A contentious proposal to change Ethereum’s gas fee structure, EIP-8037, has forced a reckoning across the ecosystem, prompting developers to audit their applications for potentially critical billing and security flaws.
The move comes after Ethereum’s core developers, including Marius van der Wijden of the Ethereum Foundation (EF), formally requested community-wide testing following the indefinite postponement of the “Glamsterdam” network upgrade, which was slated for a testnet activation on October 26, 2026.
Glamsterdam upgrade halted over Ethereum’s EIP-8037 gas breaking changes
The core of the issue lies in how EIP-8037 aims to manage the network’s ever-expanding state size. This has become a pressing concern for long-term scalability. However, the proposed solution introduces complex new accounting rules that could disrupt a wide range of decentralized applications (dApps), particularly those relying on gas-metering for transaction reimbursement.
The decision to pause the Glamsterdam upgrade, announced on October 22, 2026, underscored the severity of the concerns. The upgrade, which includes EIP-8037 and the related EIP-8038, was intended to tackle blockchain bloat but raised red flags during internal reviews and audits by security firms ChainSecurity and Dedaub. These reviews found that the changes could have unintended consequences for existing smart contracts.
In a November 6 blog post, van der Wijden acknowledged that internal audits couldn’t possibly account for every live contract on the network. He published an eight-point framework of gas assumptions and urged dApp developers and security auditors to test their systems against it.
The appeal is a call to arms for the community to help determine if the proposal itself needs to be amended before it can be safely deployed.
The feedback will be crucial in shaping the final specifications. The delay continues with no new mainnet activation date set, as developers weigh the benefits of controlling state growth against the risk of breaking the applications that make the network valuable. An Ethereum software bug discovered earlier highlighted the need for caution, as even small issues can require emergency fixes across many clients.
How new state-gas rules could break dApp billing
The primary source of disruption from EIP-8037 is its introduction of a two-dimensional gas system: “execution-gas” for computation and “state-gas” for creating new data on the blockchain. When a transaction’s gas limit exceeds 16,777,216, it funds a separate state-gas reservoir. This reservoir is used first to pay for state creation costs.
Crucially, Solidity’s widely used `gasleft()` function, which contracts use to measure gas consumption, does not account for this separate reservoir. This creates a billing blind spot. A service provider, known as a relayer, might execute a transaction for a user, measure the gas used with `gasleft()`, and find the cost is much lower than what was actually consumed from the hidden state-gas pool.
This could lead to relayers consistently undercharging users and operating at a loss. In other scenarios, attempts to calculate costs could fail entirely. If a contract performs a calculation where the final `gasleft()` reading is higher than the initial one (due to gas refunds), a standard “checked subtraction” would revert the entire transaction. An “unchecked” one could lead to a massive, erroneous reimbursement.
The official guidance suggests projects move towards pre-agreed, user-signed fees based on estimates under the new rules. But for systems that require on-the-fly metering, the proposed changes present a significant architectural challenge that developers must now solve.
The soaring cost of creating on-chain data
The driving force behind these disruptive changes is the urgent need to control Ethereum’s state growth. As of January 2026, the state database for a Geth node was already around 390 GiB.
After a previous gas limit increase, the average daily new state reportedly grew from 105 MiB to 326 MiB, an annual rate of about 116 GiB. Without intervention, this growth would only accelerate, threatening decentralization by making it more expensive to run a full node.
EIP-8037 tackles this by dramatically increasing the cost of state creation. Under the proposal, writing to a new “cold” storage slot would cost 110,020 total gas, a nearly five-fold increase from the 22,100 gas it costs under the current Osaka rules. This total is comprised of 12,100 execution gas and a hefty 97,920 state gas.
While painful for developers in the short term, the goal is to force more efficient use of on-chain storage and ensure the network remains sustainable. Managing the chain’s data footprint is a key part of understanding Ethereum’s supply dynamics and long-term economic health.
The proposal calibrates the cost to target a more manageable state growth of around 120 GiB per year, even with higher block gas limits.
The trade-off is clear: developers must now re-architect their applications to be far more conservative with data storage, or their users will face significantly higher transaction fees. This shift in economic incentives is a core feature, not a bug, of the proposal.
Unintended security risks and wallet vulnerabilities
Beyond the economic impacts, security researchers have warned that the gas repricing could introduce new vulnerabilities. An October 13 report from ChainSecurity highlighted 13 patterns in live contracts that could be susceptible to “insufficient-gas griefing,” where an attacker could exploit the new gas refund mechanics to permanently block a contract’s function.
An even more subtle risk involves EIP-7702, a proposal for smart wallet delegation. Under the new gas rules, a transaction could fail during its pre-execution phase due to insufficient gas but still be included in a block. This failure could roll back the nonce update tied to a user’s authorization but not the authorization itself.
This would leave a valid, reusable authorization exposed on-chain. An attacker could then front-run the legitimate user, replay the authorization, and potentially seize ownership of a newly initialized smart contract wallet. While authorization front-running is a known risk in public mempools, this new scenario could expose even privately submitted transactions.
These security findings add another layer of complexity to the upgrade decision. The potential for such critical vulnerabilities makes it clear why core developers are proceeding with extreme caution and demanding a thorough review from the entire community.
The path forward for EIP-8037
The fate of EIP-8037 now rests on the feedback from the dApp developer community. Projects are being asked to report what breaks and, crucially, how their systems can be repaired—whether through simple setting changes, contract upgrades, or more complex data migrations. This information will help core developers gauge the true cost of the upgrade for the ecosystem.
Some protocol-level fixes are already being discussed. One proposal would prevent the `gasleft()` function from ever increasing during a transaction’s execution, which would solve the subtraction issues for relayers. However, it wouldn’t fix the core problem of the state-gas reservoir remaining invisible to metering logic.
Ultimately, the ongoing debate is a real-world test of Ethereum’s governance and development philosophy. It balances the need for long-term sustainability and scalability against the promise of backward compatibility. The process highlights the importance of robust testing and Ethereum client diversity to secure the network.
For now, dApp developers face a period of uncertainty and necessary rework as they adapt to a future with more expensive state.