Gas repricing study: Why might transactions still fail after raising the gas budget?
The Glamsterdam repricing study published by the Ethereum Foundation in August replayed historical transactions under the new rules, distinguishing between transactions that succeeded after their gas limit was raised and those that still failed even after it was raised significantly. This is a historical replay study, not a report of a current mainnet failure.
The second category involves fixed gas assumptions in contracts—for example, a hard-coded budget for a call, or choosing different branches based on the remaining gas. My understanding is that increasing a transaction’s total budget may not change the fixed amount of gas a contract forwards to an internal call, and may not preserve the original branch behavior.
So fixes should address the point of failure. Wallets and RPC providers need to update their estimation rules; contract developers should inspect internal calls and fixed constants, and test under the new rules. Attributing every failure to “the user didn’t provide enough gas” can make it easy to overlook assumptions in the code.
One more testing detail: a successful replay may still involve changes in gas consumption or other execution details. When studying compatibility, compare the results and execution paths—not just the number of successful transactions.
This article does not claim that any particular protocol has already been affected, nor does it provide a mainnet upgrade completion time or operational instructions.
Source: https://blog.ethereum.org/2026/08/24/glamsterdam-repricing-testing
The Glamsterdam repricing study published by the Ethereum Foundation in August replayed historical transactions under the new rules, distinguishing between transactions that succeeded after their gas limit was raised and those that still failed even after it was raised significantly. This is a historical replay study, not a report of a current mainnet failure.
The second category involves fixed gas assumptions in contracts—for example, a hard-coded budget for a call, or choosing different branches based on the remaining gas. My understanding is that increasing a transaction’s total budget may not change the fixed amount of gas a contract forwards to an internal call, and may not preserve the original branch behavior.
So fixes should address the point of failure. Wallets and RPC providers need to update their estimation rules; contract developers should inspect internal calls and fixed constants, and test under the new rules. Attributing every failure to “the user didn’t provide enough gas” can make it easy to overlook assumptions in the code.
One more testing detail: a successful replay may still involve changes in gas consumption or other execution details. When studying compatibility, compare the results and execution paths—not just the number of successful transactions.
This article does not claim that any particular protocol has already been affected, nor does it provide a mainnet upgrade completion time or operational instructions.
Source: https://blog.ethereum.org/2026/08/24/glamsterdam-repricing-testing