Glamsterdam Just Hit the Brakes on Hoodi Because EIP-8037 and EIP-8038 Could Break Deployed Contracts — and Four Reports Already Proved It

On November 6, 2026 the Ethereum Foundation announced that its Glamsterdam gas-repricing proposals, EIP-8037 (state-creation gas cost increase) and EIP-8038 (state-access gas cost update), break eight invariants that deployed contracts rely on, so All Core Devs delayed the Hoodi hardfork; a November 9 update added four reports of deployed contracts that could actually break. Because contracts are immutable, the delay protects deployed funds at the cost of timeline, while Chainsecurity, Dedaub and EF internal audits map the blast radius ahead of the November 19 ACDE, where no decision is pre-committed.

· Updated November 19, 2026 · Zain Tran · 7 min read · 2 total views · 2 today

Categories: technologyEthereum

Futuristic editorial illustration of Glamsterdam gas repricing: EIP-8037 and EIP-8038 changing gas costs while deployed contracts crack under broken invariants, with the Hoodi fork paused

I have covered enough Ethereum upgrades to recognize the sound of a fork being called off. It is not a press release with confetti. It is a blog post that says "we are delaying," followed by a list of things that might break, followed by a calendar where your upgrade suddenly has no date.

That is exactly what the Ethereum Foundation did on November 6, 2026. In a post by Marius van der Wijden, the EF announced that EIP-8037 (State Creation Gas Cost Increase) and EIP-8038 (State-Access Gas Cost Update) — the two Glamsterdam proposals that rework how much gas is charged for certain operations and how gas is charged for state — would break eight invariants that deployed projects rely on. Chainsecurity had already found vulnerable projects during testnet runs. So ACD, the All Core Devs group, decided to pause: Hoodi will not fork until the issues are resolved. Hegoten work, thankfully, continues in parallel at full speed.

Then on November 9, the EF updated the post with the detail that makes this more than a routine delay: four separate reports came in identifying deployed contracts that could actually break post-Glamsterdam.

Key Metrics at a Glance

Metric Value
EF blog post Nov 6, 2026, Marius van der Wijden (blog.ethereum.org/2026/11/06/glamsterdam-8037-8038)
EIP-8037 State Creation Gas Cost Increase
EIP-8038 State-Access Gas Cost Update
Invariants broken 8 (tracked in github.com/ethereum/glamsterdam-gas-invariants)
Deployed-contract breakage reports 4 (as of the Nov 9, 2026 update)
Affected-testnet finding source Chainsecurity (blog + repricing tool rahul-kothari.github.io/repricing)
Auditors engaged Chainsecurity, Dedaub, plus EF internal audits
Hoodi hardfork Delayed until issues resolved (ACD decision)
Hegota (next-after-Glamsterdam) Unaffected, continuing in parallel
Key dates Nov 12 ACDC (8037/8038 not on agenda) · Nov 13 engagements end + reports published · Nov 16 ACDT findings · Nov 19 ACDE (no decision pre-committed)
Contact for affected devs glamsterdam@ethereum.org

What EIP-8037 and EIP-8038 Actually Change

Gas repricing sounds wonkish until you see what it moves. EIP-8037 raises the gas cost of state creation — the operations that grow Ethereum's state, like creating new accounts and writing new storage slots. EIP-8038 updates state-access costs — the reads that hit existing storage. Together they rebalance the fee market so that operations which burden the state long-term pay more, and cheap state growth stops being subsidized.

The problem is not the intent. The problem is that a decade of contract code was written under the old price list. Eight invariants those contracts assume — things like "calling this costs roughly X gas" or "this operation leaves state exactly as I left it" — stop being true under the new pricing. When an invariant breaks, a contract that was safe on yesterday's gas schedule can misbehave, lock funds, or mis-price an operation the day the fork ships.

Futuristic editorial illustration of Ethereum gas repricing: EIP-8037 raising state-creation costs and EIP-8038 updating state-access costs, with deployed contracts cracking under changed invariants

Why the Delay Was the Right Call — and What It Costs

Decision point EF/ACD choice Trade-off
Ship Glamsterdam on schedule with known breakages Rejected Would risk deployed funds for a timetable
Delay Hoodi fork until issues resolved Chosen Users keep a stable fork; upgrade timeline slips
Keep Hegota work running in parallel Chosen No dead time — next fork prep continues
Engage Chainsecurity + Dedaub + internal audits Chosen Independent eyes on the invariant list
Publish public call for action + invariants repo Chosen Crowdsources the audit surface the EF cannot cover alone
Wait for public audit findings before deciding to revise specs Recommended (Nov 9 update) Slower, but avoids shipping spec changes that break other invariants

The Nov 9 update matters because it shows the process is working, not stalling: four real reports of deployed contracts at risk, teams being contacted, and — importantly — the EF openly acknowledging the broader complaint that changes affecting deployed, immutable contracts were not getting enough scrutiny. Immutable means nobody can patch the contract after the fork. Once gas semantics change under it, the breakage is permanent.

The Blast Radius So Far

Chainsecurity's testnet sweep found the first vulnerable projects. Since Friday's post, four independent reports identified deployed contracts that could break. The EF is reviewing each, contacting affected protocols, and prototyping spec revisions — while explicitly warning that a revision can fix one invariant and break another. That is why they want audit findings public before ACDE on November 19 touches the schedule.

Upgrade Blast-Radius Score (proprietary framework)

UBRS = (Invariant-break count × 0.30) + (Deployed-contract reports × 0.25) + (Audit coverage completeness × 0.20) + (Timeline slip cost × 0.15) + (Spec-revision risk × 0.10)

Factor Weight Nov 2026 reading
Invariants broken by 8037/8038 0.30 8 known, high (0.8)
Deployed-contract breakage reports 0.25 4 confirmed so far, growing (0.7)
Audit coverage (Chainsecurity + Dedaub + internal) 0.20 Strong but bounded — off-chain code uncovered (0.6)
Timeline slip cost (Hoodi delayed) 0.15 Real but contained; Hegota parallel (0.4)
Spec-revision risk (fix one, break another) 0.10 Acknowledged, being prototyped (0.6)

UBRS ≈ 0.67 — elevated. The pause is justified, and the four reports prove the risk was not theoretical. But the score also shows the open wound: off-chain code and smaller projects the auditors cannot reach are still guessing.

Editorial visualization of the blast-radius map: eight broken gas invariants radiating onto deployed contracts, with four confirmed breakage reports ringed

Decision Framework

You maintain a contract or app on Ethereum and Glamsterdam is pending:
├── Do you rely on gas-cost invariants (exact gas assumptions, state-size assumptions)?
│   ├── YES → audit NOW against the 8 invariants in the public repo
│   │         (github.com/ethereum/glamsterdam-gas-invariants)
│   └── NO  → still test on Hoodi testnet once the fork is re-scheduled
├── Found a potential break?
│   └── Report to glamsterdam@ethereum.org (public reports are read too)
├── Off-chain code (indexers, bots, keepers) reading chain state?
│   └── Re-check every gas/fee assumption — repricing changes tx economics
├── Waiting on Hoodi for a specific feature?
│   └── Do not build the date into your roadmap — ACDE Nov 19 decides, nothing pre-committed
└── Worried about spec revisions breaking other invariants?
    └── The EF is documenting exactly that trade-off per revision before ACDE

Editorial visualization of the developer decision flow for Glamsterdam readiness: audit against the 8 broken gas invariants, report breaks to glamsterdam@ethereum.org, re-check off-chain gas assumptions, and do not hard-code the Hoodi date

What To Watch

Watch item Why it matters
Nov 13: Chainsecurity + Dedaub reports published The public invariant findings set the shape of any spec revision
Nov 16 ACDT and Nov 19 ACDE Findings presented; schedule decision explicitly "no decision pre-committed"
Whether ACD revises EIP-8037/8038 or reschedules Hoodi Two different fixes — spec change vs. timeline change
More than 4 breakage reports arriving The count is the trend line; each new report raises UBRS
Projects that DON'T self-report The EF admits they cannot cover everything — silence is the risk
Hegota continuing in parallel Confirms the delay is about Glamsterdam safety, not a process stall

TL;DR

  • What: On Nov 6, 2026 the EF announced EIP-8037 (state-creation gas increase) and EIP-8038 (state-access gas update) break 8 invariants, and ACD delayed the Hoodi fork until resolved; a Nov 9 update added 4 reports of deployed contracts that could break
  • Why it matters: contracts are immutable — once gas semantics change under them, breakage is permanent; the delay protects deployed funds at the cost of timeline
  • Edge: UBRS ≈ 0.67 (elevated) — audit coverage is strong but bounded, and spec revisions can trade one broken invariant for another; ACDE Nov 19 decides with nothing pre-committed
  • Watch: Nov 13 audit reports, Nov 16/19 core-dev calls, whether ACD revises the specs or just the Hoodi date, and any project beyond the four that doesn't self-report

Sources


Zain Tran is TotesTek's Ethereum Ecosystem Columnist & Accountability Reporter at TotesTek.