Vitalik Buterin Unveils Ethereum Roadmap to Replace EVM With RISC-V and Accelerate Development With AI

Ethereum co-founder Vitalik Buterin believes the community should redesign the application layer while preserving core principles such as decentralization, openness and censorship resistance. He is exploring combinations of AI agents, zero-knowledge proofs, or ZK, and prediction markets that could reshape how wallets, DeFi, stable assets and oracles operate.

As of July 19, 2026, Buterin had outlined three directions: replacing traditional wallets with AI agents, using ZK to protect API calls and user behavior, and turning DeFi into a general-purpose futures market. He also envisioned using AI, prediction markets and large language models to create customized stable assets and strengthen decentralized oracles. The proposals remain conceptual, with no implementation date or investment amount disclosed.

Ethereum nodes must run both consensus-layer and execution-layer software to validate transactions and track the state of the Beacon Chain. Each program has its own configuration and data interfaces, increasing the maintenance burden. Vitalik Buterin argues that making nodes easier to operate would enable more ordinary users to verify the network themselves, reduce reliance on third-party providers such as Infura, and strengthen decentralization and censorship resistance.

Buterin recently proposed integrating the backend software that connects nodes to the Beacon Chain with Ethereum's execution layer in a unified code structure, simplifying installation, synchronization and maintenance. The idea remains a technical proposal, with no formal activation date, version number or funding amount announced. Ethereum core developers must still assess its specifications, compatibility and implementation timetable.

Ethereum co-founder Vitalik Buterin proposed a new scaling roadmap on February 27, 2026, shifting the focus from the network's recent reliance on Layer 2 rollups toward simultaneously expanding its Layer 1 base layer. In the short term, Glamsterdam would use parallel validation and ePBS to make better use of each 12-second slot. Over the longer term, ZK-EVMs and blobs would increase throughput without raising the barrier to operating a node.

The Ethereum Foundation said on May 11, 2026, that Glamsterdam's multi-client development network was running ePBS reliably and that pricing under EIP-8037 had been finalized. It set a post-upgrade minimum gas-limit target of 200 million, more than triple the roughly 60 million at the time. The mainnet upgrade, originally planned for June, is now expected to be delayed until the third quarter of 2026. ePBS can extend the data-propagation window from about two seconds to roughly nine seconds.