Ethereum’s upcoming Glamsterdam upgrade will break wallets, gas estimators, and developer tools that rely on the long-standing assumption that every ETH transfer costs exactly 21,000 gas.
The Ethereum Foundation’s Protocol DevOps team issued the warning ahead of the Plataberget testnet activation on August 20, urging developers to test their applications before the changes advance to longer-running testnets like Sepolia and Hoodi. The alert focuses on a new “state gas” dimension that adds costs to transactions creating fresh Ethereum state, breaking the simple gas math that wallets and tools have relied on for years.
At the center of the disruption is EIP-8037. The proposal introduces a separate gas meter for operations that create new Ethereum state. An ETH transfer to an existing account still costs 21,000 gas. But sending ETH to an address that has never received a transaction will incur an additional state-gas charge, pushing the total to roughly 204,600 gas – nearly 10 times the current cost. Every “send max” flow and empty-wallet transaction could require far more gas than wallets currently estimate.
The reason this matters for ordinary users is straightforward. Most people copy-paste wallet addresses when sending ETH, and the first time they send to a fresh address, the transaction will cost dramatically more than expected. For developers building tools on top of Ethereum, the change means updating hardcoded assumptions throughout their codebases before the upgrade ships.
Gas repricing across the board
EIP-8037 sits within a broader bundle of gas repricing proposals tracked under EIP-8007. Together, these changes restructure how Ethereum prices storage and computation at the protocol level.
Storage writes jump from 2,800 to 10,000 gas – a 257% increase. Account writes rise from 6,700 to 9,000 gas. Cold account access increases from 2,600 to 3,000 gas. Contract creation costs go from 7,000 to 12,000 gas. The storage clear refund also rises from 4,800 to 11,616 gas, which partially offsets higher write costs for developers who clean up unused storage.
These numbers matter because many decentralized applications, smart contract wallets, and gasless transaction relayers have been optimized around current gas costs. A 257% jump in storage writes could change the economics of protocols that store large amounts of on-chain data or execute frequent storage updates. Developers working on account abstraction and smart contract wallets will need to recalculate their fee models entirely.
The repricing also affects DeFi protocols that perform frequent storage operations. Lending markets, automated market makers, and liquidation bots all depend on predictable gas costs. While the changes do not take effect until mainnet, developers are already modeling how higher storage costs will affect protocol profitability and user fees across the ecosystem.
Wallets face the most visible breakage
The practical impact hits wallet interfaces first. Most Ethereum wallets hardcode the 21,000 gas figure when estimating fees for ETH transfers. After Glamsterdam, wallets sending ETH to a new address will either produce fee estimates that are too low, causing transactions to fail, or too high, overcharging users.
The problem extends to “send max” buttons, which calculate how much ETH a user can transfer after subtracting gas. If the actual gas cost turns out higher than estimated, the remaining balance may not cover the fee and the transaction fails. Developers at several wallet teams have acknowledged the issue publicly, and testing on Plataberget has already surfaced edge cases where gas estimation returns incorrect values.
Gas estimation services and middleware providers face the same issue. Any tool that returns a single gas estimate based on the old 21,000 assumption will produce wrong numbers for transactions targeting new addresses. The Ethereum Foundation has recommended that developers test all gas-dependent logic on Plataberget before the upgrade advances to Sepolia.
What else ships with Glamsterdam
Beyond gas repricing, Glamsterdam bundles several protocol-level upgrades. Enshrined Proposer-Builder Separation (ePBS), defined in EIP-7732, restructures block production by separating proposing and building at the protocol level rather than relying on external relay infrastructure. Block-level access lists (EIP-7928) record which state locations get accessed during block execution, giving validators and MEV searchers better visibility into transaction costs.
The upgrade also raises Ethereum’s maximum deployed contract size from 24 KiB to 64 KiB and increases initcode limits from 48 KiB to 128 KiB. These changes give developers room for larger applications but require tooling updates for compilers, code analyzers, and block explorers that enforce current size limits.
Testing and timeline
Plataberget launched on August 13 as a public testing environment. Developers can submit validator and builder deposits through the testnet infrastructure. The Ethereum Foundation expects Plataberget to run for several months before changes advance to Sepolia and Hoodi, the longer-lived testnets that more closely mirror mainnet conditions.
The mainnet deployment target is the second half of 2026, with Q4 as the aspirational window. Ethereum.org lists Glamsterdam as the network’s major 2026 upgrade. The Foundation has said existing smart contracts should continue working after the fork, but developers need to review gas usage and update any tools that depend on fixed gas assumptions.
For ETH holders, no action is required. The hard fork will not change wallet addresses or balances. But anyone building on Ethereum – whether wallets, gas estimators, contract development tools, or decentralized applications – will need to update their gas calculations before Glamsterdam reaches mainnet. The months of testnet availability give developers a window, but the scope of changes means waiting until the last minute is risky.
discussion