A validity rollup on Ethereum

Aztec vs
Starknet

Starknet executes general-purpose contracts publicly; its STRK20 privacy pool gives a designated company access to viewing keys. Aztec lets applications define private and public logic without a network-wide viewing key.

QuestionsStarknetAztec

Permissioned sequencing

StarkWare runs all of Starknet's sequencers, so the company can halt the network, and has: on 5 January 2026 it paused block production for more than four hours and dropped 18 minutes of activity. L2BEAT lists permissioned sequencing and Stage 1 maturity, with an emergency upgrade path that has no exit window.

No one can stop it

Permissionless sequencers and provers run the network. Upgrades follow public governance rules with a timelock. L2BEAT assesses the rollup at Stage 2.

Audit key held by a company

STRK20 viewing keys are encrypted to a master key in trusted enclaves, with multisig backups: a designated party can read shielded activity on request.

No backdoor

Private functions execute on the user's device and the network verifies a proof. There is no network-wide decryption key, no validator that re-executes private inputs, and no co-signer. Public calls and their inputs are visible by design.

Public contracts, private pool

Cairo contracts execute publicly. STRK20 adds a privacy pool with its own disclosure model, not private contract logic.

Private + public contracts

Noir contracts combine private state and functions with public ones, and anyone can deploy them. Contracts compose on Aztec and reach Ethereum through L1/L2 messaging.

Permissionless to use

Contract deployment is open; sequencing remains permissioned.

Permissionless to use, build, and operate

Anyone can deploy contracts or participate as a sequencer or prover under the published protocol, staking, and hardware requirements.

A compliance layer built into STRK20: viewing keys are escrowed with an audit-key operator, which supports scoped disclosure for lawful requests.

Selective disclosure at the application and user level. An application decides which facts it proves or publishes, and a user can share their own viewing keys with whoever they choose. No protocol-wide audit key, screening gate, or operator sits in the path of a transaction.

What Starknet does well.

General-purpose Cairo contracts and STARK-based scaling on Ethereum, with STRK20 supporting scoped disclosure for lawful requests.

The difference. For STRK20, disclosure runs through a designated audit-key operator and there are no smart contracts with private logic. On Aztec, the application defines disclosure without a protocol-wide audit key and the platform supports smart contracts with private logic.

If something here no longer holds, tell us and it will be corrected. Tell us

Keep comparing.

  1. Aztec vs NEAR

    A sharded proof-of-stake Layer 1

  2. Aztec vs Miden

    A client-side-proving rollup, spun out of Polygon

  3. Aztec vs Zcash

    A proof-of-work Layer 1 for private money

  4. Aztec vs Zama

    An FHE confidentiality layer for public EVM chains

  5. Aztec vs Tempo

    A payments-first Layer 1, incubated by Stripe and Paradigm

Privacy shouldn't depend on permission.