Skip to content
laws.sgby aturio
Grounds publishedDecided 29 Jun 2026
Contents

Singapore International Commercial Court

Kupetz, Jonathan and others v Terraform Labs Pte Ltd and others

Feedback & support

[2026] SGHC(I) 9Case no. Originating Application No 3 of 2024

What happened, in plain English

There is no plain-English summary of this judgment yet.

Ask about this judgment

Grounds of decision

Official text from eLitigation

  1. Introduction

  2. [1] SIC/OA 3/2024 (“OA 3”) is a representative action, involving 275 claimants suing three defendants, Terraform Labs Pte Ltd (“Terraform”), Mr Kwon Do Hyeong (“Mr Kwon”), and Luna Foundation Guard Ltd (“LFG”) (collectively, the “defendants”) for losses arising from fraudulent misrepresentations made in connection with TerraUSD (“UST”), an algorithmic “stablecoin” that collapsed in May 2022. By order of the court, OA 3 was bifurcated, in the sense that issues common to the representative and represented claimants would be heard in a first tranche to be followed by several tranches dealing with issues specific to the represented claimants in these proceedings.

  3. [2] The first tranche of this civil trial, heard before me in May 2025, served two purposes. It determined issues common to all claimants including the content and scope of the actionable representations, the applicable legal principles for fraudulent misrepresentation, and the framework for the assessment of damages. It also determined the individual claims of the ten representative claimants on questions of reliance, causation and quantum – findings which bind only those ten representative claimants. Following the first tranche, the representative claimants appealed against my decision on, inter alia, the cut-off price and the quantum awarded to four of the representative claimants. The Court of Appeal allowed the appeal in part. It revised the cut-off price from US$0.8011 to US$0.60485 per UST token (“Cut-Off Price”), being the average of the opening and closing prices of UST on 12 May 2022, and revised upwards the damages awarded to the four representative claimants: Kupetz, Jonathan v Terraform Labs Pte Ltd [2026] SGCA(I) 1 (“CA Judgment”) at [148]–[149].

  4. [3] Because questions of ownership, reliance, causation and quantum necessarily turn on the particular evidence of each claimant, the remaining represented claimants must each establish these elements individually in subsequent tranches of OA 3. This judgment determines the claims of the 40 represented claimants who gave evidence before me in the second tranche from 9 to 11 March, 20 to 29 April and 18 to 21 May 2026. Mr Kwon did not appear and opted not to be represented by his lawyers at the hearing of the second tranche.

  5. Applicable principles and prior findings

  6. [4] The factual and legal background to this dispute has been set out in the first tranche judgment of Beltran, Julian Moreno v Terraform Labs Pte Ltd [2025] SGHC(I) 17 (“First Tranche Judgment”) and in the CA Judgment. I do not repeat that background here. For present purposes, I summarise the principles and findings from those judgments that govern the determination of the claimants’ cases.

  7. [5] Before the first tranche trial, the defendants conceded that the following representations were false and had been made fraudulently (collectively, “Conceded Representations”):

  8. (a) that UST was stable by design as it was pegged to a stable fiat currency (“First Representation”);

  9. (b) that the Terra protocol and its underlying token mechanics would be able to maintain the price stability of UST regardless of market size, volatility or demand through an algorithm that would allow an arbitrage process to happen with UST’s sister token, LUNA, guaranteeing that UST would always return to the 1 US dollar peg (“Second Representation”);

  10. (c) that UST holders would always be able to exchange 1 UST for 1 US dollar’s worth of LUNA on the Terra protocol (“Third Representation”);

  11. (d) that Anchor was a principal-guaranteed stablecoin savings product where UST holders could enjoy the stability of holding a stablecoin while earning passive income (“Fourth Representation”); and

  12. (e) that LFG had raised USD 1 billion through the sale of LUNA tokens to establish a reserve fund to maintain the UST peg to the US dollar (“Sixth Representation”).

  13. [6] The First to Third Representations (“Terra Representations”) were made in a white paper authored by Mr Kwon and others (“Terra White Paper”) and published on Terraform’s website (“Terra Website”): First Tranche Judgment at [11]–[20]. Importantly, the claimants pleaded that they derived the Third Representation solely from the Terra Website (specifically from the “About the Terra Protocol” section) as opposed to the Terra White Paper: First Tranche Judgment at [19]. The Fourth Representation was made in a White Paper (“Anchor White Paper”) published on the website of the Anchor Protocol (“Anchor Website”). The Anchor Protocol was a lending and borrowing platform where users could deposit or “stake” their UST: First Tranche Judgment at [9] and [21]–[23]. The Sixth Representation was made in a press release issued by LFG (“LFG Press release”): First Tranche Judgment at [27]–[31].

  14. [7] For completeness, I note that the Fifth Representation (that purchasers of UST would earn up to 20% annualised percentage yield on Anchor if they staked their UST on the protocol) and the Seventh Representation (that the defendants were deploying more capital, had a plan to rescue the US dollar peg, and that UST and the Terra ecosystem would return to form) were not conceded by the defendants. I found both not to be actionable as misrepresentations: First Tranche Judgment at [71]–[85]. There was no appeal against that finding.

  15. [8] Given the defendants’ concessions, each claimant must show that they relied on one or more of the Conceded Representations and that such reliance has caused them to suffer loss. More specifically, a claimant must prove that they: (a) read one or more pleaded sources of the Conceded Representations, (b) understood the representations that they read to bear the meanings pleaded, and (c) purchased UST with those representations present in their mind, even if such representations were not the sole or dominant cause of their loss: First Tranche Judgment at [92]–[98]. A claimant need not have acted in an objectively reasonable fashion in understanding a representation or entering into a transaction based on it. Thus, it would be no defence to a claim that a representee, as a knowledgeable or experienced investor, ought to have known better: First Tranche Judgment at [96], citing Panatron Pte Ltd v Lee Cheow Lee [2001] 2 SLR(R) 435 (“Panatron”) at [24].

  16. [9] Damages are assessed on the reliance measure, the principle being to restore each claimant to the position they would have been in had the misrepresentation not been made. The quantum is the total value of a claimant’s UST purchases less the value of any benefits received, including sales proceeds and yields. It is necessary to fix a cut-off time for the purpose of assessing damages. The cut-off would be the point at which the misrepresentations relied upon would have ceased to be operative, so that thereafter any further loss cannot be causally attributable to the defendants. Any UST still held at the cut-off time is therefore treated as having been notionally sold at that point: CA Judgment at [30]–[33].

  17. [10] The cut-off time of 12.01am UTC on 12 May 2022 (“Cut-Off Time”) established in the First Tranche Judgment binds the representative claimants, but does not invariably bind the represented claimants. The Court of Appeal clarified that the cut-off time may differ based on each represented claimant’s individual circumstances, and each claimant may adduce evidence that their individual circumstances warrant a different cut-off time: CA Judgment at [112]–[126]. In particular, the Court of Appeal noted that if a claimant did not read the 11 May 2022 tweets by the Cut-Off Time, or took all reasonable steps to sell but was unable to do so, further losses may continue to be attributed to the defendants: CA Judgment at [115] and [124]. The default Cut-Off Time nonetheless applies in the absence of countervailing individual evidence.

  18. [11] The Court of Appeal revised the Cut-Off Price from US$0.8011 to US$0.60485 per UST token, being the average of the opening and closing price of UST on 12 May 2022. This revision applies to the determination of all represented claimants’ claims: CA Judgment at [146]–[148] and [176(b)(iv)].

  19. Issues to be determined

  20. [12] For this second tranche of trial, I consider the following issues in respect of each claimant:

  21. (a) whether the claimant owns the cryptocurrency accounts or wallets through which the relevant UST transactions were conducted;

  22. (b) whether the claimant relied on one or more of the Conceded Representations to purchase and hold UST;

  23. (c) what is the appropriate cut-off date and time for the purpose of assessing the claimant's damages, and whether the claimant’s individual circumstances displace the default Cut-Off Date and Time of 12.01am UTC on 12 May 2022; and

  24. (d) what is the quantum of damages to be awarded to the claimant, including where pleaded, whether the claimant is entitled to aggravated damages.

  25. [13] Issue (a) on ownership was raised by the defendants in respect of some claimants in this tranche. As I explained in the First Tranche Judgment, it is for each claimant to decide how to discharge the burden of proving ownership. Whether the evidence adduced by a claimant establishes ownership is assessed on the balance of probabilities: First Tranche Judgment at [98(e)]. There is no single prescribed method. A claimant may establish ownership through API keys, screen recordings of them logging into their accounts, wallet access videos, transaction logs, or some combination of these. Where the defendants have challenged ownership, they have typically not adduced any positive evidence that the relevant accounts or wallets belong to someone else. In the absence of such contrary evidence, each claimant’s sworn testimony, taken together with their supporting documentary evidence, is in my assessment sufficient to establish ownership on the balance of probabilities. I address ownership briefly for each claimant in the sections below and find that all claimants in this tranche have discharged their burden on this issue. Consequently, the principal areas of dispute in this tranche are reliance and, where reliance is established, the applicable cut-off date and time.

  26. Common issues in the second tranche

  27. [14] At the outset, I address issues that recur across the claimants in this tranche. It is more efficient to resolve these common issues at the outset than to repeat the analysis for each represented claimant. The common issues are as follows:

  28. (a) the proper approach to the preparation and evaluation of witness statements in mass claims, arising from the similarity in language across a number of the claimants’ statements;

  29. (b) the applicable principles for assessing reliance on the Conceded Representations, including whether the Anchor Terms of Service (“TOS”) negated reliance, and whether the passages cited in the witness statements are sufficient;

  30. (c) the circumstances that justify a departure from the default Cut-Off Date and Time of 00.01am UTC on 12 May 2022, including the significance of the May 2021 depeg and personal extenuating circumstances;

  31. (d) the calculation of quantum of damages, including the applicable permutation from the Revised Script; and

  32. (e) the treatment of aggravated damages claims.

  33. Similarity in claimants’ witness statements

  34. [15] The first issue concerns similar, and in many instances identical, wording in substantial portions of numerous claimants’ witness statements. The defendants submit that the repeated use of similar or identical phrasing across multiple witness statements indicates that the statements have been aligned or templated, such that each claimant should not be regarded as giving their own recollection of events. The defendants say that this is a serious concern because the affected portions go to the heart of each claimant’s case, namely: (a) their understanding of and reliance on the Conceded Representations; and (b) their holding of UST beyond the Cut-Off Time. The defendants argue that the similarity in wording, coupled with other concerns as to ownership, reliance, and credibility, should lead the court to question the reliability of the claimants’ evidence.

  35. [16] Witness statements before the SICC are supposed to set out a witness’ evidence in writing: O 13 r 8(6)(a) of the SICC Rules 2021 (“SICC Rules”); Supreme Court Registrar’s Circular No 1 of 2025 (“Registrar’s Circular”) at para 13. The fundamental principle underlying that function is that a witness’s evidence must be his or her own: Ernest Ferdinand Perez De La Sala v Compañia De Navegación Palomar, SA [2018] SGCA 16 (“Compañia”) at [136]. The Registrar’s Circular states at para 4:

  36. [17] In applying this principle in practice, it is useful to consider two distinct stages in the preparation of a witness statement: first, the gathering of evidence from the witness; and second, the drafting of the statement itself.

  37. (1) Gathering evidence from the witness

  38. [18] On the gathering of evidence, the Registrar’s Circular provides:

  39. [19] The principles governing the gathering of evidence are clear in ordinary civil proceedings. The question is how they apply in a mass claim or representative action such as the present. With 275 claimants spread across multiple jurisdictions, it may not be practical to conduct individual face-to-face interviews with each claimant. It emerged in cross-examination that the claimants’ solicitors addressed this difficulty by collecting evidence through an open-ended questionnaire circulated to each claimant.

  40. [20] There is no rule or directive in Singapore governing the collection of evidence in representative actions. I have consequently looked to other jurisdictions for guidance in deciding this. The claimants’ counsel drew my attention to the Civil Procedure Rules Practice Direction (“CPR PD”) 19B which governs Group Litigation Orders in England and Wales. CPR PD 19B stipulates at para 14.3 that “the specific facts relating to each claimant on the Group Register may be obtained by the use of a questionnaire”. However, CPR PD 19B is not directly applicable. It concerns the use of questionnaires for the purpose of pleadings, which are the solicitor’s work product and are not required to be in the witness’s own words, rather than for the collection of evidence.

  41. [21] CPR PD 57AC governing trial witness statements in the Business and Property Courts in England and Wales, states at para 3.10 that if a trial witness statement is based upon evidence obtained from the witness “by other means (for example by written answers to a questionnaire or the exchange of emails or other forms of correspondence, or by the witness preparing their own draft statement), the guidance set out in this Appendix should still be followed, so far as possible and only modified as necessary.” This confirms that in the UK, the use of a questionnaire to gather evidence is permissible. But the applicable standards for preparation of witness statements continue to apply.

  42. (2) Drafting of witness statement

  43. [23] On the drafting of witness statements, O 13 r 8 of the SICC Rules provides:

  44. [24] Where a witness is not sufficiently fluent in English, O 13 r 7(1) of the SICC Rules states that a translation must be done:

  45. [25] The Registrar’s Circular elaborates that (at paras 11–17):

  46. [26] Therefore, when drafting witness statements, solicitors must, on the one hand, ensure that the witness statements (1) only contain relevant facts without arguments or inadmissible opinion evidence, (2) have an organised logical structure, and (3) are as concise as possible. On the other hand, witness statements must, as far as possible, be in the witness’s own words. The difficulty that solicitors face in fulfilling those two requirements is that a witness, particularly one without legal training, will often express their evidence in a way that is disorganised and long-winded. They may include substantial irrelevant material. The question therefore becomes how much re-writing and re-organisation a solicitor may perform to assist the court, without crossing the line from permissible editing into impermissible substitution of the solicitor’s words for those of the witness.

  47. [27] In my view, solicitors may re-arrange the witness’s evidence into a logical structure, engage in limited editing for clarity and succinctness, correct grammar, and delete irrelevant material. This may all be done with a view to assist the court in grasping the evidence. What solicitors may not do is cross over from tidying up a witness statement into authorship. A solicitor should not feed words to a witness, substitute different language for that of the witness, use one witness’s statement as a template for another, or present the solicitor’s reconstruction of events as though it were the witness’s own account.

  48. [28] Where evidence has been gathered through a questionnaire, the witness’s responses may be organised and refined, but the substance of what the witness said – and as far as possible, the words actually used by the witness – should be preserved. The questions must not be lifted wholesale into the witness statement. That risks producing an account that reflects the solicitor’s framing of the issues rather than the witness’s own account. The evidence will then be influenced by the questions that the solicitor chooses to ask, rather than around what the witness actually has to say. In other words, the witness statement must be built from the witness’s responses to the questionnaire, not around the solicitor’s questions.

  49. [29] The risk will be significantly reduced where the questionnaire is genuinely open-ended. A neutral open-ended question invites the witness to narrate their own experience in their own terms, and the response will naturally contain (and constrain) the substance of what is to go into the statement. The solicitor’s task will then be one of organisation and refinement rather than reconstruction. The resulting statement will far more likely reflect a witness’s own evidence, in contrast to a statement produced by a questionnaire that asks narrow or leading questions and generates answers that require lifting the question or substantial supplementation and rewriting by the solicitor. The quality of the questionnaire therefore has a direct bearing on the integrity of the evidence and witness statement that flows from it.

  50. [30] Solicitors must also be alert to the risk that the increasing availability of artificial intelligence (“AI”) tools will exacerbate the problem of non-individualised witness statements. There is an emerging body of case law in Singapore addressing the misuse of AI in litigation, particularly where counsel have cited fictitious authorities (so-called “hallucinations”) generated by such tools: see Tajudin bin Gulam Rasul v Suriaya bte Haja Mohideen [2025] 5 SLR 518 at [51]–[59] and Tan Hai Peng Micheal (as the executor of the estate of Tan Thuan Teck, deceased) v Tan Cheong Joo [2025] SGHC 217. Beyond the citation of fictitious authority, AI may pose concerns in the preparation of witness statements. The Guide on the Use of Generative Artificial Intelligence Tools by Court Users (“Generative AI Guide”), which applies to all matters in the Supreme Court, the State Courts and the Family Justice Courts, addresses this at page 3:

  51. [31] The Generative AI Guide therefore permits the use of AI to generate a first-cut draft of a witness statement – for instance, by assembling a claimant’s responses to a questionnaire into a structured document. That is permissible as a matter of form. What is not permissible is to use AI to generate the substance of the evidence itself, or to allow the AI-generated draft to stand, without ensuring that the resulting statement accurately reflects the witness’s own evidence in their own words. The line here is the same as that which applies to solicitor-drafted statements: AI may assist with structure, grammar and form. But the evidence and words must remain that of the witness. Solicitors are under a duty to ensure that each statement filed before this court represents the untainted and independent evidence of the witness signing it. That obligation cannot be discharged by delegating the drafting to an AI tool and asking the witness to confirm the result.

  52. Application to the facts

  53. [32] It emerged in cross-examination that the claimants’ solicitors circulated a questionnaire containing open-ended questions for each claimant to complete and prepared the witness statements on the basis of the responses received. This process likely explains why certain portions of the statements are expressed in similar or even identical terms. Where the nature of the claims is similar and the witnesses’ responses to the same open-ended questions covered similar ground, the solicitors for the claimants appear to have reproduced the same language across the statements rather than preserving each witness’s individual expression.

  54. [33] Despite serious concerns about the drafting approach adopted, I admit the evidence of each claimant and accord weight to them for the following reasons. First, when questioned in cross-examination, each claimant confirmed that the contents of their statement represented their own evidence, even if the language was not entirely their own. Second, the questions in the questionnaire circulated by the claimants’ solicitors were open-ended. Each claimant was afforded an opportunity to explain and stand by their evidence on the witness stand. I am thus satisfied that the evidence given is genuinely that of each claimant. Third, there is, as I have noted above at [17], no specific guidance in Singapore governing the preparation of witness statements in mass claims or representative actions. In these circumstances, it would neither be necessary nor fair to penalise the claimants for the drafting approach adopted by their solicitors. Order 13 r 5(4) of the SICC Rules confers on the court a discretion to refuse to admit a non-compliant witness statement. Implicitly, that discretion permits me to admit evidence where the justice of the case requires, and I exercise my discretion accordingly.

  55. [34] However, I should make clear that the approach employed thus far should not be repeated in future tranches or in other proceedings before this court. The claimants’ solicitors may have found it expedient to reproduce consistent language across the statements where the claimants’ evidence covered similar ground. But convenience is not a justification for departing from the requirement that each statement must be the witness’s own evidence in their own words. The guidance here on the gathering of evidence and the drafting of witness statements should be followed in the preparation of any forthcoming witness statements in this litigation.

  56. Assessment of reliance on the Conceded Representations

  57. [35] Regarding reliance on the Conceded Representations, the defendants raise three recurring arguments across the claimants in this tranche. First, the defendants suggest that the Anchor TOS negated the claimants’ reliance on the Conceded Representations. Second, the defendants submit that reliance on the Conceded Representations as a basis for purchasing UST must be assessed on an objective standard. Specifically, they say that the court must assess whether each claimant objectively understood the Conceded Representations in the sense that they have been admitted to be false. Thus, a claimant whose understanding of the relevant passages was so illogical or unreasonable as to be objectively untenable, cannot be said to have relied on the Conceded Representations. Third, the defendants contend that the court must assess whether each claimant truly arrived at their understanding of the Conceded Representations from the specific passages of the source documents cited in their witness statements. I address each of the recurring arguments in turn.

  58. Anchor Terms of Service

  59. [36] The defendants argue that some claimants, by reason of their professional background or approach to investment, would certainly have clicked on and read the Anchor TOS when connecting their wallets to the Anchor protocol. Having consequently read those terms, those claimants could not have relied, or continued to rely, on the Conceded Representations. In support of their submission, the defendants rely on two clauses. Clause 1 of the Anchor TOS states that terms such as “debt”, “lend”, “borrow”, “yield”, and “invest” are not to be interpreted literally. Accordingly, when users access the Anchor Protocol and enter into transactions using Smart Contracts, there are no legal agreements, promises of payment, or courts of law, and therefore no debts, loans, or other traditional finance transactions involved. Clause 14 provides that the markets for digital assets “are highly volatile due to factors including (but not limited to) adoption, speculation, technology, security, and regulation” and that users “acknowledge the risk that [their] digital assets may lose some or all of their value while they are supplied to the Protocol”. The defendants argue that, in the face of such language, any claimant reading these clauses could not have come away with the understanding that the defendants had guaranteed the UST peg regardless of market circumstances. The claimants therefore could not have relied on the Conceded Representations in the manner pleaded. The defendants clarify that they are not relying on the Anchor TOS to establish an exclusion of liability on a contractual basis.

  60. [37] I reject that argument on two grounds. First, the defendants’ argument, whatever its framing, is in substance an attempt to use the disclaimer language in the Anchor TOS to negate the legal consequences of the defendants’ own fraudulent misrepresentations. That is precisely what an exclusion of liability clause purports to do, and the law does not permit it in the context of fraud. It is well-established that a person cannot contractually exclude liability for their own fraudulent misrepresentation: Mentormorphosis Pty Ltd v Phua Raymond [2010] SGHC 188 at [21], citing S Pearson & Son Ltd v Dublin Corporation [1907] AC 351 at 353 per Lord Loreburn LC, at 356 per the Earl of Halsbury, and at 362 per Lord James of Hereford. As Tay Yong Kwang J explained in Goldring, Timothy Nicholas v Public Prosecutor [2015] SGHC 158 at [61], the reason for that principle is that a clause purporting to exclude liability for fraud is often “part of the machinery that advances and disguises that fraud”. To uphold such a clause “would be inimical to notions of justice”. The Anchor TOS contains standard disclaimer language of the kind commonly found in online platforms dealing in digital assets. To accept the defendants’ argument would be to permit such boilerplate terms to negate reliance on fraudulent misrepresentations, allowing a fraudulent party to escape the consequences of their fraud through the very instrument of that fraud. The law does not permit that.

  61. [38] Second, the contention that certain claimants could not have relied on the Conceded Representations after reading the Anchor TOS applies an objective standard to what is a subjective inquiry. As I held in the First Tranche Judgment, reliance is assessed subjectively – the question is whether the Conceded Representations in fact induced each claimant to act, not whether a reasonable person in their position would have done so: First Tranche Judgment at [96]. The defendants cannot discharge that inquiry by pointing to a claimant’s professional background and asserting that a person of such profile ought not to have relied on the representations after reading the Anchor TOS. That approach substitutes the defendants’ assessment of how a hypothetical reasonable person would have reacted, for the actual evidence of what each individual claimant in fact believed and acted upon.

  62. Unreasonable or illogical understanding of the Conceded Representations

  63. [39] The defendants argue the court must assess whether each claimant could objectively have understood the misrepresentations in the sense that they were conceded to be false. It was submitted that, consequently, a claimant whose understanding of the relevant passages in the Terra White Paper is so unreasonable or illogical as to be objectively untenable cannot be said to have relied on the Conceded Representations. The defendants rely on Chan Pik Sun v Wan Hoe Keet [2024] 1 SLR 893 (“Chan Pik Sun”) at [186] for the proposition that the meaning of a representation is assessed objectively, from the perspective of a reasonable person in the position of the representee.

  64. [40] I do not accept this argument. The applicable standard for reliance is subjective, not objective. As I held in the First Tranche Judgment at [92]–[98], and as affirmed by the Court of Appeal at [113] of the CA Judgment, the question is not what a reasonable person in the claimant’s position would have understood from the Conceded Representations, but whether the Conceded Representations in fact induced a particular claimant to purchase UST. Each claimant need only show that, having seen the relevant representations, they genuinely held the understanding that they describe and acted on it. This is a question to be resolved on the evidence of each individual claimant, not by reference to what a reasonable investor would or would not have understood from the Terra Website, Terra White Paper, or the Anchor White Paper prior to purchasing UST.

  65. [41] The defendants’ reliance on Chan Pik Sun does not advance their case. It is correct that, in determining the meaning of a representation, the court may have regard to how it would reasonably be interpreted – in other words, its objective meaning: John Cartwright, Misrepresentation, Mistake and Non-Disclosure (Sweet & Maxwell, Fifth Ed, 2019) at paras 5–18, citing Arnison v Smith (1889) 41 Ch. D. 348, CA at 368–369. But that objective inquiry goes to the threshold question of whether the words used were capable of bearing the meaning alleged. Once that threshold is crossed – and here it has been crossed, since the defendants have conceded the pleaded meanings and admitted that the Conceded Representations were false – the question of whether each representee actually understood and relied on those representations is a subjective one.

  66. [42] The defendants’ concessions have a further consequence. Since the pleaded meanings of the Conceded Representations and their falsity are no longer in dispute, a claimant establishes reliance by proving that they saw and understood the Conceded Representations to bear the pleaded meanings and that some combination of those representations was present in their mind when they transacted in UST. To insist on an objective standard at this stage of the inquiry, after the defendants have made their concessions, would impose a burden on claimants that is inconsistent with the established law on deceit and would in substance permit the defendants to resile from their admissions through the back door. I reject that approach.

  67. Analysing specific passages to prove reliance

  68. [43] The defendants argue that the court must assess whether each claimant truly arrived at their understanding of the Conceded Representations from the specific passages of the source documents cited in their witness statements. For example, many claimants cite the following passage from para 2.1 of the Terra White Paper in support of the First Representation:

  69. [44] The defendants argue that this passage does not say that UST tokens are “stable by design” as the First Representation requires, and that a claimant who relies solely on this passage cannot establish that they understood and relied on the First Representation as pleaded. In relation to the Third Representation specifically, the defendants argue that the Third Representation is pleaded as derived solely from the “About the Terra Protocol” section of the Terra Website, hence a claimant who cites only the Terra White Paper, without establishing that they accessed and read the relevant section of the Terra Website, cannot make out the Third Representation.

  70. [45] I reject the defendants’ approach of isolating the specific passages cited by each claimant in their witness statement and asking the court to scrutinise whether those passages, read in isolation, are capable of conveying the pleaded meaning of the Conceded Representations. That approach is too narrow and mistakes the purpose of the exercise. The bar for establishing reliance in cases of fraudulent misrepresentation is not a high one. The representee need only prove that the misrepresentation was actively present to their mind at the time they entered into the transaction. The misrepresentation need not be the sole or dominant cause of their decision to act: Panatron. A claimant is not required to have understood the exact operational mechanics of the UST-LUNA arbitrage mechanism or to have parsed each passage of the Terra White Paper with technical precision: First Tranche Judgment at [98(b)]. The court’s task is to assess whether each claimant, reading the relevant sources in their totality and in context, in fact formed the understanding they describe and relied on the representations as understood by them. The fact that specific passages cited in a witness statement may, read in isolation, be susceptible to a narrower or more technical interpretation does not negate the understanding that the document as a whole would have conveyed to a person. What matters is the impression the document as a whole left on a claimant.

  71. [46] On the Third Representation, the pleaded source is the Terra Website, not the Terra White Paper. But it does not follow that a claimant must specifically identify and quote the relevant passage from the “About the Terra Protocol” section in their witness statement to make out reliance on the Third Representation. Reading the Terra Website as a whole (including the “About the Terra Protocol” section), the claimant may conceivably have gleaned the Third Representation. Contrast the position where a claimant can only show that they relied on the Terra White Paper and cannot establish that they accessed the Terra Website. Such person’s claim on the Third Representation will fail because the Terra White Paper is not the pleaded source of the Third Representation: First Tranche Judgment at [19] and [148]. For example, in the first tranche, Mr Macquisten’s claim based on the Third Representation failed because the “About the Terra Protocol” section was not on the Terra Website when he accessed it in January 2021: First Tranche Judgment at [147] and [212]. I apply an analogous approach to this tranche’s claimants.

  72. Special circumstances justifying a departure from the Cut-Off Time

  73. [47] The next issue concerns what circumstances would justify a departure from the Cut-Off Time of 12.01am UTC on 12 May 2022. Many of the claimants in this tranche submitted that the default Cut-Off Time ought not to apply to them. The majority do so on the basis that they continued to believe that the Conceded Representations were true past that point and accordingly did not sell their UST. Typically, those claimants assert that they still believed that the UST-LUNA-US dollar peg would be re-established in a matter of time. As I explain below, that reason alone cannot justify a departure from the Cut-Off Time.

  74. [48] The Cut-Off Time marks the point at which the Conceded Representations ceased to be operative – that is, the point at which a reasonable investor would objectively have discovered or realised the fraud by the defendants. From that point, reliance on the representations ends and the defendants are not liable for any losses sustained thereafter. The Court of Appeal explained the underlying causation analysis (CA Judgment at [33]):

  75. [49] The question of when reliance ends involves an assessment of when the representee became aware of the falsity of the Conceded Representations: CA Judgment at [113], citing Smith New Court Securities Ltd v Citibank NA [1997] AC 254 at 266G–H. The court is concerned with when the representee must have realised that the Conceded Representations were untrue: CA Judgment at [113], citing Panatron at [24]. The Court of Appeal affirmed that the court is entitled to form a view as to how a reasonable reader would have understood the 11 May 2022 tweets (see First Tranche Judgment at [33]), since that is a common factual finding relevant to all claimants: CA Judgment at [116]. Once that common finding is made, it provides the baseline against which each claimant’s individual circumstances are evaluated. A claimant who adduces evidence of circumstances that prevented them from discovering the fraud or selling by the Cut-Off Time may depart from that baseline; one who does not is bound by it.

  76. [50] The common finding, upheld by the Court of Appeal, is that an investor reading the entirety of the 11 May 2022 tweets would have appreciated that UST’s value was not stable and that the Conceded Representations could not be true: CA Judgment at [131]–[135]. By that point, UST had failed to re-peg over four consecutive days since 7 May 2022, hitting a daily low of US$0.2998 on 11 May 2022. The content of the 11 May 2022 tweets reinforced that picture, indicating that the UST-LUNA arbitrage mechanism was struggling to restore UST’s value, that extraordinary measures including collateralisation would be required, and that the situation bore no resemblance to the Conceded Representations. A reasonable investor, within the 14-hour grace period afforded by the default Cut-Off Time, would have appreciated that the Conceded Representations pertaining to UST’s stability and its peg to the US dollar could not be true: First Tranche Judgment at [111].

  77. [51] However, the default Cut-Off Time will not invariably bind all claimants. Thus, the Cut-Off Time may be brought forward where on the evidence a claimant ceased to rely on the Conceded Representations before the default date. For example, two situations might give rise to this:

  78. (a) One situation concerns a claimant who engaged in swing trading – that is, buying when UST swings down and selling on rises – on a speculative basis after the de-peg. Such a claimant can no longer be said to have been acting in reliance on the representations and is making a free and informed decision to take on personal risk: First Tranche Judgment at [104]. The applicable cut-off is moved forward to the point at which they first began to trade on a speculative basis. The claimant may only recover net losses on their original pre-existing UST position up to that point. Any gains or losses from post-reliance swing trading are not claimable against the defendants and do not reduce the net loss recoverable from them.

  79. (b) The other situation concerns claimants who sold all or substantially all their UST before the default Cut-Off Time. The act of substantial selling is strong evidence that the claimant had by that point lost faith in the Conceded Representations. The cut-off is accordingly brought forward to the date on which the claimant last sold UST, that being the date on which reliance must have ended.

  80. [52] By similar token, the Cut-Off Time may be extended beyond the default where a claimant’s circumstances prevented them from discovering the fraud or from liquidating their UST holdings by the default Cut-Off Time. In such cases the claimant did not make a free choice to hold on to their UST. The causal link between the fraud and the claimant’s continued loss is not broken, and the defendants remain liable for losses sustained beyond the default Cut-Off Time. For instance, two situations might justify an extension. A claimant may have been prevented from learning about the de-peg by the Cut-Off Time. Alternatively, external forces beyond the claimant’s control prevented them from liquidating their holdings despite their best efforts: CA Judgment at [115].

  81. Application to the facts

  82. [53] Applying these legal principles, I make the following observations.

  83. [54] An assertion that a claimant continued to believe in the representations past the Cut-Off Time is not a sufficient basis for extending the applicable Cut-Off Time. The inquiry into when reliance on the Conceded Representations ended is subjective, but it must be assessed against the backdrop of what a claimant must have appreciated from the 11 May 2022 tweets and the market data available at the time. Where the hard facts between 7 May 2022 and the Cut-Off Time show that the Conceded Representations were plainly false, a claimant’s assertion of continuing to rely on the truth of the Conceded Representations will not generally be credible and must be approached with scepticism.

  84. [55] The Cut-Off Time incorporates a period for claimants to sell their UST once the Conceded Representations have proved to be false. Claimants arguing that the Cut-Off Time should not apply to them because they continued to believe that the peg would be restored face the difficulty that UST’s volatility between the depeg on 7 May 2022 and the Cut-Off Time made it plain (contrary to what had been represented by the defendants) that UST was not stable. The purported peg to the US dollar was not holding good over a protracted period. The arbitrage mechanism and the algorithm which were supposed to maintain the UST-LUNA-US dollar peg regardless of market size, volatility or demand, were obviously failing to perform their vaunted function. It would have further been apparent that, if indeed any Bitcoin reserves allegedly held by LFG pursuant to the Sixth Representation were being deployed, such reserves were not buttressing the peg. In short, it would have been obvious in the days between the 7 May 2022 depeg and the Cut-Off Time that the Conceded Representations were untrue. The US dollar peg was patently not always guaranteed.

  85. [56] Several claimants rely on the May 2021 de-peg as a basis for their continued confidence that the May 2022 de-peg would similarly resolve. I do not accept that argument. The two episodes were not comparable. In May 2021, UST fell to a lowest point of approximately US$0.92 and recovered within about four days. In May 2022, UST hit daily lows of US$0.7934, US$0.6841, and US$0.2998 on three consecutive days from 9 to 11 May 2022, with no recovery in the days that followed. The scale and speed of the collapse, and the sustained failure of the arbitrage mechanism to restore the peg, rendered the two situations fundamentally different. A claimant cannot credibly point to the May 2021 experience as a basis for believing that the Conceded Representations remained true and continuing to hold UST past the Cut-Off Time in May 2022: CA Judgment at [133]. There is the additional hurdle for claimants that the May 2021 depeg is not itself a source or part of the Conceded Representations. It was consequently for each claimant who became aware of the May 2021 depeg to draw their own conclusion as to the implications of the May 2021 depeg on the actions (to sell or hold to UST) that they should take once the falsity of the Conceded Representations was plain.

  86. [57] Personal difficulties – whether family-related, work-related, or otherwise – would not, without more, justify extending the Cut-Off Time. I accept that, after that point, when the Conceded Representations had been revealed to be false, claimants faced invidious choices. They could sell their UST and crystallise their loss. They could hold on to their UST in the hope of recovery based on their assessment of the situation. The hope may be based on what was being said by others or based on whether they believed the ups and downs that UST was experiencing following the 7 May 2022 depeg was a ground for optimism or pessimism. They might avoid making a decision for whatever reason. They could choose to turn a blind eye to what was happening. They may decide to prioritise other matters in their lives. Such decisions may be understandable give a claimant’s circumstances. I fully appreciate that each claimant’s response to the May 2022 depeg involved a difficult, often painful, judgment call by them. But the consequences of their decision would be at their own risk. Being patently untrue by the Cut-Off Time, the Conceded Representations ceased to be causative of their loss as a matter of law. Further loss resulting from a decision to hold UST beyond the Cut-Off Time must consequently fall on the claimant, not on the defendants. An extension to the Cut-Off Time can really only apply (1) if a claimant’s circumstances were such as to prevent them from learning about the de-peg by the Cut-Off Time or (2) if external forces beyond a claimant’s control prevented them from liquidating their holding of UST by the Cut-Off Time despite their best efforts.

  87. [58] Of the 40 claimants in this tranche, I find that the Cut-Off Time requires adjustment for eight claimants. On swing trading, two claimants have their cut-off brought forward. Mr Thor Robert Kiessling (“Mr Kiessling”) ceased to rely on the Conceded Representations from 8 May 2022, when he withdrew his full Aperture position and immediately re-entered into new higher-yielding Aperture products during the early stages of the de-peg. That re-entry was not consistent with continuing reliance on the Conceded Representations but was yield-driven swing trading. His cut-off is accordingly brought forward to 8 May 2022. Mr Gary Koh Boon Thong (“Mr Koh”) similarly engaged in post-depeg swing trading, having sold all his UST holdings during the de-peg and re-entered the market to buy UST at a lower price in the hope of profiting from a recovery. He ceased to act in reliance on the Conceded Representations from the point of that re-entry. His cut-off is brought forward accordingly. Mr Lam also engaged in swing trading from the early hours of 12 May 2022 in respect of the UST he had withdrawn from Aperture by the default Cut-Off Time, buying and selling UST in multiple cycles in the hope of accumulating more UST before the anticipated repeg. His cut-off for that portion is accordingly brought forward to approximately 4am UTC on 12 May 2022, immediately before his first swing trade commenced. Three claimants (Mr Liam Spagnol (“Mr Spagnol”), Mr Tan E-Wen Timothy (“Mr Tan”) and Mr Lim Xianlong, Kevin (“Mr Kevin Lim”)) sold substantially all of their UST by 11 May 2022. Their cut-off dates are consequently brought forward.

  88. [59] The cut-off of three claimants has been extended as they were prevented from selling by the default date. Mr Lam falls into this category in respect of a portion of his UST that remained locked in Aperture due to the mirror lock contract bug and could not be withdrawn until 11.39am UTC on 16 May 2022. He was not in a position to withdraw and sell those tokens despite his awareness of the de-peg, and accordingly had not made a free and informed choice to retain them past the Cut-Off Time. Mr Bernhard Alexander Friedrich Frey (“Mr Frey”) was at a silent meditation monastery retreat from 9 to 16 May 2022. He had surrendered his mobile phone and had no access to the internet during that period. He was unaware of the de-peg and could not have read the 11 May 2022 tweets or sold his UST in time. Mr Neo Ming Feng (“Mr Neo”)’s freely available UST is assessed at the default Cut-Off Time, but the UST locked in the OSMO-UST liquidity pool and subject to a 14-day unlock period is assessed at an extended cut-off of 26 May 2022, being 14 days from the default Cut-Off Date.

  89. [60] I address the applicable Cut-Off Date and Time for each claimant in the sections below.

  90. Calculation of damages

  91. [61] In calculating the quantum of damages for each claimant, I take as a reference point the methodology set out in the witness statements of Mr Terence Lim Zheng Wei (“Mr Lim”), who is employed by the Plan Administrator of Terraform and is authorised to give evidence on damages on behalf of both Terraform and LFG. The damages assessment script applicable to this tranche (“Revised Script”) has been revised to incorporate two principal changes from the version used in the first tranche. First, it applies the revised Cut-Off Price of US$0.60485 per UST token in place of US$0.8011, in accordance with the CA Judgment. Second, it corrects inconsistencies in the earlier script relating to the aggregation of UST balances across multiple accounts and the treatment of inflows and outflows arising from transfers between a claimant’s own accounts.

  92. [62] The Revised Script generates two main permutations for the assessment of damages. Permutation 1 excludes all purchases of UST made after 7 May 2022, the date of the de-peg, and all UST transactions after 12 May 2022. This reflects the approach applicable to a claimant who is found to have ceased relying on the Conceded Representations from the commencement of the de-peg on 7 May 2022. Permutation 2 excludes only UST transactions after 12 May 2022, and reflects the approach applicable to a claimant who is found to have relied on the Conceded Representations up to and including the Cut-Off Time. For the reasons explained above, the applicable permutation for the majority of the claimants in this tranche is Permutation 2.

  93. [63] Under Permutation 2, damages are calculated using the following formula: Total Purchase Value – (Actual Sales Proceeds + Notional Sale Proceeds) = Net Loss where:

  94. (a) Total Purchase Value is the aggregate value of all the claimant’s purchases of UST up to and including the Cut-Off Time, including UST earned as yield, whether on Anchor or on some other platform. The value of each purchase is determined using the hourly price of UST on CoinGecko, matched to the time of each transaction.

  95. (b) Actual Sales Proceeds is the aggregate value of all the claimant’s sales of UST up to and including the Cut-Off Time, similarly valued by reference to CoinGecko data.

  96. (c) Notional Sale Proceeds is the value of any UST still held by the claimant as at the Cut-Off Time, treated as notionally sold at the Cut-Off Price of US$0.60485 per UST token.

  97. [64] Where the Net Loss is positive, the claimant has a prima facie claim for that sum of damages. Where it is negative, the claimant has made a profit on their UST transactions and has no claim for damages.

  98. [65] The claimants’ Permutation 2 figures adopt a different pricing methodology. Rather than using CoinGecko’s hourly price of UST to value transactions, the claimants use the price of the cryptocurrency or fiat currency actually exchanged. Where fiat currency was used to purchase or sell UST, the claimants use the actual value in US dollars. Where cryptocurrency was used, the claimants value the exchanged cryptocurrency using the average of its daily opening and closing price on CoinGecko. The claimants submit that this approach more accurately reflects the actual price paid for UST and the actual benefit received on sale.

  99. [66] There is a large degree of convergence between the two sets of Permutation 2 figures for the majority of claimants, with differences typically attributable to the differing pricing methodologies described above, rounding and minor differences in transaction fee treatment. Where the parties’ Permutation 2 figures are the same or differ only marginally, I adopt the higher of the two. This is with regard to the principle that, in cases of fraud, a flexible approach should be taken that is more generous to the injured party as opposed to the fraudster: CA Judgment at [147], citing Vita Health Laboratories Pte Ltd v Pang Seng Meng [2004] 4 SLR(R) 162 at [93]. Where there is a significant discrepancy between the parties’ figures, I address the reasons for the discrepancy and the appropriate figure in the individual sections below.

  100. [67] I do not apply the Inference Permutation described in the Revised Script in respect of the claimants in this tranche. The Inference Permutation was developed as a best-efforts approximation for claimants who did not provide deposit and withdrawal logs for their centralised exchange accounts. The Inference Permutation used blockchain wallet transfer records, read together with centralised exchange trading logs, to infer the balance of UST held on those accounts. The claimants in this tranche have since disclosed their deposit and withdrawal logs. Mr Lim confirmed in cross-examination that once all deposit and withdrawal logs are provided, the Inference and the non-inference Permutations should converge to produce the same result, and the Inference Permutation becomes unnecessary. Given that the relevant deposit and withdrawal logs are now in evidence, there is no need to resort to inference, and I have assessed each claimant’s damages on the basis of their actual disclosed records.

  101. [68] Five claimants’ Permutation 2 figures are disputed: Mr Charles Lee Teague (“Mr Teague”), Dr Tobias Speerschneider (“Dr Speerschneider”), Dr Yanni Lin (“Dr Lin”), Mr Clarence Khoo (“Mr Khoo”), and Mr Michael Eli Leidner (“Mr Leidner”). I address the applicable figure for each of those claimants in their individual sections below. Additionally, for the seven claimants whose Cut-Off Time departs from the default, Permutation 2 does not apply. I address the applicable calculation of damages individually in their respective sections below.

  102. Aggravated damages

  103. [69] Of the 40 claimants who gave evidence in this tranche, six seek aggravated damages, saying that they suffered mental and physical health problems as a consequence of the defendants’ fraudulent misrepresentation: Ms Karen Soo Kyung Moon (“Ms Moon”), Mr Cocosatu Mihai (“Mr Cocosatu”), Mr Neo, Ms Ren Qingwen (“Ms Ren”), Dr Aidan Lindsay Fenner Rich (“Dr Rich”) and Mr Leidner. Aggravated damages are compensatory in nature and they augment the basic award where the normal measure of general compensatory damages is insufficient to compensate the plaintiff for the totality of the injury suffered: Noor Azlin bte Abdul Rahman v Changi General Hospital Pte Ltd [2022] 1 SLR 689 (“Noor Azlin”) at [261]. To establish an entitlement to aggravated damages, the plaintiff must show that (1) there is “contumelious or exceptional conduct or motive on the part of the defendant” and (2) the plaintiff suffered “an intangible loss, injury to personality or mental distress”: Noor Azlin at [265].

  104. [70] The claimants argue that medical or psychiatric evidence is not required to establish mental distress, and the court may accept a claimant’s own testimony as sufficient. They rely on Li Siu Lun v Looi Kok Poh [2015] 4 SLR 667, where the High Court affirmed an award of aggravated damages based principally on the claimant’s own testimony of mental anguish and distress, notwithstanding the absence of a formal psychiatric diagnosis (at [183]). I accept that the absence of a medical diagnosis is not, without more, fatal to a claim for aggravated damages. The question is whether the claimant’s evidence of mental distress is credible and sufficient to establish, on the balance of probabilities, that they suffered the intangible loss or mental distress they claim.

  105. [71] For the reasons I explain in the individual sections below, I find that none of the five claimants have sufficiently established an entitlement to aggravated damages. While I accept that each of them suffered financial loss and the distress that naturally accompanies such loss, the evidence in each case falls short of establishing the kind of intangible loss, injury to personality, or mental distress of a character and severity that is distinct from and beyond the financial loss compensated by the general damages award. In particular, where a claimant has not adduced medical or professional evidence and relies solely on their own testimony, the court must scrutinise that testimony with care to determine whether it establishes harm beyond the ordinary consequences of financial loss – which is not, without more, compensable by way of aggravated damages.

  106. Analysis of each claimant’s situation

  107. [72] With the foregoing principles in mind, I turn to analyse each claimant’s situation.

  108. Mr Liam Spagnol

  109. [73] Mr Spagnol claims damages of US$69,106.59. He purchased UST between 9 November 2021 and 11 May 2022, depositing substantially all of it into the Anchor protocol for staking purposes. His case is that he relied on the First to Fifth Representations in purchasing and holding UST. He lost faith in the representations by 11 May 2022 and sold almost all of his holdings that day, realising only a fraction of his purchase price.

  110. [74] On ownership, Mr Spagnol had conducted his UST transactions through a Binance account, an FTX account, an ACryptoS wallet, and three Terra wallets. He submitted a screen recording of him logging in to his Binance account, completing two-factor authentication, and exporting transaction records. He also submitted a screen recording of him logging into his three Terra wallets. I accept his evidence of ownership of all the accounts.

  111. [75] The defendants argue that the screen recordings of Mr Spagnol logging into his Terra wallets are of poor resolution and the wallet addresses cannot be clearly read. On that basis, the defendants invite me to draw the adverse inference that Mr Spagnol was concealing the wallet addresses because they were not his. I decline to draw that inference. I accept Mr Spagnol’s explanation during cross-examination that the lower quality arose from having to use a camera rather than screen-recording software. I also accept Mr Spagnol’s explanation that he can no longer access his FTX account because the FTX customer portal has been disabled. He testified that he attempted to contact FTX support a few weeks before giving evidence but was unsuccessful. In the absence of any positive evidence that the wallets and accounts belong to someone else, I accept Mr Spagnol’s sworn evidence that he had full access to and control over all his accounts at the time the relevant transaction was conducted.

  112. [76] Mr Spagnol relies on the First to Fifth Representations. I need not analyse his evidence on the Fifth Representation as it has been held to be unactionable. Mr Spagnol says that he read the Terra White Paper and the Terra Website on or around 15 October 2021, and the Anchor White Paper and Anchor Website at or around 9 November 2021. He was unable to produce browsing histories as he resets his browsing history regularly and no longer has the same laptop. I accept that explanation.

  113. [77] I accept Mr Spagnol’s reliance on the First Representation. He relied on section 2.1 of the Terra White Paper and understood from it that UST was designed to be pegged to the US dollar. I also accept Mr Spagnol’s reliance on the Second Representation. He relied on sections 2.3, 2.4, and 2.5 of the Terra White Paper. In cross-examination, he explained that he understood the arbitrage and miner stability mechanisms as two distinct but complementary systems that together gave him confidence that UST’s peg would be maintained in all market conditions. I also accept Mr Spagnol’s reliance on the Fourth representation. His evidence is that he read the Anchor White Paper and the Anchor Website before making his first purchase, and understood from the relevant passages that the Anchor protocol was a principal-guaranteed savings product that would protect his principal while generating passive income. I also accept Mr Spagnol’s reliance on the Third Representation. Mr Spagnol’s evidence is that he accessed the Terra Website on or around 15 October 2021 before his first purchase. Although his witness statement explains his understanding of the Third Representation primarily by reference to section 2.4 of the Terra White Paper rather than the “About the Terra Protocol” section of the Terra Website, this alone does not defeat reliance.

  114. [78] Overall, I am satisfied that some combination of the First to the Fourth Representations was actively present in Mr Spagnol’s mind when he purchased and held UST. I do not accept the defendants’ argument that Mr Spagnol, as a banker familiar with the concept of credit risk, could not have understood the representations in the manner pleaded. It is Mr Spagnol’s actual understanding that is relevant, not what someone of his professional background ought to have known or ought not to have relied upon.

  115. [79] On the cut-off time, Mr Spagnol lost faith in the representations by 11 May 2022, when he sold substantially all of his UST holdings. The act of selling is evidence that Mr. Spagnol by that point had lost faith in the representations. The cut-off is accordingly brought forward to 11 May 2022, being the date on which Mr Spagnol’s reliance on the representations ended and his loss crystallised. I reject the defendants’ first argument that reliance ceased from 7 May 2022 merely because Mr Spagnol regarded a 1% price deviation as unusual. I also reject the defendants’ second argument that his purchase of UST on the morning of 11 May 2022 was speculative. Mr Spagnol explained in cross-examination that the only available trading pair for LunaX at that time was LunaX/UST, and that he converted his LunaX to UST as a risk management decision, moving from a more volatile asset to one he regarded as relatively more stable. I accept his explanation. His subsequent sale of his entire holdings later that afternoon on 11 May 2022 is the point at which I find him to have lost faith in the representations.

  116. [80] On quantum, since the cut-off date is brought forward to 11 May 2022, Permutation 2 does not apply. Mr Spagnol’s total eligible UST is 110,458.86 UST, comprising 107,739.31 UST purchased at a total purchase value of US$107,655.26 and 2,719.55 UST earned as yield from staking on the Anchor protocol. By 11 May 2022, Mr Spagnol had sold 108,509.46 UST, realising total sale proceeds of US$41,273.13. A residual balance of 1,949.40 UST remained unsold at the end of 11 May 2022. The residual balance will be treated as having been notionally sold at the average of the opening and closing prices of UST on 11 May 2022 as recorded on CoinMarketCap, being US$0.80025. That yields notional sale proceeds of US$1,560.01. The net damages are therefore US$107,655.26 less US$41,273.13 less US$1,560.01, giving US$64,822.12. I award Mr Spagnol that sum in damages for misrepresentation.

  117. Mr Mark Eric Burrell

  118. [81] Mr Mark Eric Burrell (“Mr Burrell”) claims damages of US$294,843.64. He purchased UST between 16 November 2021 and 17 December 2021, depositing substantially all of it into the Anchor Protocol through his Terra wallet for staking purposes. He sold only a small quantity of UST in December 2021 and held the remainder through the depeg period, leaving the majority of his holdings unsold at the Cut-Off Time. His evidence is that he relied on the First to Fifth Representations to purchase UST. His case is that he only read Mr Kwon’s 11 May 2022 tweets on 13 May 2022, and that his cut-off should accordingly be extended to 13 May 2022, 2pm UTC, being the end of the 14-hour grace period from that point.

  119. [82] On ownership, I accept that Mr Burrell owns both the KuCoin account and the Terra wallet through which his UST transactions were conducted. He adduced a screen recording of himself logging into his KuCoin account via a VPN (KuCoin having restricted access for US citizens), completing two-factor authentication, and exporting transaction records matching those exhibited in his witness statement. He was unable to generate API keys on his KuCoin account because KuCoin no longer offers KYC services to US-based users, which is the prerequisite for generating API keys. I accept that explanation. He also adduced a screen recording of himself logging into his Terra wallet through his Terra Station Chrome extension and displaying the wallet address. The defendants did not seriously challenge the ownership of either account in their submissions. Nor did they adduce positive evidence to suggest that the accounts or wallets belonged to someone else.

  120. [83] On reliance, Mr Burrell’s evidence is that he accessed the Terra Website at or around October 2021 and continued to access and read it on a daily basis between his first purchase of UST on 16 November 2021 and his last purchase on 17 December 2021. He was unable to produce browsing histories because his security settings automatically erase all web and app activity after 90 days. I accept that explanation. His evidence is that he read the Terra White Paper and the “About the Terra Protocol” section of the Terra Website, including the updated version published in November 2021. He also accessed the Anchor White Paper and the Anchor Website in or around October 2021 before deciding to purchase UST.

  121. [84] In cross-examination, Mr Burrell accepted that he was unable to recall exactly which parts of the Terra Website he read on each of the 31 days he visited it. He maintained, however, that he had read the First to Third Representations prior to making his first purchase on 16 November 2021. His position, which I accept as credible, is that he went through the whole of the relevant material at some point over those 31 days. Given that he was committing approximately US$300,000, it is wholly believable that he read the source documents carefully and repeatedly before committing that sum. The defendants argue that Mr Burrell had imprecise recollection of which specific part of the website he read on which specific day, and that suggests that he may have only read the representations after his first purchase of UST in November 2021. I do not accept that argument. The question is whether, by the time he made his first purchase, he had seen the relevant representations. I accept his evidence that he had.

  122. [85] I am therefore satisfied that Mr Burrell read and understood the First, Second, Third, and Fourth Representations from the Terra White Paper, the Terra Website, the Anchor White Paper, and the Anchor Website respectively. Although Mr Burrell also says he relied on the Fifth Representation, that representation has been held in the First Tranche Judgment not to be actionable in misrepresentation.

  123. [86] On the cut-off time, Mr Burrell’s case is that he did not read Mr Kwon’s 11 May 2022 tweets until on or around 13 May 2022. He explains that he still believed in the representations before that point, and considered the price increases between 10 and 11 May 2022 to be an early indicator of the stability mechanism taking hold. He also believed that Terraform would deploy the reserve fund as part of a recovery plan. It was only after reading Mr Kwon’s 13 May 2022 tweets, from which he understood that the attempts to address the de-peg had been unsuccessful, that he lost faith in the representations and began taking steps to sell.

  124. [87] I do not accept Mr Burrel’s argument. Mr Burrell was, by his own evidence, an attentive and diligent investor who read the Terra Website on a daily basis over 31 days before committing US$300,000 to UST. He was monitoring the UST price closely throughout the depeg. I find it difficult to accept that he would not have seen Mr Kwon’s 11 May 2022 tweets, which were widely circulated in the crypto community, until two days later. Mr Burrell has provided no positive evidence to explain what he was doing on 11 and 12 May 2022 that would have prevented him from seeing those tweets. A bare assertion that he did not see them until 13 May 2022 is insufficient to displace the default Cut-Off Time.

  125. [88] I also note paras 52–53 of Mr Burrell’s second witness statement, which originally set out an elaborate explanation of why he was unable to transfer UST out of his Terra wallet during this period, were withdrawn at the commencement of his evidence, after Mr Burrell accepted that the basis for those paragraphs was factually wrong. He had in fact transferred 315,830 UST from his Terra wallet to his KuCoin account on 12 May 2022 at 3.37am UTC+8, which is consistent with him having discovered the fraud and taken steps to sell by the default Cut-Off Time. This further undermines his evidence that he was unaware of the scale of the crisis until 13 May 2022. The default Cut-Off Time of 12.01am UTC on 12 May 2022 therefore applies.

  126. [89] On quantum, since the default Cut-Off Time applies, the applicable figure is Permutation 2. The parties’ Permutation 2 figures are close: the claimants’ figure is US$108,053.11 and Mr Lim’s figure is US$104,551.87. I take the higher of the two figures. I award Mr Burrell US$108,053.11 in damages for misrepresentation.

  127. Mr Cocosatu Mihai

The judgment continues past this page. Read the rest at eLitigation.