Ethereum's next hard fork, Glamsterdam, cleared its final engineering milestone on May 2, 2026, when over 100 core contributors concluded the Soldøgn Interop in Svalbard, Norway. The upgrade targets a 3.3x increase in Layer 1 execution capacity — from the current 60 million gas limit to 200 milli...
"By Friday, the group had delivered on its three core goals: alignment on a post-Glamsterdam gas limit floor of 200M, stable ePBS implementations running with external builders, and final EIP-8037 repricing numbers locked in." — Ethereum Foundation, Soldøgn Interop Recap (May 2, 2026)
Ethereum's next hard fork, Glamsterdam, cleared its final engineering milestone on May 2, 2026, when over 100 core contributors concluded the Soldøgn Interop in Svalbard, Norway. The upgrade targets a 3.3x increase in Layer 1 execution capacity — from the current 60 million gas limit to 200 million — while enshrining proposer-builder separation (ePBS) directly into the protocol and recalibrating gas costs via EIP-8037 and EIP-7904.
Mainnet activation, initially targeted for June 2026, has slipped to Q3 2026 pending further devnet validation. The fork's economic implications are material: a projected 78.6% reduction in L1 gas fees, elimination of the network's 80-90% dependence on off-chain MEV-Boost relays, and a structural repricing of state creation to prevent unbounded state growth at higher throughput. For an ecosystem generating $1.42 billion in fee revenue during Q1 2026 while processing 200.4 million transactions, Glamsterdam represents the most significant execution-layer overhaul since the Merge.
Glamsterdam bundles three interdependent protocol changes into a single hard fork:
| Component | EIP | Function | |-----------|-----|----------| | Enshrined PBS | EIP-7732 | Moves block building into protocol, replaces MEV-Boost relays | | Block-Level Access Lists | EIP-7928 | Enables parallel transaction execution | | State Creation Repricing | EIP-8037 | Raises state-write costs to cap state growth at ~60 GiB/year | | General Repricing | EIP-7904 | Recalibrates opcode gas costs using modern hardware benchmarks | | State-Access Repricing | EIP-8038 | Increases cold-read gas costs to reflect I/O reality |
The Ethereum Foundation's May 11, 2026 Protocol Cluster update confirmed all three core goals from the Soldøgn Interop are complete: 200M gas limit alignment, stable multi-client ePBS devnet operation, and finalized EIP-8037 parameters. The upgrade is running on bal-devnet-6 with external builders tested end-to-end.
Leadership transition: Tim Beiko, Barnabé Monnot, and Alex Stokes are stepping away from protocol cluster leadership. Will Corcoran, Kev Wedderburn, and Fredrik are assuming lead roles.
Raising the gas limit from 60M to 200M without repricing would produce unbounded state growth — the central risk that delayed previous gas limit increases. EIP-8037 solves this with a fixed cost_per_state_byte model calibrated to cap state growth at approximately 60 GiB per year at a 300M gas block limit.
Specific cost increases under EIP-8037:
EIP-7904 General Repricing delivers offsetting reductions:
EIP-8038 completes the picture by raising cold account and storage read costs, better reflecting the actual I/O expense of accessing infrequently used state. The net result: state creation and cold reads become expensive (reflecting real costs), while execution of common opcodes becomes cheap (reflecting modern CPU throughput).
| Metric | Current (May 2026) | Post-Glamsterdam (Projected) | |--------|--------------------|-----------------------------| | Base fee | ~0.05 gwei | Further reduction expected | | Simple ETH transfer | ~$0.01 | ~$0.002 | | Average transaction fee | $0.16–$0.22 | Sub-$0.05 projected | | Gas limit | 60M | 200M | | Daily transactions (record) | 2.89M (Feb 2026) | Capacity for 9.5M+ |
Ethereum's Q1 2026 fee revenue totaled $1.42 billion across 200.4 million transactions — a 43% quarter-over-quarter growth rate. The question Glamsterdam raises: does 3.3x capacity expansion compress per-unit fees faster than volume growth can compensate?
Under the current MEV-Boost architecture:
This concentration exists because MEV-Boost is an off-chain, opt-in system. Validators must trust relays, relays must trust builders, and builders with more exclusive order flow win more blocks, entrenching their position.
ePBS replaces the trusted relay model with an in-protocol commit-reveal flow:
The multi-client Glamsterdam devnet has demonstrated stable ePBS operation with external builders, according to the Ethereum Foundation's May update. Relay infrastructure (MEV-Boost, Flashbots relays) becomes optional rather than structurally necessary.
ePBS does not eliminate builder dominance. According to a Web3 Developer Forum analysis, exclusive order flows — particularly those channeled through Flashbots' MEV-Share — still give certain builders disproportionate transaction access. Unless exclusive order flow agreements are dissolved or made public, builder concentration may persist even post-ePBS.
Current Ethereum execution is serialized: transactions execute one after another because the EVM cannot know in advance which state slots a transaction will touch. EIP-7928 changes this by requiring each block to declare which accounts and storage slots it will access.
Mechanism:
Throughput target: Combined with ePBS and the 200M gas limit, the Ethereum Foundation projects L1 throughput approaching 10,000 TPS — up from an effective ~15-30 TPS today.
This represents a fundamental architectural shift. Ethereum moves from a single-threaded execution model (inherited from the original 2015 design) to multi-core parallel processing, while maintaining deterministic state transitions.
Layer 2 networks now account for approximately 95% of Ethereum's total transaction throughput, with Base and Arbitrum holding 77% of rollup liquidity (per existing webthreepedia analysis). L2 daily transactions exceed 2 million, surpassing Ethereum mainnet.
The upgrade affects L2 economics through two channels:
1. Blob capacity expansion: Glamsterdam targets an increase from the current blob allocation to 72+ blobs per block. More blob space means lower data availability costs for rollups, with settlement costs projected to fall ~70%.
2. L1 fee competition: A 78.6% reduction in L1 fees narrows the cost advantage that L2s currently hold. If L1 simple transfers cost $0.002, the economic argument for routing transactions through L2s weakens for low-complexity operations.
| L2 Cost Component | Current | Post-Glamsterdam | |-------------------|---------|-----------------| | Data posting (blobs) | Primary cost driver | ~70% reduction | | Settlement/proof verification | Secondary | Reduced via repricing | | L1 execution alternative | ~$0.16-$0.22 avg | ~Sub-$0.05 |
The implication: L2s must compete on features (speed, specific application support, ecosystem) rather than pure cost advantage. Networks that have built independent economic moats — Base's Coinbase integration, Arbitrum's DeFi ecosystem — are better positioned than pure fee arbitrage L2s.
Applying the economic value framework to Glamsterdam reveals structural shifts in who captures value:
Winners:
Losers:
Ethereum's current inflation rate stands at 0.35% annualized, with 35.86M ETH staked (28.91% of supply) earning approximately 3.08% nominal yield. Post-Glamsterdam, if per-unit fees fall faster than volume grows, net issuance could remain inflationary for extended periods — a headwind the market has not fully priced given current ETH valuations.
| Milestone | Status | Date | |-----------|--------|------| | Soldøgn Interop (Svalbard) | Completed | May 2, 2026 | | Protocol Cluster Update | Published | May 11, 2026 | | Multi-client devnet (bal-devnet-6) | Running | Active | | Mainnet activation (original target) | Slipped | Was June 2026 | | Mainnet activation (revised estimate) | Pending | Q3 2026 | | Hegotá (follow-up fork) | Planning | Late 2026 |
FOCIL (Fork-Choice Enforced Inclusion Lists) — the censorship-resistance mechanism — was removed from Glamsterdam to avoid scope creep and is now confirmed as Hegotá's headline feature. This means Glamsterdam ships performance without censorship guarantees; validators can still exclude transactions post-upgrade.
Glamsterdam represents Ethereum's attempt to resolve the fundamental tension between decentralization and performance that has defined its roadmap since 2015. The 200M gas limit, parallel execution, and enshrined PBS collectively target 10,000 TPS on L1 — a figure that would have been dismissed as incompatible with decentralization three years ago.
The economic tradeoffs are material. Lower per-unit fees may reduce ETH burn rates, extending net inflation. ePBS removes relay operators from the value chain but may not eliminate builder concentration. Cheaper L1 execution narrows the cost moat protecting L2 operators. And the deferral of FOCIL means performance arrives without censorship guarantees — a deliberate prioritization by the Ethereum Foundation.
For an ecosystem where 85-90% of total value flows remain subsidy-driven (per prior webthreepedia analysis), Glamsterdam's capacity expansion is necessary but not sufficient. The question is whether 3.3x more throughput generates proportionally more fee revenue, or whether it simply moves Ethereum further down the commoditization curve where execution capacity is abundant and value accrues elsewhere in the stack.
Q3 2026 activation remains the target. Devnets are running. The leadership is changing. The economics are shifting. What has not changed: Ethereum still generates approximately $13.7 billion in identifiable on-chain income against $55-71 billion in annual subsidies. Glamsterdam does not fix that ratio. It bets that scale will.