ZEC as Collateral: Non-Custodial Strategy Insurance on Zcash

About Us

We are the team behind Nabu (https://nabu.pro/) and we are exploring possibility to build strategy insurance collateralised in ZEC, which will provide a way for ZEC holders to earn on their coins without them leaving Zcash. We would like to find out whether the community would be interested in such a product. We are also building https://pelagos.network/ and execution layer leveraged by Nabu

Team: https://pelagos.network/team/
X: Nabu (@nabupro) / X
X: Pelagos Network (@PelagosNetwork) / X
Docs: https://docs.nabu.pro/; https://docs.pelagos.network/

About Nabu

Nabu is an AI driven solution built to achieve people’s financial goals. A user states a goal in plain language, for example reaching a target balance by a certain date at a given level of risk. Nabu drafts several candidate plans, tests them against market history, and compiles the chosen one into deterministic code with hard limits attached: a loss cap, a whitelist of venues and a kill switch. Only that compiled code is executed, on the user’s own accounts, and every action leaves a record that can be checked afterwards. Execution runs on Pelagos, an execution layer our team built for this purpose. The team includes former Ethereum Foundation, Polygon and Erigon engineers.

Today Nabu is focused on trading, with spot and perps live on Hyperliquid and Binance. Trading is the first market we cover. The same approach of a stated goal, a verified plan, bounded risk and round-the-clock execution applies to any market that software can operate. Yield and lending are on roadmap, and tokenised stocks, along with prediction markets. Each new market is an integration on the same system rather than a new product.

The longer-term aim is that anyone with a financial goal can hand it to Nabu and have it achieved within limits they set. Crypto markets offer yield and run around the clock, but most people cannot use them. Almost everyone already talks to an AI, and it still cannot operate their money. Nabu brings these together, with the model proposing the plan and the compiled code enforcing the limits whether or not the user is watching.

What we are thinking of building for Zcash Ecosystem

Zcash has no native yield, so ZEC earns nothing unless you move it somewhere else. Traders have the opposite problem: they have a strategy they believe in, but risking all of their own capital on it is hard, and taking outside money means managing it for others.

The idea is to connect the two. A trader runs a strategy on Nabu with only their own capital, on their own Hyperliquid account for example. ZEC holders lock ZEC at a Zcash address as cover against that strategy’s losses for a fixed period. If the period ends in a loss, the trader receives an agreed share of that loss in ZEC at the closing price, capped at the collateral and taken from each holder in proportion to what they locked. If it ends in profit, the holders receive an agreed share of the profit, converted to ZEC and sent to their addresses, and get their collateral back in full.

Every term is public before anyone commits: the period, the share of losses covered, the share of profit paid out, and the strategy’s drawdown limit, which stops the strategy early if it is hit. Holders can also see the strategy’s history, backtest and paper-run results. They find strategies in their Zcash wallet or on the insurers page, then join and pay from their wallet, without creating a Nabu account. Before the period starts they can withdraw. Once it starts, nobody can exit until the period ends, which can happen early if the drawdown limit is hit or the trader stops the strategy. The result is computed from the trader’s public trades and fees, so anyone can recompute it.

The ZEC never leaves the Zcash network. It is locked, returned and paid at Zcash addresses. In the first version the collateral would sit at addresses under a Nabu key. The target design holds it under a script with several spending paths, where our operators can only co-sign according to the result and cannot move the collateral on their own.The Zcash components would be open source: reading the Zcash chain and signing transactions in Go, accepting a ZEC payment for a specific deal without an account, and the collateral script.

Your Inputs Matter :slight_smile:

Before we build it, we would like to hear from you:

  • Would you be interested in locking ZEC as cover for a strategy in return for a share of its profit?
  • What would you need to see before doing it?
  • For ecosystem teams / builders: would you be open to showing insurable strategies inside your wallets / dApps?

We would truly love to collaborate with Zcash community on this and looking forward to your constructive feedback!

1 Like

interesting idea and good that every term is public before anyone commits. a few things i’d want to see before locking ZEC:

the hedge problem; the cover only looks at the trader’s insured account, so a trader can open the opposite position somewhere else. if the insured side loses, holders pay a share of the loss while the hedge wins. if it wins, the trader pays a share of the profit and loses on the hedge. whenever the loss share is bigger than the profit share, that’s money taken from holders, and nobody can see the hedge. the same goes for stopping the strategy early: the trader can pick the moment the loss is biggest. how do you plan to handle that?

custody; in v1 the collateral sits under a Nabu key, so it’s custodial until the target design ships. on zcash a script with several spending paths only exists for transparent addresses, which would make every lock and payout public. is the target transparent multisig, or FROST over shielded funds?

the closing price; the payout uses ZEC at the closing price, so which source sets it, and can it be checked afterwards like the trades?

1 Like

Hey Igor,

Many thanks for your questions and apologies for taking time as we were trying to provide a mathematically sound design:

A holder never loses more than the figure fixed up front: the loss share multiplied by the drawdown limit, and never a cent more than their posted collateral. On top of that, holders only lose when the trader has lost at least as much. The loss share is capped at half (50%), meaning the trader’s own loss will always equal or exceed the combined loss of all holders. That loss is real, not engineered, and here is why.

We can’t monitor outside hedges, so the math is set up to make hedging pointless. The period result is calculated strictly as the strategy’s exposure multiplied by the move in the market price. Trading fees, slippage, and funding paid are the trader’s own operational costs and are excluded from the result. Funding received is included. This means a loss cannot be manufactured on purpose: it only triggers when the market moves against the strategy, which an outside hedge cannot force. Combined with the rule that the loss share can never exceed the profit share, a hedged trader receives, on average, zero or less from holders. If funding paid weren’t excluded, someone could go long here, short on another venue to pocket the funding, and leave holders to cover their synthetic loss. We’ve accounted for that.

A period ends either on the predetermined end date or via the automated drawdown stop locked in before launch. If the trader decides to shut the strategy down manually, they still have to pay the premium if they are in profit, but they receive zero compensation if they are in drawdown. There is no timing window worth gaming.

The holder is buying a share of the strategy’s return, not selling insurance, and that is exactly what we will call it. Strategy risk still exists, an algorithm can simply underperform, just like any shared outcome. To protect against that, the strategy’s source is open, the backtest is public, the trader’s historical record is visible, and both the share and the term are capped.

For custody, there will never be a version locked under a Nabu key. From day one, collateral sits in a script with three distinct spending paths: holder plus operators, trader plus operators, or the holder acting alone after a specific block height. No threshold cryptography is required. Using a transparent address is deliberate: Zcash’s shielded pool doesn’t support timelocks, so running FROST over shielded funds would leave holders unable to recover their collateral on their own if the operators vanish. A shielded version will follow once there is an independent operator set and mature FROST. Locks and payouts remain public; every deal has its own address, and holders can use a fresh address every time.

For closing price, this is locked into the public terms before the start: the median ZEC/USD trade price across three designated exchanges over the final hour of the period. All three publish raw trade histories, meaning anyone can independently recalculate the exact number the same way the result is audited. Operators simply verify the price.

Hope this helps and do let me know if you have additional questions :slight_smile:

Do let us know if you have questions / feedback :slight_smile:
We truly would love to work with the Zcash community!
Thank you all!