- Ethereum’s clear path
- Three paths
Ethereum’s next evolution will be shaped by a set of upcoming invariants and protocol caps that Vitalik Buterin outlined. These are deep structural changes that harden the network, streamline clients and block entire classes of DoS vectors.
Ethereum’s clear path
Zooming out over the past few years reveals a distinct pattern: Ethereum continues to move toward stringent, predictable limitations on the capabilities of a single transaction or block. And that trend is going to pick up speed.
The groundwork was already in place. In 2021, EIP-2929 and 3529 increased SLOAD gas costs and slashed refunds, reducing disk I/O abuse and preventing refund-based spam loops. One of the most exploitable instructions in the EVM was eliminated in 2024 by Dencuns SELFDESTRUCT nerf, which closed significant complexity gaps. Now, in 2025, the 16777216 gas-per-transaction hard cap finishes the cycle: no more super-dense monster transactions capable of locking up nodes or stressing clients in unpredictable ways.
Each of these constraints trims the attack surface while pushing Ethereum closer to a system where worst-case behavior is strictly bounded.
Three paths
The first path is a cap on the number of contract code bytes accessed per transaction. In the short term, this means it becomes more costly to call large contracts. In the medium term, it standardizes contract scaling and eliminates pathological situations where a single call thrashes through megabytes of bytecode, pushing the ecosystem toward binary trees and per-chunk pricing.
Second, ZK-EVM prover cycle bounds are required by Ethereum. Repricing proofs are becoming more and more important as ZK-based layer 2s become more common. Without restrictions, block builders could create consensus-layer bottlenecks by packing proofs with excessive computational overhead. Bound it, and the network benefits from safer L2 growth and predictable verification costs.
Third, there will be changes to memory prices. Although EVM memory expansion is currently quasi-bounded, attackers can still push clients into uncomfortable areas. Every client team can easily handle worst-case modeling, and execution engines are made simpler with a more transparent hard cap on memory usage.
Source: https://u.today/3-crucial-ethereum-updates-in-the-future-revealed-by-vitalik-buterin
