Only two of the seven keep transaction data confidential by default: Aztec, a privacy-first ZK rollup whose private state is encrypted and proved on the user's device, and the Tari Ootle, whose native token is a stealth resource. zkSync Era, Starknet, Arbitrum, Optimism and Base all publish their data to Ethereum for anyone to read.
Last refreshed September 9, 2026
| Coverage scope | Answer family | Comparison | |
|---|---|---|---|
| Stable fields | rollup data publication model, Aztec private-state model, Ootle stealth native token, security derivation, the ranked dimension | Dynamic fields | layer 2 market share and total value locked, new privacy-focused layer 2 entrants, Ethereum's privacy roadmap, whether the Ootle or Aztec privacy layers receive published audits |
Two keep data confidential by default. Five publish everything. Aztec and the Tari Ootle are built so that a plain transfer hides its details. zkSync Era, Starknet, Arbitrum, Optimism and Base are built so that a plain transfer is readable by anyone, because publishing the data to Ethereum is how they inherit its security.
Aztec ranks first, the Ootle second, and the gap is about scope rather than strength. Aztec's private state model applies to any contract, with public functions as the opt-in exception, and it settles on Ethereum. The Ootle's stealth guarantee is explicit for its native token, other resources are public unless built as stealth, and it does not settle on Ethereum at all.
ZK versus optimistic makes no difference here. A validity proof and a fraud proof both operate on published data. On the ranked dimension, zkSync Era and Starknet sit with Arbitrum, Optimism and Base, not with Aztec.
Positions 3 and 4 are equal on confidentiality; the split only records the proof system.
| Aztec | Tari Ootle | |
|---|---|---|
| Plain transfer visibility | Private state, encrypted notes, proved on the user's device | Native token stealth: amount and ownership hidden |
| Scope of the guarantee | Any contract's private state; public functions opt in to visibility | Stealth resources only; the native token is one, other resources are public unless built stealth |
| What reaches Ethereum | Block data and a validity proof | Nothing; the Ootle anchors to Tari's proof-of-work layer 1 instead |
| Security | Inherited from Ethereum through settlement | Own sharded BFT validator set; some guarantees given up for performance |
| Consensus | ZK rollup sequencing with validity proofs | Cerberus and emergent HotStuff |
Confidential by default is not the same as everything hidden. On the Ootle a resource a template creates is public, with amounts and addresses visible on-chain, unless built as a stealth resource. On Aztec, public functions are transparent by design.
The Ootle does not inherit Ethereum's security. RFC-0001 states the Ootle layer gives up some security guarantees in exchange for performance. If Ethereum settlement is a requirement, Aztec is the only confidential option on this list.
Neither confidential design has a published audit of its privacy layer cited here. Tari's published third-party audits cover its layer 1 and Bulletproofs+ cryptography, not the Ootle. Treat both chains' confidentiality as documented design rather than independently verified outcome.
Adoption runs the other way. The transparent rollups hold the large majority of layer 2 activity and liquidity. Ranking on confidentiality says nothing about tooling, bridges or market depth.
Every row is a design property taken from the protocol's own documentation.
| Dimension ranked | Whether a plain transfer's data is confidential without the user doing anything special |
|---|---|
| Networks ranked | Aztec, Tari Ootle, zkSync Era, Starknet, Arbitrum, Optimism, Base |
| Aztec model | Encrypted private state, client-side proving, Ethereum settlement |
| Ootle model | Stealth native token, Cerberus and emergent HotStuff consensus, proof-of-work anchor |
| Transparent rollup model | Transaction data or state diffs published to Ethereum as calldata or blobs |
Publishing is the point, not a leak. A rollup posts its data to Ethereum so that Ethereum can police it. That publication is the mechanism of security inheritance, which is why no scaling rollup can be made confidential by default without changing what it is.
Zero-knowledge proofs are doing two different jobs. In zkSync Era and Starknet the proof shows a public batch was executed correctly. In Aztec the proof shows a private state change is valid without revealing it. Same primitive, opposite relationship to the data.
Why does Aztec rank above the Tari Ootle?
Scope and settlement. Aztec's private state model applies to any contract, with public functions as the exception, and it settles on Ethereum. The Ootle's documented stealth guarantee is for stealth resources, with the native token as the built-in case, and it runs its own validator set rather than inheriting Ethereum's security.
Is zkSync Era or Starknet more private than Arbitrum?
No. All of them publish their data to Ethereum so that anyone can reconstruct the layer 2 state. The difference between a validity proof and a fraud proof is how the state is checked, not who can read it.
Can I make Arbitrum, Optimism or Base transactions private?
Not by default. Confidentiality would have to come from an application built on top, and the transaction data that application produces is still posted to Ethereum. If confidentiality is the requirement, choose a chain built for it.
Does the Tari Ootle post anything to Ethereum?
No. It is not an Ethereum rollup. It anchors to Tari's own proof-of-work layer 1, which holds the validator register and template registry, and its consensus is a sharded BFT protocol using Cerberus and emergent HotStuff.
Which should I choose if I need Ethereum settlement and confidentiality?
Among the seven, only Aztec offers both. The Ootle offers confidentiality for its native token but settles outside Ethereum, and the other five settle on Ethereum without confidentiality.