Ethereum's Glamsterdam hard fork — combining the Amsterdam execution-layer and Gloas consensus-layer upgrades — has locked its final EIP bundle across Devnet-5 and is moving toward public testnet deployment. Sepolia testnet activation is tentatively scheduled for August 3, 2026, Hoodi for August ...
"Glamsterdam is the most ambitious protocol overhaul since the Merge. We're fundamentally reorganizing how Ethereum builds blocks, verifies transactions, and prices permanent storage." — Ethereum Foundation, Glamsterdam Upgrade Documentation
Ethereum's Glamsterdam hard fork — combining the Amsterdam execution-layer and Gloas consensus-layer upgrades — has locked its final EIP bundle across Devnet-5 and is moving toward public testnet deployment. Sepolia testnet activation is tentatively scheduled for August 3, 2026, Hoodi for August 17, with mainnet activation targeting a September 16 window, pending stable multi-client results across both testnets.
The upgrade ships 10 Ethereum Improvement Proposals tracked under Meta EIP-7773. Two headline changes — EIP-7732 (Enshrined Proposer-Builder Separation) and EIP-7928 (Block-Level Access Lists) — address structural bottlenecks that have defined Ethereum's architecture since the Merge. Supporting EIPs repricing gas costs, expanding contract size limits, and capping state growth complete the bundle. If executed on schedule, Glamsterdam clears the path for validators to signal a gas-limit increase from 60 million to 200 million per block, a 233% expansion in available block space.
Ethereum L1 already set a daily transaction record of 2.9 million on June 8, 2026, according to on-chain data reported by CoinDesk, surpassing the previous high of 1.7 million set during 2021. Q1 2026 logged 200.4 million L1 transactions, up 43% from Q4 2025's 145 million. Average transaction fees have dropped to $0.21, with some periods reaching $0.15 — down over 50% year-over-year. Glamsterdam aims to push these numbers further by enabling parallel transaction processing and eliminating relay middleware from block construction.
The single largest architectural change in Glamsterdam is EIP-7732, which moves Proposer-Builder Separation (PBS) from an off-chain relay system into the Ethereum consensus layer itself.
Today, over 90% of Ethereum blocks are constructed using external relay infrastructure, primarily MEV-Boost, developed by Flashbots. Under this system, validators (proposers) outsource block construction to specialized builders who compete to assemble the most profitable block. The winning block passes through a trusted relay that mediates between builder and proposer. This relay layer — operated by a handful of entities — represents a centralization vector. Between late 2023 and early 2024, three builders produced nearly 80% of all Ethereum blocks. Approximately 46% of blocks were produced by actors enforcing OFAC-based censorship policies, according to research from the Ethereum community.
EIP-7732 replaces this off-chain relay dependency with protocol-native messaging. Under ePBS, bids, commitments, and payments between proposers and builders are exchanged through protocol-defined messages verifiable on-chain. The relay is eliminated as a trust intermediary. The specification introduces a new beacon-chain payload type where builders commit to an execution payload header, and proposers include that header without needing to see or trust the full block contents until after commitment.
Two academic papers published in January 2026 flagged design considerations. A paper titled "Enshrined Proposer Builder Separation in the presence of Maximal Extractable Value" (arXiv:2601.12989) identified formal limitations in the ePBS design using mathematical analysis. A second paper, "The Free Option Problem of ePBS" (arXiv:2509.24849), found that builders gain a short-dated option to prevent their committed execution payload from becoming canonical without incurring additional penalty — a design gap core developers are aware of and monitoring through devnet testing.
The second headline EIP introduces Block-Level Access Lists (BALs). EIP-7928 requires block producers to declare in advance which accounts and storage slots a block will touch. This declaration allows validator nodes to distribute non-conflicting transactions across multiple CPU cores for parallel verification.
Current Ethereum execution is serial — transactions in a block are processed one after another, regardless of whether they touch overlapping state. This creates a throughput ceiling determined by single-core CPU performance. BALs break this constraint by enabling parallelism at the block level, a capability Ethereum has lacked since launch.
The performance target, according to developer estimates cited by multiple sources including CoinDesk and The Defiant, is a path toward 10,000 TPS on L1 — roughly 10x current throughput. This target is aspirational rather than immediate; the full TPS gain depends on subsequent validator adoption of higher gas limits and client optimization, not solely the Glamsterdam fork itself.
Beyond the two headline changes, eight additional EIPs complete the Glamsterdam bundle:
EIP-8037 (State Growth Pricing): Finalized in May 2026, this EIP sets a fixed cost per byte of new state creation and carves out a separate gas reservoir for state growth. According to The Defiant, this gives client teams a sustainability ceiling under which a 200M gas limit can be raised without bloating the state database past 120 GiB per year — a critical constraint for node operators running consumer hardware.
EIP-7976 (Calldata Floor Repricing): Raises the floor cost of calldata, aligning economic incentives so that rollups and data-posting protocols shift more aggressively toward blob space (introduced in Dencun's EIP-4844) rather than consuming L1 calldata.
EIP-7954 (Contract Size Limit): Raises the maximum deployable contract size from approximately 24 KiB to 32 KiB, giving developers additional deployment headroom for complex smart contracts.
EIP-7981 (Access List Cost Adjustment): Raises EIP-2930 access list costs to reflect actual computational overhead.
Additional EIPs in the bundle — EIP-7778, EIP-7843, EIP-8024, and EIP-8061 — address validator exit queue improvements, ETH transfer log formatting, and various operational refinements.
Glamsterdam does not enforce a 200M gas limit directly. Instead, the upgrade removes technical barriers — primarily state bloat risk (addressed by EIP-8037) and serial execution bottlenecks (addressed by EIP-7928) — that currently make a gas-limit increase unsafe.
The gas limit is set by validators via a signaling mechanism. Validators currently coordinate around a 60M gas limit. After Glamsterdam, the community expectation is that validators will begin signaling a gradual increase toward 200M as node operators confirm their infrastructure handles larger blocks without degraded propagation times.
The projected impact of a 200M gas limit, combined with parallel execution, includes:
These projections remain conditional on validator adoption of higher gas limits and stable client performance post-fork.
| Milestone | Target Date | Status | |---|---|---| | Devnet-5 (final devnet) | June 2026 | Active — multi-client testing with full EIP bundle | | Sepolia testnet fork | August 3, 2026 | Tentative | | Hoodi testnet fork | August 17, 2026 | Tentative | | Mainnet activation | ~September 16, 2026 | Tentative, pending testnet stability |
The upgrade was originally targeted for H1 2026 but was pushed back after the Soldøgn interop devnet concluded on May 2, 2026, requiring additional testing cycles. Devnet-5 is the final devnet stage before public testnet deployment. Core developers will set the mainnet activation date only after multi-client stability holds for several epochs across both Sepolia and Hoodi testnets.
Based on the cadence of recent forks (Dencun required approximately two months of public testnet seasoning), a September–October 2026 mainnet window represents the base case. Slippage into Q4 2026 remains possible if testnet instability surfaces.
Three EIPs directly affect validators and institutional staking operators, according to analysis from Figment, one of the largest non-custodial Ethereum staking infrastructure providers:
EIP-7732 (ePBS): Validators no longer need to run MEV-Boost relay software. Block construction is handled at the protocol level, reducing operational complexity. However, validators must upgrade their consensus clients to support the new ePBS message types.
EIP-8061: Improves the validator exit queue mechanism. For large, concentrated exit requests — the kind institutional operators are more likely to encounter — processing times could drop from weeks to days.
EIP-8080: Addresses additional validator operational concerns (specifics pending final devnet confirmation).
Staking providers including Everstake and Figment have indicated they will deploy Glamsterdam client updates across their infrastructure ahead of mainnet activation. Institutional stakers operating self-managed validators will need to coordinate client upgrades across both execution and consensus layers.
ETH is trading at approximately $1,857 with a market capitalization of $224.1 billion, according to MetaMask price data. Polymarket bettors give Ethereum a 17% probability of reaching $3,000 in 2026, per a report from Bitcoin.com.
The free-option problem: Academic research (arXiv:2509.24849) has identified that under ePBS, builders can effectively walk away from committed blocks without penalty, creating a design gap that could be exploited. Core developers are monitoring this through devnet testing but have not proposed a specific mitigation within Glamsterdam's scope.
Gas-limit governance: The 200M gas target depends on voluntary validator signaling. There is no guarantee validators will coordinate a rapid increase, and some node operators may resist higher limits if their hardware struggles with larger blocks.
State growth sustainability: While EIP-8037 caps state growth at 120 GiB per year, this ceiling assumes the gas repricing holds under real-world usage patterns. Edge cases could emerge once the higher gas limit goes live on mainnet.
Testnet delays: The upgrade has already slipped from H1 to H2 2026. Further delays during Sepolia or Hoodi testing would push mainnet activation into Q4 2026 or later, overlapping with the planned Hegotá fork, which focuses on state management and Verkle Trees.
L2 impact: Lower L1 gas costs could cannibalize some L2 rollup activity. Rollups derive economic value from the spread between their own fees and L1 data-posting costs. A 78% L1 fee reduction narrows that spread.
Glamsterdam represents a structural overhaul of Ethereum's block production and execution architecture. By enshrining PBS at the protocol layer and enabling parallel transaction processing, the upgrade addresses two of the longest-standing criticisms of Ethereum L1: MEV relay centralization and serial execution throughput limits.
The economic implications are material. A 200M gas limit, if adopted by validators, would expand available L1 block space by 233% and could reduce per-transaction costs by up to 78%. This would make Ethereum L1 competitive with some L2 rollups on cost while retaining higher security guarantees. The question is whether this pulls activity back to L1 or simply expands the total addressable market for Ethereum's ecosystem.
Execution risk remains. The upgrade has already slipped by one half-year, and academic research has flagged an unresolved design gap in the ePBS mechanism. The September mainnet target is conditional on clean testnet runs across Sepolia and Hoodi — a process that historically takes two to four months.
For validators and institutional stakers, Glamsterdam simplifies operations by removing MEV-Boost dependencies but requires coordinated client upgrades across both execution and consensus layers. The exit-queue improvements in EIP-8061 offer a tangible operational benefit for large staking operations.
The data is clear on one front: Ethereum L1 usage is at all-time highs, with 200.4 million transactions in Q1 2026 and fees at historic lows. Glamsterdam's bet is that removing execution bottlenecks will convert latent demand into sustained throughput growth.