CIGNA'), Mabb v Merriam 129 Cal 663, 62 P 212 (Cal 1900) and Western Federal Savings & Loan Association v Heflin Corporation 797 FSupp 790 (ND Cal 1992) ('Heflin'). Having read those decisions it is apparent that the equitable power of reformation is akin to what in Australian law is known as rectification. For example, in Heflin, the District Court referred at 792 to a provision of the California Civil Code (s 3399) under which a written contract may be revised to express the true intention of the parties when through fraud or mistake the contract as written does not express their true intentions. The court noted that 'The party seeking relief must prove the true intent of the parties by clear and convincing evidence'. In CIGNA the Supreme Court approved a statement in Hearne v Marine Insurance Co 87 US 488 (1874) that 'The reformation of written contracts for fraud or mistake is an ordinary head of equity jurisprudence'. 140 Whilst I accept, therefore, Judge Fogel's evidence that state and federal courts in Santa Clara County would have an equitable jurisdiction to reform a contract, this would only arise where the written contract did not reflect the true intent of the parties by reason of fraud or mistake. As such I do not accept that it is a close substitute for the power in s 87(2) which may be exercised without proof of fraud or mistake. 141 In relation to the severing of unconscionable provisions, Judge Fogel referred me to the US Apple Proceedings where Judge Gonzalez Rogers said at [128]: In California, "where a single contract provision is invalid, but the balance of the contract is lawful, the invalid provision is severed, and the balance of the contract is enforced." For example, when a contract is held to be unconscionable, "the strong legislative and judicial preference is to sever the offending term and enforce the balance of the agreement." (citations omitted) 142 Again, I do not accept that the ability to sever unconscionable terms from a contract is equivalent to the power of the Court under s 87(2) to vary the terms of a contract. The power under s 87(2) is not pre-conditioned on any finding that a term of the contract is unconscionable. Instead, the power may be exercised where the contract is one which offends Pt IV. Unconscionability is not a necessary concept under Pt IV. 143 Consequently, I accept Epic's submission that if it is required to litigate its claims in Santa Clara County it will lose the juridical advantage of being able to seek the revision of the contracts it impugns. The availability of a remedial power to reform written contracts where they do not reflect the intentions of the parties or to sever from them unconscionable terms falls far short of the power conferred on this Court by s 87(2). 144 Google submitted that in the present context it was not sufficient for Epic to demonstrate that there were differences between the remedies which were available to a court sitting in California and the remedies which were available in this Court. It needed to be shown instead that the differences were such that the party seeking to resist enforcement would be denied justice in the foreign jurisdiction: Spiliada Maritime Corporation v Cansulex Ltd [1987] AC 460 at 482-483; Re Harrods (Buenos Aires) Ltd [1992] Ch 72 at 117, 122, 126. I accept Google's submission that an asserted juridical advantage generally needs to have an impact on substantial justice: Wigmans v AMP Ltd [2021] HCA 7; 388 ALR 272 at [42]; Allstate Life Insurance Co v Australia and New Zealand Banking Group (No 1) (1996) 64 FCR 1 at 37G-38A. However, I am satisfied that requiring Epic to prove that damages are not an adequate remedy before it can obtain an injunction to restrain the contraventions of Pt IV which it alleges constitutes a relevant denial of justice. An assessment of the justice of the situation cannot put to one side the fact that the remedy of a s 80 injunction is available precisely so that the public good may be served by ensuring that the requirements of Pt IV are observed. Part IV protects the competitive integrity of the economy and the justice of the situation is not to be examined in a blinkered way as if the issues in the case were solely of a private nature. So viewed, it seems to me that the public nature of Pt IV means that the injunction remedy available in California will not be sufficient to ensure that the public interests involved are protected. 145 Insofar as the revision of contracts is concerned, it appears to me that a court sitting in California has no power approaching in scope or breadth the power in s 87(2). I am satisfied that this juridical advantage in this Court is one the denial of which would be an injustice if the proceeding were to be heard in California. 146 For completeness, for the same reasons I have given in relation to the remedy of injunction under s 80, I likewise reject Google's contention that s 87(2) is inseparable from the causes of action conferred by Pt IV of the CCA and s 21 of the ACL.

The Platform Provisions Issue 147 Epic drew attention to ss 83 and 87(1A) of the CCA which it referred to together as the 'Platform Provisions'. The thinking behind that nomenclature is that they erect third party remedial rights on the platform of any judgment delivered by this Court. For example, if this Court finds that certain conduct by a respondent is a contravention of Pt IV of the CCA, a third party may rely upon that finding in a subsequent suit against that respondent as evidence of that contravention.

Section 83 148 Section 83 of the CCA provides (relevantly): 83 Findings and admissions of fact in proceedings to be evidence (1) In a proceeding against a person under section 82 or in an application under subsection 51ADB(1) or 87(1A) for an order against a person, a finding of any fact made by a court, or an admission of any fact made by the person, is prima facie evidence of that fact if the finding or admission is made in proceedings: (a) that are proceedings: (i) under section 77, 80, 81, 86C, 86D or 86E; or (ii) for an offence against section 45AF or 45AG or subsection 56BN(1) or 56CC(1); and (b) in which that person has been found to have contravened, or to have been involved in a contravention of: (i) a provision of Part IV or IVB; or … (2) The finding or admission may be proved by production of: (a) in any case - a document under the seal of the court from which the finding or admission appears; or (b) in the case of an admission - a document from which the admission appears that is filed in the court. 149 It will be seen that this provision makes a sealed copy of this Court's judgment prima facie evidence of any fact found in that judgment in a proceeding under s 82 - that is to say, in a proceeding for damages. In such a suit it therefore overcomes the application of the hearsay rule in s 59 of the Evidence Act which ordinarily excludes evidence of representations as to facts contained in a judgment where tendered to prove their truth (although not when tendered to prove the fact of the judgment). Depending on the litigation involved, this is potentially a significant provision. For example, in this case if this Court were to conclude that there was a market for the distribution of apps on Android devices and that Google Asia entered into the DDA with the purpose of substantially lessening competition in that market, s 83 would relieve any party making a similar allegation against Google Asia from having to prove that fact in chief. It would be sufficient to tender a sealed copy of the judgment containing the finding of fact. The relevance of s 83 in circumstances such as the present may fluctuate. For example, in a small claim under Pt IV in which the only affected parties are those before the Court it may tend to be of little significance. Examples of cases of that kind are afforded by the decision of this Court in Nicola v Ideal Image Development Corp Inc [2009] FCA 1177; 215 FCR 76 ('Nicola') and Jagot J's decision in Casaceli v Natuzzi S.p.A [2012] FCA 691; 292 ALR 143 ('Casaceli'). On the other hand, in a case such as the present, the possibility of a judgment in this action being used as a platform by developers and consumers in class actions against Google is real. 150 If this proceeding is stayed and Epic is forced to litigate in Santa Clara County, any judgment which ensues from a court in that place will not enliven s 83. Consequently, requiring Epic to litigate there prevents s 83 from operating in this case. Unlike Nicola and Casaceli, therefore, it is apparent that any stay of this proceeding will have a significant impact on the operation of s 83. In the Full Court Apple Decision, the Full Court considered this to be a significant matter tending against the grant of a stay: see [108]. I do not think that its weight is any different in this case.

Section 87(1A) 151 Section 87(1) and (1A)(b), (c) and (d) provide: 87 Other orders (1) Without limiting the generality of section 80, where, in a proceeding instituted under this Part, or for an offence against section 45AF or 45AG or subsection 56BN(1) or 56CC(1), the Court finds that a person who is a party to the proceeding has suffered, or is likely to suffer, loss or damage by conduct of another person that was engaged in (whether before or after the commencement of this subsection) in contravention of a provision of Part IV or Division 2 of Part IVB, or of section 55B, subsection 56BO(1) or 56BU(1), section 56CD, 60C or 60K or a civil penalty provision of the consumer data rules, the Court may, whether or not it grants an injunction under section 80 or makes an order under section 82, 86C, 86D or 86E, make such order or orders as it thinks appropriate against the person who engaged in the conduct or a person who was involved in the contravention (including all or any of the orders mentioned in subsection (2) of this section) if the Court considers that the order or orders concerned will compensate the first-mentioned person in whole or in part for the loss or damage or will prevent or reduce the loss or damage. (1A) Without limiting the generality of sections 51ADB and 80, the Court may: … (b) on the application of the Commission in accordance with subsection (1B) on behalf of one or more persons who have suffered, or who are likely to suffer, loss or damage by conduct of another person that was engaged in in contravention of Part IV (other than section 45D or 45E), Division 2 of Part IVB, subsection 56BN(1), 56BO(1), 56BU(1) or 56CC(1), section 56CD, 60C or 60K or a civil penalty provision of the consumer data rules; or … (c) compensate the person who made the application, or the person or any of the persons on whose behalf the application was made, in whole or in part for the loss or damage; or (d) prevent or reduce the loss or damage suffered, or likely to be suffered, by such a person. 152 The effect of these provisions is that if this Court delivers a judgment in this case against Google finding that it contravened Pt IV of the CCA then the Australian Competition and Consumer Commission ('the Commission') may apply on behalf of persons who have suffered damage as a result of that contravention to compensate them. In effect, this gives rise to the same issues as s 83 does. I therefore accept that a stay of this proceeding will prevent s 83 being applicable, and that in this case, this is a significant matter weighing against a stay.

Other Provisions 153 Epic also submitted that if the proceeding were heard in Santa Clara County ss 84 and 87CA of the CCA would be frustrated.

Section 84 154 Section 84 provides: 84 Conduct by directors, employees or agents (1) If, in: … (b) a proceeding under this Part in respect of conduct engaged in by a body corporate, being conduct in relation to which section 45AJ, 45AK, 46 or 46A, Part IVB, section 55B, Part V, subsection 56BN(1), 56BO(1), 56BU(1) or 56CC(1), section 56CD or a civil penalty provision of the consumer data rules applies; … it is necessary to establish the state of mind of the body corporate, it is sufficient to show that: (c) a director, employee or agent of the body corporate engaged in that conduct; and (d) the director, employee or agent was, in engaging in that conduct, acting within the scope of his or her actual or apparent authority; and (e) the director, employee or agent had that state of mind. (2) Any conduct engaged in on behalf of a body corporate: (a) by a director, employee or agent of the body corporate within the scope of the person's actual or apparent authority; or (b) by any other person at the direction or with the consent or agreement (whether express or implied) of a director, employee or agent of the body corporate, where the giving of the direction, consent or agreement is within the scope of the actual or apparent authority of the director, employee or agent; shall be deemed, for the purposes of this Act and the consumer data rules, to have been engaged in also by the body corporate. 155 This provision is only relevant to Epic's claims under s 46. Section 46 prohibits a corporation with a substantial degree of market power from engaging in conduct for the purpose of substantially lessening competition in that or certain other markets, or which has or is likely to have the effect of doing so. 156 Two questions arise in relation to s 84. First, assuming that it would not be applied by a court sitting in Santa Clara, would the approach adopted by a court sitting there be burdensome on Epic? Secondly, if so, would a court sitting in Santa Clara in fact apply s 84 to the claim under s 46 on the basis that it was a substantive part of the cause of action or would it not apply it on the basis that it was merely procedural?

Would the non-application of s 84 be burdensome to Epic? 157 Judge Larson was of the opinion that the approach a court in California would take would be more burdensome on Epic than the approach exhibited in s 84. I was not taken to any evidence of Judge Fogel to the contrary and indeed Google did not join issue with this point in its submissions. I conclude therefore that the approach of US courts to proving the state of mind of a corporation would be more burdensome on Epic to the extent that it needed to prove Google's state of mind.

Would a court in Santa Clara apply s 84? 158 Both Judge Fogel and Judge Larson agreed that s 84 would be applied by a court sitting in Santa Clara County if it was a substantive part of the cause of action under s 46 but would not be applied if it was merely procedural. Here, however, the trail runs cold since the parties were successful in excluding from the evidence the views of both judges on the issue of whether s 84 would be treated by a court sitting in Santa Clara County as procedural or substantive. That said, Judge Larson did say that he thought that a party could elect not to rely on s 84 but I do not think that Judge Larson is qualified to express that view. 159 In that circumstance, I will proceed on the basis that the law applying in California is the same as the law applying in Australia and consider the issue myself: see Nielson v Overseas Projects Corporation of Victoria Ltd [2005] HCA 54; 223 CLR 331 ('Neilson') at [125] per Gummow and Hayne JJ, [249] per Callinan J and [275] per Heydon J. 160 Section 84(1) of the CCA is not a provision which is concerned with proof of a body corporate's state of mind generally. Rather, it is directed towards the proof which is required to establish that a body corporate held a particular state of mind in certain, select proceedings (those which are identified in s 84(1)(a)-(ba)). Relevantly, these select proceedings include those where a body corporate is alleged to have engaged in conduct which contravenes s 46 such as in the present case. If a question about the state of mind of a corporation arises outside of the circumstances identified in s 84(1)(a)-(ba), therefore, then the rule erected by it does not apply. For example, if in a case under s 46 it became necessary to prove that a corporation had a particular state of mind but that state of mind did not relate to the conduct which was said to impugn s 46 then s 84(1) would not be engaged. This suggests that s 84 is not a procedural rule about proof of states of mind but a substantive rule about corporate liability. This is consistent with the observation of Morling J in Trade Practices Commission v Queensland Aggregates Pty Ltd (No 3) (1982) 44 ALR 391 at 404 that the legislative intent underlying s 84(1) was 'to extend, rather than limit, the liability of corporations for the actions of others'. It is also consistent with the fact that s 84(1) ties the rule that it establishes back into specified causes of action. 161 In John Pfeiffer Pty Ltd v Rogerson [2000] HCA 36; 203 CLR 503 ('Pfeiffer') the High Court held that matters which affect the existence, extent or enforceability of rights or duties of parties are matters which on their face are substantive: at [99] per Gleeson CJ, Gaudron, McHugh, Gummow and Hayne JJ. In McKain v R W Miller & Co (SA) Pty Ltd (1991) 174 CLR 1, Mason CJ suggested a dichotomy between rules which are directed to governing or regulating the mode or conduct of court proceedings (which were to be seen as procedural) and everything else (which was to be seen as substantive): at 26-27. That observation was referred to by the Court in Pfeiffer. 162 Relevant too is the fact that s 84(1) does not supplant the common law of corporate liability but is by way of supplement. This matters because the pre-existing law of corporate liabilities was not, I think, part of the law of evidence but rather part of the substantive law governing corporations. On the other hand, the Full Court in the Full Court Apple Decision described s 84(1) as 'adjectival' (at [61]), nomenclature adopted also by Mr Hutley upon the hearing of the present application (T18.19), which would ordinarily mean procedural although this statement was not the result of any analysis by the Full Court and was not I think intended to be definitive. So too it seems to me that a party does not have to rely upon s 84(1) if it does not wish to do so, which suggests that it may be waived (although a contrary view would be that it is not waived, merely not utilised). 163 On balance, I would prefer the view that s 84(1) is substantive although I think that the question is a difficult one on which minds might legitimately differ. 164 I conclude therefore that s 84(1) is a substantive part of s 46. On the basis of Nielson, I conclude that a Court sitting in Santa Clara County would apply s 84(1) to any proceeding before it brought under s 46. I do not accept therefore that requiring Epic to litigate there causes it to lose any juridical advantage in relation to s 84(1). 165 If I were wrong in that conclusion, however, I was not taken to any concrete example whereby the differences between the US position and s 84(1) might have any practical significance. Whilst the evidence shows that the US position is more burdensome, I was not taken to any aspect of the case where the effect of this burden upon Epic might be usefully assessed (for example, by way of examination of how the purpose cases under ss 45 and 46 are intended to operate).

Section 87CA 166 Section 87CA provides: 87CA Intervention by Commission (1) The Commission may, with the leave of the Court and subject to any conditions imposed by the Court, intervene in any proceeding instituted under this Act. (2) If the Commission intervenes in a proceeding, the Commission is taken to be a party to the proceeding and has all the rights, duties and liabilities of such a party. 167 This provision authorises the Commission to intervene with the leave of the Court whereupon it becomes a party. The most significant aspect of being a party is the right to lead evidence and to appeal. Undoubtedly, the Commission may seek to do that in this case. Indeed, in the Epic v Apple litigation, the Commission sought leave to intervene in the appeal and was granted that leave. I do not therefore regard the prospect of the Commission's intervention in this litigation as unlikely although it has not yet sought to do so. 168 Section 87CA will not apply in any proceeding in Santa Clara County since it is a procedural rule. Google submitted that in any such proceeding the Commission could seek leave to file an amicus brief. This submission was based on the evidence of Judge Fogel. In his report in chief he expressed the view that an agency of the Australian government such as the Commission could seek leave to file an amicus brief in a US court. Whilst it was true that such briefs could not be filed as of right, Judge Fogel thought that applications to file such briefs by foreign governments were often granted: see First Report of Judge Fogel at s108. I would note that the Commission cannot intervene as of right under s 87(1C) either. 169 Judge Larson thought that amicus briefs were often received from foreign governments on the meaning of foreign law and he thought it likely that the Australian government (and I would infer the Commission) would be permitted to file an amicus brief in this proceeding if it were pursued in a court sitting in Santa Clara County. This amicus brief could express the Commission's views on the merits of the parties' arguments. On the other hand, I also accept Judge Larson's evidence that the role of an amicus is more limited than that of an intervener (which appears to mirror the position here). In particular, the amicus is confined to the record between the parties and to matters of practice and may not appeal. 170 In that circumstance, I conclude that if the proceeding were pursued in California, the Commission would likely be permitted to file an amicus brief but it would not be permitted to take an active role in the proceeding as a party. It would also have no right of appeal. It is unlikely to be the case therefore that if the proceeding is pursued in California that the Commission will be denied an opportunity to participate in the proceeding although I accept that its right to take an active role in the proceedings will be curtailed. 171 On the other hand, given the nature of the parties in this litigation, I doubt whether the Commission would intervene in this Court for the purpose of giving evidence or supplementing the record. A much more likely scenario is that the Commission will intervene at trial (if it intervenes at trial at all) to make submissions about the operation of Pt IV. Further, whilst it is true in that circumstance that the Commission could appeal in this proceeding as a party, it seems to me unlikely that it would do so. This is because the losing party to this litigation will almost certainly appeal and it would be easier for the Commission simply to seek to intervene in any appeal. Thus the fact that it could not appeal in a Californian proceeding seems to be a minor matter given that it would very likely be given leave to file an amicus brief on the appeal. The position in the United States and the practical realities in this Court do not appear much to differ.

The Remaining Issues 172 There are a number of remaining matters which were raised which do not fit tidily under the headings above. 173 First, Epic submitted that if the matter proceeded in California it would take place through the lens of expert evidence and possibly before a jury. Both Judge Fogel and Judge Larson agreed that questions of foreign law would be treated as ruling on a question of law (which differs from the Australian position) and that they would be examined de novo on appeal. In the Full Court Apple Decision at [110], the Full Court drew attention to the fact that subtleties in meaning and context could be overlooked or misconstrued. In relation to s 21 of the ACL I accept the submission that what is regarded as unconscionable within the meaning of that provision is likely to be a function of local values and that a foreign court may have some difficulty in identifying its content with precision: cf Dialogue Consulting Pty Ltd v Instagram, Inc [2020] FCA 1846 at [181] per Beach J holding that an assessment of s 21 unconscionability may require 'an assessment of standards in Australia and the value and norms recognised by the ACL'. On the whole, however, whilst I regard this as a matter going against the grant of a stay, it appears to me to be a minor one given the resources and enthusiasm of the parties. 174 Secondly, I accept that if the case proceeds in California it will not contribute to the corpus of Australian competition law and that the High Court will be denied the last word on the topic. The same is true, however, in any area of Australian law which becomes the subject of foreign proceedings. The only thing which may take this case outside that observation is the difficulties of interpretation posed by Pt IV and s 21 together with the wide significance of the case from the perspective of Australian consumers interacting with the Android platform. I regard this as a matter going against the grant of a stay. 175 Thirdly, it was submitted that any determination by a court sitting in California would lack the sanction of this Court. No doubt this is true where sanction is understood to mean the expression of this Court's opinion (as opposed to the imposition by it of a sanction). I regard this as a minor matter. 176 Fourthly, picking up on what was said in the Full Court Apple Decision at [104], it was submitted that this Court has specialist competition judges and this advantage would be lost if the proceedings were heard in California. Whilst there is force in Google's submission that there is no reason to think that competition judges in California are any less expert in competition law (and perhaps some reason to think that they are more expert), I do not think that I should depart from the Full Court's view. I therefore accept that this is a matter which favours the refusal of a stay. 177 Fifthly, it was said that s 86(4) evinced an intention that only this Court should hear cases under Pt IV. Section 86(4) provides: (4) The jurisdiction conferred by subsection (1) on the Federal Court is exclusive of the jurisdiction of any other court other than: (a) the jurisdiction of the Federal Circuit Court under subsection (1A); and (b) the jurisdiction of the several courts of the States and Territories under subsection (2); and (ba) the jurisdiction of the Supreme Courts of the States under subsection (3A); and (bb) the jurisdiction of the Supreme Courts of the Territories under subsection (3B); and (c) the jurisdiction of the High Court under section 75 of the Constitution. 178 I rejected the same argument in the Apple Decision holding that s 86(4) was concerned with the distribution of jurisdiction and was unlikely to speak to the position of foreign courts. This conclusion was not interfered with by the Full Court. Since then Stewart J has reached the same conclusion that I did in Karpik v Carnival plc (The Ruby Princess) (Stay Application) [2021] FCA 1082. I do not propose to depart from what I said in the Apple Decision. Accordingly, I do not think this is a matter which is relevant to the question of whether a stay should be granted. 179 Sixthly, it was submitted that if the stay were granted it would disrupt the MDL Action by causing the adjournment of that proceeding from September 2022 to some later date. I have already indicated that I accept that this is likely. I accept that the disruption of that proceeding is a matter which militates against the grant of a stay (although I also note that the risk of fragmentation if that course is not taken militates in the opposite direction). 180 Seventhly, I accept that Epic's proceeding raises a question of considerable public interest. Epic's allegations concern the way in which Google operates its Android platform in Australia and a large number of Australian residents are affected by the outcome of the litigation. Epic's suit does not just concern its own position but explicitly relies on the conduct of Google towards all Australian developers of apps on Android devices. It also raises the question of whether the use of forum selection clauses by Google (and hence other such entities) may itself be a contravention of Pt IV. Further, the markets involved are important markets in the Australian economy and the result of the litigation will have a significant impact in the Australian economy. There are therefore powerful policy reasons militating towards the conclusion that this suit should be decided in Australia. 181 Eighthly, I accept that if the matter proceeds in California there is a risk that it will be tried by a jury. Google has already sought to requisition a jury in the MDL Action: Exhibit 4. To overcome that difficulty, Google proffered an undertaking to this Court (Exhibit 5) not to seek a jury trial and to agree jointly with Epic to submit the issues to a bench trial. To my mind, this creates the fragmentation risk to which I have already adverted. As Google would have it, there would now be a trial of the issues in the MDL Action by jury but this proceeding would be tried by judge alone. The same problem arises of conflicting fact finding by the judge and the jury. 182 Ninthly, Epic submits that it was accepted in the Full Court Apple Decision at [99] that Pt IV exhibits a legislative policy that claims pursuant to Pt IV should be determined in Australia: T31.43-32.11. I am bound by the Full Court's reasoning and therefore proceed on the basis that Pt IV does exhibit such a legislative policy, although I would record my respectful dissent from the Full Court's conclusion on this aspect of the matter. 183 Tenthly, I do not accept Epic's submission that the presence of a non-exclusive jurisdiction clause in the GPAL terms of service assists it. Clause 11.3 provides: 11.3 Jurisdiction; Governing Law: These Terms of Service shall be governed by the laws of California, except for California's choice of law rules, and applicable federal United States laws. Each party agrees to submit to personal and non-exclusive jurisdiction of the courts located in Santa Clara County, California. The parties specifically exclude from application to the Terms of Service the United Nations Convention on Contracts for the International Sale of Goods and the Uniform Computer Information Transactions Act. 184 There was no evidence from Judge Fogel or Judge Larson as to whether the present dispute would be governed by this clause which is materially different to cl 16.8. It is therefore necessary to apply Nielson. Epic assumed in its submissions that cl 11.3 would apply to this dispute although it proffered no reasons for that assumption. Google made no submissions about it. The difficulty is that cl 11.3 does not identify the disputes to which it applies. In the absence of any submissions as to how this problem is to be overcome I do not accept that I should conclude that cl 11.3 does in fact apply to this dispute. Further, even if cl 11.3 does apply as a matter of Californian law, there is left unanswered a problem which then arises under that law: how the inconsistency between cl 11.3 and cl 16.8 should be resolved. Neither Judge Fogel nor Judge Larson gave evidence on that question and the answer is not obvious. In those circumstances, I do not accept that cl 11.3 assists Epic.

Whether a Stay Should be Granted 185 In the Apple Decision, Epic International was a party to the agreement containing the exclusive jurisdiction clause. That is not so in this case. Further, as I have explained above, Epic International is not bound, as Google submitted it should be, by cl 16.8 on the basis of the Manetti-Farrow rule. In that circumstance, it is not appropriate to start with a prima facie disposition in favour of a stay of proceedings: Australian Health at [90] as cited in the Full Court Apple Decision at [69]. 186 The factors tending against the grant of the stay are these. First, it is unclear whether a court sitting in Santa Clara County would in fact apply Pt IV of the CCA and s 21 of the ACL to the dispute. Whilst it is true that Google has proffered an undertaking to the Court not to rely upon cl 16.8 I am not satisfied that this undertaking is sufficient to achieve that outcome. The failure of Google to establish that these mandatory laws of the forum would be applied by the foreign court is a significant matter: Full Court Apple Decision at [87], citing Akai at 445 per Toohey, Gaudron and Gummow JJ. 187 Secondly, the ability to obtain an injunction in California to restrain the contraventions Epic alleges Google has engaged in will be subject to the stricter requirement that Epic show that it will suffer irremediable harm if the injunction is not granted so that damages will not be an adequate remedy. Under s 80 it faces no such requirement. Although I think that the requirements of both legal systems will probably be met in the case of the injunction by which Epic seeks to have Fortnite restored to the Google Play Store with its own in-app payment feature, I am not satisfied that this is so in relation to the injunctions which Epic seeks to prevent the other contraventions of which it complains. 188 Thirdly, Epic will not able to revise the agreements it impugns under s 87(2) of the CCA. The equitable remedies available to courts sitting in Santa Clara County of reforming the contract and severing unconscionable terms are far more limited. 189 Fourthly, if the case proceeds in California, neither consumers (under s 83) nor the Commission (under s 87(1A)) will be able to rely upon the findings of any judgment as prima facie evidence whereas if the proceeding continues here they will be able to do so. The Full Court has emphasised the public interest importance of this consideration: see the Full Court Apple Decision at [60]. 190 Fifthly, if the case proceeds in California, the legal issues will likely be tried through the lens of expert evidence. However, as I have explained above I regard this as a matter only marginally favouring the refusal of a stay. 191 Sixthly, if the case proceeds in California, it will not be heard by this Court's specialist competition judges: Full Court Apple Decision at [104]. I propose to treat it as a matter marginally favouring the refusal of the stay. 192 Seventhly, if the case proceeds in California, this Court's competition law jurisprudence will not be developed. Given the subject matter of this case and its wide import for Australian consumers who use the Android platform I regard this as a matter favouring the refusal of a stay in this case although I would generally doubt its significance in everyday Pt IV litigation. 193 Eighthly, if the case proceeds in California, any decision will not have the imprimatur of this Court. I regard this as a minor matter favouring the refusal of a stay. 194 Ninthly, there are powerful public policy reasons why the suit should be heard in Australia (explained above) including the legislative policy referred to by the Full Court in the Full Court Apple Decision at [99]. 195 There is only one matter which favours the grant of a stay which is the fact that its refusal will lead to the fragmentation of the proceeding. However, the fragmentation issue is complex. Since I have concluded that cl 16.8 does not justify a stay of Epic International's suit, granting a stay of Epic Games suit would not, on the face of it, reduce the fragmentation. Indeed, it would increase it. There would then be a risk of inconsistent fact finding between this Court's findings in the Epic International suit, the factual findings of the judge conducting the bench trial of the Epic Games suit and the jury's findings of fact in the MDL Action. 196 That risk would be reduced if I then stayed the Epic International suit on discretionary grounds. But there would then still be the risk of inconsistent findings by the judge hearing this suit and the jury hearing the MDL Action. I do not think therefore that staying this proceeding reduces the risk of inconsistent fact finding in light of Google's undertaking to seek a bench trial of this matter regardless of whether I stay the Epic International suit. I do accept, however, that there will be a cost saving in that course (related to travel and the retention of two sets of lawyers) and that the witnesses will not have to give their evidence twice. That is a matter which favours the grant of a stay to which I propose to give weight. However, it lacks the usual weight that the fragmentation argument has since it must be adjusted for the fact that the risk of inconsistent findings will still exist. 197 There are a number of matters which I regard as neutral. Whilst the Commission will not be able to intervene by leave in any case in Santa Clara County, I am satisfied that it will receive a grant of leave to file an amicus brief. Whilst it will not be able to participate as a party, I think it unlikely that it would substantively participate as a party at trial if the matter remained in this Court even if it were granted leave to intervene. The same is true on appeal. 198 Whilst the question is difficult, on balance, I think that the state of mind provision, s 84(1), will be applied if the case proceeds in California. 199 I do not think that s 86(4) which limits jurisdiction under Pt IV of the CCA to this Court is relevant to the issues at hand. I also regard cl 11.3 of the GPAL terms of service as neutral. 200 Taking each of those matters into account, I do not think that a stay should be granted. In reaching that conclusion I have taken into account each and every matter I have discussed in these reasons. 201 If I am wrong about the significance of Epic International's status as a stranger to cl 16.8 then it will follow that Epic must show strong grounds as to why a stay should not be granted. Viewed through that lens, I am satisfied that the stay should still be refused. Given the emphasis given to the platform provisions in the Full Court Apple Decision those are matters which should be given significant weight. Whilst the refusal of the stay will undoubtedly cause the complex fragmentation problem I have previously described, I am bound to regard the loss of the platform provisions as being a graver peril in this case. Again, I take into account all of the matters to which I have referred in these reasons in reaching that conclusion including those referred to in the Full Court Apple Decision. 202 On either view, the stay application should be refused.

The Challenge to Clause 16.8 203 In its further amended concise statement Epic directly challenges the validity of cl 16.8. It says that the choice of law clause and the exclusive jurisdiction clause are part of a series of restraints by which Google, in contravention of s 46, has misused its market power. Epic's case is not therefore that the DDA is invalid because of contraventions of Pt IV and therefore that incidentally, cl 16.8 is invalid as a collateral consequence. Rather, it is said that cl 16.8 is itself part of the way in which Google has contravened s 46. 204 I do not regard that case as trivial. Epic says that Google has made it difficult for developers to do anything but use the Google Play Store and to submit themselves to the 30% commission. One of the ways it says that it has done that is by making developers litigate in California under Californian law. I regard it as sufficiently arguable that the practical impact that has on many Australian developers will be to secure their obedience to the other restraints in the DDA. It is not to the point in such an analysis that the clause may be invalid under Pt IV (as Google now contends against its own agreement) because the argument takes as its point of departure that many developers will be deterred from making such a challenge by cl 16.8. On this view, the competitive impact of the restraint emerges not from its legal efficacy but from its existence and status as a deterrent. 205 Likewise, there is no substance in Google's contention that Epic has not challenged the exclusive jurisdiction clause directly. Here the point was that the Further Amended Concise Statement challenged the choice of law clause and the exclusive jurisdiction clause together. Since they are both part of the same clause this is an underwhelming submission. In any event, even if there were anything in the argument I do not think that the exclusive jurisdiction clause can be fully understood without the choice of law clause by which it is accompanied. This is not just a clause requiring developers to litigate in California. It is a clause requiring developers to litigate in California under the substantive law applying in California which on the face of it excludes Australian law. It would be unrealistic to require these clauses to be challenged separately. 206 In any event, I am satisfied that Epic has challenged cl 16.8 in a sufficiently direct and substantial fashion that it would be inappropriate to enforce cl 16.8 pending the trial of this case. This was the conclusion reached by Byrne J in Commonwealth Bank of Australia v White [1999] VSC 262; 2 VR 681 ('White') e.g. at [11], [89]. Google submitted that White was distinguished in Telesto Investments Limited v UBS AG [2013] NSWSC 503; 94 ACSR 29 ('Telesto'). However, in Telesto there was no direct challenge to the exclusive jurisdiction clause and at [278] Sackar J cited White for the proposition that it may be a factor militating against the stay that the plaintiff attacks the exclusive jurisdiction clause itself but it is not enough for the plaintiff merely to attack the validity of the agreement containing it. In any event, this case appears to me to be on all fours with White. 207 Even if I had not refused the stay on discretionary grounds I would therefore have refused it on the basis that there was a distinct and direct challenge to the exclusive jurisdiction clause. 208 I have not found it necessary in that circumstance to consider Epic's parallel argument against cl 16.8 based upon s 21 of the ACL.

Forum Non Conveniens 209 I reject this application for the reasons I gave in the Apple Decision at [65].

Epic's Amendment Application 210 During the course of argument, Epic submitted that one reason for refusing any stay was that it alleged that cl 16.8 was both part of the anti-competitive behaviour at which its Pt IV case was aimed and part of the unconscionable conduct at which its claim under s 21 was aimed. It submitted that where there is an independent challenge to the validity of an exclusive jurisdiction clause it cannot be relied upon in support of a stay. I have dealt with that submission above. 211 In partial response to that submission, Google submitted that the Amended Originating Application and Amended Concise Statement did not in fact challenge the choice of law clause but had only challenged the exclusive jurisdiction clause. It was also said that the amended originating application did not challenge cl 16.8 on the grounds of unconscionability under s 21. 212 I accept the correctness of both of these criticisms as did Epic who then propounded the Further Amended Originating Application and Further Amended Concise Statement that are the subject of the present amendment application. 213 I did not apprehend that Google opposed the amendment application on the basis that the two problems to which it had adverted had not been resolved. Instead, it raised four reasons why the amendment should be rejected. 214 First, it was said that the explanation proffered as to why the amendments were being made at this time was insufficient. Mr Poddar, who gave affidavit evidence on the amendment application, said that the current form of the pleadings had been the result of efforts on his part to achieve concision. Google says that Mr Poddar does not say that he made a mistake or a slip in the earlier version of the pleading. No doubt a party seeking the exercise of an indulgence such as an amendment needs to proffer an explanation as to why the indulgence is needed. But the nature of the explanation needed is a function of the situation and the circumstances: see Tamaya Resources Limited (in liq) v Deloitte Touche Tohmatsu (A Firm) [2016] FCAFC 2; 332 ALR 199 at [154] per Gilmour, Perram and Beach JJ; Cement Australia Pty Ltd v Australian Competition and Consumer Commission [2010] FCAFC 101; 187 FCR 261 at [51] per Keane CJ, Gilmour and Logan JJ. Here Google has not put on a defence, there have been no interlocutory steps apart from the present applications and Epic's ex parte application to serve its proceeding on Google out of the jurisdiction and no trial date has been set. Mr Poddar's explanation more than suffices given that context. 215 Secondly, it was said that the proposed amendments relied upon White at [89] which I have discussed above. Google says that that decision was distinguished by Sackar J in Telesto at [280] where is his Honour said that 'the authorities favour the approach that something more is need that the mere inclusion of an exclusive jurisdiction clause which have the effect of displacing the application of the misleading and deceptive conduct laws'. I have rejected this argument above. Whilst it is true that Sackar J did distinguish White he did not do so in a way which is of any assistance to Google. 216 Thirdly, Google submitted that the effect of the choice of law clause and the forum selection clause was not as Epic now alleged. It was said that because Pt IV and s 21 are mandatory laws of the forum, it was common ground that the forum selection clause could not be enforced if the effect was to prevent those claims being pursued. It therefore followed that the clause could not have the anti-competitive effect for which Epic contends. I do not accept this argument. Epic says that Google imposed a series of restraints upon developers (not just Epic): the 30% commission requirement, the requirement not to distribute apps outside the Google Play Store, the requirement not distribute any product which might facilitate the distribution of apps outside the Google Play Store, and the requirement that any dispute be resolved in California under the substantive law applying in that place. It is said that this was a misuse of market power by Google contrary to s 46. The evident import of the allegation is that cl 16.8 is part of the mechanism by which Google has misused its market power. The effect of requiring developers to litigate under Californian law in California is part of the machinery of that abuse. In the context of that argument the contention that the provision could not have that effect because it is not enforceable misses the point about the clause. That point is not concerned with whether the clause is valid but rather with whether, when imposed on the class of all Australian developers, it has or is likely to have the real world restraining effect for which Epic now contends, or was done with the purpose of bringing about such a result. Assuming in Google's favour that the clause is invalid that does not diminish the real world effect the clause has on those developers who unlike Epic do not have the mettle to challenge Google. I reject the argument. 217 Fourthly, it was submitted that even if the combined effect of the forum selection clause and the choice of law clause was such as to amount to a contravention of s 46, the validity of the forum selection clause would not be impugned. Here the point was that cl 16.8 contained two clauses - the choice of law clause and the forum selection clause - and each had to be distinctly challenged. This was because an exclusive jurisdiction clause was to be regarded as a separate or severable contract from the balance of the contract. I do not accept this argument. In my view the choice of law clause and the forum selection clause are to be considered together for two reasons. First, the full effect of the forum selection clause can only be appreciated when it is read with the choice of law clause. Secondly, as a matter of textual reality there is only one clause, cl 16.8. The challenge to cl 16.8 under s 46 is direct and explicit. Epic says that forcing developers to litigate in California under Californian law is a misuse of Google's market power. This is not a case like Telesto where the challenge was to the contract containing an exclusive jurisdiction clause but no direct challenge to the clause itself. It is a case like White where the challenge is to the clause itself. 218 Finally, it was said that the principle of finality in litigation was relevant. Here the point was that the argument on the stay application had been argued on a particular basis and therefore that the principle of finality prevented an amendment to the pleadings on which the argument had been conducted. If there had been some kind of determination this argument might fall for consideration. However, there has not and I reject the argument. 219 For those reasons, I propose to allow the amendments.

Conclusions 220 The Respondents' stay application will be dismissed with costs. Epic's application for leave further to amend its Amended Originating Application and Amended Concise Statement is granted. In my view, Google's position on that application was sufficiently unreasonable to justify a departure from the ordinary approach to the question of costs on an amendment application. I will order Google to pay Epic's costs of the amendment application although not on an indemnity basis. However, if Google wishes to discharge that order it may file a written submission not exceeding 1,000 words within 7 days which I will consider before troubling Epic. I certify that the preceding two hundred and twenty (220) numbered paragraphs are a true copy of the Reasons for Judgment of the Honourable Justice Perram.