The Glamsterdam protocol represents the largest architectural upgrade to the network’s base layer, aimed at radically increasing throughput, reducing costs, and improving censorship resistance. Originally scheduled for the second quarter, the rollout was postponed to the third quarter due to the need for additional cross-client compatibility debugging and the completion of testing for key consensus mechanisms. The upgrade builds on the foundation established by the previous hard fork and focuses on structurally reorganizing the processes used to create and verify blocks.
Unlike previous iterations, which primarily optimized data availability for Layer 2 solutions, this upgrade directly addresses the scalability trilemma at the base layer. Development is focused on three interconnected areas:
Key Success Indicator: The network’s developer foundation has officially identified a post-upgrade gas limit target of 200 million units. This represents more than a threefold increase from the current baseline of 60 million and will require the new synchronization mechanisms to operate flawlessly.
The upgrade combines changes at both the Execution Layer and the Consensus Layer. Several proposals have been selected as the protocol’s priority headline features.
Under the current architecture, interaction between validators, which propose blocks, and block builders relies on trusted third-party middleware such as MEV-Boost. This out-of-protocol dependency creates a critical bottleneck: validators are required to broadcast and execute transactions within a compressed two-second window, severely limiting the size of the block payload.
EIP-7732 embeds proposer-builder separation directly into the protocol core, eliminating the need for trusted relays. To address the two-second bottleneck, it introduces the Payload Timeliness Committee (PTC) and a dual-deadline mechanism. This allows validators to attest separately to the consensus block and the execution payload, extending the data propagation window from two seconds to approximately nine seconds. The additional time is essential for safely processing larger blocks without overwhelming network nodes.
The traditional transaction-processing model requires validators to execute operations strictly one after another because the network does not know which data a transaction will access until it is actually executed. This creates the equivalent of a single-lane road and limits throughput.
Block-Level Access Lists radically change this paradigm by providing an advance map of all transaction dependencies within a block. The execution layer stores a complete access list detailing every account affected by the transactions, together with the final results of those changes. The block header contains a compact cryptographic commitment to this list.
Technical Advantage: This map allows validators to process non-conflicting transactions in parallel. In addition, nodes can synchronize much faster by reading the final results without having to re-execute every transaction, using a mechanism known as executionless sync. This lays the foundation for exponential growth in Layer 1 throughput.
Raising the gas limit is useful only if the network can handle the resulting state growth without suffering performance degradation. EIP-8037 introduces more intelligent pricing for data-storage operations by increasing the cost of operations that create new state while keeping computation relatively inexpensive.
This reform is directly aligned with the 200 million gas target. Without it, increasing the limit would cause excessive state growth, making it prohibitively expensive for ordinary participants to operate and maintain a full node. Aligning gas costs with real long-term hardware expenses ensures that scaling remains sustainable.
Reaching a gas limit of 200 million units is not guaranteed as an immediate launch-day setting, but rather represents a credible post-upgrade capacity target. The gas limit determines how much computation can fit into a single block. Increasing it allows more transactions and smart contract operations to be processed in each block, but it also makes blocks larger and more resource-intensive to validate.
The synergy between ePBS, Block-Level Access Lists, and revised gas pricing is specifically designed to make this increase safe. ePBS provides more time for propagation, access lists accelerate verification, and EIP-8037 controls long-term storage costs.
The postponement to the third quarter is driven by the complexity of integrating ePBS. Ensuring seamless compatibility between different consensus and execution client implementations requires exhaustive testing under conditions that closely resemble the production environment. Upgrade-specific devnets are already operational, while interoperability testing has been actively conducted during dedicated developer events.
The primary scheduling risk remains the possibility of additional issues being discovered during public testnet deployments, particularly in relation to cross-client consistency when processing blocks with parallel execution. Mainnet activation will occur only after every stage of test deployment has been completed successfully.
While finalizing the current upgrade, the development team is already defining the outline of the next major iteration, provisionally named Hegotà. Its expected implementation window is the end of the year or later, depending on the pace of development.
A key component of Hegotà will be the introduction of Verkle Trees, which will replace the network’s current data structure with a more efficient one and fully enable stateless clients. This will allow nodes to validate blocks without storing enormous amounts of historical state data, representing the final step toward making node operation more accessible.
EIP-8141, known as Frame Transactions, is also under consideration. It would introduce native account abstraction, make transaction validation programmable, and allow users to select post-quantum signature schemes instead of being restricted exclusively to ECDSA. This work forms part of the network’s long-term strategy for achieving quantum resistance.
“The ultimate goal of scaling is to create a system capable of processing thousands of transactions per second while preserving full decentralization and security. We are not sacrificing one for the other; we are changing the architecture so that we can achieve both.”
— Vitalik Buterin, co-founder of the network
