Beyond Payments: A Proposal to Bring Shielded Assets to Ironwood for Data Protection

Prepared by: @Zecoshi
Date: 13 November 2026
Status: Proposal for discussion


Abstract

Zcash is known as a private way to pay. This proposal argues that it can be something larger: the first public network able to protect records themselves, not only payments. Zcash Shielded Assets would let anyone issue an asset whose holdings and transfers are private while its total supply stays publicly provable.

The specifications exist as drafts, but they are written for the Orchard pool, which no longer accepts new value, and they are not scheduled for any network upgrade. This proposal argues that shielded assets should be scheduled for a named upgrade, retargeted to the Ironwood pool, and covered by the same formal verification as Ironwood itself. It argues that cheap machine analysis has ended the protection that obscurity once gave to sensitive records, so the records that matter most should exist only in a form that cannot be read in bulk.

It sets out the security case for users and businesses, the adoption case for Zcash, the objections, and four recommendations.


1. Problem statement

Data breaches rarely depend on clever attacks. They depend on readable records being held in one place. Every organisation that keeps a register of who holds which licence, credential, title, membership or right of access holds exactly that kind of record. Public blockchains make the problem worse, because whatever is written to them is readable by anyone, permanently.

Zcash already solves this for one kind of record. ZEC held in a shielded pool exists on chain only as commitments and encrypted notes. There is no table of who owns what, so there is nothing to steal in bulk. The gap is that this protection covers ZEC only. An organisation that wants the same protection for its records, such as a register of licences, a set of credentials, a list of members or a record of titles, has no shielded option on Zcash mainnet today. This proposal is about those records. It is not a proposal for new coins or currencies.


2. Why artificial intelligence makes this urgent

The argument in this section is reasoning, not measurement. It does not depend on any one statistic. It depends on what cheap machine analysis does to data that is merely hard to read.

2.1 An asset is data.
In this protocol an asset is simply a record that a holder is entitled to something: a licence, a credential, a certificate, a title, a membership, a ticket or a right of access. This proposal is about protecting those records. It is not about coins, and it does not propose any new currency.

2.2 Obscurity is no longer protection.
A great deal of sensitive data has been safe only because it was expensive to analyse. A public ledger, a leaked table or a set of pseudonymous accounts could be connected to real people, but only with skilled human effort. Machine analysis removes that cost. Patterns across millions of records can be linked, clustered and matched to outside data by anyone. Data that is only obscure should now be treated as readable.

2.3 Every leaked record is worth more.
A stolen ledger used to be raw material that needed work. Models can now sort it, join it to other leaks and write a convincing, personal approach to every person in it. The damage from one breach grows with the quality of the tools that process it.

2.4 Flaws are found faster.
The same tools that help defenders review code help attackers search it. Any system that keeps readable records behind a perimeter is betting that its perimeter has no flaw that a tireless automated search can find.

2.5 What is public stays public.
A record published on a transparent chain can be collected today and analysed by better models in every year that follows. Privacy that depends on the limits of present analysis expires.

2.6 The only durable answer is to hold nothing readable.
Encryption inside a company still leaves the keys and the complete ledger in one place. In a shielded pool the record exists only as a commitment and an encrypted note, the key is held by the owner alone, and correctness is shown by proof and not by exposure. There is no aggregate for a model to mine and no perimeter whose failure reveals everything.

2.7 Software agents will hold and move assets.
As automated agents begin to transact for people and businesses, each one becomes a new place where an owner’s records can leak. An agent that can prove an entitlement without revealing the holder, the amount or the history exposes far less when it is compromised.

2.8 Beyond payments.
Seen this way, a shielded pool is not only a payment system. It is a public ledger on which a record can be proven correct, transferred and audited without ever being readable in bulk. Payments were the first use because money was the first record worth protecting. Credentials, licences, titles and entitlements are records of the same kind.

A network that can protect them offers something no database and no transparent chain can: data protection that does not depend on trusting whoever holds the data.


3. Background

3.1 Shielded assets are specified but not scheduled.
ZIP 226 covers transfer and burn and ZIP 227 covers issuance. Both have Draft status (Kogan et al., 2022a; Kogan et al., 2022b). ZIP 226 states that the protocol is not scheduled for deployment in any network upgrade (Kogan et al., 2022a).

3.2 The drafts target a pool that is closed to new value.
Both ZIPs are written as an extension of the Orchard protocol (Kogan et al., 2022a). After network upgrade NU6.3, no new value may enter the Orchard pool (Hopwood, 2026).

3.3 Ironwood is the successor pool.
Ironwood is the shielded pool introduced in NU6.3. A formal verification of the protocol, written in Lean 4 over Mathlib, is published in the open and is described by its own repository as in progress, intended to cover at least the soundness of the Action circuit (Zcash, 2026).

3.4 Ironwood carries a quantum recovery path.
ZIP 2005, which has Final status, makes notes in the Ironwood pool recoverable through a future Recovery Protocol. The same ZIP states that this does not by itself make the protocol secure against quantum adversaries.

Notes in the Sprout, Sapling and Orchard pools are not recoverable (Hopwood and Grigg, 2025; Hopwood, 2026).


4. Proposal

Schedule Zcash Shielded Assets for a named network upgrade, retargeted from Orchard to Ironwood, with the asset logic covered by the same formal verification effort as the Ironwood circuit.


5. The case

5.1 Security of user and business data

  1. Holdings and transfers are private by design. ZIP 226 keeps the amounts and the identifiers of assets private, and a transfer does not reveal which assets, or how many distinct assets, are being moved (Kogan et al., 2022a).

  2. Every asset shares one anonymity set. Note commitments for custom assets are indistinguishable from those for ZEC (Kogan et al., 2022a). A new asset does not start with a small, traceable set of users. It inherits the whole pool on its first day.

  3. There is no register to breach. Lists of holders, members, licensees and credential owners exist only as encrypted notes. A compromise of the issuing company’s servers does not expose them.

  4. The foundation is machine checked. Automated analysis makes finding flaws cheaper for attackers every year. A pool whose core circuit is formally verified leaves fewer flaws to find. Assets built on Ironwood inherit that work (Zcash, 2026).

  5. Assets gain a quantum recovery path. Assets issued on Ironwood would be recoverable into a future protocol under ZIP 2005. Assets issued on Orchard would not (Hopwood and Grigg, 2025; Hopwood, 2026).

5.2 Adoption of Zcash

  1. Private holdings with public supply. ZIP 227 enables transparent issuance only. The issuer key, the amounts issued and a global balance per asset are public, and an issuer can finalize an asset so that no more can ever be issued (Kogan et al., 2022b). This combination, privacy for holders and public proof of supply, is what a regulated issuer needs.

  2. A use for ZEC beyond payment. Fees for shielded asset transfers follow the existing fee mechanism of ZIP 317 (Kogan et al., 2022a; Hopwood, Nuttycombe and Grigg, 2022). Every asset transfer is paid for in ZEC.

  3. New builders and institutions. ZIP 227 provides for assets of limited or unlimited supply and for non fungible assets, which is what one of a kind records such as a title or a certificate require. ZIP 226 names application developers and asset custody institutions as the intended audience (Kogan et al., 2022a; Kogan et al., 2022b).

5.3 Illustrative uses

  • Professional licences and certifications
    Private: who holds which licence. Public: the number issued by the authority.

  • Identity and age credentials
    Private: the holder and when it is used. Public: that the issuer is genuine.

  • Memberships and access rights
    Private: the member list and activity. Public: the number of memberships issued.

  • Titles and certificates of ownership
    Private: who holds which title. Public: that each title exists once.

  • Tickets and entitlements
    Private: holder identity and transfers. Public: the number issued.

  • Software and content licences
    Private: the customer list. Public: total licences in existence.


6. Objections and responses

The drafts are written for Orchard.
Correct. Retargeting to Ironwood is the first piece of work. It likely needs a new transaction format, because version 6 is now used for Ironwood (Hopwood et al., 2026).

Ironwood is not quantum secure.
Correct, and ZIP 2005 says so itself. The claim made here is recoverability, not immunity (Hopwood and Grigg, 2025).

The formal verification is not finished.
Correct. The repository describes it as in progress (Zcash, 2026). Asset logic adds circuit changes, which is the reason to include it in that effort rather than add it afterwards.

Issuance is public.
By design. It is what makes supply auditable (Kogan et al., 2022b). Issuers who need a private identity must manage that through their issuing key.

Artificial intelligence helps defenders too.
It does, but defence must hold everywhere and attack needs one gap. Removing the readable ledger removes the prize, whichever side has better tools.

Privacy does not protect stolen keys.
Correct. Shielding removes the bulk ledger as a target. It does not remove the need for key security or for protecting identity data held off chain.

Issue notes are unencrypted, so the first recipient is public.
Correct (Kogan et al., 2022b). An issuer who mints directly to holders reveals them at mint. Issuers of sensitive records should mint to their own address and distribute by shielded transfer under ZIP 226. The retargeting work should state this as required practice, or consider shielded issuance.


7. Recommendations

  1. Assign shielded assets to a named network upgrade.
  2. Revise ZIP 226 and ZIP 227 to target the Ironwood pool.
  3. Extend the Lean formal verification to cover the asset circuit logic before activation.
  4. Confirm that every asset note on Ironwood is recoverable under ZIP 2005.

8. Conclusion

Zcash has a live shielded pool with a verified core in progress and a quantum recovery path, and a drafted asset protocol that keeps holdings private while supply stays provable. They are not yet joined.

Joining them changes what Zcash is for. A private payment network serves people who want to pay privately. A network that can hold any record of ownership or entitlement in a form that cannot be read in bulk, yet can still be proven and audited, serves every business that holds sensitive data and fears the day it leaks.

In an age when machines can read everything that is merely hidden, that is not a feature of a currency. It is a new kind of data protection, and the pieces Zcash needs to deliver it already exist.


References

Hopwood, D.-E. (2026) ZIP 326: NU6.3 Consequences for Wallets. Zcash Improvement Proposals. Available at: ZIP 326: NU6.3 Consequences for Wallets

Hopwood, D.-E. and Grigg, J. (2025) ZIP 2005: Ironwood Quantum Recoverability. Zcash Improvement Proposals. Available at: ZIP 2005: Ironwood Quantum Recoverability

Hopwood, D.-E., Nuttycombe, K. and Grigg, J. (2022) ZIP 317: Proportional Transfer Fee Mechanism. Zcash Improvement Proposals. Available at: ZIP 317: Proportional Transfer Fee Mechanism

Hopwood, D.-E., Nuttycombe, K., Ojha, D. and Bowe, S. (2026) ZIP 229: Version 6 Transaction Format. Zcash Improvement Proposals. Available at: ZIP 229: Version 6 Transaction Format

Kogan, P., Arte, V., Hopwood, D.-E. and Grigg, J. (2022a) ZIP 226: Transfer and Burn of Zcash Shielded Assets. Zcash Improvement Proposals. Available at: ZIP 226: Transfer and Burn of Zcash Shielded Assets

Kogan, P., Arte, V., Hopwood, D.-E. and Grigg, J. (2022b) ZIP 227: Issuance of Zcash Shielded Assets. Zcash Improvement Proposals. Available at: ZIP 227: Issuance of Zcash Shielded Assets

Zcash (2026) ironwood: Formal Verification of the Zcash Protocol and Documentation for the Ironwood Pool. GitHub repository. Available at: GitHub - zcash/ironwood: Formal Verification of the Zcash Protocol and Documentation for the Ironwood Pool · GitHub

All sources were read on 13 November 2026. Years are the creation dates shown on each proposal.

1 Like

is this good for radbro?

1 Like

Good for anyone with a record they’d rather not publish to the whole internet. So yep, radbro included.

Automatically removed quote of whole previous post.