May 2022. A stablecoin stops being stable.
The legal question in Terra-Luna was never just: did the algorithm fail?
Software fails. Markets fail. Models fail. A failure alone does not prove fraud.
The sharper courtroom question is: what did Terraform Labs and Do Kwon say the mechanism did, what did the system actually do, and who intervened when the “algorithmic” promise came under pressure?
TerraUSD, or UST, was marketed as an algorithmic stablecoin designed to maintain a one-dollar peg through its relationship with LUNA. In the sales story, that was not just a product feature. It was the proof of decentralization, automated stability, real-world demand, and a financial system that supposedly worked because the code and incentives made it work.
(The word “algorithmic” did a lot of legal labor here. More than it could carry.)
The artifact: the peg mechanism, not the slogan
The artifact in the case was not one source-code file.
It was the whole operating record of the peg:
- the mint/burn economics between UST and LUNA
- the on-chain transaction history during stress events
- the May 2021 de-peg and re-peg
- the May 2022 collapse
- the role of reserve-control decisions
- the alleged undisclosed external trading support used to restore the peg in 2021
- the investor-facing statements that described the recovery as proof that the algorithm worked
That is where a software expert stops being decorative and becomes necessary. A lawyer can read the marketing deck. A software and data expert can test whether the represented mechanism matches the operational record.
What the SEC and DOJ focused on
The SEC charged Terraform and Kwon in February 2023 and later won a civil fraud verdict. In April 2024, a Manhattan jury found Terraform Labs and Kwon civilly liable for defrauding investors in crypto asset securities. In June 2024, the SEC announced a more than $4.5 billion resolution after that verdict.
The core allegation was not simply that UST collapsed. It was that investors were misled about the stability of UST and about whether a popular Korean payment application actually used the Terraform blockchain to settle transactions.
The DOJ later framed the same technical story in criminal terms. Kwon pleaded guilty in August 2025, and in December 2025 SDNY announced that he was sentenced to 15 years in prison. The DOJ described misrepresentations about the Terra Protocol, the Luna Foundation Guard, Mirror Protocol, Chai, and the use of genesis stablecoins. Its account of the May 2021 re-peg is especially important: the peg, according to DOJ, was restored only after a high-frequency trading firm strategically purchased large amounts of UST, not because the Terra Protocol had independently solved the problem.
That distinction is the case.
If the public story was “the algorithm restored confidence,” but the operational story was “outside trading support propped up the peg,” then the artifact is not just code. It is code plus transactions plus control plus disclosure.
The evidence stack
For a litigation team, Terra-Luna is a clean map of what to collect in a software-heavy financial dispute:
Public claims: white papers, investor decks, interviews, tweets, docs, website copy.
System behavior: blockchain transactions, oracle inputs, liquidity conditions, smart-contract events, parameter changes.
Human control: who controlled reserves, who approved interventions, who could direct support, who could alter governance assumptions.
Crisis conduct: what happened during prior de-pegs, what was disclosed, what was hidden, and what story was told afterward.
Reliance link: whether technical claims were used to attract users, liquidity, capital, or investor confidence.
The “bug or fraud” line is usually not inside the code alone. It sits in the gap between what the product was said to do and what the technical record shows it actually did.
Why Israeli counsel should care
Israeli fintech, crypto, AI, and SaaS teams often use technical adjectives as business shorthand: autonomous, decentralized, stable, risk-managed, self-learning, algorithmic.
Those words are not harmless when customers, investors, or counterparties rely on them.
In a dispute, the question becomes evidentiary: can the company prove the system behaved as described? If not, was the description cautious enough? Were known failures disclosed accurately? Were manual interventions documented? Did the logs support the investor-facing story?
For lawyers, the practical move is simple: do not wait for a collapse to ask for the architecture. Ask early for the mechanism, the logs, the control points, and the incident history.
Terra-Luna shows why. The case was not built on “crypto went down.” It was built on the difference between a promised autonomous mechanism and the mechanism that actually operated when the peg was under stress.
That difference is where the evidence lives.