Vitalik Buterin Introduces Ultra-Fast ZK-EVM GKR Protocol

On Oct 20, 2025 at 10:10 am UTC by · 2 mins read

Ethereum co-founder Vitalik Buterin has introduced GKR, which he described as a family of protocols for fast proving.

Ethereum co-founder Vitalik Buterin introduced the GKR protocol, which comprises innovations that facilitate very fast proof systems. He described these solutions as a family of protocols that is behind the extreme speed of many operational proving systems. Buterin focused on showing GKR implementation for proving Poseidon hashes.

GKR Is Not Zero-knowledge

Buterin started his post by acknowledging the existence of ultra-fast ZK-provers. These are ZK-EVM provers that manifest the Ethereum Layer-1 in real-time and do this with just 50 consumer GPUs.

“People proving 2 million Poseidon hashes per second on consumer laptops, and zk-ML systems proving LLM inference with increasing speed,” Buterin noted.

Here is where the GKR protocol comes in. It serves as the powerhouse of these provers. This solution is designed to enable fast ZK-EVM and zk-ML proofs. It is also optimized for computations with multiple low-degree layers and many repeated function applications, including hashing and neural networks.

All that it requires are commitments to inputs and outputs, not intermediate layers. In the long run, it ends up improving efficiency significantly. By “commitments,” Buterin was referring to putting the data into a cryptographic data structure. This could be achieved either through a KZG or a Merkle tree, as long as it allows a user to prove queries about specific things about that data.

Users must be careful not to mistake the GKR for a zero-knowledge protocol, seeing that it handles succinctness and not privacy. However, zero knowledge can still be achieved with the GKR, as all that needs to be done is to wrap the proof in a ZK-SNARK or ZK-STARK.

Ethereum Co-founder Criticizes Agentic AI

Among the many protocols that Buterin has been discussing in recent times is Artificial Intelligence (AI) models. Two months ago, he criticized the AI industry for focusing intensely on “agentic” models. In his opinion, there are too many autonomous systems, and this is not good.

He noted that systems allowing greater human input produce better and safer results, aligning with earlier comments by former Tesla AI director Andrej Karpathy. Apparently, the Ethereum co-founder is more enthusiastic towards open-weight AI models, which he claims have strong editing capabilities.

In September, he also emphasized that there is great danger attached to over-reliance on AI for governance.

Share:

Related Articles

Ethereum Price Prediction: ETH Traders Quietly Lost Millions in ‘Sandwich Attacks’ – New Signal Hints at a Safer, More Bullish ETH

By December 5th, 2025

A new study has revealed that Ethereum traders lost millions to sandwich attacks over the past year, quietly draining value from everyday users and favoring stealthy exploiters.

ETH to $62,000? Tom Lee Is Ultra Bullish but Traders Watch These Numbers

By December 5th, 2025

BitMine’s Tom Lee predicts an ultra-bullish climb toward $62,000 for Ethereum following the Fusaka upgrade on Dec. 3.

Ethereum Price Holds $3,000 Level as Vitalik Confirms Fusaka Upgrade Success

By December 4th, 2025

Ethereum broke above $3,000 following the Fusaka upgrade’s successful deployment of PeerDAS, marking progress toward long-anticipated sharding capabilities.

Exit mobile version