Ethereum Glamsterdam Upgrade Exposes Unexpected Payment Behavior
The Ethereum Glamsterdam upgrade is facing renewed scrutiny after developer testing uncovered an unusual problem: a blockchain transaction can appear successful even when an important payment inside it has failed.
The findings emerged from simulations involving Juicebox V6, a smart contract system designed to manage project funding, payments, and token distributions.
Tests conducted on November 14, 2026, identified four execution paths where the overall transaction succeeded but an internal payment did not complete as intended.
That distinction matters because decentralized applications often rely on multiple operations happening within a single transaction. A user might purchase an NFT, distribute project revenue, or withdraw funds while the application simultaneously processes fees and payments to other parties.
If one of those internal operations fails without reversing the overall transaction, the financial result can differ from what the application was designed to guarantee.
The discovery does not establish that Ethereum's live network has been exploited. Instead, it highlights a potential compatibility problem with proposed changes to how Ethereum calculates gas costs.
What Did Developers Discover During Testing?
The issue was reported through developer-published simulations of Juicebox V6 contracts running under proposed Glamsterdam execution rules on the Sepolia test network.
Developer mejango used Ethereum's functionality to examine how deployed smart contract code behaves under different transaction gas limits.
The simulations identified several concerning outcomes.
1. NFT purchases without the expected payment
One test demonstrated a situation where a buyer could retain an NFT even though a related payment to another project failed.
This does not mean every NFT marketplace is vulnerable. The behavior was observed in a specific tested contract path.
Nevertheless, it raises questions about applications that assume completing an NFT-related transaction also guarantees delivery of associated payments.
2. Project payouts that fail internally
Another scenario involved project revenue distributions.
Under the tested conditions, a transaction could complete while a payment intended for another project failed. The amount remained with the paying project rather than reaching its intended recipient.
For applications managing multiple beneficiaries, this could create unexpected accounting differences.
3. Reserved-token distribution problems
The tests also identified a reserved-token distribution path in which the intended cross-project payment failed, while the beneficiary received raw tokens instead.
This illustrates how changes to execution costs can influence more than transaction fees. They can also alter the final outcome of application-specific financial operations.
4. Withdrawals completing without fees
Perhaps the clearest example involved a cash-out operation.
In the developer's reported Sepolia simulations, gas limits between approximately 679,622 and 817,622 allowed a withdrawal to complete while its associated 2.5% fee payment failed.
The withdrawal succeeded, but the fee was effectively forgiven.
These findings concern specific simulated execution paths rather than a confirmed production exploit.
How EIP-8037 and EIP-8038 Are Connected to the Problem
Two Ethereum Improvement Proposals are central to the current discussion.
EIP-8037 proposes changes to the cost of creating persistent blockchain state, including separating aspects of state creation from execution gas accounting.
EIP-8038 addresses gas costs associated with accessing and modifying blockchain state.
These proposals are intended to improve Ethereum's resource accounting and help the network scale more efficiently.
However, changing the relative costs of smart contract operations can have unintended consequences.
An application developed under one set of gas assumptions might behave differently after those costs change, particularly when its internal calls have limited gas and its error-handling logic permits partial completion.
That is why the Juicebox findings are more than a conventional software bug report. They raise a question about backward compatibility: should an Ethereum upgrade preserve the practical behavior of existing contracts, even when those contracts rely on complicated gas-sensitive execution paths?
The Ethereum Foundation has acknowledged the need to review compatibility findings before finalizing the relevant specifications and testing plans.
Ethereum Mainnet vs Glamsterdam Sepolia: What the Tests Show
The developer also compared the outcomes with sampled transactions under existing mainnet execution rules.
Test scenario | Glamsterdam Sepolia | Mainnet comparison |
|---|---|---|
Project payout | Split payment could be skipped | Reverted or fully delivered |
NFT transaction | NFT retained despite failed split | Reverted or fully delivered |
Reserved tokens | Cross-project payment could fail | Reverted or fully delivered |
Cash-out | 2.5% fee could be skipped | Reverted or fully delivered |
The results are based on the developer's published simulations and sampled gas ranges, not an independent comprehensive audit.
Importantly, the mainnet comparison did not reproduce the skipped-payment behavior in those sampled scenarios.
This suggests the proposed gas changes may be an important factor in the observed differences. It does not establish that all existing contracts are safe or that every Glamsterdam implementation would exhibit the same behavior.
Why the Discovery Matters for Ethereum's DeFi Ecosystem
Ethereum supports decentralized finance applications, NFT marketplaces, payment protocols, and other services that depend on smart contracts executing predictable financial operations.
For those applications, an unexpected difference between transaction completion and payment delivery could affect revenue distributions, fee collection, or contractual obligations.
The implications extend beyond the Juicebox contracts used in the tests.
For example, a decentralized application that distributes revenue between several participants may need to verify that each required payment completed, rather than relying exclusively on the status of the overall transaction.
Similarly, an NFT marketplace or token distribution platform may need additional checks to ensure that successful transactions produce the expected financial outcomes.
These are potential design implications, not evidence that the broader Ethereum ecosystem has already suffered such failures.
The discovery also arrives as Ethereum continues pursuing network improvements.
Bitnxt previously covered Ethereum's Glamsterdam Hoodi testing plans, including efforts involving a 200 million gas limit. The current compatibility concerns illustrate why performance improvements must be evaluated alongside the behavior of applications already deployed on the network.
Ethereum's expanding financial role makes these questions increasingly important.
Recent Bitnxt coverage of Ethereum staking activity and Ethereum market developments provides additional context for the network's growth and the importance of maintaining reliable infrastructure.
Bitnxt Take: Ethereum's Biggest Challenge Is Preserving Trust While Scaling
At Bitnxt, we believe the most important part of this discovery is not the possibility of avoiding a 2.5% fee. It is what the simulations reveal about the relationship between protocol upgrades and existing smart contracts.
Ethereum's long-term success depends on more than processing additional transactions or reducing execution costs.
It also depends on developers being able to trust the behavior of applications they have already deployed.
An immutable smart contract presents a particularly difficult challenge. Unlike conventional software, its original code may not be directly replaceable after deployment.
If a network upgrade changes the conditions under which that contract executes, developers may have limited options for correcting unexpected behavior.
This is why we view the Juicebox simulations as an important warning rather than evidence of an Ethereum-wide security crisis.
The testing process appears to be doing something valuable: identifying potential problems before the proposed rules become part of a production network upgrade.
There is also a broader lesson for the industry.
A successful blockchain transaction is not necessarily proof that every intended financial obligation was fulfilled.
Applications that depend on multiple internal payments should verify those outcomes explicitly and consider how their contracts behave when individual operations fail.
For Ethereum, the challenge is finding a balance between scaling the network and preserving the assumptions on which existing decentralized applications were built.
What Happens Next for the Ethereum Glamsterdam Upgrade?
Ethereum developers are reviewing the compatibility concerns as part of ongoing Glamsterdam discussions.
The November 16 All Core Devs Testing agenda included the Juicebox findings within the investigation of EIP-8037 and EIP-8038.
The Ethereum Foundation's published guidance also identified November 19, 2026, as a date for further discussion of findings and potential next steps.
Several possible responses remain open, including modifying specifications, adjusting testing plans, or determining whether certain application behaviors should be addressed at the contract level.
No final resolution to these specific findings had been established in the reviewed material as of November 17.
For developers, the immediate priority is to understand whether their contracts depend on similar gas-sensitive behavior.
For ordinary Ethereum users, the important distinction is that the reported results come from simulations of proposed execution rules. They do not demonstrate that Ethereum's live network currently allows users to bypass payments generally.
Conclusion
The Ethereum Glamsterdam upgrade tests have highlighted a potentially significant compatibility issue involving smart contract payments.
Juicebox V6 simulations demonstrated four scenarios where transactions completed despite failed internal operations, including skipped project payments and a forgiven withdrawal fee.
Although the findings do not establish a live mainnet exploit, they provide Ethereum developers with concrete examples of behavior that may need to be addressed before the upgrade proceeds.
For Bitnxt, the broader takeaway is clear: Ethereum's scaling roadmap must deliver not only greater capacity but also predictable execution for the financial applications already relying on the network.































