Ethereum’s plan to triple network speed could silently break millions of existing smart contracts

Ethereum’s upcoming “Glamsterdam” upgrade aims to increase base‑layer throughput roughly threefold by aligning gas charges with the actual resources each operation consumes. While the proposal promises higher capacity, it also raises gas costs for state‑creation actions, which could cause failures for existing contracts that rely on current gas assumptions.

How the upgrade changes gas pricing

The two pending EIPs (8037 and 8038) introduce a separate “state‑gas” dimension. Under the candidate schedule, creating a new account, storage slot, or deploying contract code will cost significantly more gas than today. For example, the charge to create a new account would rise from 25,000 gas to 183,600 gas, and deploying 24 KiB of code would increase from 4,947,200 gas to 37,784,880 gas.

Impact on state growth

EIP‑8037 estimates that, at a 200 million gas‑limit scenario, annual growth of persistent state could reach about 387 GiB, potentially crossing a 650 GiB performance threshold within a year. The proposal targets an average growth of 120 GiB per year at a 150 million gas limit, with a worst‑case estimate of 160 GiB per year.

Replay testing results

Researchers replayed 929,731,274 transactions from Dec 2024 to Jun 2026 against both the current and candidate schedules. Findings include:

• Under EIP‑8037, 174,473,898 transactions failed at their original gas limit but succeeded with higher limits; 2,687,652 were classified as potentially broken.
• Under EIP‑8038, 84,708,228 transactions were fixable with higher limits; 3,036,537 were potentially broken.

The “potentially broken” group consists of transactions that were successful under the current schedule but failed even when allowed up to ten times the original gas limit under the new schedule. These failures often involve fixed gas stipends, hard‑coded gas forwarding, or logic that depends on gasleft().

Who needs to prepare

Wallets, RPC providers, indexers, node tooling, and gas estimators must update to incorporate the new gas rules. Front‑end applications, bundlers, and ERC‑4337 smart‑account stacks (e.g., ZeroDev, Alchemy) also need to revise assumptions about gas limits and internal call behavior. Developers will need to resimulate transactions, review fixed stipends, and test affected code paths against the candidate schedule.

Timeline

The Glamsterdam fork was initially scheduled for Aug 20 and reported live with the repricing schedule by Aug 24. The upgrade will progress from stable devnets to Sepolia and Hoodi testnets before a mainnet deployment targeted for the fourth quarter of 2026, though exact dates remain unannounced.

Conclusion

Ethereum can increase throughput by pricing state growth more accurately, but the change creates a narrow window for builders to adapt. Contracts and tooling that rely on legacy gas assumptions must be audited and possibly migrated to avoid silent failures after the upgrade.