Ethereum co-founder Vitalik Buterin has laid out a multi-year plan called “Lean Ethereum,” framing it as the network’s third major evolution after the Merge. He said core protocol components will be replaced step by step over the next three to four years, following discussions with researchers in Berlin and client teams in Svalbard.
Recursive STARKs replace direct transaction re-execution
The roadmap calls for Ethereum to adopt recursive STARKs as a native verification system instead of relying on direct transaction re-execution. Buterin also said the chain will swap out remaining cryptography that is vulnerable to quantum attacks and move toward quantum-safe alternatives. The scope is broad, but developers intend to keep disruption to existing applications as low as possible during the transition.
Consensus changes are also part of the plan. Ethereum aims to reach faster finality with a simpler design, while adding multidimensional gas and updating client architecture. Buterin said H-star, also known as Hegota, will likely be the last major upgrade before releases focused on Lean Ethereum begin.
Privacy and formal verification move to the center
Buterin said privacy should be treated as a primary protocol objective rather than an optional feature. That shifts attention to privacy requirements in transaction processing and in future protocol designs. He also described formal verification as a central security target. Work on evm-asm, he said, supports that goal by building a canonical proof system for the Ethereum Virtual Machine.
Those priorities show the roadmap is not limited to raw throughput. It also revisits how Ethereum verifies computation, secures its cryptographic base, and structures its execution layer.
Expanded state model sets long-range capacity targets
Another major proposal is a broader state model. Instead of depending only on today’s dynamic state, developers are exploring new scalable state types. According to Buterin, Ethereum could support roughly 2 terabytes of dynamic state and 100 terabytes of scalable state by 2030.
He said that design could cut transaction costs for ERC-20 tokens, NFTs, and many DeFi applications by more than 10x. That cost reduction is tied to a mix of changes across verification, state management, and execution architecture.
A second virtual machine is under consideration
Buterin also said Ethereum will eventually need another virtual machine alongside the EVM. Current candidates include RISC-V and leanISA. He said those architectures could improve scalability while supporting programmable privacy.
On nearer-term network changes, Ethereum will keep raising gas limits, expanding blob capacity, and reducing slot times in the coming years. Buterin added that the upcoming Glasterdam upgrade is expected to deliver a significant increase in the chain’s gas limit.