Ethereum devs warn Glamsterdam upgrade could break wallets and gas tools

The Ethereum Foundation has warned that wallets, indexers and gas estimators could break as Glamsterdam changes the 21,000-gas assumption for some ETH transfers, with the upgrade due to activate on the Platåberget testnet on Aug. 20.

The Ethereum Foundation’s Protocol DevOps team said on Aug. 17 that any tool relying on a hardcoded maximum gas limit “will break,” naming wallets, indexers and gas estimators among the software likely to be affected. The team urged application and infrastructure developers to test their systems on Platåberget, a public testnet designed to stay online for several months.

Forkcast data shows Platåberget launched on Aug. 13, giving developers an early environment for Glamsterdam before the upgrade moves to Sepolia and Hoodi. The Glamsterdam fork is scheduled to activate on Platåberget on Aug. 20, according to the foundation, with public validator and builder deposits available as part of the testing process.

Glamsterdam could break hardcoded gas assumptions

Under the planned gas repricing package, the foundation said software can no longer safely assume that Ethereum has a single fixed gas ceiling or that common operations will continue to cost the same amount in every case. The changes are designed around a roughly 200 million gas floor and alter the price of individual operations as well as assumptions tied to the block gas limit.

For application developers, the immediate issue is software that sets fixed boundaries when estimating transaction costs. The Protocol DevOps team said such systems need to be reviewed before Glamsterdam reaches mainnet because the repricing touches wallets, indexers and gas estimators across the network.

The warning follows June 17 Glamsterdam upgrade coverage from crypto.news, which reported that Ethereum developers were already testing the full set of planned EIPs on development networks. At the time, Ethereum Foundation developer Parithosh Jayanthi said the upgrade would change the cost of actions on Ethereum, with high-level computation becoming cheaper while state becomes more expensive.

During the same testing phase, Jayanthi said developers had made “massive progress” but noted that no fixed mainnet timeline had been set. Deployment would depend on the results of testing and whether Ethereum client teams were ready to support the new rules.

EIP-8037 changes how new state is priced

A central part of the warning concerns EIP-8037, which introduces a separate state-gas dimension for operations that create new state. The foundation said creating an account, deploying code or writing a new storage slot will be metered at a fixed cost per state byte and charged at runtime.

Because of that change, a basic ETH transfer will not always carry the same gas cost. Sending ETH to an account that already exists will continue to cost 21,000 gas, with the amount broken into the base transaction cost, cold account access and the value-transfer cost. Sending funds to an address that does not yet exist will also incur a state-gas charge tied to creating the new account.

Developers should therefore revisit applications that treat 21,000 gas as sufficient for every ETH transfer, the Protocol DevOps team said. Gas estimators built around only one gas dimension may also return incorrect estimates once new state is metered separately.

EIP-8037 had already moved close to its final form by May. A May 11 protocol development report said the proposal had reached final-draft status and was being parameterised on a Glamsterdam development network. At the time, its cost-per-state-byte model was designed around limiting annual state growth to roughly 60 GiB at a 300 million gas block limit.

Under the parameters reported in May, new account creation could become roughly 8.5 times more expensive, while contract deployment costs could rise about tenfold. Separate metering for code deposits was designed to keep large contracts deployable, including code-heavy decentralised finance applications.

The state-gas model also changes where Ethereum accounts for the cost of permanent state. The foundation said account creation, new storage slots and deployed code will incur charges based on the amount of new state created, making applications that frequently add permanent data particularly important targets for testing before mainnet deployment.

Ethereum Glamsterdam upgrade also changes block production

Gas repricing is only one part of Glamsterdam. The foundation said the fork also includes enshrined proposer-builder separation, or ePBS, which changes how blocks are built, proposed and validated inside Ethereum’s core protocol.

Under ePBS, the split between the block-building process and the proposer role is incorporated into the protocol, alongside a new builder API flow and payload-timeliness checks. Infrastructure tied to Ethereum’s block-production and validation pipeline should expect to be affected, according to the foundation.

With Platåberget open for public participation, the Protocol DevOps team has encouraged solo stakers, distributed validator technology projects, custom software operators and large staking providers to test their infrastructure. The testnet allows users to deposit new validators and experiment with validator and builder-deposit workflows before the same changes move to longer-lived networks.

Block-Level Access Lists form another major component of the fork. The foundation said the lists will record state locations accessed during execution and post-transaction state changes, with BAL data stored separately from the block body and exchanged between execution-layer peers through the eth/71 networking protocol.

Earlier June reporting said the access-list design gives Ethereum clients advance information about which accounts and smart-contract data a block will use. The system can allow nodes to preload required data and process transactions in parallel when transactions do not access the same state, according to Ethereum.org.

Glamsterdam will also increase size limits for deployed contracts and initialisation code. The foundation said the maximum deployed contract size will rise from 24 KiB to 64 KiB, while the maximum initcode size will increase from 48 KiB to 128 KiB. Forward-compatible consensus data structures are also included in the planned fork.

Platåberget gives developers a longer testing window

Unlike the shorter development networks used during earlier Glamsterdam work, Platåberget is intended to remain available for several months. The Protocol DevOps team said the longer lifespan should give developers time to test post-Glamsterdam behaviour and identify failures before the changes reach Sepolia and Hoodi.

Its validator set is relatively small but open to public participation. For the initial testing period, the foundation has listed container images for consensus clients including Lighthouse, Lodestar, Nimbus, Prysm, Teku and Grandine, alongside execution clients including Besu, Geth, Erigon, Nethermind, Reth, NimbusEL and Ethrex. Tagged client releases remain optional while development teams prepare their own builds.

After feedback from Platåberget has been incorporated into specifications and client software, a non-finality devnet is expected to follow within the month to test difficult consensus scenarios, according to the foundation. Sepolia and Hoodi are due to receive Glamsterdam after the development networks remain stable, while Ethereum mainnet activation will follow successful upgrades on the long-lived testnets.

Development on Ethereum’s next scheduled fork is also proceeding separately. An Aug. 16 Hegotá planning report said developers were considering 66 proposals for the 2027 upgrade, although Fork Choice enforced Inclusion Lists was the only EIP formally scheduled for inclusion at the time.

Several proposals under review for Hegotá concern future gas and state pricing as Ethereum increases Layer 1 capacity. EIP-8368, for example, would recalibrate state-creation pricing if the block gas limit rises beyond the reference level used by Glamsterdam, while developers have discussed preparing Ethereum for a possible path towards a 600 million gas limit.