News analysis ·
Ethereum upgrade trade-off: Glamsterdam vs 15,000 contracts
Ethereum developers weigh shipping EIP-8037 and EIP-8038 in Glamsterdam even if up to 15,000 contracts break. How the gas change breaks code and who decides.

Ethereum developers are now openly weighing whether the Glamsterdam upgrade should ship two gas proposals that could leave up to 15,000 smart contracts broken. Decrypt reported the estimate from a core developer call on Monday, where no decision was taken. This article looks at why a pricing change can break working code, who carries the cost, and what the developers have to choose between.
The change in one paragraph
Glamsterdam is Ethereum's next planned hard fork. Two of its proposals, EIP-8037 and EIP-8038, change how much gas certain operations cost. Gas is the unit Ethereum uses to measure computing work, and every block has a cap on how much gas it can hold. The aim is to lift that cap, today 60 million gas, so a block can hold up to 200 million gas, a jump Decrypt put at roughly 3.3x. Developers say this would make most transactions cheaper while keeping the growth of chain state, the data every node must store, under control. The catch is that the repricing changes rules some deployed contracts quietly depend on.
Under the hood: how a gas change breaks working code
A smart contract is code that cannot easily be changed once it is deployed. Its authors write it against the protocol as it exists at that moment. Some of those authors treat certain behaviours as permanent. Developers call these invariants: assumptions that the protocol is expected to keep true. A fork that changes them can turn a correct contract into a broken one without a single line of its code changing.
Gas costs can be a hidden invariant. These are general illustrations of how a repricing can break code, not findings from the audits:
- Fixed gas budgets: a contract forwards a set amount of gas to another call. If the operations inside that call become more expensive, the same budget no longer covers them and the call fails.
- Cost-based safety checks: some contracts assume an operation is too expensive to repeat many times in one transaction, and build their protections on that. Changing the price changes the math behind the protection.
- Immutability: a contract with no upgrade path cannot be patched. Its users can only move away from it.
This is why the audit findings matter. Security firms ChainSecurity and Dedaub ran AI-assisted reviews and, together, found roughly 200 distinct classes of affected contracts. According to Decrypt, most of them break because operations get more expensive, not because of the new mechanics the proposals add.
Marius van der Wijden, a developer at the Ethereum Foundation and part of its scaling work, put it directly: "Multiple contracts were relying on invariants in the protocol that are modified by 8037 and 8038".
Who is affected, and how much
Van der Wijden's highest estimate is 10,000 to 15,000 contracts. He called that a small fringe group out of about 90 million contracts on Ethereum. He also gave a loose figure for the value at risk: low hundreds of thousands of dollars, counting only contracts whose teams have not yet committed to upgrading.
Jango, who builds the Juicebox protocol, argued that the change weakens composability, the ability of contracts to plug into each other and rely on each other's behaviour. In his view it also makes builders less sure that immutable protocols are safe to build. A developer from Uniswap said EIP-8037 should be reconsidered. For most users the intended effect is cheaper transactions; the risk sits with users of contracts that rely on old gas costs.
Why the problem appeared so late
The invariant issue surfaced late in development. Van der Wijden said the team had intended wide public outreach to contract developers, but timelines slipped while the Ethereum Foundation went through reorganizations and budget cuts. As the last scaling lead still at the foundation, he accepted much of the blame.
A proposal can be sound on its own terms and still collide with code written years earlier. Finding those collisions means searching deployed contracts and warning their owners.
Ethereum upgrade status and next steps
The Ethereum Foundation had already postponed the Glamsterdam testnet fork because of these gas-related security flaws. On Monday, developers stressed that nothing has been decided. The question goes to the execution-layer call on Thursday.
Several developers want to ship the fork as planned. They argue that pulling the two EIPs could delay Glamsterdam by four to six months, and with it the post-quantum security roadmap, and they said Ethereum cannot afford that delay given growing quantum and AI threats. On the other side, Jango argues the change damages developer confidence in building immutable protocols, and a Uniswap developer said EIP-8037 should be reconsidered. Whatever the call decides, teams running contracts that depend on gas costs should check their exposure now rather than after the fork date is set.
Sources
- Decrypt, 17 Nov 2026
AI wrote this summary from the primary documents and reports cited above. It went live automatically after checks that figures match those sources and nothing is copied; there was no human review. It explains a protocol or rule change and is not investment advice.