Ethereum Developers Weigh Scaling Upgrades Over Invariant Risks to 15,000 Contracts
Proposals EIP-8037 and EIP-8038 offer 200M block gas limits, but broken smart contract invariants have delayed testnets and divided core developers.
At a glance Proposed
- What happened
- Ethereum core developers delayed the Glamsterdam upgrade on the Hoodi testnet after audits revealed EIP-8037 and EIP-8038 break invariants in up to 15,000 smart contracts.
- Why it matters
- Excision of the repricing proposals would protect existing contracts but delay Ethereum's 3.3x scaling push and post-quantum roadmap by four to six months.
- Who is affected
- Ethereum core developers, node operators on testnets, and application developers with contracts relying on fixed gas assumptions.
- What's next
- Core developers will discuss whether to modify, retain, or drop EIP-8037 and EIP-8038 during the next execution-layer developer call.
- Primary source
- Ethereum Foundation Blog: EIP-8037/EIP-8038 Glamsterdam
Key points
- Core developer Marius van der Wijden estimates EIP-8037 and EIP-8038 could break between 10,000 and 15,000 smart contracts due to gas-repricing invariant shifts.
- The Oct. 26, 2026, Glamsterdam upgrade on the Hoodi testnet was canceled during ACDE #248, rolling back client gas schedules targeting 200M gas limits.
- Developers face a trade-off: stripping the repricing EIPs prevents contract breakage but threatens to delay Glamsterdam and post-quantum roadmaps by 4 to 6 months.
Ethereum core developers face an escalating technical dilemma over the upcoming Glamsterdam hard fork [1, 5]. At the center of the dispute are two gas repricing standards—EIP-8037 and EIP-8038—that are designed to raise Layer-1 throughput but risk breaking assumptions inside thousands of live smart contracts [1, 5].
At the time of writing, market sentiment remains cautious, with $ETH trading down 2.0% over 24 hours at $1,789 amid an Extreme Fear reading of 11 on the Fear & Greed index.
The tension highlights an ongoing trade-off in protocol development: whether to preserve absolute backward compatibility for existing decentralized applications or execute aggressive architectural revisions to scale Layer-1 capacity 5.
The Technical Mandate Behind Glamsterdam
Glamsterdam is designed to advance Ethereum's base-layer scaling roadmap by merging the Amsterdam execution layer upgrade with the Gloas consensus layer upgrade under meta-specification EIP-7773 [3, 4]. The hard fork brings two major structural components: enshrined proposer-builder separation (ePBS via EIP-7732) and block-level access lists (BALs via EIP-7928) [3, 4].
To complement these structural changes, the upgrade re-engineers block economics to expand capacity from 60 million gas to a targeted 200 million gas per block—roughly a 3.3x throughput increase [3, 5]. Achieving this higher limit safely requires recalibrating execution and storage costs 3. EIP-8037 introduces higher, separately metered gas costs for creating state, while EIP-8038 recalibrates state-access costs [3, 4]. These updates are paired with EIP-7778, which removes refunds from block gas accounting, and EIP-7976, which raises calldata floor pricing [3, 4].
Following its successful activation on the Sepolia testnet on Oct. 6, 2026, at epoch 353,024, the upgrade appeared on track for further deployment [3, 4]. Readers can review earlier milestones in Ethereum Glamsterdam Upgrade Activates on Sepolia Testnet.
Audits Reveal Invariant Failures
The scaling timeline encountered significant friction after security audits scrutinized the interactions of EIP-8037 and EIP-8038 [1, 2]. Security firm ChainSecurity alerted developers that several live protocols relied on specific execution invariants that these repricing proposals break 1.
Subsequent AI-assisted audits conducted jointly by ChainSecurity and Dedaub identified approximately 200 distinct classes of affected smart contracts 5. Ethereum Foundation developer Marius van der Wijden, speaking on a Nov. 16, 2026, core developer call, estimated that the changes could disrupt between 10,000 and 15,000 smart contracts 5. Van der Wijden characterized this as a "fringe subset" within Ethereum’s broader landscape of roughly 90 million contracts, with affected value loosely pegged in the low hundreds of thousands of dollars across deployments not yet committed to upgrading 5.
The root issue is that contracts relying on fixed gas stipends, strict assumptions about remaining gas, or hardcoded gas allowances fail when state access and creation costs increase [1, 3, 5]. On Nov. 6, 2026, the Ethereum Foundation issued an open call for action, releasing a repository cataloging eight invariants broken by Glamsterdam and providing prompts for developers to test both on-chain and off-chain code 1.
Testnet Delays on Hoodi
The discovery of these contract vulnerabilities triggered an immediate halt to the testnet rollout [1, 2]. During All Core Developers Execution call #248 (ACDE #248), client engineers agreed to cancel the Glamsterdam deployment on the Hoodi testnet, which had been scheduled to activate on Oct. 26, 2026, at 17:42:48 UTC (epoch 132,352) 2.
The decision rolled back the testnet configuration, removing both the Amsterdam and Gloas triggers alongside the gas schedule that would have raised Hoodi's gas limit to 200M [2, 8]. Across client implementations, pull requests unscheduled the fork [7, 8]. Besu reverted its scheduling commit in pull request #11709 7, while the upstream Hoodi testnet repository merged pull request #33, which set the fork epoch to a far-future value 8. Consensys and execution client teams, including Lodestar, Ethrex, Nethermind, and Besu, subsequently issued client releases stripping out the Oct. 26 activation instructions [2, 8].
The Ethereum Foundation clarified that the Hoodi postponement did not alter mainnet state, where an activation date has never been scheduled [2, 3].
Composability Versus Execution Velocity
The debate has exposed philosophical divides within Ethereum's engineering and application ecosystems 5. Application developers have strongly criticized pushing forward with changes that break existing code 5. Jango, builder of the Juicebox protocol, warned that breaking invariants damages protocol composability and shakes developer confidence in building immutable systems 5. A developer from Uniswap similarly urged that EIP-8037 be reconsidered 5.
Van der Wijden acknowledged during the Nov. 16 call that planned public outreach efforts had slipped following organizational restructurings and budget reductions at the Ethereum Foundation, noting he remains the last scaling lead at the organization 5.
Conversely, core protocol engineers have voiced concerns regarding the cost of inaction 5. Removing or overhauling EIP-8037 and EIP-8038 from the Glamsterdam hard fork would set back the scaling timeline by four to six months 5. Developers argue that such delays would cascade across Ethereum's broader engineering roadmap, notably pushing back vital research and implementation milestones for post-quantum cryptographic defenses at a time when quantum and artificial intelligence threats are advancing 5. More background on this debate is detailed in Ethereum Devs Debate Glamsterdam Upgrade Over Risk to 15,000 Smart Contracts.
What to Watch
Protocol participants are tracking several near-term junctures that will determine the fate of EIP-8037 and EIP-8038:
- Execution Layer Consensus: Core developers plan to revisit whether to retain, adjust, or abandon EIP-8037 and EIP-8038 during their next execution-layer developer call 5.
- Audit Feedback: The Ethereum Foundation continues to solicit data from application teams and independent security auditors through its dedicated Glamsterdam repository and review channel 1.
- Hoodi Testnet Rescheduling: A replacement activation epoch for the Hoodi testnet will remain unscheduled until core teams agree on a revised specification for gas accounting [1, 2].
Sources
- [1] EIP-8037/EIP-8038 Glamsterdam — Ethereum Foundation Blog, November 6, 2026
- [2] Glamsterdam on Hoodi Postponed — Ethereum Foundation Blog, October 22, 2026
- [3] Glamsterdam Hoodi Testnet Announcement — Ethereum Foundation Blog, October 16, 2026
- [4] Glamsterdam Testnet Announcement — Ethereum Foundation Blog, September 28, 2026
- [5] Ethereum Devs Weigh a Scaling Upgrade That Could Break 15,000 Apps — decrypt.co, November 17, 2026
- [6] Ethereum Developers Cancel Oct. 26 Glamsterdam Upgrade on Hoodi — thedefiant.io, October 22, 2026
- [7] Unschedule Amsterdam on Hoodi by daniellehrner · Pull Request #11709 · besu-eth/besu — github.com, October 22, 2026
- [8] chore: unschedule Hoodi Glamsterdam activation by barnabasbusa · Pull Request #33 · eth-clients/hoodi — github.com, October 22, 2026
Written by Basis Desk's newsroom system from the primary sources above and machine-verified against them before publication. Market figures marked "at the time of writing" come from live exchange data. Report an error: corrections@basisdesk.news · corrections policy.
The Daily Brief, in your inbox at 07:00 ET
Five stories, the numbers that moved, what to watch. Three minutes. No hype, no advice, unsubscribe in one click.
Get the big crypto stories first
A few alerts a day at most: major breaking news and the morning brief. Switch off anytime.
Not financial advice. Basis Desk publishes information, not recommendations. Crypto assets are volatile and you can lose what you invest.