Ethereum Pauses Glamsterdam Testnet Rollout Over Smart Contract Gas Flaw

Ethereum core developers have indefinitely paused the testnet rollout of the major Glamsterdam network upgrade after independent security auditors discovered a critical accounting flaw that could trap funds inside live smart contracts.

By Amir Hassan | November 4, 2026

If you hold Ether (ETH) in your crypto wallet or earn yield in decentralized finance applications, the sudden pause on testnet development is actually reassuring news for your portfolio. With Ethereum trading at 2,423 USD today, everyday investors do not need to move their tokens, update their digital bank accounts, or respond to any third-party prompts. Instead, the delay demonstrates that protocol safeguards worked exactly as intended, stopping an algorithmic bug in the testing phase before it could jeopardize real money on the main blockchain.

The setback centers on the scheduled activation of the Glamsterdam upgrade on the public Hoodi testnet. Originally slated to go live on October 26, 2026, the trial deployment was abruptly stripped from client software configurations during All Core Devs Execution call 248. The move followed a technical disclosure from blockchain security auditing firm ChainSecurity, which discovered that an experimental fee-restructuring proposal could break error handling in automated applications, potentially freezing user balances permanently.

The Architecture

The Glamsterdam upgrade—tracked officially under meta-proposal EIP-7773—combines two major software updates: Amsterdam on the transaction-processing layer and Gloas on the network coordination layer. For the past two years, Ethereum focused heavily on scaling through Layer 2 express lanes. Glamsterdam marks a return to base-layer performance, aiming to make the underlying blockchain significantly faster and cheaper to use directly.

The primary engine driving this transformation is EIP-7928, known as Block-Level Access Lists. Historically, Ethereum processed transactions sequentially, like a grocery store operating a single checkout line. Every transfer, swap, or loan had to be verified one by one because the network could not predict in advance which digital balances a transaction would touch. EIP-7928 solves this bottleneck by requiring block creators to publish an upfront map of every account and storage slot involved. By knowing these dependencies beforehand, validator computers can process transactions in parallel across multiple processing cores.

I checked the All Core Devs execution repository and the client release notes for the Hoodi testnet: the activation code for late October was completely removed, confirming that the promised 200 million gas limit remains on slide decks rather than running code.

Alongside parallel processing, developers planned to ramp Ethereum’s block capacity—measured by its gas limit—from roughly 60 million up toward a target of 200 million. In simple terms, this would allow more than triple the transaction volume into each block, lowering baseline network fees for everyday users.

Consensus Mechanisms

While the processing layer handles parallel transactions, the coordination layer introduces EIP-7732, known as Enshrined Proposer-Builder Separation (ePBS). To understand why this matters to retail investors, consider how transactions currently travel from your wallet onto the blockchain.

Presently, the computers that secure Ethereum rely on third-party software relays to organize and bundle user transactions into blocks. While this setup prevents individual validators from exploiting trade order for profit, it has concentrated significant power into a handful of private relays. That middleman layer creates potential points of failure and censorship risks.

EIP-7732 moves this entire auction system directly into Ethereum’s core code. Rather than relying on private brokers, the blockchain natively divides duties between the packing crews that assemble transactions and the drivers who deliver them to the network ledger:

• Enshrined block builders — compete transparently to arrange transactions efficiently without centralized middleman relays.
• Network block proposers — validate and append blocks without needing to trust external third-party software.
• Censorship resistance — prevents private relays from filtering transactions or manipulating everyday decentralized trades.
• Faster data propagation — widens the communication window across the network, reducing orphaned blocks.

Network Health

Ethereum’s overall network health remains robust, but the transition toward higher transaction capacity revealed subtle technical cracks. The successful activation of Glamsterdam on the Sepolia testnet on October 6, 2026 proved that parallel execution and native block separation could function smoothly under simulated traffic. However, higher throughput introduces severe data bloat, requiring higher fees for creating permanent entries in the network ledger.

To combat database bloat, core developers introduced EIP-8037 and companion proposal EIP-8038. These proposals increase transaction fees specifically for writing new storage slots and introduce a dual-reservoir accounting structure that meters state data separately from normal computing instructions. While intended to prevent the network from becoming too heavy for standard hardware to run, this fee split triggered unintended side effects.

The audit by ChainSecurity revealed that separating state creation costs altered how the network measures remaining fuel during execution. Smart contracts work like automated digital vending machines: you insert funds, press a button, and the code dispenses a token. If something goes wrong, the machine is designed to safely cancel the order and return your money. Under EIP-8037, unexpected fuel exhaustion inside complex nested transactions would not revert cleanly. Instead, the contract could record the failure as complete, locking user funds permanently inside the contract.

Developer Ecosystem

The immediate reaction across Ethereum’s developer ecosystem was swift. ChainSecurity identified at least 13 live smart contracts on the Ethereum mainnet that would fail under the proposed gas changes. Because decentralized applications are immutable—meaning their original code cannot simply be overwritten or updated—changing base-layer rules can destroy protocols that have run reliably for years.

Ethereum Foundation researchers and protocol contributors agreed that protecting existing decentralized capital must take priority over meeting arbitrary upgrade deadlines. The All Core Devs team pulled the Glamsterdam code from the Hoodi testnet to give decentralized finance teams time to stress-test their protocols using updated developer tooling.

• 13 live mainnet contracts — identified by ChainSecurity as vulnerable to permanent fund freezes under the new fuel rules.
• Indefinite testnet delay — Hoodi activation on hold until researchers redesign fuel refund mechanics.
• Developer tool updates — testing frameworks updated to simulate the dual-reservoir state-accounting structure.
• Long-term mainnet roadmap — pushed past initial late-2026 targets while client teams resolve contract backward compatibility.

Final Assessment

The delay of Ethereum’s Glamsterdam upgrade is a practical victory for decentralized security over premature speed. While competing networks frequently advertise massive throughput figures on paper, pushing base-layer capacity requires balancing transaction speed against application safety and long-term hardware accessibility. Glamsterdam promises substantial improvements—enabling parallel processing through EIP-7928 and removing centralized relays via EIP-7732.

For everyday crypto investors, the pause on the Hoodi testnet ensures that when higher speed and lower fees finally arrive on the mainnet, they will not come at the expense of frozen savings. With broader market sentiment sitting in Fear at 34 on the Fear and Greed Index and Bitcoin trading at 80,366 USD, Ethereum’s core developers chose stability over haste, demonstrating why protocol conservatism remains the primary anchor of institutional confidence.

Amir Hassan’s take

Catching an error on a testnet before it freezes immutable money is the system working as intended, but pushing base-layer speed while protecting older apps is a brutal trade-off. I am watching the next All Core Devs meeting to see whether researchers drop EIP-8037 entirely or rewrite its gas counters. If developers rush the 200 million gas target before verifying legacy contracts, the danger shifts from simple delay to permanent capital loss.

The cryptocurrency market remains highly volatile. This article is for informational purposes only and does not constitute financial advice.