Ethereum's next hard fork, Glamsterdam, reached a stable multi-client devnet in May 2026 after more than 100 core contributors convened above the Arctic Circle in Svalbard, Norway, for the Soldøgn interop event. The upgrade targets a 3.3x increase in the network's gas limit — from 60 million to 2...
"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 2026
Ethereum's next hard fork, Glamsterdam, reached a stable multi-client devnet in May 2026 after more than 100 core contributors convened above the Arctic Circle in Svalbard, Norway, for the Soldøgn interop event. The upgrade targets a 3.3x increase in the network's gas limit — from 60 million to 200 million per block — alongside enshrined proposer-builder separation (ePBS), gas repricing via EIP-8037, and block-level access lists (EIP-7928) enabling parallel transaction execution.
If deployed as planned, Glamsterdam would raise Ethereum's base-layer throughput from approximately 1,000 TPS to 10,000 TPS while cutting gas costs by an estimated 78.6% for both simple transfers and complex smart contract interactions. Originally targeted for H1 2026, mainnet activation has slipped toward Q3 2026 according to multiple observers. A second upgrade, Hegotá, is scheduled for late 2026 and will introduce Verkle Trees and FOCIL censorship-resistance mechanisms deferred from Glamsterdam's scope.
The upgrade arrives at an inflection point for Ethereum's economics. Layer-2 rollups have successfully absorbed most transaction volume, pushing L1 fee revenue to multi-year lows. ETH supply is now mildly inflationary at ~0.23% annually as daily burns fall below validator issuance of approximately 1,700 ETH per day. Glamsterdam's fee reduction could further compress L1 revenue, intensifying the value-accrual debate that has weighed on ETH's price — trading near $2,022 with a market cap around $233 billion as of late May 2026.
Glamsterdam bundles three workstreams into a single hard fork, each targeting a different structural limitation of the Ethereum execution layer:
Enshrined Proposer-Builder Separation (ePBS) moves block-building mechanics from third-party relays into the protocol itself. Today, approximately 85.9% of Ethereum blocks are produced by the top three builders, with 86.8% of MEV value concentrated among them, according to empirical research published in 2026. Nearly 46% of blocks are produced by actors enforcing OFAC-based censorship policies through the relay layer. ePBS restructures slot timing with explicit deadlines for block construction, payload reveal, and attestations, reducing reliance on the Flashbots-dominated relay infrastructure.
Gas repricing via EIP-8037 introduces a cost_per_state_byte pricing model that recalibrates gas costs to match actual computational resource consumption on modern hardware. Many current gas prices were set years ago and no longer reflect execution costs. The recalibration reduces fees by 78.6% for standard operations while simultaneously making state-heavy contracts 8–10x more expensive to deploy, targeting approximately 60 GiB of state growth per year at the new 300M gas ceiling.
Block-Level Access Lists (BALs) via EIP-7928 pre-declare the accounts and smart contracts each block will interact with, enabling clients to process non-conflicting transactions concurrently across multiple CPU cores. This transforms Ethereum's sequential execution model into a parallelized one.
Together, these changes target a gas limit increase from 60 million to 200 million per block and throughput of approximately 10,000 TPS — roughly 10x the current base-layer rate.
The current MEV supply chain operates through an off-protocol relay system pioneered by Flashbots. Validators outsource block building to specialized builders who optimize transaction ordering for MEV extraction, then submit blocks through relays. The system works, but it introduces centralization risk at the relay layer and gives a small number of builders outsized influence over block content.
The concentration metrics are stark. According to a 2026 empirical study of proposer-builder separation effects, the Herfindahl index for builder market share stands at 3,186 — 2.5x more concentrated than miner distribution before Ethereum's Merge to proof-of-stake and a 300% increase from the month following the Merge.
ePBS addresses this by moving builder selection and block construction into the consensus protocol. Validators no longer need to trust external relays; the protocol itself manages the handoff between proposers and builders with cryptographic commitments and explicit timing windows. The Soldøgn interop confirmed that ePBS is now running stably on a multi-client devnet, with the external builder pipeline tested end-to-end across nearly all client implementations.
The censorship dimension is equally significant. With 46% of blocks currently subject to OFAC-based filtering through relay-level policies, protocol-enshrined building creates a neutral construction layer. However, ePBS alone does not solve censorship — that requires FOCIL (Fork-choice Inclusion Lists), which has been deferred to Hegotá.
Ethereum's gas limit has been a community-governed parameter since the network's launch. The current effective limit sits at 60 million following a community-driven increase in early 2025. Glamsterdam proposes a floor of 200 million — a 3.3x increase that would be the largest single gas limit expansion in Ethereum's history.
The increase is not a simple parameter change. Without guardrails, a higher gas limit enables faster state growth, which increases storage requirements for node operators and threatens decentralization. EIP-8037 serves as the counterbalance.
The proposal introduces a fixed cost_per_state_byte that reprices state-creation operations — creating new accounts, deploying contracts, and writing new storage slots. Under the revised model, these operations become 8–10x more expensive in gas terms, offsetting the inflationary effect of a higher gas limit on state growth. The target: no more than 60 GiB of annual state growth at 300 million gas, comparable to current growth rates at 60 million gas.
For users, the net effect is lower fees. The recalibration of gas costs for computation-heavy but state-light operations — which comprise the vast majority of DeFi interactions — results in a 78.6% reduction in costs for both simple ETH transfers and typical smart contract calls. The pricing adjustment reflects a core principle: gas should track real resource consumption, and computation has gotten cheaper while storage has not.
EIP-8037 reached finalization at Soldøgn, with the fixed cost_per_state_byte adopted and full repricing numbers delivered on bal-devnet-6.
EIP-7928, the Block-Level Access Lists specification, represents a structural change to Ethereum's execution model. Currently, the EVM processes transactions sequentially — each transaction must complete before the next begins, because any transaction could theoretically read or write any part of the global state.
BALs change this by requiring block producers to declare, in advance, which accounts and storage slots each transaction will access. This pre-declaration serves as a dependency graph. Clients can identify non-conflicting transactions and execute them in parallel across multiple CPU cores. For transactions that do conflict, sequential execution is preserved.
The performance implications are substantial. Parallel execution, combined with batched I/O and parallel state-root computation, is a key enabler of the throughput jump from ~1,000 TPS to ~10,000 TPS at the 200 million gas limit. Without BALs, simply raising the gas limit would produce blocks that take longer to process, potentially threatening the 12-second slot time.
The BAL track at Soldøgn ran on separate devnets from the ePBS chains, isolating optimization benchmarks from consensus-layer stabilization. Each optimization sat behind its own feature flag, allowing measurement in isolation. This methodical approach reflects the technical risk involved: BALs require changes to the block structure and engine API that are not backwards-compatible.
The Soldøgn interop, held in Longyearbyen, Svalbard, during early May 2026, produced three concrete outcomes according to the Ethereum Foundation's May 2026 Protocol Cluster update:
cost_per_state_byte pricing model was locked in with full repricing numbers.Infrastructure support came from three Ethereum Foundation teams: EthPandaOps shipped ethIQ and a "panda MCP server" for agentic developer workflows; Protocol Support maintained soldogn.xyz as the coordination hub; and the EF Digital Studio team documented the event.
The Foundation also announced leadership changes. Barnabé Monnot and Tim Beiko are departing, and Alex Stokes is taking a sabbatical. Will Corcoran, Kev Wedderburn, and Fredrik have been named as new Protocol Cluster leads — a significant reshuffle at the protocol governance layer.
Mainnet activation was originally targeted for H1 2026. Multiple observers now cite Q3 2026 as more realistic. The Soldøgn interop demonstrated technical readiness for the core components, but the transition from devnet to testnet to mainnet involves additional rounds of security auditing, client release cycles, and community coordination.
Glamsterdam's fee reduction arrives at an awkward moment for Ethereum's economic model. The network processed a record 2.6 million transactions in a single day in January 2026 and "barely generated revenue from it," according to industry analysis. Layer-2 rollups — Arbitrum, Base, Optimism, zkSync — have successfully absorbed the majority of transaction volume, exactly as Ethereum's scaling roadmap intended.
The consequence: L1 fee revenue has cratered. With validator issuance running at approximately 1,700 ETH per day and daily burns well below that level, Ethereum's supply is growing. The annual inflation rate stands at roughly 0.23%, according to Binance data — low by historical standards, but positive rather than deflationary. The EIP-1559 burn mechanism, which made ETH deflationary in high-demand periods of 2021-2023, is now structurally underperforming issuance.
A 78.6% reduction in gas fees through Glamsterdam could further compress L1 revenue per transaction, even as higher throughput increases transaction count. The math depends on demand elasticity: if 10x more transactions occur at 78% lower cost each, aggregate fee revenue rises modestly. If L2s continue to capture marginal demand growth, L1 revenue could remain flat or decline.
ETH traded near $2,022 as of late May 2026, with a market capitalization around $233 billion. The existing report in webthreepedia's archive documents this value-accrual tension in detail. Glamsterdam does not resolve the economic question — it deepens it by making the base layer cheaper and faster, which is good for users and L2s but ambiguous for ETH as a fee-generating asset.
Features deferred from Glamsterdam have been staged into Hegotá, Ethereum's second 2026 hard fork, targeted for late 2026. Key components include:
Verkle Trees: A replacement for Ethereum's current Merkle Patricia Trie data structure. Verkle Trees could reduce node storage requirements by up to 90%, enabling "stateless clients" that verify blocks without maintaining full state locally. This directly addresses decentralization concerns around rising hardware requirements.
FOCIL (Fork-choice Inclusion Lists): A censorship-resistance mechanism that lets validators propose mandatory transaction inclusion lists, preventing builders from systematically excluding transactions. A runnable prototype exists, according to the May 2026 update.
Native Account Abstraction: The scope of account abstraction requirements for Hegotá has been defined, with multi-client devnet validation as the next phase.
The twice-a-year hard fork cadence — established with Fusaka in December 2025 — represents a structural acceleration of Ethereum's upgrade velocity. Previously, major upgrades landed roughly once per year. The faster cadence allows features to be split across smaller forks rather than bundled into monolithic releases, reducing deployment risk but increasing coordination overhead.
Verkle Trees have been delayed multiple times in Ethereum's history. Their inclusion in Hegotá is not guaranteed; the Ethereum Foundation has acknowledged the technical complexity involved.
Glamsterdam's three core components — ePBS, EIP-8037 gas repricing, and EIP-7928 block access lists — reached stable multi-client devnet status in May 2026 following the Soldøgn interop in Svalbard.
The gas limit floor rises to 200 million, a 3.3x increase from 60 million, targeting ~10,000 TPS on the base layer. Gas fees for standard operations drop an estimated 78.6%.
ePBS moves block building on-chain, reducing reliance on the relay layer where the top three builders control 85.9% of block production and 46% of blocks enforce OFAC censorship.
EIP-8037 counterbalances state growth by making state-creation operations 8–10x more expensive, capping annual growth at ~60 GiB even at the higher gas limit.
Mainnet activation has slipped from H1 to Q3 2026. Devnet stability is confirmed; testnet and audit phases remain.
The fee reduction compounds Ethereum's L1 revenue problem. Supply is mildly inflationary at 0.23% annually as burns underperform issuance. Whether higher throughput offsets lower per-transaction fees depends on demand elasticity.
Leadership turnover at the Ethereum Foundation — including departures of Barnabé Monnot and Tim Beiko — adds a governance variable to the execution timeline.
Glamsterdam is the most ambitious execution-layer upgrade in Ethereum's post-Merge history. The technical progress is real: a stable multi-client devnet, finalized EIPs, and a clear path from interop to testnet. The upgrade addresses three persistent criticisms — high fees, low throughput, and centralized block building — with quantifiable targets.
The tension lies in what these improvements mean for Ethereum's economic model. A cheaper, faster base layer is unambiguously good for users and for the L2 ecosystem that depends on Ethereum for settlement and data availability. Whether it is good for ETH as an asset is a separate question. Lower fees mean lower burns, which means higher net issuance, which means the deflationary narrative that supported ETH's valuation in 2021-2023 continues to weaken.
Ethereum is executing its technical roadmap. The economic roadmap remains unwritten.