Ethereum Researchers Detail "Lean" Rebuild, Betting Throughput and Quantum Resistance Can Coexist

Over the past several days, Ethereum's core researchers published the technical designs for what Vitalik Buterin, the network's co-founder, has described as its biggest rebuild since the 2022 transition to proof of stake. The community now calls the program Lean Ethereum. The plan reorganizes the protocol into three parts: a redesigned consensus layer, a leaner data layer with post-quantum features, and a minimal execution engine built to be easily verified. CoinDesk reported that Buterin describes the effort as Ethereum's third major version.

A post titled The Extremely Lean Chain describes a consensus design built around single-slot finality and recursive signature aggregation using STARKs, a cryptographic proof method (a Scalable Transparent Argument of Knowledge) intended to make validator signatures resistant to future quantum computers. A companion piece on Lean Execution argues for a smaller, verifiable virtual machine and pricing computation across multiple dimensions. A separate thread on native unspent-transaction-output accounting addresses the problem that every address on Ethereum today adds permanent data the network must store.

The stated ambition is large. Public reporting on the roadmap cites targets of roughly 10,000 transactions per second on the base layer and around one million across layer-2 networks. The first concrete step, the Glamsterdam upgrade, is scheduled for the second half of 2026. Several researchers have said publicly that the proposed three-to-four-year timeline is too slow, and the plan comes as the Ethereum Foundation restructures its staff and budget.

Measured against Ethereum's history, what matters most is not the throughput target but the order of the work. Ethereum is proposing to combine scaling, verifiability, and quantum resistance in a single architecture rather than adding them separately later. The premise is that simpler, mathematically provable components scale better than accumulated complexity.