Ethereum's Glamsterdam hard fork — the network's most significant upgrade since The Merge — is approaching mainnet activation with a target window of Q2–Q3 2026. The upgrade centers on three protocol-level changes: EIP-7732 (enshrined proposer-builder separation), EIP-7928 (block-level access lis...
"This [ePBS] ensures that block builder centralization does not creep into staking centralization, but it leaves the question: what do we do about block builder centralization?" — Vitalik Buterin, Ethereum Co-Founder
Ethereum's Glamsterdam hard fork — the network's most significant upgrade since The Merge — is approaching mainnet activation with a target window of Q2–Q3 2026. The upgrade centers on three protocol-level changes: EIP-7732 (enshrined proposer-builder separation), EIP-7928 (block-level access lists enabling parallel execution), and EIP-8037 (state creation repricing). Combined, these changes raise Ethereum's gas limit from 60 million to 200 million per block, targeting approximately 10,000 TPS on Layer 1 — a 500x increase from the current 15–20 TPS — and a projected 78% reduction in gas fees.
A week-long interoperability workshop in Svalbard, Norway, concluded on May 2, 2026, with over 100 core contributors validating a stable multi-client devnet. The event also produced leadership transitions within the Ethereum Foundation's Protocol Cluster, with Tim Beiko and Barnabé Monnot departing and Alex Stokes beginning a three-month sabbatical. Features not included in Glamsterdam — including Verkle Trees, FOCIL, and native account abstraction — have been formally scoped into Hegotá, a second 2026 hard fork targeting late in the year.
Between April 28 and May 2, 2026, just over 100 Ethereum core contributors convened in Longyearbyen, Svalbard (78°N, above the Arctic Circle) for the Soldøgn Interop — a single-track, multi-client development sprint. According to the Ethereum Foundation's May 2 recap, the group delivered on three primary goals:
The devnet initially targeted a 4 EL × 4 CL client configuration on Monday. A shift to Tuesday yielded a stable 4×3 setup that ran through the week. By Friday, cross-client edge cases in execution-layer bid pathways had been identified and resolved.
Additional EIP decisions made during the sprint: EIP-8061 (exit/consolidation churn increase) was included. EIP-8080 (consolidation queue exits) was declined. EIP-8045 (slashed-validator duty removal) was scoped to proposer duties only. EIP-7688 (SSZ stable containers) was retained but deferred from devnet-1. EIP-8237 was deferred from Glamsterdam entirely.
The headline feature. EIP-7732 moves proposer-builder separation (PBS) from off-chain middleware into the consensus protocol itself.
Current state: MEV-Boost, the dominant off-chain relay system, processes the vast majority of Ethereum blocks. According to relayscan.io data from April 2026, the top three block builders — Titan (52.16% of blocks), BuilderNet (24.63%), and Quasar (15.06%) — produce over 91% of MEV-boosted blocks. On the relay side, relay.ultrasound.money handles 33.92% of payloads, followed by titanrelay.xyz at 24.19%.
What changes: Under ePBS, the block proposer commits to a block header and a separate builder constructs the execution payload. The handoff is enforced at the protocol level, eliminating the need for trusted relay intermediaries. A Payload Timeliness Committee validates builder submissions. Builders gain approximately 7 additional seconds per slot (expanding the data propagation window from ~2 to ~9 seconds) to assemble more complex blocks.
Tradeoffs: Research from arXiv (2506.18189) warns that ePBS could amplify profit centralization among builders. The Gini coefficient for builder profits rises from 0.1749 under standard PoS to 0.8358 under ePBS, according to the paper. The "free option problem" — where a builder can choose not to furnish execution data after a proposer commits to the header — remains an active area of research.
EIP-7928 introduces Block-Level Access Lists (BALs), requiring each block to declare upfront which state accounts it will read and write. This structural change enables three concurrent optimizations:
A BAL root is embedded in each block header, enabling verification. According to BlockEden.xyz's analysis, BALs could deliver a 10–30x execution throughput boost. A benchmark dashboard was deployed during the Svalbard interop for cross-client performance comparison.
This puts Ethereum in direct competition with parallel-execution chains. Monad targets 10,000+ TPS with ~400ms block times; Solana achieves 1,000–1,500 real-world TPS with ~400ms slots. Ethereum's approach achieves comparable throughput while maintaining its 12-second block time and 500,000+ validator set — compared to Solana's approximately 1,500 validators.
EIP-8037 increases the gas cost of creating new on-chain state (accounts, storage slots). The mechanism shifted during the Svalbard sprint from dynamic per-state-byte pricing to a fixed-cost model, simplifying testing and benchmarking. Future adjustments would occur only at hard-fork boundaries.
Accounting was also restructured: state creation costs now settle at end-of-call-frame rather than mid-execution. Supporting changes include EIP-7904 (targeting a 78.6% gas fee reduction through broader repricing) and EIP-8038 (increased cold storage access costs).
The net effect: routine transactions become cheaper while state-expanding operations — which impose long-term storage costs on the network — carry higher fees. This aligns economic incentives with Ethereum's persistent storage burden, an issue identified in the network's growing state size, which currently exceeds 300 GB.
The current Ethereum gas limit sits at 60 million per block, a figure that rose from 36 million in early 2025 via validator voting. Glamsterdam's target of 200 million represents a 3.3x increase.
Context on current fee conditions: Ethereum L1 gas prices averaged approximately 0.50 gwei in January 2026 and dropped to historic lows of 0.16 gwei by April 2026, according to Blocklr. A Uniswap swap that cost $200 in May 2021 now costs roughly $0.03 at these gas levels. L2 networks operate at 90–99% lower cost still: OP Mainnet charges approximately $0.0007 for a simple ETH transfer.
The 200M target is enabled by ePBS (which decouples execution from consensus timing), BALs (which allow safe throughput scaling), and EIP-8037 (which prevents state bloat from cheap computation). Without all three, a raw gas limit increase would risk degrading node performance and raising hardware requirements — the same concern that has historically constrained Ethereum's scaling.
ePBS formalizes the MEV supply chain at the protocol level, replacing off-chain trust assumptions with on-chain enforcement. This carries significant structural consequences.
For validators: Solo stakers benefit most. Under the current system, validators without MEV-Boost access forfeit block-building revenue to sophisticated operators. Under ePBS, all validators — including home stakers — receive builder bids through the protocol. The computational burden on validators decreases, as they no longer need to construct blocks themselves.
For builders: The top three builders currently control 91.85% of MEV-boosted blocks. ePBS does not eliminate this concentration. Instead, it contains the centralization to the builder role, preventing it from spilling into validator/staker centralization. Builders gain more construction time per slot but face protocol-enforced transparency requirements.
For relays: MEV-Boost relays — the trusted intermediaries matching proposers and builders — become redundant at the protocol level. Over 90% of Ethereum validators currently rely on these off-chain relays. Removing them eliminates a systemic trust dependency but also removes a coordination layer that the ecosystem has relied on since late 2022.
Unresolved issues: Builder signature commitment requirements remain contentious. A proposal requiring 1 ETH builder staking for P2P Sybil resilience raised concerns during the Svalbard sprint. These items are expected to be resolved on subsequent AllCoreDevs calls.
Approximately 30 EIPs were declined from Glamsterdam's scope, with several high-profile features explicitly deferred to Hegotá, Ethereum's second planned 2026 hard fork targeting late in the year:
According to the Ethereum Foundation's May 11 Protocol Cluster update, FOCIL is confirmed as the consensus-layer headliner for Hegotá.
The Svalbard sprint coincided with significant personnel changes in Ethereum's protocol governance. According to the Ethereum Foundation's May 11 update:
Departures: Tim Beiko and Barnabé Monnot have left the Ethereum Foundation. Alex Stokes began a three-month sabbatical immediately following the event.
New Protocol Cluster Leads: Will Corcoran, Kev Wedderburn, and Fredrik.
Interim coordination roles: Pari (interim ACDC coordinator), Barnabas (interim ACDT coordinator). ACDE co-leads confirmed as Nixo and Ansgar. ACDT rotation includes Mario, Barnabas, and Danceratopz.
The transitions represent the most significant leadership turnover in Ethereum's protocol development since the Foundation's restructuring in early 2025. The departures of Beiko (who coordinated AllCoreDevs calls for years) and Monnot (a key mechanism design researcher) occur at a critical implementation phase.
Glamsterdam represents Ethereum's most ambitious L1 scaling effort since the network transitioned to proof-of-stake in September 2022. The upgrade addresses a structural tension: Ethereum's gas fees have already fallen to historic lows (0.16 gwei in April 2026), but L1 throughput remains at 15–20 TPS — inadequate for the network's stated ambition of becoming the global settlement layer.
The technical architecture is sound. ePBS, BALs, and gas repricing form an interlocking system: ePBS creates the slot structure that allows higher gas limits, BALs make those limits safe through parallel execution, and repricing prevents cheap computation from inflating state. The Svalbard interop demonstrated multi-client stability under these combined changes.
The open questions are governance and centralization. Three builders producing 91% of blocks is a structural risk that ePBS acknowledges but does not resolve — that falls to FOCIL in Hegotá. The departure of core governance figures during a critical implementation window introduces execution risk that new leadership must navigate. And the 200M gas limit target, while technically validated on devnets, will require sustained validator coordination to implement via on-chain voting after the fork.
The data suggests Ethereum is executing on its roadmap. Whether the market rewards that execution remains a separate question — ETH/BTC hit a 10-month low at 0.028 in recent weeks, and L1 fee revenue continues to decline as activity migrates to L2s. Glamsterdam may be the upgrade that determines whether Ethereum's L1 remains economically relevant or becomes primarily a settlement and data-availability layer for its L2 ecosystem.