Hyper Unit

Our research assessment is adverse. Hyper Unit lets a user deposit native BTC, ETH, SOL, and other assets from their home chain and mints a matching balance on Hyperliquid, distinct from the separately reviewed Hyperliquid Bridge, the Arbitrum-to-HyperCore USDC path. Custody runs through a “Guardian Network” using MPC/threshold-signature cryptography rather than a conventional multisig, but the network is reported to have only three total operators with a 2-of-3 signing threshold. That is a smaller, more concentrated trust set than the already rejected Hyperliquid Bridge’s 27-validator set. Cryptographic sophistication does not offset having only three parties, any two of whom can move funds, and no independent, Unit-specific security audit was confirmed.

Review triggers: events that reopen the verdict
• Guardian count expands materially, for example to seven or more, with published, named operator identities
• An independent, Unit-specific security audit, not the general Hyperliquid L1 audit, is published with no critical findings outstanding
• Withdrawal timing and dispute or challenge mechanics are independently confirmed and pass a proposed-size stressed-exit test
• Twelve consecutive months with no Guardian-key or MPC-ceremony compromise

Watched nightly: a warning on its venues or files, or a cited document that changes, reopens the memo. The first confirmation is due 2026-11-17.

The research file

Mechanism

Lock-and-mint. A user sends a native asset on its home chain. Unit locks it and mints a matching token to the user’s Hyperliquid spot balance, where it can be traded or used as collateral, and withdrawn back to the origin chain. Supported origin chains include Bitcoin, Ethereum, Solana, and others on DefiLlama’s tracked chain list for this protocol.

Control and governance

Custody and signing use an MPC/threshold-signature scheme, so no single party ever holds a complete private key. Unit markets this as the “Guardian Network.” The reported operator count is three, with a 2-of-3 signing threshold, meaning any two of three parties can authorize a movement of funds. That is a materially smaller trust set than this registry’s already rejected Hyperliquid Bridge entry, with 27 validators, and WBTC’s named 2-of-3 custodian structure across three identifiable institutions. Unit’s guardians were not found named in public documentation. This review confirmed no Unit-specific third-party security audit. A Zellic audit sometimes cited in connection with Hyperliquid appears to cover the core L1 and bridge, not Unit specifically, and should not be treated as a Unit audit.

Incident record

No exploit of the legitimate protocol was identified. A documented phishing lookalike site impersonates Unit and is designed to drain connected wallets. This is a user-facing scam risk rather than a protocol-level failure, but it is worth flagging for client communication regardless of this entry’s verdict.

Exit

Unit’s API reference documents a withdrawal-lifecycle process, but this review could not confirm specific withdrawal timing, fees, or evidence of a stress-tested redemption at scale. Together with the unconfirmed audit status, that leaves both the custody model and the exit path short of what this registry requires to approve a position.

Comparison

The already rejected Hyperliquid Bridge has 27 validators and closed-source software. Unit’s 3-guardian, 2-of-3 signing set is smaller and uses a more novel but less independently audited design. WBTC has a 2-of-3 multisig across three named, identifiable institutional custodians and is also rejected in this registry for a different reason. Unit’s guardians are unnamed. Neither comparison supports approval. Unit also depends on the underlying Hyperliquid L1, which this registry’s stHYPE memo separately flags for reported validator-set concentration around Hyperliquid’s own team. This was most visible in the March 2025 episode when the exchange intervened on a contested position rather than letting on-chain mechanics resolve it. That adds a systemic risk beneath Unit’s own custody model, and Unit does not resolve it.

Sources

The claims above trace to these. Where a number could not be independently verified, the thesis says so.

• Unit Docs — architecture quickstart · primary · accessed 2026-08-17
  Supports: mechanism, Guardian Network
• OAK Research — Hyper Unit: the hidden engine powering Hyperliquid · secondary · accessed 2026-08-17
  Supports: guardian count, trust model
• Messari — analyst note: introducing Unit · secondary · accessed 2026-08-17
  Supports: product overview
• PCrisk — fake HyperUnit website scam · secondary · accessed 2026-08-17
  Supports: phishing risk, user-facing scam
• DefiLlama — Unit protocol record · secondary · accessed 2026-08-17
  Supports: tracked TVL, chain perimeter

Inherited controls

The research above describes the protocol layer. Every position also inherits the asset it holds and the chain it settles on. The layer with the most administrative power sets the position’s effective control; that describes control, not quality or suitability.