Glamsterdam Upgrade May Break Wallets, Gas Tools — Test Now

Why Ethereum Developers Are Urging Immediate Testing

The Ethereum Foundation has issued a clear advisory to developers and infrastructure operators: the upcoming Glamsterdam upgrade could invalidate assumptions baked into wallets, indexers and gas estimation tools. With a major change to how state creation is metered under EIP-8037, some ETH transfers and other operations will no longer fit the old fixed-gas model. The foundation has opened the Platåberget public testnet ahead of the planned activation so teams can test software behavior before the changes propagate to Sepolia, Hoodi and eventually mainnet.

Quick summary

• Glamsterdam revises gas accounting and introduces state-gas metering (EIP-8037), shifting costs for operations that create new state.
• A simple ETH transfer to an existing account still uses 21,000 gas; transfers to brand-new accounts will include extra state-gas charges.
• Tools that assume a single fixed gas ceiling or hardcoded costs—wallets, indexers and gas estimators—may produce incorrect results or fail.
• Developers should run thorough tests on Platåberget, which is live and accepting validator and builder deposits, before the fork moves to longer-lived testnets.

What Glamsterdam changes about gas and state

Glamsterdam bundles a gas repricing package that redraws several assumptions about Ethereum’s per-operation costs and the effective block gas limit. Central to the change is EIP-8037, which creates a distinct state-gas dimension: operations that add permanent state (new accounts, storage writes, or deployed code) are metered by the bytes of state created and charged at runtime. This separates transient computational work from lasting state additions and allows the protocol to better account for long-term storage pressure on Layer 1.

The immediate implication is practical: software that assumes a single gas figure for all ETH transfers—or that relies on fixed ceilings and one-dimensional gas models—can return wrong estimates or fail to construct valid transactions. The Foundation’s Protocol DevOps team warned that any tool with hardcoded maximum gas assumptions "will break," explicitly naming wallets, indexers and gas estimators as at-risk components.

How costs change in practice

• Existing-account transfers: remain at 21,000 gas, composed of the base tx fee, cold-account access cost and the value-transfer cost.
• Transfers to new addresses: will carry extra state-gas charges for account creation, making those transfers significantly more expensive than 21,000 gas.
• Contract deployment and storage writes: subject to per-byte state charging; parameters tested in devnets suggest large multipliers compared with legacy costs (roughly an order of magnitude increase for some operations under earlier test parameters).

Developers should not assume that a wallet or gas estimator that returned safe fees before Glamsterdam will behave the same after the fork. Gas estimators that model only one gas dimension will underprice transactions that create state and may lead to failed or stuck transactions.

Platåberget testnet: the public sandbox

To give teams a meaningful window for testing, Platåberget launched as a public testnet on Aug. 13, with Glamsterdam slated to activate there on Aug. 20. The network is intended to remain live for several months so application teams, infrastructure providers and staking operators can exercise post-Glamsterdam behavior and surface issues before the upgrade moves to Sepolia and Hoodi.

The Protocol DevOps team has invited solo stakers, distributed validator technology projects, major staking providers and custom operators to try validator and builder-deposit workflows on Platåberget. Public validator and builder deposits are part of the testing process, and the testnet’s open participation model aims to catch a broad set of real-world deployment patterns.

Clients and builds available for testing

To reduce friction for testers, container images and tagged releases are available for multiple consensus and execution clients. Among the consensus (beacon) clients: Lighthouse, Lodestar, Nimbus, Prysm, Teku and Grandine. Among execution clients: Besu, Geth, Erigon, Nethermind, Reth, NimbusEL and Ethrex. Teams are encouraged to use client releases that fit their operational posture while client teams finalize builds for mainnet compatibility.

Protocol-level changes beyond gas repricing

Glamsterdam is not limited to gas changes. The fork also enshrines proposer-builder separation (ePBS) into protocol rules, formalizing the split between block builders and block proposers. ePBS introduces a builder API flow and timeliness checks for payloads, altering how blocks are proposed and validated. Infrastructure that participates in block production, relays, or MEV-related flows should evaluate compatibility and monitor payload-timeliness logic carefully.

Another core innovation is Block-Level Access Lists (BAL). BAL records the state locations a block will access during execution and the post-transaction state changes, with this data stored separately from the block body and shared between execution-layer peers via the eth/71 networking protocol. By giving clients advance visibility into which accounts and storage slots a block will touch, BAL enables preloading of required data and helps parallelize processing when transactions touch disjoint state—improving throughput under the new pricing regime.

Contract size and forward-compatible consensus data

Glamsterdam raises the maximum deployed contract size from 24 KiB to 64 KiB and increases allowed initcode from 48 KiB to 128 KiB. These changes aim to keep large, code-heavy applications (including DeFi systems) deployable despite higher state costs. The fork also adds forward-compatible data structures at the consensus layer to smooth later upgrades.

Recommended developer actions

1. Test on Platåberget now: deploy apps, submit ETH transfers to new addresses, deploy contracts, write storage paths and observe gas metering changes.
2. Audit gas estimators: ensure tools model both execution gas and the new state-gas dimension introduced by EIP-8037.
3. Update wallets and user flows: wallets must detect when a transfer creates a new account and surface the correct fee to users; guardrails for insufficient gas must be tightened.
4. Inspect indexers and explorers: indexers that rely on assumptions about fixed costs or block gas limits should be verified against the new block gas floor and per-operation pricing.
5. Exercise block-production stacks: validators, builders and relays should validate ePBS behavior and payload timeliness under stress scenarios.
6. Monitor client releases: coordinate with client teams and use tagged or containerized client builds recommended for Platåberget testing.

These steps are vital to avoid user-facing regressions when Glamsterdam reaches long-lived testnets and, eventually, mainnet.

Parameters, performance and the path to mainnet

EIP-8037 and surrounding repricing proposals were near final-draft status earlier in the development cycle. Parameter tuning has been focused on limiting annual state growth while preserving deployability for important smart contracts. Under some May parameters, new account creation became approximately 8.5x more expensive and contract deployment roughly 10x higher in state-related charges; these numbers emphasize the importance of testing particularly for applications that frequently write permanent state.

The Foundation has outlined a staged rollout: development networks (like the initial Platåberget instance), followed by a non-finality devnet to exercise tougher consensus scenarios, then longer-lived testnets such as Sepolia and Hoodi, with mainnet activation contingent on stable client implementations and robust test coverage.

Future forks and state-pricing recalibration

Work on the next scheduled fork (Hegotá) is already in progress. Developers have tabled dozens of proposals for 2027, many of which revisit gas and state pricing as Layer 1 capacity goals increase. For example, EIP-8368 proposes a recalibration of state-creation pricing if block gas limits grow beyond the reference level used by Glamsterdam, and other proposals explore paths toward higher block gas limits (e.g., a 600 million gas limit scenario). These efforts signal that state-pricing is a multi-step journey rather than a one-off adjustment.

Conclusion: Plan, test, and adapt

Glamsterdam represents a significant evolution in how Ethereum balances computation, storage and long-term state growth. The change is deliberate: by charging for permanent state at a per-byte rate, the network can better align costs with the long-term resource burden of Layer 1. However, the practical effect will ripple across wallets, gas tooling, indexers and block-production infrastructure.

Teams should treat Platåberget as a critical opportunity to discover regressions and harden their stacks. The combination of EIP-8037 state metering, ePBS, BAL and larger contract-size limits means both risk and potential—if teams update gas estimation logic and adapt wallet UX, users will get clearer fee signals and infrastructure will be better prepared for future throughput improvements. If teams don’t act, wallets and gas tools that rely on old assumptions risk breaking when Glamsterdam moves to longer-lived testnets and mainnet.

Developers and infrastructure operators: run comprehensive tests now, coordinate with client teams, and monitor protocol updates closely as Glamsterdam progresses through Platåberget, Sepolia and Hoodi toward mainnet activation.