Vitalik Buterin highlights need for new cryptographic tools in blockchain

Vitalik Buterin stood before attendees at the 12th Global Blockchain Summit in Shanghai and delivered a message that was equal parts technical roadmap and warning siren: the privacy assumptions that blockchain was built on are broken, and fixing them requires an entirely new cryptographic toolkit.

The Ethereum co-founder’s core argument is straightforward. Blockchains currently require every node to re-execute every computation, a design choice that worked fine in 2015 but now looks increasingly like a bottleneck. The alternative, Buterin said, is to lean hard into cryptographic proofs, specifically SNARKs, STARKs, fully homomorphic encryption (FHE), and indistinguishability obfuscation (iO), to verify computations without re-running them.

Why pseudonymity is no longer enough

He argued that traditional pseudonymity, the idea that using a wallet address instead of your name keeps you private, is now essentially a polite fiction. Advances in AI and data analytics have made de-anonymizing blockchain users significantly easier. If a sufficiently motivated actor wants to figure out who’s behind a wallet, the tools to do so already exist and are getting better fast.

Buterin’s proposed solution is what he called “programmable privacy,” a framework where modern cryptographic primitives bake privacy directly into the protocol layer rather than relying on users to obscure their own tracks.

The numbers behind the proposal

EIP-8288 is one of the more concrete examples. The proposal aims to use recursive STARKs for processing mempool transactions, which could slash the gas cost of quantum-safe private transactions from roughly 10 million gas to somewhere in the low tens of thousands.

Buterin also shared his own probability estimates for the broader cryptographic roadmap. Based on an analysis from September 6, 2026, he placed a 60% likelihood that SNARKs, FHE, and iO will achieve less than 10 times the computational overhead of standard processes by the end of the decade. Among these technologies, he expects SNARKs to hit that efficiency threshold first.

The broader initiative falls under what Buterin described as “Lean Ethereum,” Ethereum’s third major iteration. This version of the protocol would integrate post-quantum cryptography and recursive STARKs directly into the base layer, rather than bolting them on as afterthoughts.

A quick primer on the cryptographic alphabet soup

SNARKs (Succinct Non-interactive Arguments of Knowledge) let one party prove to another that a computation was done correctly, without revealing the underlying data or requiring the verifier to redo the work. They’re already used in several blockchain projects and are the most mature of the technologies Buterin discussed.

STARKs (Scalable Transparent Arguments of Knowledge) are similar to SNARKs but don’t require a trusted setup ceremony, making them arguably more decentralized. They also hold up better against quantum computing attacks, which is why they feature prominently in Ethereum’s forward-looking plans.

Fully homomorphic encryption allows computations to be performed on encrypted data without ever decrypting it. It’s powerful but currently very expensive computationally.

Indistinguishability obfuscation is a technique for making a computer program’s code unintelligible while preserving its functionality. Practical implementations remain elusive, but if achieved, iO could enable entirely new classes of privacy-preserving applications.

Buterin’s 60% confidence estimate is worth taking seriously precisely because it’s not 95%. He’s signaling that these technologies are promising but not guaranteed to hit their targets on schedule.