JONATHAN KUPETZ & 274 Ors v TERRAFORM LABS PTE. LTD. & 2 Ors

Outcome: The Singapore International Commercial Court (“SICC”) awarded damages to 40 represented claimants in the second tranche of a representative action against Terraform Labs Pte Ltd, Mr Kwon Do Hyeong, and Luna Foundation Guard Ltd (collectively, the “defendants”) for fraudulent misrepresentations in connection with TerraUSD (“UST”), an algorithmic stablecoin that collapsed in May 2022.

1 SIC/OA 3/2024 is a representative action involving 275 claimants who purchased UST, an algorithmic stablecoin that was designed to maintain a 1:1 peg to the US dollar through a mechanism involving its sister token LUNA, but which collapsed in May 2022, wiping out billions of dollars of investor value. The claimants allege that they purchased UST in reliance on representations made by the defendants as to UST’s stability, UST’s peg to the US dollar, and the principal-guaranteed nature and high returns available on the Anchor protocol from staking UST. The defendants conceded that those representations were false and had been made fraudulently.

2 The first tranche, heard in May 2025, determined issues common to all claimants and resolved the individual claims of the ten representative claimants. The Court of Appeal subsequently revised the cut-off price for damages (that is, the notional price at which claimants’ remaining UST holdings are treated as having been sold for the purposes of assessing loss) from US$0.8011 to US$0.60485 per UST token. This second tranche determines the individual claims of 40 represented claimants on questions of ownership, reliance, the applicable cut-off time for the assessment of damages, and quantum.

3 The SICC found that each of the 40 claimants had established ownership of the relevant cryptocurrency accounts and wallets, and that each had relied on one or more of the conceded representations in purchasing and holding UST. The SICC rejected three recurring arguments advanced by the defendants. First, the Anchor Terms of Service could not negate reliance on the conceded representations, since a party cannot contractually exclude liability for fraudulent misrepresentation: at [37]. Second, reliance is assessed subjectively. The question is whether the representations induced each claimant to act, not whether a reasonable person would have been induced. Since the defendants conceded the pleaded meanings of the representations and their falsity, the claimants only needed to establish that they subjectively understood and acted on those representations: at [40]–[42]. Third, a claimant’s professional background does not raise the threshold for establishing reliance. It is no defence that a claimant acted incautiously or failed to take the steps a more prudent person would have taken: at [8] and [37].

4 On the applicable cut-off time for the assessment of damages, the SICC held that the default cut-off of 12.01am UTC on 12 May 2022 did not invariably apply to all claimants: at [51]–[52]. It may be brought forward where a claimant ceased to rely on the representations before the default date. This arises where a claimant engaged in post-depeg swing trading, in which case the cut-off is the date when speculative trading commenced: at [51(a)]. Another example is where a claimant sold substantially all of their UST before the default cut-off. The act of selling was evidence that reliance on the Conceded Representations had ended and the cut-off was brought forward to the date of sale, with any residual UST being treated as notionally sold at the average of the daily opening and closing prices of UST on the latter date: at [51(b)]. The cut-off may be extended beyond the default where a claimant was prevented from discovering the fraud or liquidating their holdings due to circumstances beyond their control, such as a technical lock-up on a yield platform or the protracted absence of internet access: at [52]. The SICC also held that a claimant’s belief that UST would repeg, whether drawn from a reading of Mr Kwon's tweets or the opinions of third parties, reference to an earlier depeg in May 2021, or a deliberate choice to disengage from market developments, could not be a basis for departing from the default cut-off time. By the time the default cut-off time had passed, it would have been apparent that the Conceded Representations were untrue. Any further loss resulting from a decision to hold UST would be at the claimant’s own risk and could not be attributable in law to the defendants’ fraud: at [55]–[57].

5 Six claimants sought aggravated damages for mental distress. The SICC declined to award aggravated damages to them, holding that testimony of emotional distress, without supporting evidence, was insufficient to establish intangible loss or mental distress of a severity distinct from the financial loss compensated by general damages: at [69]–[71].

6 The SICC provided guidance on the preparation of witness statements in representative actions with multiple claimants at [16]–[34]. The use of open-ended questionnaires to collect evidence from a large number of geographically dispersed claimants is permissible as a practical necessity in proceedings of this scale. But the resulting statements must reflect each witness’s evidence in their own words: at [22]. Solicitors may not use a witness’s statement as a template for other statements: at [27]. AI tools may assist with structure and form. But AI should not be used to generate the substance of a witness’s evidence: at [31]. Despite concerns about similarity or even identity in language across numerous statements, the SICC admitted the claimants’ evidence, given that each claimant was able to explain their evidence in cross-examination and that there is currently no specific guidance in Singapore governing witness statements in mass claims: at [33]. The SICC stressed that this approach should not be repeated in witness statements to be filed in future tranches: at [34].