There is no contract address.
That is the finding, not the caveat. The New York Attorney General is seeking roughly $1.37 billion in restitution and penalties from a crypto trading operation whose entire public on-chain footprint resolves to a null set. No mainnet deployment. No testnet repository. No audit report. No reserve attestation. No qualified custodian. No exchange address cluster traceable to a treasury wallet.
Four promoters settled. Combined value: approximately $6 million.
Divide. Six million against one point three seven billion is 0.4379%.
That ratio is the most informative number in the file. It is not a rounding artifact. It is the executable boundary of the enforcement action — the layer where jurisdiction still touches flesh, bank accounts, and a home address. Everything above that layer has already crossed a border.
I have audited systems that failed. In 2018 I spent four hundred hours inside the EOS delegation logic and found three integer overflow paths before listing. Those were fixable defects. This is different. There is no defect to fix, because there is no code to defect.
A row of zeros is an audit result. Most readers treat an empty column as missing data. It is not. When a system claims to manage institutional-scale capital and leaves no artifact, the absence is the artifact.
The public record on NovaTech contains five load-bearing information points. Four are enforcement outputs: a restitution demand, an absconded founder, a promoter settlement, and a distribution channel. One is the channel itself — prayer groups used as the acquisition funnel.
None of the five describes a system.
So I do not run a token model here. Running a supply schedule against a project with no disclosed token would generate false precision. I run a gap ledger instead. For any operation claiming crypto execution and custody, a fixed set of artifacts must exist. Then I query for each one.
One data-hygiene note before the analysis. Media summaries of this case circulate two figures: $1.3 billion and $1.37 billion. That 5.4% spread appears in the reporting itself. When a headline number drifts by five percent between summaries, treat the primary filing as the only valid citation. Figures that move in retelling do not belong in a damage model.
The query is deliberately boring.
