The Race for Private Ethereum | Zac Williamson

Aztec is trying to build a “private Ethereum”: a smart-contract network where users can create assets and use composable applications without exposing sensitive data. Zac Williamson says Aztec uses zero-knowledge proofs to let contracts work with private data, identities, and credentials. He argues that third parties can make private transactions easier to operate, but also gain powers to monitor or censor users. Aztec’s protocol has no viewing keys, although individual applications can add them. Williamson describes Aztec as the only privacy-preserving smart-contract blockchain besides Zcash without protocol backdoors or centralized actors and auditors able to control or censor users, while acknowledging that others may dispute the claim.

Privacy could help bring regulated assets into open DeFi without making users’ identities and portfolios public. Tokenized mortgages and bonds often rely on custodians to identify users and enforce compliance; proprietary networks and other L2s can offer those checks but sacrifice open composability. Aztec’s aim is to keep compliance data encrypted on chain so applications can use it without revealing it. Williamson says the necessary identity processes will largely have to be built from scratch, so this will take time. A stablecoin seeking integration with Circle, for example, might be asked to include a viewing key for compliance monitoring through TRM.

Aztec’s zk passport app reads a passport chip, matches its photo to a live face, and generates a proof without sending passport data off the phone. A smart contract can use that proof to check a limited claim, such as citizenship or absence from a sanctions list, without learning other personal information. Williamson says credentials could be queried across applications without linking them to a user’s crypto accounts. He contrasts this with conventional KYC apps that transmit passport information to third parties, where it may be exposed in a hack. Private credentials could support compliant tokenized assets while keeping users’ holdings and trading activity out of public analytics.

Williamson expects privacy to become necessary for many real-world assets and traditional finance activities, but says users often choose cheaper products and better experiences over privacy, in part because they may not understand how their data is harvested. He argues that vertically integrated apps can silo users and liquidity, while privacy can protect data without giving up composability. Potential uses he names include private deals, AI inference payments, identity and reputation systems, undercollateralized lending, and a stablecoin with private issuance and settlement. Private transactions may still require some public information for price discovery; Aztec is exploring multiparty computation as a way to address that constraint.

Aztec is focused first on Web3 users and hardening the network. Williamson says institutional adoption depends on a turnkey, white-label product that meets traditional security expectations, and that more engineering work was needed. He describes Aztec as intended to launch, and now live, as a Stage 2 rollup; the host characterized it as a fully decentralized Stage 2 Ethereum L2 with a proof-of-stake token. Aztec launched in April, and the Sierra upgrade was expected to have landed by the time the episode aired. Williamson says major protocol components have been formally verified and that he wants the entire protocol verified, with the upgrade also adding formal verification and AI code audits.

Privacy adds prover entities that generate block proofs, but Williamson characterizes the current overhead as relatively light: a transaction proof takes about five seconds in a web browser on his laptop and about one second in a native app. A governance change corrected an overpayment to provers, and planned work aims to lower proving costs and privacy-related operating expenses, potentially enabling faster, near-real-time block production. The network provides soft finality in roughly 6–12 seconds; hard finality takes about 40 minutes, with a five-minute target. For withdrawals, Aztec plans to let users pay an extra 10 or 20 cents for an immediately generated proof to be put on chain. Its token pays gas, secures the network through validator staking, and supports governance.

Williamson warns that AI is making it cheaper to analyze visible on-chain data, increasing the risks of public transaction histories, copy trading, and data exposure. He sees blockchains as built for adversarial conditions but stresses that they remain vulnerable to attacks, and calls security and formal verification essential. David hopes private transactions grow from about 1% today to at least 10% by the end of 2027. Williamson’s broader bet is that privacy will win where public ledgers cannot support sensitive data, though adoption depends on making private applications usable and affordable.