Vitalik Buterin unveils GKR: a faster, more scalable zero-knowledge protocol

GKR reduces ZK-proof costs by only committing to inputs and outputs and skipping all the heavy intermediate steps.

Ethereum co-founder Vitalik Buterin has introduced a cryptographic protocol called GKR, short for Goldwasser-Kalai-Rothblum, to make zero-knowledge (ZK) proofs much faster and more scalable.

The protocol will reduce the cost of computing ZK proofs, which are the foundation for privacy and scalability solutions across the Ethereum ecosystem.

How GKR accelerates proofing without heavy obligations

In a detailed tutorial on his personal blog, Buterin described how GKR works to make computations involving more than one data layer, such as cryptographic hashing or neural network inference, more efficient.

According to him, traditional systems like STARKs require users to commit to every step in a computation, which means performing hundreds of hash operations on every byte of data. However, GKR skips these interim commitments entirely, instead focusing only on the inputs and outputs, vastly reducing both time and resources.

The core mechanism uses a mathematical process known as sumcheck. This allows the prover and verifier to check complicated calculations by looking at just a few randomly chosen points rather than the entire data set.

While GKR itself does not provide privacy, it can be built into existing ZK-SNARK or ZK-STARK systems to provide zero-knowledge guarantees. In his tutorial, Buterin demonstrated an implementation where GKR proved millions of Poseidon2 hash functions in parallel, a structure common in blockchain operations and even AI workloads.

The theoretical efficiency of the protocol is said to be very high, with the Ethereum co-founder estimating that it could save fifteen times the overhead compared to regular STARKs. Real-world testing has shown improvements of just under 10x, meaning GKR-based systems can audit massive cryptographic calculations for far less money and hardware than is currently required.

You might also like:

What GKR means for Ethereum and beyond

GKR’s launch comes just weeks after Buterin praised the progress of the Lean Ethereum team, which is developing a minimal zkVM for long-term scalability.

His new protocol could have a major impact on how blockchain scales and how cryptographic computing works. For Ethereum, it strengthens the foundation for ZK rollups, which are networks that batch transactions off-chain and post small proofs to the main chain, making them faster and cheaper to execute.

Beyond blockchain, Buterin noted that GKR’s design is suitable for machine learning proofs, where large models need to demonstrate correct outputs without revealing the underlying data. Researchers are already exploring how GKR can be used in zk-ML inference and general proof systems.

The programmer’s recent defense of Ethereum’s 45-day queue also reflects his general belief that long-term network security and trust should come before short-term convenience, a principle used in GKR’s design to keep it both robust and scalable.