Hegota marks the end of pre-Lean Ethereum

Vitalik Buterin identifies Hegota as the final thematically "pre-Lean" fork before the network undergoes a three to four year rebuild. This "Lean Ethereum" effort aims to replace almost every major piece of the protocol, including how the network confirms transaction validity, how the chain stores data, and how it achieves quantum safety. Quantum safety shifted up in priority, requiring a new design for blobs to protect Layer 2 networks. Hegota, the second upgrade of 2026, has a proposal deadline of October 4, 2026, and a selection date of November 3, 2026. The PFI deadline passed on August 6, 2026, while the CFI deadline arrives on December 10, 2026. I recommend following this timeline because it separates the current architecture from the massive rebuild. The roadmap contains a large gas limit increase in the Glamsterdam upgrade to provide more room for transactions in each block. This new storage tier holds 50 times more data than the current system, allowing tokens and DeFi apps to use a cheaper tier where transaction fees fall by more than 10x. The plan keeps the existing system for complex applications like Uniswap while adding the cheaper tier for simpler ones.

EOF implementation dependency struggles

The EOFv1 rollout (EIP3540) combined with EIP5450 (StackCode Validation) provides a path to improve backward compatibility. To lower dependencies, researchers suggest launching EIP3540 and EIP5450 together without requiring EIP4200 or EIP4750. This strategy allows the network to update EOFv1 later through non-breaking changes. However, the EIP5450 code validation algorithm remains a struggle because it relies on static branching. Because the current code validation rules under EIP5450 rely on static branching, the stack validation algorithm requires that unreachable code is not permitted, which means that EIP3540 cannot launch without including both EIP4200 and EIP4750 in the same fork. This dependency means EOFv1 cannot launch without including static branching opcodes like RJUMP or subroutine calls like CALLF and RETF. A better approach involves a single linear scan where every byte in a code section is either a valid opcode or part of the immediate operands of an instruction. I find the push for a single linear scan more logical than the current complex worklist method. If the network adopts the proposal to permit unreachable instructions, EIP4200 or EIP4750 can deploy as non-breaking changes later.

Development shifts and roadmap delays

Glamsterdam has not yet activated despite an original expectation for the first half of 2026. This delay occurs while Ethereum faces a narrowing lead in the developer ecosystem. As of July 2026, Ethereum maintains 7,700 monthly active developers, but Solana possesses 2,500. Arbitrum has 999 monthly active developers and Optimism has 637. You know how quickly these numbers shift when developers choose higher performance or better programming models like Rust. The Ethereum Foundation’s June 2026 restructuring cut 54 staff members, which represents 20% of the workforce. This reduction in personnel might slow the implementation of the Lean Ethereum roadmap. The transition involves replacing the current EVM with alternative computing formats like RISC-V or leanISA. This massive rebuild involves a new storage tier. The roadmap also addresses fragmentation from Layer 2 solutions, such as the 22 rollups and seven validiums listed on L2BEAT. The foundation wants to make cross-L2 transactions invisible or efficient to stop the developer exodus. Will the smaller team meet the ambitious goals of the three to four year rebuild?